Aristote et l'âme humaine: Lectures de 'De anima' III offertes à Michel Crubellier [1 ed.] 904294045X, 9789042940451, 9789042940468

Citation preview

ARISTOTE TRADUCTIONS ET ÉTUDES

ARISTOTE ET L’ÂME HUMAINE LECTURES DE DE ANIMA III OFFERTES À MICHEL CRUBELLIER

ÉTUDES RÉUNIES PAR

GWELTAZ GUYOMARC’H CLAIRE LOUGUET CHARLOTTE MURGIER

LOUVAIN-LA-NEUVE

PEETERS 2020

ARISTOTE ET L’ÂME HUMAINE

ARISTOTE TRADUCTIONS ET ÉTUDES

ARISTOTE ET L’ÂME HUMAINE LECTURES DE DE ANIMA III OFFERTES À MICHEL CRUBELLIER

ÉTUDES RÉUNIES PAR

GWELTAZ GUYOMARC’H CLAIRE LOUGUET CHARLOTTE MURGIER

Ouvrage publié avec le concours de l’UMR 8163 Savoirs Textes Langage (CNRS, Université de Lille) et de l’UMR 7219 SPHERE (CNRS, Université de Paris) – équipe GRAMATA (Université Paris 1 Panthéon-Sorbonne).

PEETERS LEUVEN - PARIS - BRISTOL, CT

2020

ARISTOTE TRADUCTION ET ÉTUDES COLLECTION DIRIGÉE PAR P. DESTRÉE ET PUBLIÉE PAR LE

CENTRE DE WULF-MANSION RECHERCHES DE PHILOSOPHIE ANCIENNE ET MÉDIÉVALE À L’INSTITUT SUPÉRIEUR DE PHILOSOPHIE DE L’UNIVERSITÉ CATHOLIQUE DE LOUVAIN

A catalogue record for this book is available from the Library of Congress. ISBN 978-90-429-4045-1 eISBN 978-90-429-4046-8 D/2020/0602/104

© 2020, Peeters, Bondgenotenlaan 153, B-3000 Leuven Tous droits de reproduction, de traduction ou d’adaptation, y compris les microfilms, réservés pour tous pays.

TABLE DES MATIÈRES AVANT-PROPOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LISTE DES

CONTRIBUTEURS

................................

VII XI

Pierre PELLEGRIN : « Si un sens nous manquait ». De Anima III 1, 424b22-425a13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David CHARLES: Aristotle on the Perception of Objects . . . . . . . . Annick STEVENS : La conscience de sentir n’est pas une sensation : De Anima III 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dorothea FREDE: The Two Meanings of phantasia in De Anima III 3 Annick JAULIN : De Anima III 4, 429a10-430a9 . . . . . . . . . . . . . . . Stephen MENN: From De Anima III 4 to De Anima III 5 . . . . . . . Sylvain DELCOMMINETTE : De Anima III 6 . . . . . . . . . . . . . . . . . . . Klaus CORCILIUS: De Anima III 7: The Actuality Principle and the Triggering of Mental Episodes . . . . . . . . . . . . . . . . . . . . . Michel CRUBELLIER : L’âme comme la main : traduction et commentaire du chapitre III 8 . . . . . . . . . . . . . . . . . . . . . . . . . Pierre-Marie MOREL : Mouvement animal et théorie des facultés en DA III 9. Les paradoxes de l’élimination . . . . . . . . . . . . . . Christof RAPP: Joints and Movers in the Cliffhanger Passage at the end of Aristotle, De Anima III 10 . . . . . . . . . . . . . . . . . . . Jennifer WHITING: The Mover(s) of Rational Animals. De Anima III 11 in context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robert HOWTON: Why De Anima Needs III 12-13 . . . . . . . . . . . .

303 329

BIBLIOGRAPHIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

351

INDEX DES NOMS ANCIENS ET MÉDIÉVAUX . . . . . . . . . . . . . . . . . . . . . INDEX DES NOMS MODERNES ET CONTEMPORAINS . . . . . . . . . . . . . . . INDEX DES LIEUX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

361 363 367

1 19 39 53 75 95 157 185 221 255 273

AVANT-PROPOS

La division des traités d’Aristote en livres est connue depuis l’Antiquité. En l’état actuel de nos connaissances, il est tout à fait plausible que cette composition des traités en livres soit de la main même d’Aristote. L’unité de ces livres est souvent manifeste. Dans de nombreux cas, un livre s’ouvre par l’énoncé de son objet, comme en témoignent certaines premières phrases célèbres : « Il existe une science qui étudie l’être en tant qu’être » (Métaphysique Γ 1) ; « Tout enseignement et tout apprentissage rationnel proviennent d’une connaissance préexistante » (Seconds Analytiques I 1). L’énoncé de l’objet peut aussi être accompagné d’un plan, le cas échéant. Les livres I et II du Traité de l’âme offrent, en ce sens, deux exemples de choix. Au début du livre I, après avoir justifié l’intérêt d’étudier l’âme, Aristote expose plus en détail son projet et ses difficultés inhérentes, avant de lister des apories qui guideront l’examen général de l’âme. Le début du livre II entérine le chemin parcouru à travers l’examen des opinions des devanciers sur l’âme et annonce un nouveau départ pour « essayer de définir ce qu’est l’âme ». Dans ce même traité, pourtant, le livre III constitue, à cet égard, un cas à part. En s’ouvrant, en III 1, sur l’idée qu’il n’existe pas d’autres sens que ceux qui ont été préalablement étudiés, il s’inscrit dans la continuité parfaite des analyses entamées en DA II 5, sur la sensation. Et c’est aux deux tiers du livre, en III 9, qu’apparaît l’articulation principale : après l’étude des puissances cognitives, ou « critiques », de l’âme – conduite, donc, depuis DA II 5 –, Aristote se penche alors explicitement sur la puissance motrice, qui l’occupera jusqu’à la fin du traité. La distinction entre les fonctions cognitives et motrices de l’âme n’a rien qui doive étonner un lecteur du De Anima : elle a été établie dès le premier livre (I 2, 403b25-27) et réaffirmée en III 3, 427a17-19. La continuité et l’ordre de l’enquête conduite en DA III ne font donc guère difficulté. C’est son unité, qui, dès l’Antiquité, a posé question.

VIII

AVANT-PROPOS

Les commentateurs antiques ont ainsi proposé de lire ce livre III comme une enquête sur l’âme « qui fait des choix », l’âme « rationnelle »1, en un mot : l’âme humaine. Assurément, cette interprétation est discutable. La sensation commune, étudiée aux chapitres 1-2, n’est pas réservée à l’être humain, pas plus que l’imagination du chapitre III 3, qu’ont en partage au moins une partie des bêtes (428a9-11). Et l’étude des mécanismes de la motricité fait droit au mouvement des animaux non-humains (par exemple, III 9, 433a10-12 et 11, 433b31 sq.). Cette ampleur des analyses menées au livre III, ampleur que les commentateurs antiques ont tenté de maîtriser et de délimiter, en fait aussi tout l’intérêt. L’interprétation des commentateurs antiques s’explique quant à elle surtout par les pages dévolues à l’intellect, parmi les plus discutées du corpus aristotélicien (III 4-7). Il demeure que, par-delà cette simplification, placer le livre dans la perspective de l’âme humaine n’est pas dénué de fondement. Contrairement à une légende tenace, Aristote évite généralement de brandir une définition en bonne et due forme de l’être humain. Mais, comme l’a noté Michel Crubellier2, il existe un passage du corpus qui s’en approche : en Éthique à Nicomaque VI 2, 1139b4-5, quand Aristote définit la décision ou choix réfléchi (προαίρεσις) comme « une intelligence qui désire ou un désir réfléchi ». Aristote ajoute alors, laconiquement : « et un tel principe, c’est l’homme ». Michel Crubellier commente : « L’homme est principe de ses actions, et il l’est parce qu’il est à la fois un être désirant et un être intelligent »3. Le livre III du De Anima, en explorant, dans la connaissance, la continuité entre sensation et intellection, et, dans l’action, l’articulation entre désir et intellect, prolonge cette caractérisation elliptique de l’humanité de l’homme. L’être humain n’est ni un animal tout à fait comme les autres, ni totalement différent, et c’est cette singularité que le livre III du De Anima contribue à éclairer. Les études qu’on va lire ci-dessous explorent patiemment les treize chapitres du livre III et ses principaux thèmes. Elles souhaitent fournir un guide pour éclairer les enjeux d’un texte difficile mais majeur, datant très probablement de la maturité d’Aristote4. 1

Voir par exemple Ps-Simplicius, In DA 172, 4-11 ; Ps-Philopon, In DA 446, 6. M. Crubellier (2004), p. 9. 3 Ibid. 4 Dans sa Paraphrase du traité de Théophraste sur l’âme, Priscien nous décrit un Théophraste bien perplexe à la lecture du De Anima d’Aristote. Théophraste s’est engagé 2

AVANT-PROPOS

IX

Ces études sont le fruit d’un atelier de lecture organisé à Lille en l’honneur de Michel Crubellier en Mars 2016. Cet événement a bénéficié du soutien conjugué de l’UMR 8163 Savoirs Textes Langage (CNRS, Université de Lille), de l’Institut de Recherches Philosophiques de Lyon (EA 4187), de l’EA 4395 Lettres Idées Savoirs (Université Paris-Est Créteil Val de Marne), de la Faculté des Humanités de l’Université de Lille et de la MESHS Lille Nord de France. Le volume issu de cet atelier de lecture a été publié avec le concours de l’UMR 8163 Savoirs Textes Langage (CNRS, Université de Lille) et de l’UMR 7219 SPHERE (CNRS, Université de Paris) – équipe GRAMATA (Université Paris 1 Panthéon-Sorbonne). Les éditeurs du présent volume souhaitent remercier les institutions partenaires de cet événement, ainsi que toutes celles et ceux qui ont collaboré à cet atelier, dont, au premier chef, les intervenants : David Charles, Klaus Corcilius, Sylvain Delcomminette, Dorothea Frede, Robert Howton, Annick Jaulin, Jean-Louis Labarrière, Stephen Menn, Pierre-Marie Morel, Pierre Pellegrin, Christof Rapp, Annick Stevens, Jennifer Whiting. En compagnie de Michel Crubellier, les discussions ont été aussi nourries qu’amicales. Dans l’esprit de l’École de Lille, tout effort d’interprétation naît de l’expérience de l’incompréhension, tant il est vrai, comme le dit le proverbe grec, que « les choses belles sont difficiles ». Mais cette expérience parfois douloureuse vaut d’autant mieux qu’elle est vécue dans le compagnonnage d’une amitié intellectuelle. C’est pour nous avoir appris cela, et bien d’autres choses, que nous adressons nos remerciements à Michel Crubellier. Gweltaz Guyomarc’h Claire Louguet Charlotte Murgier

dans un travail de commentaire des thèses aristotéliciennes comme s’il avait été privé de l’opportunité d’en discuter directement avec son maître (cf. P. Huby (2015), p. 15). Si l’on ajoute à cela l’inachèvement manifeste du livre III, tout plaide en faveur d’une datation tardive du Traité de l’âme.

LISTE DES CONTRIBUTEURS David CHARLES (Yale University) Klaus CORCILIUS (Universität Tübingen) Michel CRUBELLIER (Université de Lille, UMR 8163 STL) Sylvain DELCOMMINETTE (Université Libre de Bruxelles) Dorothea FREDE (Universität Hamburg) Robert HOWTON (Koç University) Annick JAULIN (Université Paris 1 – Panthéon Sorbonne, UMR 7219 Gramata-Sphere) Stephen MENN (McGill University) Pierre-Marie MOREL (Université Paris 1 – Panthéon-Sorbonne, UMR 7219 Gramata-Sphere) Pierre PELLEGRIN (CNRS, Paris, UMR 7219 Gramata-Sphere)

XII

LISTE DES CONTRIBUTEURS

Christof RAPP (Ludwig Maximilians Universität München) Annick STEVENS (Indépendante) Jennifer WHITING (University of Pittsburgh)

« SI UN SENS NOUS MANQUAIT » DE ANIMA III 1, 424b22-425a13 Pierre PELLEGRIN

Le but du premier chapitre du livre III du De Anima est assez clair : il s’agit de montrer que les animaux ont au mieux cinq sens. Mais son argumentation a été un calvaire pour les commentateurs. En fait, personne ne l’a trouvée à la fois cohérente, appropriée à ce qu’elle entend démontrer et convaincante, et je ne prétends pas faire exception. Je me limiterai à la première partie allant de 424b22 à 425a13. Le raisonnement se déploie à partir d’une thèse qui se trouve en 424b26 : ἀνάγκη τ’, εἴπερ ἐκλείπει τις αἴσθησις, καὶ αἰσθητήριόν τι ἡμῖν ἐκλείπειν, que l’on traduit ordinairement par quelque chose comme : « si une perception fait défaut, il est nécessaire que nous fasse également défaut un organe sensoriel ». On peut faire deux remarques. D’abord, il y a dans les Seconds Analytiques (I 18, 81a38-39) un passage qui semble être un parallèle du nôtre : Φανερόν δὲ καὶ ὅτι, εἴ τις αἴσθησις ἐκλέλοιπεν, ἀνάγκη καὶ ἐπιστήμην τινὰ ἐκλελοιπέναι, ἣν ἀδύνατον λαβεῖν. Il est clair que, si une perception manquait, nécessairement un savoir scientifique manquerait aussi, qu’il nous serait impossible d’acquérir.

Ce que veut dire Aristote dans ce dernier cas, c’est qu’il ne saurait y avoir de science dans un domaine dont les objets ne pourraient pas être perçus et que si nous n’avions, par exemple, aucune saisie perceptive des scarabées dorés, une partie de l’entomologie nous manquerait. Le terme αἴσθησις n’a donc pas le même sens dans les deux passages, puisque dans notre chapitre du De Anima il ne désigne pas la capacité de saisir tel objet, laquelle dépend de conditions qui nous sont extérieures (il y a sans doute des insectes que nous ne connaissons pas encore), mais la capacité à saisir une espèce de sensible, par exemple la capacité de saisir les objets colorés,

2

PIERRE PELLEGRIN

qui sont les objets propres de la vue. Aristote semble, en effet, sousentendre que sa thèse s’applique à la perception des sensibles propres. Dans les deux cas, en revanche, αἴσθησις signifie la capacité à percevoir et non la perception au sens de l’acte de percevoir. Ensuite, sous des airs d’évidence (tout senti suppose un sens), cette thèse doit être précisée, puisque dans le cas des sensibles communs, elle est fausse telle qu’elle est exprimée ici. En effet, ceux-ci ne sont les sensibles propres d’aucun des cinq sens énumérés à la deuxième ligne du chapitre. Ce qui veut dire que si l’un ou plusieurs des cinq sens nous manquaient, nous pourrions néanmoins percevoir sinon tous les sensibles communs, du moins certains d’entre eux. La vue, par exemple, n’a pas la pluralité ou le mouvement comme sensibles propres, et un aveugle pourra saisir la pluralité par le toucher et le mouvement par l’ouïe. Or, comme sans la saisie des sensibles communs il n’y a pas de perception possible, la règle « tout senti suppose un sens » ne s’applique pas au domaine entier de la perception. En effet, on ne peut pas dire que la pluralité suppose plutôt la vue que le toucher pour être perçue. C’est peut-être aussi parce qu’Aristote s’était fait cette objection à lui-même qu’il a abordé le problème des sensibles communs dans la seconde partie de ce chapitre. Suivons donc le cheminement de cette première partie de chapitre. Le raisonnement des lignes 424b24-27 prend la forme d’une inférence. Si nous percevons tous les tangibles par le toucher (car toutes les qualités tangibles d’un sensible en tant que tangible sont perçues par le toucher), alors il est nécessaire que, si une capacité à percevoir nous manquait, c’est qu’un organe sensoriel nous ferait défaut. Il va donc s’agir, pour Aristote, de montrer qu’il n’existe pas d’objet sensible que nous sommes incapables de percevoir. Ici il faut sans doute (comme Georges Rodier) considérer qu’Aristote, après avoir donné l’exemple du toucher, sousentend qu’il en va de même pour tous les autres sens. On peut signaler deux points. D’abord, répétons-le, Aristote, en disant que le toucher ne perçoit que ce qui est tactile, se cantonne sans le mentionner au cas de la saisie des sensibles propres, car la vue peut saisir le rugueux, mais par accident. Tout cela a été dit dans le livre II du De Anima. Ensuite, il passe subrepticement de la proposition « le toucher ne perçoit que ce qui est tactile » (424b24 : παντὸς οὗ ἐστὶν αἴσθησις ἁφὴ καὶ νῦν αἴσθησιν ἔχομεν : « de tout ce dont le toucher est perception nous avons une perception effective », il faut sous-entendre « par le

« SI UN SENS NOUS MANQUAIT »

3

toucher »1) à « tout ce qui est tactile est perçu par le toucher » (424b25 : πάντα γὰρ τὰ τοῦ ἁπτοῦ ᾗ ἁπτὸν πάθη τῇ ἁφῇ ἡμῖν αἰσθητά ἐστιν). Or, si on définit le tactile par un ensemble de propriétés (dur / mou, lisse / rugueux, etc.), on peut concevoir qu’il existe des objets doués de ces propriétés (et qui donc sont tactiles) et qui sont imperceptibles, notamment à cause d’une question de seuil. La science moderne décèle des sensibles imperceptibles (comme les infrarouges), par exemple à travers leurs effets, c’est-à-dire par le biais d’un acte de l’esprit, ce qu’Aristote veut absolument éviter. C’est précisément à la fois pour échapper au pur phénoménisme, selon lequel la vue, par exemple, ne perçoit qu’un ensemble de taches colorées, et à l’obligation d’un détour par un acte de l’intellect (comme celui que l’on trouve dans le Théétète) qu’Aristote introduit sa doctrine des sensibles communs. Et, de fait, on ne sait trop si l’on doit admirer Aristote pour cette performance d’avoir construit une théorie de la perception entièrement sensorielle, ou si on doit le plaindre d’avoir été obligé de se forger une théorie aussi difficile à concevoir et à manier que celle des sensibles communs… Pour nous sortir de la difficulté des sensibles ayant des caractéristiques d’objets sensibles (lisse / rugueux, coloré, etc.) mais que nous ne percevons pas parce qu’ils sont imperceptibles, il faut donner une valeur forte au ἡμῖν de 424b26 et comprendre quelque chose comme : « toutes les propriétés du sensible en tant que sensible sont perçues par le toucher dans la mesure où nous les percevons ». En tout cas, il me semble ici y avoir une différence très importante, sur le point qui nous occupe, entre les sensibles propres et les sensibles communs. Pour que nous ayons une perception complète du monde, il suffit de montrer, non pas que nous n’avons pas de sixième sens, mais qu’il n’y a aucun objet qui ne tombe pas sous les cinq que nous avons. Alors Aristote pourrait en conclure que, de fait, nous n’avons que les cinq sens que nous avons au nom du principe téléologique selon lequel « la Nature ne fait rien en vain ni de superflu ». Pour les sensibles communs, en revanche, il est crucial de montrer qu’ils ne sont pas perçus par un sens spécial, ce que fera la seconde partie du chapitre. 1 Certains (Madvig) ont voulu, en changeant la ponctuation, faire dire à Aristote que le toucher intervenait dans tous les sens, ce qui éviterait de sous-entendre, comme Rodier, qu’en 424b27, ce qu’Aristote dit du toucher doit être étendu à tous les sens. Cette lecture est peut-être possible d’un point de vue aristotélicien, mais elle entrerait difficilement dans l’argumentation en cours.

4

PIERRE PELLEGRIN

Ensuite Aristote pose une grande distinction selon laquelle il y a deux sortes de perception sensible, celle, immédiate, qui se fait par contact et celle qui se fait à travers un medium. Mais la question récurrente que pose ce chapitre, c’est celle des rapports entre les quatre éléments – feu, air, eau, terre – et la perception, et Aristote la pose d’abord à propos de la perception médiate. Après avoir parlé de la perception immédiate pour en dire qu’elle relève du toucher, comme nous venons de le voir, Aristote passe à la perception médiate. Il reviendra ensuite, assez brièvement, à la perception immédiate. Aristote étudie la perception médiate en considérant deux situations. La première est examinée en 424b31-34 : ἔχει δ’ οὕτως ὥστ’ εἰ μὲν δι’ ἑνὸς πλείω αἰσθητὰ ἕτερα ὄντα ἀλλήλων τῷ γένει, ἀνάγκη τὸν ἔχοντα τὸ τοιοῦτον αἰσθητήριον ἀμφοῖν αἰσθητικὸν εἶναι (οἷον εἰ ἐξ ἀέρος ἐστὶ τὸ αἰσθητήριον, καὶ ἔστιν ὁ ἀὴρ καὶ ψόφου καὶ χρόας). Les choses se présentent ainsi : si plusieurs sensibles différant génériquement sont perçus à travers un seul medium, il est nécessaire que celui qui possède un tel organe sensoriel [c’est-à-dire qui perçoit à travers ce medium] puisse percevoir les deux (par exemple si l’organe sensoriel se compose d’air et si l’air est le medium des perceptions du son et de la couleur).

Il faut donc suppléer à la fin de ce passage : « l’organe sensoriel en question percevra les deux ». Depuis l’Antiquité, les commentateurs ont relevé l’absurdité comique de ce texte, qui dit, si on le prend à la lettre, que l’oreille pourrait percevoir le son et la couleur, puisque son et couleur peuvent se percevoir à travers l’air et que l’oreille est composée d’air. Aristote doit vouloir dire que si deux sensibles, même s’ils diffèrent génériquement, sont perçus à travers un seul medium, l’animal aura dans les deux cas un organe des sens qui sera capable de percevoir chacun des deux à travers ce medium. Ainsi, si la couleur et le son se perçoivent à travers l’air, il y aura dans les deux cas un sens pour les percevoir qui les percevra à travers l’air. Nous avons donc là une maladresse d’expression qui va au-delà de l’habituelle désinvolture syntaxique d’Aristote et qui peut avoir plusieurs causes, par exemple un défaut dans la transmission du texte, mais qui reste tout de même étrange. Mais même si on lit ce passage de cette manière, seule une partie du problème est résolue. Certes, l’audition se fait à travers l’air parce que l’ouïe est de l’air. Mais qu’en est-il de la vision qui se fait à travers l’air alors que l’œil (du moins l’œil des animaux qui nous ressemblent sur ce point) est composé d’eau ?

« SI UN SENS NOUS MANQUAIT »

5

C’est ce à quoi répond la seconde situation, plus sensée, qui se trouve en 424b34-425a3 : εἰ δὲ πλείω τοῦ αὐτοῦ, οἷον χρόας καὶ ἀὴρ καὶ ὕδωρ (ἄμφω γὰρ διαφανῆ), καὶ ὁ τὸ ἕτερον αὐτῶν ἔχων μόνον αἰσθήσεται τοῦ δι’ἀμφοῖν, τῶν δὲ ἁπλῶν ἐκ δύο τούτων αἰσθητήρια μόνον ἐστίν, ἐξ ἀέρος καὶ ὕδατος. Si plusieurs milieux concernent le même senti, par exemple, pour la couleur, l’air et l’eau (car tous deux sont diaphanes), alors le sens qui contient l’un seulement des deux milieux [c’est-à-dire qui contient l’élément qui est également présent dans l’un des deux milieux, air ou eau] sentira ce qui se sent à travers les deux ; or, parmi les corps simples, il n’y a que ces deuxlà qui constituent les organes sensoriels, l’air et l’eau.

Il faut donc comprendre que si la couleur peut se percevoir soit à travers l’air soit à travers l’eau, alors il suffit que l’organe des sens contienne l’un de ces deux éléments seulement, air ou eau, pour qu’il puisse la percevoir. La raison donnée est que l’air et l’eau sont tous deux diaphanes. Un problème délicat est celui de l’odorat. En Histoire des animaux IV 8, 534a13, en effet, Aristote rapporte que les poissons ne mordent pas tous aux mêmes appâts, mais à des appâts spécifiques qu’ils identifient par l’odeur2. Mais Aristote ne dit rien de la faculté des autres animaux (nous par exemple) de sentir les odeurs sous l’eau. D’ailleurs, dans notre chapitre du De Anima, Aristote laisse le cas de l’odorat dans le vague en disant en 425a4 que « la pupille est faite d’eau, l’ouïe d’air et l’odorat de l’un des deux ». Il semble que ce qu’Aristote veut établir soit simplement ceci : étant donné que mon œil est formé d’eau, qui est diaphane, je peux voir à travers l’eau, qui est diaphane, mais aussi à travers l’air, qui est aussi diaphane. De même pour l’ouïe et l’olfaction, la première parce que l’eau aussi bien que l’air portent les sons, la seconde étant un cas moins sûr, mais sans doute identique à la première sur ce point. Je serais tenté de comprendre que, dans le cas de l’olfaction, l’expression « perceptible dans l’air et dans l’eau » doit être prise en un sens distributif : les animaux terrestres (du moins certains d’entre eux) sentent dans l’air et les poissons (du moins certains d’entre eux) dans l’eau. Mais il n’est pas sûr qu’aucun ne puisse sentir des odeurs à la fois dans l’air et dans l’eau. 2 Ce texte est, quant à lui, difficile à concilier avec le passage du De Sensu cité dans une note ci-dessous, car si ce qui est perçu par l’olfaction c’est une « exhalaison fumeuse » qui est de la nature du feu, on imagine difficilement qu’elle puisse être perçue sous l’eau.

6

PIERRE PELLEGRIN

Donc, loin d’être un obstacle, le medium dilate la faculté de percevoir : l’œil qui est de l’eau peut aussi percevoir dans l’air, parce que ce medium, l’air, est diaphane, comme l’eau. Si la vue se faisait par contact direct entre le sensorium et l’objet perçu, un œil comme le nôtre ne pourrait percevoir que des objets composés, au moins en partie, d’eau. En fait, il ne pourrait rien percevoir du tout, comme on va le voir. Disons les choses autrement. Ce qu’Aristote a « établi » se résume à ceci. Dans le cas des perceptions à travers un medium, il y a parfois, première situation, correspondance directe entre le milieu de la perception et l’élément composant l’organe sensoriel. Ainsi le son se transmet dans l’air et l’ouïe est de l’air, le visible se transmet dans l’eau pour les poissons et l’œil est de l’eau. Ceci fait difficulté, car chez nous aussi la vue (la pupille) est de l’eau, et nous voyons dans l’eau. Mais, seconde situation, nous voyons aussi, et habituellement, dans l’air alors que notre œil n’est pas devenu de l’air. Alors, de trois choses l’une : ou nous voyons aussi bien dans l’air que dans l’eau, ou mieux dans l’air, ou mieux dans l’eau. Aristote nous renseigne sur ce point, mais dans le cas de l’audition : « on entend dans l’air et dans l’eau, mais moins » (De Anima II 8, 419b18), ce qui s’applique certainement à nous. Mais qu’en est-il des poissons ? Sans doute la question est oiseuse, parce qu’ils n’ont pas, du moins pour la plupart d’entre eux et la plupart du temps (exception faite des exocets, dont il est question en Histoire des animaux IV 9, 535b26, et qui ne séjournent pas longtemps dans l’air), l’occasion d’entendre dans l’air. En tout cas, rien n’établit clairement que la perception est meilleure quand c’est le même élément qui constitue le medium et l’organe sensoriel. La conclusion de cette analyse en deux temps de la perception médiate avait été donnée avant même l’analyse des deux situations dans lesquelles cette perception peut intervenir : ὅσα δὲ διὰ τῶν μεταξὺ καὶ μὴ αὐτῶν ἁπτόμενοι, τοῖς ἁπλοῖς, λέγω δ’ οἷον ἀέρι καὶ ὕδατι. Tout ce que nous percevons à travers un intermédiaire et non en le touchant, nous le percevons par les corps simples, j’entends en l’occurrence l’air et l’eau. (424b29-30)

On sait que la théorie aristotélicienne de la perception s’oppose à deux modèles concurrents. Le premier, dont il y a des traces chez Platon, c’est que les sens se projettent vers les objets qu’ils saisissent. Laissons-le de côté. Le second, c’est que la meilleure perception c’est celle qui est

« SI UN SENS NOUS MANQUAIT »

7

immédiate, au sens où le perçu imprime directement sa marque sur le sens. Ainsi Démocrite pensait que l’air est un obstacle à la vision et la déforme, et que si, entre notre œil et une fourmi qui serait sur la voûte céleste, il y avait non pas de l’air, mais du vide, nous la verrions distinctement (cf. De Anima II 7, 419a15). La position d’Aristote, c’est que le sensible ne peut pas affecter directement le sens, mais qu’il doit imprimer un mouvement à un intermédiaire entre le sens et lui et que c’est cet intermédiaire qui va frapper l’organe sensoriel et l’altérer. Aristote l’a établi fortement pour la vue : si l’objet coloré est plaqué sur l’œil, celui-ci ne voit rien. La couleur peut mouvoir plusieurs milieux différents (l’air et l’eau), pourvu que ces milieux aient assez d’affinité avec le sensorium lui-même, à savoir l’œil. Dans l’eau, cette affinité est immédiate, puisque l’œil est de l’eau. Dans l’air, cette affinité s’appelle la diaphanéité : elle suffit à ce que l’air puisse porter à l’œil le mouvement qu’il a reçu de la couleur. C’est sans doute la description du processus de la vue par Aristote qui comporte à la fois le plus d’informations et le plus de difficultés. Du fait de son « réalisme », Aristote ne pouvait pas penser que c’est la structure de l’œil des animaux qui donne leur couleur aux choses colorées : même si la couleur est « ce qui meut le diaphane en acte » (De Anima II 7, 418a31), elle n’en reste pas moins une propriété des choses3, que le diaphane, qui devient lumière quand une source ignée le fait passer à l’acte, rend visible. Plus exactement, la couleur est « l’extrémité du diaphane dans les corps » (De Sensu 439a27). Je ne suis pas de ceux qui croient comprendre complètement la doctrine aristotélicienne du diaphane, mais je pense que l’on peut affirmer que le diaphane est à la fois ce qui, quand il est limité dans un corps (le diaphane est invasif et compénètre tous les corps, cf. De Sensu 3, 439a23), donne sa couleur à ce corps, et ce qui, quand il est sans limite dans le medium (air ou eau), rend les corps colorés visibles en les éclairant grâce à une source lumineuse extérieure. La communauté entre les objets perçus et le medium est ici obvie : c’est le diaphane, car les objets visibles ne sont visibles que parce qu’ils ont en eux du diaphane. 3 Alexandre d’Aphrodise (Quæstiones I 13) recense au moins trois théories concernant la couleur. Des atomistes comme Épicure pensent que les atomes ne sont pas colorés ; certains supposent des atomes colorés ; pour « ceux pour qui il y a quatre éléments » à la racine des choses (I 13, 25, 23) ces éléments sont colorés, ce qui explique la couleur des choses qu’ils composent, position encore plus « réaliste » que celle d’Aristote, qu’Alexandre connaissait fort bien (cf. Quæstiones I 2).

8

PIERRE PELLEGRIN

À partir de 425a5, Aristote réintroduit les éléments autres que l’air et l’eau, et, à ce propos, revient rapidement sur la perception immédiate. Or il s’avère que l’opposition exprimée au début du chapitre, et résumée dans le passage de 424b29-30 que j’ai cité – à savoir que, contrairement aux objets du toucher qui sont saisis immédiatement, les objets des sens médiats sont saisis « par les corps simples » – est une fausse opposition, puisqu’Aristote avait déjà appliqué à la perception immédiate le modèle explicatif de la perception à travers un medium. Ainsi en De Anima II 11, 423b1-4, on lit : Πότερον οὖν πάντων ὁμοίως ἐστὶν ἡ αἴσθησις, ἢ ἄλλων ἄλλως, καθάπερ νῦν δοκεῖ ἡ μὲν γεῦσις καὶ ἡ ἁφὴ τῷ ἅπτεσθαι, αἱ δ’ ἄλλαι ἄποθεν. τὸ δ’ οὐκ ἔστιν. Est-ce que tous les objets perçus le sont de la même façon, ou bien les uns d’une façon les autres d’une autre, à savoir, comme on le pense effectivement, le goût et le toucher par contact et les autres à distance ? Mais il n’en va pas ainsi.

Et, un peu plus bas (423b17-19) : Ὅλως δ’ ἔοικεν ἡ σὰρξ καὶ ἡ γλῶττα, ὡς ὁ ἀὴρ καὶ τὸ ὕδωρ πρὸς τὴν ὄψιν καὶ τὴν ἀκοὴν καὶ τὴν ὄσφρησιν ἔχουσιν. D’une manière générale, il semble bien que la chair et la langue sont comme l’air et l’eau par rapport à la vue, l’audition et l’olfaction.

C’est-à-dire que, dans le toucher (le goût), la chair (la langue) sert de medium et, donc, la question du rapport entre le medium et les éléments se ramène à celle du rapport entre les éléments et les sensoria. Ceci permet aux éléments autres que l’air et l’eau, c’est-à-dire le feu et la terre, de « revenir dans le tableau », même si c’est de manière à la fois rapide et peu nette. Pour ce qui est du feu, il rend la perception, et d’ailleurs la vie, possibles par la chaleur : Τὸ δὲ πῦρ ἢ οὐθενὸς ἢ κοινὸν πάντων (οὐθὲν γὰρ ἄνευ θερμότητος αἰσθητικόν). Le feu, pour sa part, ou bien n’est constitutif d’aucun , ou bien il est commun à tous, puisqu’il n’y a pas d’animal en mesure de percevoir sans chaleur. (425a5-6)4 4 Je laisse à plus savant que moi le soin de décider si la divergence qu’il semble y avoir entre la doctrine ici exposée et celle que l’on trouve à la fin du chapitre 2 du De Sensu est due à une évolution chronologique. Le De Sensu, en effet, dit que « l’odorat

« SI UN SENS NOUS MANQUAIT »

9

Quant à la terre, elle trouve ici un traitement aussi cavalier que celui du feu : « la terre ne se trouve dans aucun organe des sens, ou c’est surtout dans le toucher qu’elle entre en composition » (425a6), ce qui, je crois, signifie qu’elle ne compose pas seule l’organe du toucher, mais qu’elle y est unie à d’autres éléments5. De sorte que quand le De sensu déclare que « l’organe du toucher est fait de terre » (2, 438b30-31)6, il faut sans doute entendre que l’organe du toucher est aussi fait de terre. D’autres passages du De Anima montrent moins de réticence. Ainsi en II 5, 417a4-5 : « Le feu, la terre et les autres éléments sont contenus dans les organes sensoriels ». Les organes sensoriels étant des objets physiques, ils ne peuvent être composés que des quatre éléments, si l’on excepte le cas de corps célestes, qui sont bien physiques, mais composés de la « cinquième essence ». Mais de la proposition évidente « les sensoria sont composés d’éléments », Aristote passe à une proposition au contenu plus riche : « les sensoria sont composés des éléments »7. Tout acte perceptif s’accomplit donc à travers un medium, y compris quand ce medium, dans le cas de la perception par contact, coïncide avec l’organe sensoriel et quand le medium a, si l’on peut dire, une existence conceptuelle plus que physique : la chair, par exemple, d’un certain point de vue est le sensorium et, d’un autre point de vue, est le medium. Cette distinction n’est pas seulement verbale. On le voit notamment quand, en De Anima II 5, Aristote se demande pourquoi les sensoria ne sont pas eux-mêmes perçus alors que « ils contiennent du feu, de la terre et des autres éléments » (417a4). La réponse est : parce que le sensorium « n’est pas en acte, mais seulement en puissance » (417a6). Mais est du feu (en effet, ce que l’odorat est en acte, la faculté olfactive doit l’être en puissance, car le sensible fait passer le sens à l’acte, de sorte que, nécessairement, ce qu’il est alors, il l’était auparavant en puissance. D’autre part, l’odeur est une sorte d’exhalaison fumeuse, et toute exhalaison fumeuse vient du feu) » (438b20-25). 5 Ainsi dans le passage III 13, 435a21 disant que « le corps de l’animal ne peut pas être γήϊνον », ce dernier terme signifie « entièrement formé de terre ». On peut même sans doute comprendre « entièrement ou principalement formé de terre ». Quand, en effet, Aristote nous explique que ni l’os, ni le cheveu, ni la plante n’ont la sensation parce qu’ils sont « de la terre (ὅτι γῆς ἐστιν) » (435b1), il n’entend sans doute pas dire que toutes ces réalités ne laissent place à aucun autre élément. De toute façon, il semble bien que pour Aristote aucun élément n’existe dans la nature à l’état pur. 6 Le texte continue : « et le goût est une espèce du toucher », ce qui confirme ce que j’ai dit plus haut. Il faut aussi noter l’expression remarquable de « toucher lingual » (ἡ ἐπὶ τῆς γλώττης ἁφή, II 11, 423a17). 7 C’est notamment l’objet du chapitre 2 du De Sensu.

10

PIERRE PELLEGRIN

on peut aussi comprendre qu’une telle hypothèse viole le principe selon lequel toute perception est médiate, car le sensorium, pour devenir objet de perception (par lui-même), devrait recevoir lui-même une impression de lui-même à travers un medium qui est lui-même. Or c’est bien le fait qu’ils sont composés des quatre éléments qui rend les objets du monde perceptibles. Dans un passage déjà cité du De Anima, Aristote mentionne, comme quelque chose d’évident, que ce sont les « éléments en eux-mêmes ou leurs accidents » qui sont objets de perception (II 5, 417a5). Dans son commentaire sur le De Anima, Simplicius (118, 26, Hayduck) interprète cela comme l’affirmation qu’en euxmêmes les éléments sont essentiellement tactiles, mais qu’ils ont aussi une couleur, une taille, etc. Et, de fait, Aristote a toujours considéré le toucher comme le sens basique que tout animal doit posséder. Il est ici intéressant de rappeler que les qualités tactiles que sont le chaud / froid, sec / humide, dur / mou, visqueux / friable, rugueux / lisse, épais / fin, lourd / léger dérivent des deux couples fondamentaux que sont le chaud / froid et sec / humide, lesquels sont les constituants des quatre éléments (Cf. De la génération et la corruption II 2, 329b34). La communauté élémentaire entre le système sensorium-medium et les objets perçus fait que tous les sens, qu’ils le fassent à travers un medium ou directement par contact, perçoivent par l’intermédiaire des quatre éléments. Si on creusait la question, on trouverait même que, dans l’exercice d’un sens médiat comme la vue, les éléments autres que l’air et l’eau sont aussi mobilisés à travers la délicate doctrine aristotélicienne du diaphane. La lumière, qui est du diaphane en acte, est invisible, mais elle est condition de la visibilité des objets colorés. Mais les objets colorés eux-mêmes ont la couleur qu’ils ont, située entre le blanc et le noir, sans doute du fait de la proportion des différents éléments qu’ils renferment. Il faut donc conclure de tout cela que les media, aussi bien dans le cas de la perception à distance que dans celui de la perception immédiate, sont capables de déchiffrer le message envoyé par les objets sensibles parce qu’ils sont aptes à recevoir toutes les qualités perceptibles de ces objets qui découlent du fait que ces objets sont composés des quatre éléments. C’est ce message déchiffré que les media impriment dans les sens sous la forme d’une altération. Donc, la déclaration d’Aristote, citée plus haut, selon laquelle « tout ce que nous percevons à travers un intermédiaire et non en le touchant, nous le percevons par les corps simples,

« SI UN SENS NOUS MANQUAIT »

11

j’entends en l’occurrence l’air et l’eau » (424b29-30), doit être étendue à tous les sens et à tous les éléments. Autrement dit, les media par lesquels les sens perçoivent étant ce qu’ils sont, aucun objet sensible ne peut échapper à la perception, au moins à celle des animaux qui nous ressemblent. Je reviendrai sur ce point. Aristote nous avait donc, en quelque sorte, promis une démonstration que Edwin Wallace, dans sa magnifique édition de 1882, formalisait dans le syllogisme hypothétique suivant : – si une αἴσθησις nous manque, un αἰσθητήριον nous manque nécessairement ; – or aucun αἰσθητήριον ne nous manque ; – donc aucune αἴσθησις ne nous manque. Mais, en fait, Aristote établit la mineure (« or aucun αἰσθητήριον ne nous manque ») sous une forme « décalée » : « les αἰσθητήρια que nous avons ‘font le travail’ pour tous les sensibles possibles ». Les deux premières lignes du chapitre, οὐκ ἔστιν αἴσθησις ἑτέρα παρὰ τὰς πέντε (λέγω δὲ ταύτας ὄψιν, ἀκοήν, ὄσφρησιν, γεῦσιν, ἁφήν) (424b22-23), doivent donc s’entendre comme signifiant : « il n’y a pas besoin d’un autre sens outre les cinq que nous avons (vue, ouïe, olfaction, goût, toucher) pour percevoir tous les objets du monde ». Or, malgré l’indéniable souplesse grammaticale d’Aristote, on ne peut pas traduire cette première phrase ainsi. Finalement, Aristote aboutit à la position suivante. Notre accès aux objets du monde (qui sont tous composés des quatre éléments) se fait par le truchement des quatre éléments, que ce soit dans le cas de la perception par contact direct, ou dans celui de la perception à travers un medium. Aristote a donc établi que nos cinq sens fonctionnent, à eux tous, à travers les quatre éléments. Mais il y a deux choses qu’il n’a pas établies. D’abord, d’après les passages cités plus haut, aussi bien ceux de notre chapitre du De Anima que ceux venus d’ailleurs, un animal qui ne serait doué que du toucher (sens de base indispensable à tout animal) ainsi que de la vue et de l’audition percevrait le monde à travers les quatre éléments. Le feu, avons-nous vu, est condition de toute sensation, la terre est commune à l’organe qui touche et à ce qui est touché, et l’œil et l’oreille sont formés d’eau et d’air et perçoivent à travers l’eau et l’air. Peut-être même le toucher et l’olfaction suffiraient-ils. Il est cependant clair qu’un animal dépourvu d’odorat

12

PIERRE PELLEGRIN

(ce qui existe, y compris sous forme contre nature, dans l’espèce humaine par exemple), même s’il aurait le moyen de n’ignorer l’existence d’aucun des objets présents dans le monde, n’en appréhenderait pas toute la richesse perceptive puisqu’il n’en saisirait pas les odeurs, pour ceux d’entre eux qui ont une odeur. Aristote n’a donc pas établi que, même si la communauté d’éléments entre le monde et les media de la perception fait que rien ne nous échappe dans le monde, nous saisissons toutes les propriétés des objets perçus. On ne trouve pas chez Aristote la prudence de Spinoza disant que, même si l’être humain ne peut pas concevoir d’autres attributs de la substance que la pensée et l’étendue, cela ne veut point dire que ce sont les deux seuls attributs possibles et réels. Ensuite, on voit qu’Aristote n’a pas établi non plus qu’il n’y a pas de sixième sens, contrairement à ce qu’il avait annoncé dans la première phrase du chapitre. Au moins a-t-il assuré l’essentiel : décrire, à travers la communauté élémentaire entre ce qui fait que le percevant perçoit et le perçu, une connaissance sensible qui fonctionne sans le secours d’une opération intellectuelle de mise en ordre et / ou de rectification. Il n’y a donc rien dans le monde qui soit statutairement imperceptible, du moins tant qu’il s’agit de perception sensible. Il n’en reste pas moins, évidemment, que des objets peuvent être imperceptibles parce qu’ils émettent des stimuli trop faibles (corps trop petit pour être vu, son trop faible pour être entendu, etc.), mais ils ne sauraient être imperceptibles du fait de leur nature. De la seconde partie du chapitre, que je n’entends pas considérer ici, j’ai une lecture qui a beaucoup évolué au cours du temps, qui reste lourdement influencée par celle de Jacques Brunschwig8 et dont je me laisse encore quelque temps pour livrer une version définitive. Mais le résultat de cette seconde partie, du moins celui qui m’intéresse ici tout en étant incontestable, est le suivant. Les réalités comme le mouvement, la grandeur, le repos, le nombre, l’unité ne sont assurément pas des réalités composées des quatre éléments. Mais il y en a néanmoins une saisie purement sensorielle à travers les cinq sens et eux seuls, et non à travers un sixième sens9. 8

J. Brunschwig (1991) et (1996). À ce propos, j’accepte deux des interprétations de Jacques : (i) les αἰσθητά des κοινὰ αἰσθητά (syntagme qui, je crois, n’apparaît pas dans le De Anima) sont des « sentis » et non des « sensibles » ; (ii) le terme κοινὰ doit être pris en un sens distributif : Aristote n’entend pas dire qu’un sensible commun doit être perçu par tous les sens propres, mais que, pour percevoir tous les sensibles communs, il faut faire appel à tous les sens propres. 9

« SI UN SENS NOUS MANQUAIT »

13

Quand on sait la place du savoir sensible chez Aristote, on réalise l’importance de ces remarques. On voit que cette homogénéité entre le sujet connaissant et le monde extérieur sera le garant le plus sûr de la justesse de l’appréhension du monde par ce sujet. L’homme aristotélicien est quelqu’un qui est de plain-pied dans la vérité dès qu’il exerce ses sens, parce que ses sens et le monde sont, d’une certaine façon, la même chose. Lecture téléologique Comme je l’ai laissé entendre plus haut, c’est peut-être le recours, qu’Aristote ne fait pas explicitement dans notre chapitre du De Anima, à la téléologie qui peut seul le sauver sinon de l’incohérence, du moins de l’accusation de ne pas tenir ses promesses. Je dois d’abord rappeler mon interprétation des formules finalistes aristotéliciennes « la Nature réalise toujours le meilleur, si rien ne l’en empêche », « la Nature ne fait rien en vain ni de superflu ». C’est une lecture non absolue, tant il est vrai que, bien des fois, la Nature eût pu faire mieux et que, bien des fois, elle a donné aux vivants des organes ou des comportements inutiles, voire nuisibles. Ainsi la bile ne sert à rien, elle est même nocive, violant ainsi les deux principes téléologiques à la fois. Dans le rapport entre traits utiles et traits nuisibles à une espèce, « réaliser le meilleur » veut dire que, de toute éternité, la Nature a fait que ce rapport soit assez favorable pour assurer la survie éternelle de l’espèce. C’est ce que montre cette construction remarquable, et bien connue (ce qui me dispense de donner des références textuelles) de la zoologie aristotélicienne, qui pourrait se décrire comme un ordre décroissant d’excellence : un organe ou une propriété (i) est nécessaire, ou du moins suffisant, à l’accomplissement excellent d’une fonction (les poumons pour la réfrigération de l’organisme) ; (ii) n’est pas nécessaire mais fait que la fonction est mieux remplie si il / elle existe (les reins font que la fonction de la vessie est mieux accomplie) ; (iii) est inutile mais non nuisible (les cornes des bisons leur sont inutiles pour se défendre, mais ne leur nuisent pas) ; (iv) pourrait être nuisible, mais l’animal (ou la Nature) trouve le moyen de neutraliser cet aspect nuisible (certains bœufs à cornes très grandes broutent en reculant) ; (v) est nuisible mais son aspect négatif est largement contrebalancé par un aspect positif (les serres empêchent les rapaces de marcher, mais leur sont évidemment

14

PIERRE PELLEGRIN

d’une grande utilité pour se nourrir) ; (vi) est nuisible mais compensé par un autre dispositif (l’absence de deux rangées de dents est dommageable, mais est compensée par la rumination) ; (vii) est nuisible mais pas suffisamment pour renverser le ratio avantages / désavantages (l’étroitesse de leurs tentacules empêche une certaine variété de poulpes d’avoir deux rangées de ventouses, mais cela n’est pas suffisant pour menacer la survie éternelle de l’espèce). On ne peut tout simplement pas maintenir à la fois que la Nature donne de la bile à certains animaux et réalise toujours le meilleur, si on prend « meilleur » au sens absolu, car il eût été meilleur pour les êtres humains de ne pas avoir de bile. C’est d’ailleurs ce qu’Aristote reconnaît en disant que certaines parties sont en vue du pire : C’est pourquoi la même différenciation [que celle entre sang épais et chaud et sang froid et subtil] existe pour les parties supérieures par rapport aux inférieures (…). Et il en va de même pour les autres parties, homéomères comme anoméomères : il faut considérer qu’elles diffèrent, les unes en vue du meilleur ou du pire, les autres relativement aux fonctions et à l’essence de chacun des animaux (τὰ μὲν πρὸς τὸ βέλτιον ἢ χεῖρον, τὰ δὲ πρὸς τὰ ἔργα καὶ τὴν οὐσίαν ἑκάστῳ τῶν ζῴων)10. (Parties des animaux II 2, 648a11-16)

Pour ce qui est de l’expression « la Nature ne fait rien en vain »11, elle peut se prendre en au moins deux sens. Sous sa forme développée (« la Nature ne fait rien en vain ni de superflu »), elle exprime un principe de non-redondance. Qui a des cornes pour se défendre n’a nul besoin de griffes, qui lui seraient donc données en vain. Mais on peut aussi, je crois, dans cette expression « la Nature ne fait rien en vain », comprendre « en vain » comme se rapportant à quelque chose que la Nature selon l’essence aurait pu faire, la nature matérielle nécessaire étant donnée, mais qui aurait renversé ou, au moins, aurait diminué le ratio favorable entre avantages et désavantages qui assure la survie de l’animal. C’est en vain que les poissons auraient des pattes, parce que cela, en rendant leur progression plus difficile, ne pourrait que nuire à leur survie. Mais cela ne signifie nullement que tous les organes et caractères de tous les animaux soient indispensables, ou même utiles. 10 Une partie de la tradition manuscrite inverse les deux membres de phrase « les unes … les autres ». 11 Le De Anima emploie la fameuse formule « la Nature ne fait rien en vain » deux fois en III 9, 432b21 et III 12, 434a31.

« SI UN SENS NOUS MANQUAIT »

15

D’autre part, le schéma central bien connu de l’explication téléologique aristotélicienne, c’est que la Nature donne des moyens à qui peut en tirer profit. L’exemple le plus fameux, développé dans les Parties des animaux, est celui de la main : l’homme n’est pas plus intelligent que les autres animaux parce qu’il a des mains, comme le pensait Anaxagore, mais la Nature lui a donné des mains parce que, étant plus intelligent que les autres animaux, il est apte à s’en servir, alors qu’eux ne le seraient pas. D’une certaine manière, nous avons là une illustration de la seconde figure du principe selon lequel « la Nature ne fait rien en vain », puisque, si les vaches avaient des mains (ce que la Nature eût sans doute pu réaliser), non seulement elles ne sauraient pas les utiliser, du moins les utiliser pleinement, mais, du fait qu’elles seraient privées de sabots, cela rendrait leur survie plus difficile. Dans ma lecture de notre chapitre du De Anima, on ne peut donc pas dire : « aucun objet du monde n’échappe à nos cinq sens, or la Nature ne fait rien en vain, donc il n’y a pas de sixième sens ». Et cela pour deux raisons que j’ai données : d’abord le spectacle même du monde vivant nous interdit de prendre « en vain » en un sens fort ; ensuite les cinq sens eux-mêmes sont pléthoriques pour saisir tous les objets possibles, puisque, si l’on s’appuie sur la démonstration d’Aristote par la communauté élémentaire entre les organes sensoriels, les media pour saisir les objets du monde extérieur et ces objets, deux ou trois sens suffisent pour saisir tous les sensibles possibles, même si toutes les caractéristiques n’en sont pas nécessairement saisies, comme nous l’avons vu. À mon avis, donc, Aristote échoue à donner une structure téléologique forte à la possession des sens par les animaux, car il ne démontre pas qu’il ne peut pas y avoir de sens en dehors des cinq connus. Mais il établit la thèse extrêmement forte suivante : rien ne peut échapper aux cinq sens, ni les objets matériels, ni les propriétés immatérielles saisies par le sens commun. La Nature a donc équipé les animaux des moyens qu’il faut pour leur survie à travers leur système sensoriel, et cela aussi bien les animaux suffisamment proches de nous pour avoir cinq sens que les autres. Aristote affirme que tous les animaux ont des organes sensoriels, mais tous ne les ont pas tous12, ni doués des mêmes performances. 12 « Il faut maintenant parler des organes des sens. Car ils ne se présentent pas de la même manière chez tous les animaux, mais certains les ont tous, d’autres en moins grand nombre. Ils sont au maximum (et à côté de ceux-là il n’y en a manifestement aucun autre qui soit caractérisé) au nombre de cinq : vue, ouïe, odorat, goût, toucher » (Histoire des animaux IV 8, 532b29-33).

16

PIERRE PELLEGRIN

Cependant il ne fait guère de doute que les grands mammifères, par exemple, n’aient des performances sensorielles comparables à celles des êtres humains, certains surclassant les êtres humains sur un point (les chiens pour l’odorat), mais tous étant surclassés par eux sur d’autres points (les êtres humains ont le toucher le plus fin). Aristote était trop bon zoologiste pour ne pas avoir remarqué les différences importantes qui existent entre les systèmes perceptifs des différents animaux. Du coup, la téléologie refait son apparition. Il y a deux textes fort instructifs sur le sujet : Si l’on traite maintenant de chaque sens en particulier, le toucher et le goût accompagnent nécessairement tous les animaux, le toucher pour la raison alléguée dans le traité De l’âme, et le goût à cause de la nutrition (…). Quant aux sensations externes, chez les animaux capables de marcher, comme l’odorat, l’ouïe et la vue, leur présence assure la sauvegarde de tous ceux qui les possèdent, en leur permettant de rechercher leur nourriture en fonction d’une sensation antécédente et de se détourner de ce qui est mauvais et dangereux ; mais chez ceux qui possèdent en outre l’intelligence (φρονήσεως), ces sensations existent en vue du bien-. Elles leur annoncent en effet un grand nombre de différences, d’où proviennent à la fois l’intelligence (φρόνησις) des choses intelligibles (νοητῶν) et celle des actions à accomplir (πρακτῶν). (De Sensu 1, 436b12-437a3, trad. P.-M. Morel) Si, parmi les animaux, c’est pour ainsi dire l’être humain qui possède, par rapport à sa taille, le moins d’exactitude dans la perception à distance, pour ce qui est des différences, c’est lui qui, de tous les animaux, en a la sensation la plus fine. La cause en est que son organe sensoriel est pur, le moins constitué de terre et le moins corporel, et, par nature, l’être humain est celui des animaux qui possède la peau la plus fine par rapport à sa taille. (Génération des animaux V 2, 781b17-22, trad. D. Lefebvre)

Ce dernier passage, en disant que la peau humaine est la plus fine, veut signifier que le toucher des êtres humains est le plus subtil. Or le toucher est le sens qui perçoit le plus de différences (Cf. De Anima II 11, 422b23 ; Parties des animaux II 1, 647a16). Bref, les êtres humains sont globalement supérieurs aux autres animaux, en ce qui concerne la perception sensible, parce que leurs performances perceptives l’emportent, non pas en acuité, mais dans la saisie des différences. Cela leur vient en dernier ressort de la pureté de leurs organes sensoriels. Il faut remarquer que le passage du De Sensu cité ci-dessus respecte la structure explicative signalée plus haut : c’est parce qu’ils possèdent l’intelligence, que la Nature donne à certains animaux une capacité sensorielle supérieure pour qu’ils en fassent un usage intelligent. La phrase en 437a2-3 (πολλὰς

« SI UN SENS NOUS MANQUAIT »

17

γὰρ εἰσαγγέλλουσι διαφοράς, ἐξ ὧν ἥ τε τῶν νοητῶν ἐγγίνεται φρόνησις καὶ ἡ τῶν πρακτῶν) est, de ce point de vue, très intéressante : c’est la saisie de ces différences qui permet l’usage intelligent des intelligibles et des actions à accomplir. Nous avons donc là une échelle des vivants, qui se conforme aux principes téléologiques aristotéliciens. Tous les animaux n’ont pas besoin de cinq sens et les huîtres, par exemple, posséderaient la vue ou l’ouïe en vain. Ce serait, en effet, une charge bien inutile pour les huîtres d’avoir des systèmes perceptifs aussi complexes dont elles ne pourraient pas avoir l’usage13. C’est donc un signe de perfection pour les animaux que d’avoir cinq sens. Et c’est le sommet de la perfection que de pouvoir se servir de ses cinq sens pour accéder à la φρόνησις et à la « vie heureuse ». Les autres animaux y ont un accès si l’on peut dire métaphorique, puisqu’il existe une φρόνησις animale qui est une copie imparfaite de la φρόνησις humaine, tout en étant une véritable φρόνησις. Mais ce sont évidemment les êtres humains qui ont globalement le meilleur système perceptif, qui leur permet d’avoir un accès plus différencié au monde. La « vie heureuse » est donc le signe qu’un vivant possède le plus parfait des systèmes sensoriels, tout comme elle est le signe que les hommes sont arrivés au stade suprême de l’association humaine, à savoir la cité. De ce point de vue, on pourrait peut-être trouver enfin la « preuve » du fait qu’il n’y a pas plus de cinq sens. Si l’animal le plus parfait (« le plus conforme à la nature », Locomotion des animaux 4, 706a19), le seul à atteindre la vraie vie heureuse, n’a que cinq sens, cet état « penta-sensoriel » est un état parfait. En fin de compte, la capacité de connaître que donnent les sens est la chose du monde la mieux partagée. Mais il y a deux cas. Pour les animaux qui n’ont pas cinq sens, Aristote ne nous dit pas s’ils peuvent (ou si certains d’entre eux peuvent) saisir tous les objets sensibles. Ceci n’est pas exclu, puisque nous avons vu qu’il suffisait peut-être de deux ou trois sens pour pouvoir percevoir tous les objets, même si l’animal n’en percevrait alors pas toutes les propriétés. Pour les animaux qui, comme nous, ont cinq sens, leur appareillage sensoriel fait qu’aucun sensible ne 13 Cf. De Anima III 12, 434b24-27 : « les autres sens [que le toucher et peut-être le goût] servent au bien- et n’importe quelle famille animale n’en est pas pourvue mais seulement certaines, à savoir celles qui possèdent la locomotion et qui les ont nécessairement. Car pour qu’un animal de cette sorte assure sa sauvegarde, il lui faut non seulement percevoir par contact, mais aussi de loin ».

18

PIERRE PELLEGRIN

peut leur échapper du fait de sa nature. Parmi ces animaux qui ont cinq sens, la prééminence humaine est clairement une question de degré. Alors pourquoi Aristote pose-t-il une frontière infranchissable entre l’homme et les animaux sur la question de la possession du logos  ? En fait il faut ajouter un codicille de taille à ce qu’on (moi compris) dit souvent sur Aristote, à savoir que, contrairement à Platon, il pose une véritable connaissance sensible. Parce que, même si la science prend son départ dans la sensation (nihil est in intellectu quod non prius fuerit in sensu), ce n’est pas l’exercice de la sensation qui donne l’intelligence, de même que ce n’est pas l’usage de la main qui rend les hommes intelligents. L’intelligence qui va permettre aux êtres humains de faire un usage théorétique de leurs sens est une donnée de leur équipement biologique et c’est cette donnée innée qui permet aux humains de faire un usage intelligent de leur faculté de percevoir.

ARISTOTLE ON THE PERCEPTION OF OBJECTS David CHARLES

1. Introduction In this essay I shall consider a general philosophical problem and Aristotle’s distinctive solution to it. The general problem is this: how does our understanding or knowledge of ordinary objects and their properties arise out of our perceptual experience? I shall call the thesis that our understanding of the concepts of ordinary objects and their properties and our knowledge of such objects and properties is grounded in our perceptual experience “experientialism”. The philosophical question is: how does perceptual experience provide the basis for our concepts of and knowledge about ordinary mindindependent objects and properties? Since Aristotle’s requirements for scientific knowledge are particularly high, I shall investigate only his account of the first step on the road to scientific knowledge. A great deal depends, in addressing this general problem, on how perceptual experience itself is to be understood. Bishop Berkeley conceived of such experience as the experience of mind-dependent entities (of which sense-data would be an example). His puzzle was this: how do we move to knowledge of mind-independent objects on the basis of our experience of mind-dependent objects? To this question, he returned his famous negative answer: we cannot do so1. Structurally similar problems arise if one thinks of perceptual experience as being of colours, tastes, or sounds, what I shall call “modalspecific properties”. Here, the question is: how do we move from our 1 This issue is the subject of excellent discussion in Berkeley’s Puzzle: J. Campbell and Q. Cassam (2014).

20

DAVID CHARLES

awareness of modal-specific properties to knowledge of objects? If perception is of properties, such as being red, being green, being noisy, being bitter or being yellow, how do we move from perceiving properties of this type to knowledge of objects? This question is particularly acute if the objects in question are (what I shall call) “cross-modal”, accessible to different senses. Aristotle’s starting point, I shall argue, is distinctive and worthy of serious consideration. My suggestion is that, in his account, visual or tactile perceptual experience is of cross-modal objects and of some of their cross-modal properties (such as size, shape and movement). We see, using the special sense of vision alone, cross-modal objects and properties of this type as well as modal-specific ones, such as colour. Consider a traditional view of Aristotle’s discussion of the common sense: our perception of cross-modal objects and properties. According to this interpretation, Aristotle held that: (A) Special senses, such as vision and hearing, perceive, colour and sound (modal-specific properties) and nothing else; (B) The common sense perceives: (i) one cross-modal object as, for example, bitter and yellow, and (ii) two different cross-modal objects as different. (C) Although the common sense goes beyond the special senses but, in some way, arises out of them. Anna Marmodoro, in her recent book Aristotle on Perceiving Objects2, develops a sophisticated version of this view. She writes: [Common sense] empowers and enriches its constituents with functionalities their union alone could not secure. Its operation results in the generation of new content (e.g., awareness of the common sensibles) over and above unifying the existing content from the special senses (e.g., discerning white from sweet). The generation of new perceptual content, and the unification of existing perceptual content into further contents, is the nature and role of the common sense. (p. 275)

The question she addresses is: how does perception of modal-specific properties, such as colours and sounds give rise, in Aristotle’s view, to the perception of cross-modal objects (objects perceivable by more than one type of perception)? 2

A. Marmodoro (2014).

ARISTOTLE ON THE PERCEPTION OF OBJECTS

21

There are two issues to address. First, did Aristotle accept claim (A): are the special senses confined to seeing colours and tasting flavours, the detection of modal-specific properties? Or do we also see and taste the objects to which the colours and flavours belong? If we do, are the objects we see and taste modal-specific or cross-modal objects, ones accessible to more than one sense? The second issue is this: how does common sense, in Aristotle’s account, arise out of the special senses and issue in the perception of one cross-modal object (such as bile) as bitter and yellow (to use his example)? These two issues are the principal focus of this paper. 2. What do the special senses sense? In De Anima 418a24-25 Aristotle writes: Tῶν δὲ καθ’ αὑτὰ αἰσθητῶν τὰ ἴδια κυρίως ἐστὶν αἰσθητά, καὶ πρὸς ἃ ἡ οὐσία πέφυκεν ἑκάστης αἰσθήσεως. Of the two types of sensibles (sensibilia) which are perceived in their own right, it is the objects special to different senses that are in the paradigmatic way perceptible and it is to these which the essential character of each sense is naturally directed.3

There are several ways to understand this remark. According to one traditional interpretation, sight perceives colour and nothing else, hearing sound and nothing else, taste flavour and nothing else. Even when the colours, sounds and tastes are the colours, sounds and flavours of objects, sight only perceives colours, hearing only sounds and taste only flavours4. On this view, even if colour is always a property of an object, the object itself is not something which is seen. The subject takes a major step in moving from seeing colour to perceiving (in some other way) the objects which are coloured. On the second account, by contrast, 3 I translate καθ’ αὑτὰ αἰσθητά as “perceived in their own right” to capture the thought that they are per se causes of perception. I understand κυρίως as “paradigmatically” or “principally” in line with the use in Cat. 2a11-13, where it is conjoined with “firstly” and “most of all” (μάλιστα). See also Nicomachean Ethics 1098b14, where it is conjoined with “most of all” (μάλιστα) and 1157a31, where is used together with “firstly” (πρώτως). 4 Those inclined to this view interpret Aristotle as saying that colours and sounds (etc.) are not just principally but strictly speaking perceptibles: the only objects of the special senses.

22

DAVID CHARLES

what is seen is a coloured object, what is heard is a noisy object, what is tasted is a sweet object. Even if colour is what is seen in the paradigmatic way directly, direct seeing is not confined to seeing colour5. There is a second distinction. There are two ways to understand the claim that what is seen is a coloured object, not simply some specific colour. The object in question could be a cross-modal object, open to more than one sense, or a modal-specific coloured-object or auditoryobject (defined as an object accessible to only one sense). On the latter view, there will be still a major step from the visual discrimination of a modal-specific object (this coloured-object) to the perception of cross modal-objects. On the former view, by contrast, visual perception is, at the outset, of cross-modal objects such as large moving coloured objects. No further step is required to arrive at the perception of cross-modal objects. From the very beginning, we directly see and visually discriminate objects of this type. In De Anima II 6, Aristotle says that movement is perceived both by vision and by touch (418a18-20). Movement, together with size, shape and number is a common sensible: a feature common to all the senses. Since shape, size, movement and number are also perceived by different senses, what is discriminated by vision or by touch is a cross-modal object, such as one large moving object. Such objects can be visually discriminated, even though their oneness or movement is not the type of feature to which vision alone is targeted. Vision alone may be the only sense directed to grasp the colour of the object. However, while other senses grasp other features of the same object, the object itself is nonetheless a per se cause of the visual discrimination of one coloured object. What you see is one coloured moving cross-modal object (such as a red cricket ball coming towards you). Here are some further considerations in favour of this suggestion. (i) In 418a11 Aristotle writes that a specific (ἴδιον) type of object of perception cannot be perceived by any other sense. Indeed, what makes something a specific object or feature of this type is that it 5 Those who take this view understand Aristotle to say that the special sensibles are the principal or basic type of perceptibles: those to which the nature of each sense is specifically directed (perhaps taking the “and” in 418a25 as epexegetic). So understood, a person can (non-accidentally) see the common sensibles, even though vision is not specifically targeted as them. They are accessible to other senses as well.

ARISTOTLE ON THE PERCEPTION OF OBJECTS

23

cannot be perceived by any other sense. By contrast, common objects can be perceived by more than one sense (perhaps by all). If such objects are common to all senses, each sense will be able to have access to them in their operation as those senses. In particular, visual and tactile perception will have access to the same cross-modal objects and features (such as shape and movement), not to different modal-specific objects and features (such as visual-movement). They will perceive the same moving object, even though sight perceives it visually, touch in a different tactile way (418a19-20). (ii) In 418a15 Aristotle writes that each sense (such as visual perception) cannot be deceived with regards to whether it sees, for example, a colour or hears a sound, but can be deceived as to which object is coloured and where it is. If so, it seems that visual perception discriminates (and can be deceived about) a determinate object at a specific location: it visually discriminates, for example, that red moving object over there. While Aristotle’s remark at 418a15, taken in isolation, could be about modal-specific objects, his next claim about movement, size and number, suggests that the objects discriminated by sight and touch are crossmodal, accessible to both these senses. There is certainly no sign of the visual-size, visual-movement or visual-objects required if the entities discriminated are to be modal-specific ones. It is against this background that we should interpret Aristotle’s remark in 418a24-25, already mentioned: Of the two types of sensibles (sensibilia) which are perceived in their own right, it is the objects special to different senses that are in the paradigmatic way perceptible and it is to these which the essential character of each sense is naturally directed.

The essential character of vision may be naturally directed to colour in that it alone perceives colour. Seeing colours is the unique contribution of vision. But this is compatible with vision by itself seeing both colour and cross-modal objects (which are coloured). Perhaps vision has a complex essence: one part unique to it, another shared with other senses. Alternatively, and equally compatibly with this remark, seeing cross-modal objects could be a necessary feature of visual experience, even though its essential character is limited to what is unique to it (seeing colours). On both

24

DAVID CHARLES

readings, vision can see moving objects as well as colours, even if its essence is the capacity to see colours. Either way, there will be (nonaccidental) seeing of colours and of moving objects. That what is seen is not confined to colour alone is confirmed by Aristotle’s remark at the beginning of the next chapter, De Anima II 7: Οὗ μὲν οὖν ἐστιν ἡ ὄψις, τοῦτ’ ἐστὶν ὁρατόν, ὁρατὸν δ’ ἐστὶ χρῶμά τε καὶ ὃ λόγῳ μὲν ἔστιν εἰπεῖν, ἀνώνυμον δὲ τυγχάνει ὄν (…), τοῦτο δ’ ἐστὶ τὸ ἐπὶ τοῦ καθ’ αὑτὸ ὁρατοῦ (…) ὅτι ἐν ἑαυτῷ ἔχει τὸ αἴτιον τοῦ εἶναι ὁρατόν. The object of sight is the visible. What is visible is colour and that which can be described but has no name (…), and colour is that which is on that which is in itself visible (…) because it has in itself the cause of being visible. (418a26-31)

If what is in itself visible, the cause of visibility, is the transparent on the surface of objects, what is seen (and so visually discriminated) will include this as well as colour. But does Aristotle consistently include the object as visible also (as suggested by 418a15-16, cited above)? The case for an affirmative answer is strengthened by his subsequent remarks in De Anima II 7. He lists, at 419a2 sq., among the objects of visual sense cross-modal objects such as fungus which are not seen in the light, and contrasts them with coloured objects seen at a distance. The per se objects of sight are the colour of the object (the fungus) and the object insofar it is coloured, even if it is the colour of the object which is a per se cause of the movement in the transparent mentioned here. This is why Aristotle says later that what is seen is the colour or that which has the colour (425b18-19). Nor are these remarks confined to his discussion of sight. In 422a8 sq., Aristotle refers to the object of taste as a type of tangible object and as a body. Since bodies are cross-modal objects, the object of taste must be an object with a given taste (or the taste of that object) and not only some modal-specific feature or object. The objects of taste include what Aristotle calls the drinkable and the undrinkable (422a31-34), where what is drinkable and undrinkable are cross-modal objects such as juices, water, beer, and soda6. 6 There is, it should be noted, a persistent ambiguity in Aristotle’s use of phrases such as “the white” or “the bitter”. They can refer either to the whiteness of the object or the white object and need to be disambiguated in context.

ARISTOTLE ON THE PERCEPTION OF OBJECTS

25

Nor are these remarks confined to De Anima II. In De Anima III 1 Aristotle writes that if only one sense were available to us, the common sensibles (sensibilia), such as oneness and movement, would tend to escape notice. (425b7 sq.) This comment suggests that the common sensibles, the objects of the common sense, would be present to visual perception, even if we had only one sense and so tended not to notice them. It implies that the objects in question are cross-modal, even though their cross-modal status could easily escape our attention if we only had one sense. In that case, we would, in fact, be discriminating a cross-modal object, even if we were not aware that we were doing so. Before we turn to consider Aristotle’s discussion in De Anima III 1, we should note that the traditional view requires careful statement. What is the object of vision in this account? Some speak, somewhat uncritically, of redness, a universal, as what is discriminated by perception. However, Aristotle is clear that it is the particular and not the universal that is perceived (as in, for example, Posterior Analytics 87b37-39)7. He makes similar comments in De Anima: what is perceived is a particular, not the universal (De Anima 417b22-28; 429a15-18). Other interpreters have suggested that the object of vision is a patch in one’s visual field or some quantity of redness, perhaps redness here and now. But, one must ask, what for Aristotle determines the relevant “here and now”? Surely this must be, in his account, an object, since he describes colours as a type of quality (ποιόν τι8), and – as such – a quality of an object. Indeed in De Anima II 6, as already noted, Aristotle talks of the coloured object as a determinate and spatially locatable particular. A third suggestion is that the object of vision is modal-specific: a coloured-object, not a cross-modal object. However, this too is problematic as Aristotle notes that movement is accessible both to sight and to touch. What is seen and touched must be a cross-modal object (418a19-20). 7 What is perceptually discriminated is a particular coloured object even if (de re) we are perceiving a universal (Posterior Analytics II 19,100b3 sq.). The per se cause of colour perception may be universals, such as redness, even if what is visually discriminated is a particular coloured object. 8 Topics, 103b32, 120b35 sq.

26

DAVID CHARLES

A fourth suggestion runs as follows: Aristotle held that what makes a colour one colour is its being the colour of one cross-modal object but denied that the cross-modal object itself is seen. However, here too there are problems: he talks elsewhere, as already noted, of the object itself being seen at a given place and of vision being deceived as to which object is seen and where it is. But these phrases indicate that it is crossmodal objects that are seen and visually discriminated. Deception is a way of going wrong about where the object is and which object it is. If so, seeing (when all goes well) will involve, in the case of common sensibles, the successful discrimination of the object in question and its location. Both the object and its location will be part of what is seen. Further, given that in De Anima II 6 the object has features (such as movement) which are open to several senses, what is discriminated must be a cross-modal object. Nor has any reason been given to deny that the objects seen, which are per se causes of visual perception, are cross-modal. (This is, of course, consistent with colours being per se causes of visual perception in the paradigmatic way in that they are the target of vision alone.) To conclude: I have presented some considerations in favour of the view that, for Aristotle, vision and touch (along with the other special senses) individually perceive cross-modal objects. They are not confined to modal-specific properties (as in the traditional view) nor to modalspecific properties and modal-specific objects (as in empiricist theories of perception). In the next section, I shall consider Aristotle’s discussion in De Anima III 1 which contains his discussion of the case when a perceiver perceives one bitter yellow cross-modal object. My aim is to interpret this chapter in such a way that it is consistent with the view, just sketched: that vision and the other special senses by themselves have perceptual access to cross-modal objects. 3. How the perceiver perceives one bitter yellow object Aristotle begins his discussion of common sensibles in De Anima III 1 as follows: Ἀλλὰ μὴν οὐδὲ τῶν κοινῶν οἷόν τ’ εἶναι αἰσθητήριόν τι ἴδιον, ὧν ἑκάστῃ αἰσθήσει αἰσθανόμεθα κατὰ συμβεβηκός, οἷον κινήσεως, στάσεως, σχήματος, μεγέθους, ἀριθμοῦ· ταῦτα γὰρ πάντα κινήσει

ARISTOTLE ON THE PERCEPTION OF OBJECTS

27

αἰσθανόμεθα, οἷον μέγεθος κινήσει (ὥστε καὶ σχῆμα· μέγεθος γάρ τι τὸ σχῆμα), τὸ δ’ ἠρεμοῦν τῷ μὴ κινεῖσθαι, ὁ δ’ ἀριθμὸς τῇ ἀποφάσει τοῦ συνεχοῦς, καὶ τοῖς ἰδίοις (ἑκάστη γὰρ ἓν αἰσθάνεται αἴσθησις). There is no special sense organ for the common sensibles (sensibilia) which we perceive with each [special] sense accidentally: such as movement, being at rest, shape, size, being one or number. All these we perceive by movement – as we perceive size by movement (and shape – a type of size), being at rest by not being moved and number by the discreteness of the object – and by the individual sense organs (for each sense perceives one thing). (425a13-20)

The first sentence, as I understand it, states the view Aristotle wishes to reject: that there is no special sense organ directed at the common sensibles (movement, being at rest, being one) which we perceive by each other sense accidentally. He intends to reject this because, in his view, the common sensibles are all, in fact, perceived by the special sense organs (425a19). As in De Anima II 6, vision is non-accidentally of the common sensibles, such as one moving cross-modal object. While sight is the only sense which detects the colour of an object, it also discriminates the object which is coloured: it sees one coloured, moving, object. Aristotle argues that the special sense organs do not detect oneness and movement accidentally in the way they would if there was (as there is on the hypothesis to be rejected) a special organ directed at the common sensibles. For, if that were the case, vision would only see colours accidentally. His present position, so understood, is fully consistent with the one he adopted in De Anima II 6: while the essence of vision may be targeted at colours alone, we do, in fact, also see (non-accidentally) moving objects. Their number and movement is a per se cause of visual experience, even if seeing them is not what is distinctive of seeing. After all, senses other than vision can also discriminate them. Even if the essence of seeing is the discrimination of colours, it also necessarily discriminates moving (or stationary) objects9. 9 This use of “accidentally” in 425a15 has caused problems. Torstrik suggested adding a “not” to read “not accidentally” so as to make it consistent with Aristotle’s remarks in De Anima II 6. But his radical move is unnecessary. One could interpret this term as indicating only (as Christopher Shields notes: C. Shields (2016), p. 261) the view of Aristotle’s opponent (and not Aristotle’s own). The opponent would suggest that (on his – the opponent’s – hypothesis) we detect common sensibles accidentally with each of the special sensibles. This interpretation dates back to Themistius, Philoponus and Simplicius

28

DAVID CHARLES

His remark in the final sentence quoted suggests the same account: we perceive the common sensibles by each of the special sense organs (τοῖς ἰδίοις: 425a19), a comment which Aristotle explains by noting that “each sense perceives one thing”10. In the immediate context, this comment is best taken as suggesting that each of the senses perceives one (cross-modal) object when it perceives a coloured or sweet object. This is why individual senses (and their organs) perceive common sensibles. They all perceive cross-modal objects. (I shall return below to Aristotle’s telegrammatic claim that we perceive common sensibles “by movement”: 425a17). Aristotle next argues that common sensibles (sensibilia), such as one moving object, are not grasped by a further sense apart from vision, taste, hearing, touch, and smell. If they were, he suggests, several undesirable consequences would follow. He reasons as follows: Ὥστε δῆλον ὅτι ἀδύνατον ὁτουοῦν ἰδίαν αἴσθησιν εἶναι τούτων, οἷον κινήσεως· οὕτω γὰρ ἔσται ὥσπερ νῦν τῇ ὄψει τὸ γλυκὺ αἰσθανόμεθα· τοῦτο δ’ ὅτι ἀμφοῖν ἔχοντες τυγχάνομεν αἴσθησιν, ᾗ ὅταν συμπέσωσιν ἅμα γνωρίζομεν. εἰ δὲ μή, οὐδαμῶς ἂν ἀλλ’ ἢ κατὰ συμβεβηκὸς ᾐσθανόμεθα (οἷον τὸν Κλέωνος υἱὸν οὐχ ὅτι Κλέωνος υἱός, ἀλλ’ ὅτι λευκός, τούτῳ δὲ συμβέβηκεν υἱῷ Κλέωνος εἶναι)· τῶν δὲ κοινῶν ἤδη ἔχομεν αἴσθησιν κοινήν, οὐ κατὰ συμβεβηκός· οὐκ ἄρ’ ἐστὶν ἰδία· οὐδαμῶς γὰρ ἂν ᾐσθανόμεθα ἀλλ’ ἢ οὕτως ὥσπερ εἴρηται τὸν Κλέωνος υἱὸν ἡμᾶς ὁρᾶν. (see J. Owens (1982)), who took Aristotle to be presenting a hypothesis which he did not himself accept. Alternatively, if one were to take Aristotle himself as endorsing the claim in 425a15, he need only be saying (as Hicks suggests) that the seeing of common sensibles is not part of the essence of vision – even though it is still a necessary part of the operation of seeing (as he suggested – on one reading – in De Anima II 6, 418a24-5). In scholastic terms: seeing an object’s cross-modal features would be a necessary accident of seeing (see, for the latter notion, Metaphysics 1025a31). While Shields’ proposal is the more elegant, it is not necessary for our present purposes to adjudicate between these two interpretations. (I am indebted to Christopher Shields for permitting me to see a prepublication draft of his acute discussion of this passage.) 10 “Sense organ” is the nearest relevant noun (425a14) where it is qualified, as here, by “specific”. On this reading, the next phrase “for each sense senses one thing”, is taken to refer to the cross-modal object they sense and so to explain how individual senses grasp one common object and its properties. Some commentators have supplied the term “objects of sense” (αἰσθητά) to qualify “specific”, even though there is no noun in the immediate context to underpin this reading. They support their suggestion by understanding the phrase “each sense senses one thing” to mean that each sense senses one type of sense object (vision coloured objects of sense etc.). However, this clause, so understood, seems irrelevant to Aristotle’s claim (which is required in these lines) that the special senses discriminate movement, size and shape.

ARISTOTLE ON THE PERCEPTION OF OBJECTS

29

It is clear that there cannot be a special sense for these common sensibles, such as movement; for then the situation will be as it is when we now perceive what is sweet by sight. This happens because we have perception of both and grasp on this basis when they co-occur. If this was not so [that is, if there were a special sense for common sensibles and we perceived objects moving not in the way just mentioned], we would never perceive the common sensibles except accidentally in the way in which we see Kleon’s son not because he is Kleon’s son, but because he is pale and what is pale happens to be Kleon’s son11. But we have, in fact, a common [or shared] perception of the common sensibles non-accidentally. Thus they are not specific to one sense; if they were we would not perceive them otherwise than in the way in which it has been said we see Kleon’s son. (425a20-30)

Aristotle is still concerned to argue that our perception of common sensibles is not the work of a further special sense, dedicated to perceiving just them. If it were, either we would perceive a moving yellow object in the way we perceive by sight a sweet object or in the way in which we perceive Kleon’s son. The two cases are different. In the former, we see some yellow object, taste something sweet and, when they co-occur, grasp them simultaneously. Applied to the case of seeing a moving red object: we perceive something moving by the common sense and see something red and, when they co-occur, grasp that they do so (when the red object is the moving object). However, in Aristotle’s own account, we see one moving object without beginning from these two distinct initial preliminary and distinct acts of perceiving plus a further act of grasping that the red object is the moving one. This case differs from that of grasping that the yellow object is the sweet one because, as I have suggested, we see immediately one moving red object. There is no further step required to identify the red object as moving. Nor is the situation like that in which we see Kleon’s son accidentally. This is because the object’s moving and its being white (unlike its being Kleon’s son) are both directly given to us in perception. They, unlike being Kleon’s son, are both per se causes of visual experience12. Aristotle makes a more positive point: we have, non-accidentally, a shared (or common) perception of common sensibles (425a27-28). Each 11 In 425a26, I follow those MSS who read τούτῳ (referring to the pale to which it belongs to be Kleon’s son) rather than those which read τοῦτο (where the presence of εἶναι in 425a27 looks cumbersome). 12 Movement etc. could be a per se cause of visual perception even if seeing it is a necessary accident of visual perception.

30

DAVID CHARLES

sense has access to the same common sensibles. They are not the unique preserve of a sense dedicated only to perceiving one moving object. As in De Anima II 6, one such object is visually perceived, touched, smelled or tasted. If so, the object in question is cross-modal, open to these different senses. As before, there is no suggestion in De Anima III 1 that he is thinking of the mental construction of a cross-modal object out of modal-specific ones, such as a coloured-object or a taste-object. The objects perceived are accessible to sight, taste and touch. Aristotle next attempts to say more about what actually happens. He writes: Τὰ δ’ ἀλλήλων ἴδια κατὰ συμβεβηκὸς αἰσθάνονται αἱ αἰσθήσεις, οὐχ ᾗ αὐταί, ἀλλ’ ᾗ μία, ὅταν ἅμα γένηται ἡ αἴσθησις ἐπὶ τοῦ αὐτοῦ, οἷον χολὴν ὅτι πικρὰ καὶ ξανθή. But the different senses do perceive accidentally each other’s specific objects, not as the senses they are, but as one sense when there is a simultaneous perception of the same object, as when, with respect to bile, they perceive that it is yellow and bitter. (425a30-b2)13

This passage raises two problems: (a) How does perception of this yellow object and of this bitter object yield the perception with regard to one object that is yellow and bitter14? How does Aristotle understand this transition? Why did the step seem so unproblematic for him? How do vision and taste function as one when targeted on one object which is both yellow and bitter? One suggestion runs as follows: STEP: Vision and taste are both individually in perceptual contact at the same time with the same cross-modal object: one yellow bitter object, such as bile. Since perception involves discrimination, both senses perceptually discriminate the same object with the same shape at the same place: that object here with that shape and size. 13 At 425b1, I read with E and L, χολὴν ὅτι. Ross’ χολῆς ὅτι has no MSS support. It would be possible (without major change of sense) to take χολὴν not as an accusative of respect but, with Hicks, as the direct object of “perceive”. 14 Indeed he seems to treat the move as unproblematic. At a similar point in De Sensu he writes: “The soul perceives numerical identity on the basis only of simultaneously perceiving the objects” (447b24-25). However, once again, he does not spell out how this happens, treating the phenomenon as relatively unproblematic.

ARISTOTLE ON THE PERCEPTION OF OBJECTS

31

On the basis of each sense perceiving the same cross-modal object at the same time with the same cross-modal properties, perceivers perceive one yellow bitter object. This happens because their two senses perceptually discriminate the same object at the same spatial position. There is no need for any further perception of the common sensibles beyond what is given to these two senses. They both focus on one and the same object at the same place. This is the way in which the different senses, acting as one, discriminate one and the same cross-modal object albeit with differing modal-specific properties, when it is yellow and bitter15. It is important to note that while vision perceives accidentally the bitter object (e.g. the bile), it does not follow that it perceives the crossmodal object itself accidentally. Aristotle notes that the senses perceive each other’s objects accidentally when they are directed to the same object. This will be true (on the account just offered) when perceivers perceptually discriminate (with their differing sense modalities) one and the same cross-modal object at one spatial position. They visually discriminate a bitter object only accidentally because its bitterness is not a property directly accessible to vision. It is not a per se cause of vision. By contrast, its being one moving object is something which is directly accessible to vision. It is, as I have emphasized, a per se cause of vision, even if the essence of vision is targeted at colour alone. (b) Aristotle describes the case as follows: with regard to bile, the senses as one perceive that it is bitter and yellow (425a31-b2). If perceivers perceive one cross-modal object (such as bile) at the same place, they perceptually discriminate one bitter yellow object. In doing so, they clearly perceive one object as bitter and as yellow. But do they also, in Aristotle’s account, perceive it as both bitter and yellow – as one object with both features? That is: do they perceive that these two features both belong to one object? Aristotle continues: “For it is not the task of a further sense to say that one object is both [yellow and bitter]” (425b2-3)16. How is this remark to be interpreted? Is he saying that it is not the task of a further sense to say that one object is both yellow and bitter because we perceive 15

For a modern view of this general form, see J. Campbell (1987). Or, perhaps, to say that that yellow = that bitter (as Hicks suggests). I am inclined to follow Ross’ translation in which one (thing) is the subject and “both” refers to being yellow and bitter. But nothing of significance for this essay depends on this decision. 16

32

DAVID CHARLES

one object as both yellow and bitter (perceiving both these features as co-predicated of one object), when the senses, acting as one, focus on the same cross-modal object at the same place. That is, these senses, as one, say that one object is both yellow and bitter. No further sense is required beyond these senses acting in this way. Or, alternatively, is he suggesting that it is not the task of any other sense to say that one object is both yellow and bitter because perception only goes so far as to perceive one object as yellow and as bitter? No other sense is required because the task of co-predicating yellowness and bitterness of one and the same object belongs to thought and not to perception. It is not any sense that says that one object is both yellow and bitter. While both interpretations may be possible, the first seems preferable. Aristotle, so understood, is clarifying his immediately preceding remark that the senses acting as one take bile to be yellow and bitter by saying that in this condition – without any further sense (such as the separate common sense rejected earlier: 425a13 sq.) – say that one object is both yellow and bitter (co-predicating yellowness and bitterness of one object). By contrast, the suggestion that thought is what says that one object is both yellow and bitter would do nothing to clarify the preceding remark about the senses taking bile to be yellow and bitter. Consider Aristotle’s next remark: “This is why mistake occurs and when the object is yellow it is thought to be bile” (425b3-4). What is his explanation of the mistake involved in thinking that something yellow is bile? Does it arise because the senses (as one) perceive, on some occasions, one object as both yellow and bitter? On this understanding, when we perceive something as yellow, we make the mistake of thinking – because of past co-perceptions of this type – that it is also bitter (and so bile)17? Or does the error arise simply because we, on occasion, have perceived one object as bitter and as yellow and so now on perceiving a yellow object think it bitter? On this view, one will mistakenly associate yellowness with bitterness (and bile), without the support of any (past) perception of them as co-belonging to one object. It will be a case of over-hasty generalisation from past experience. 17 There is a grammatical issue here: is the subject of “is mistaken” the sense as one (425a31) or impersonal subject (the perceiver)? I leave this issue open, suspecting that it may be irresoluble.

ARISTOTLE ON THE PERCEPTION OF OBJECTS

33

While, once again, both interpretative options may be possible, the first seems preferable. It appears to use the idea of perceiving one object as both yellow and bitter as the basis for the explanation of error18. The past perception of two features as belonging to one object is what leads us astray. It is not simply that we associate in thought yellowness and bitterness on the basis of past experience. For then it would not be faulty reasoning and not past experience that misled us. This said, my preferred interpretation of 425b3-4 has significant consequences. If to focus on one and the same cross-modal object as yellow and bitter is, in Aristotle’s view, to perceive it as the common subject of both yellowness and bitterness, there has to be some one thing to which both the yellowness and bitterness of the object are presented. And what is this – since it is neither vision nor taste? What is it that is perceptually sensitive to both the bitterness and the yellowness of one object at one time? What, we might say, is the unified perceptual consciousness to which both features of the object are present? The question is pressing but unanswered in De Anima III 1. His brief remarks in 425a30-b3 may serve to set up questions to be answered (if at all) later. In the next chapter, Aristotle considers how we perceive that the whiteness of an object is different from its sweetness (426b17 sq.) – a question that presupposes, it seems, that both features are given to, and perceived by, one perceiver. Later, in De Anima III 7, he talks, admittedly in a particularly obscure passage, of one (perceptual?) mean receiving both visual and auditory information (431a17-20) and as discriminating – as something one in number (431a22-23) – some one thing which is white and sweet (431a29-b2). It seems that his introductory remarks in De Anima III 1 (425b2-4) generate problems which need to be tackled elsewhere. They set the agenda for subsequent investigation rather than showing how to resolve them. 4. Questions about this interpretation We should consider some exegetical objections to the reading just proposed. Aristotle is clear that there is no single organ of perception which grasps the common sensibles. So how, one might ask, given there 18 On this reading, the “this is why” (διὸ: 425b3) refers back to the claim: the sense perceives with regard to bile that is yellow and bitter.

34

DAVID CHARLES

is no single common organ, can common sensibles impact on us in the way suggested? In the account developed above, the perceiver is, I assume, causally assimilated in every relevant case to one stationary or moving, threedimensional object: this is what the visual and tactile perceiver is assimilated to when seeing or tasting. Against the background of Aristotle’s realist metaphysics of three-dimensional moving objects, there is causal assimilation of this type in any successful case of seeing or touching one moving object19. The critic will not be satisfied. What is needed, she will insist, is an account of how information about the location of a cross-modal object is carried to our special senses and grasped by them. How does vision see one moving object? When Aristotle speaks of sight, he talks of the medium being affected by colour, but not by certain other features of object which is coloured (424a21-24). How do cross-modal features of objects impact on the relevant medium or the visual perceiver? We are clearly sensitive, in Aristotle’s account, to oneness and movement in the case of seeing and hearing, when we see one object moving from left to right or hear a sequence of notes. There must, therefore, in the case of sight be a distinctive impact on us made by large triangular objects as opposed to small circular ones. In Aristotle’s terminology, the objects in question will have different forms. Further, the objects will be seen as occupying differing spatial positions, as moving in front of and occluding some objects and, in turn, being occluded by them. Objects as they move towards us will look bigger and smaller as they move away. There must, if Aristotle’s account is to be sustained, a distinctive type of impact caused by these differing common sensibles when we perceive them. Aristotle comments in 425a16-17: “we perceive all the common sensibles by movement (κινήσει)”. The Greek commentators took this phrase to refer to the movements set up in us by cross-modal objects and / or their cross-modal properties20. Their idea was that we perceive an object’s size because it moves us continuously (as one unit), we perceive its being at 19 There can, of course, be cases where we are misled and see a two-dimensional array as three dimensional (as in the cinema). While these cases require further discussion, they do not by themselves call into question the claim that in good cases we see three dimensional objects. For further discussion of Aristotle’s views on the role of causal assimilation in accounting for perception and thought, see D. Charles (2000), chapter 5. 20 Hicks summarizes their views in his note on 425a17 (R.D. Hicks (1907), p. 428).

ARISTOTLE ON THE PERCEPTION OF OBJECTS

35

rest because it does not move us at all and we perceive number (e.g., of two different objects) because there is some discontinuity in the movements they set up in us. In support of their view, one might cite Aristotle’s talk of contrary movements set up in us by sameness and difference (see 426b30). Aristotle, so understood, is only saying here that there are types of movement caused in perceivers by cross-modal objects and / or their cross-modal properties, distinct from those caused by their modal specific properties21. However, this is not the only alternative: Aristotle may be suggesting that we perceive all the common sensibles by perceiving motion. So understood, he claims that we perceive an object’s size by perceiving an object moving continuously (with all its parts): we perceive it being at rest because we perceive it not moving and we perceive number because we perceive the discontinuous movement of (what are in fact) two objects. On this reading, the basis for our perception of all common sensibles will be our perception of movement in the objects we perceive22. Since Aristotle did not develop this reference to “by movement” further, it is difficult to determine which of these interpretations is to be preferred. However, it is an advantage of the second interpretation that it takes “movement” to refer to the movement of objects perceived in all its uses throughout the whole passage: 425a15-21. The first reading, by contrast, takes the first and last occurrence of “movement” to refer to the external movements of objects, the intervening three to internal movements in the sense organ, the last to external movement of objects. The second interpretation has a further advantage: since all per se sensibles 21 Hicks, commenting on this passage, rejects this view, thinking (I believe, mistakenly) that it blurs the distinction between special and common sensibles. But this needs not be the case: the former can have effects specific to one sense even though that sense is affected by other different movements as well (to which other senses also are subject). 22 Aristotle says that “we perceive all the common sensibles by movement (…) and by the individual sense organs (for each sense perceives one object)” (425a16-20). Does he mean: we perceive all common sensibles, including oneness, by perceiving movement etc. via the individual sense organs? Or is he saying: we perceive them all by (a combination of) perceiving movement etc. and the individual sense organs (The latter, unlike the former, does not require that we perceive them all by perceiving movement. The oneness of an eternally static object, for example, could be perceived, without perceiving movement.) While the text leaves both options open, I am inclined to think that Aristotle intended the former. If so, rather than offering an account of how we perceive oneness by perceiving movement, he chose to emphasize that we perceive it by each of the individual senses.

36

DAVID CHARLES

set up movements in the sense organ, why should Aristotle make this comment now specifically about the common sensibles? By contrast, the claim that the distinctive movements in our sense organs (in the case of common sensibles) result from our perceiving the objects moving is a distinctive (and substantial) claim about our perception of common sensibles. Further, so understood, an object’s movements – as well as its colour – can be given to us visually. If the perception of its movement is the basis of our perception of size, shape, unity and location (distance), we can perceive all these on the basis of what reaches us through the transparent – if what is there varies as objects move. Why think that what is present in the transparent is insensitive to whether or not a red (cricket) ball is moving or stationary? Or to whether it is extended in space or unitary? Of course, if the transparent were confined to registering momentaneous snap-shots (“Blicks”) of static colours and unextended sensibles, the situation would be radically different. But why think that Aristotle was attracted to this particular notion23? While Aristotle’s account does not engage with issues concerning perception of distance, dimensionality or depth (understandable since these topics remain matters of controversy even today), there is no reason to believe that his account of the relevant processes in the medium or the eye precludes the possibility of our seeing three-dimensional objects extended in, and moving through, space. Maybe, for example, he was inclined to think that we see one object as close to us when we see – through the relevant medium – something moving in a continuous way against a background of stationary objects. (Consideration of these matters would take us far away from De Anima III 1.) 5. More general philosophical issues Aristotle has, I have argued, a distinctive view of perception: the special senses, such as sight or touch, are, at the outset, of cross-modal objects and their features (some of which are cross-modal). It is because special senses individually perceive objects of this kind that we (and 23 More generally: it should not be assumed that the medium, in Aristotle’s account, yields only two-dimensional information on which we as perceivers project a threedimensional array. If we see size (μέγεθος), we will see three dimensional objects extended in, and moving through, space. Indeed, there is reason to think that this was why he argued against the possibility of unextended perceptibles in De Sensu 7, 449a21-22.

ARISTOTLE ON THE PERCEPTION OF OBJECTS

37

other animals) perceptually discriminate cross-modal objects and their features and can discriminate one object as yellow and as bitter. Further, it is because our special senses individually have access to these objects and features that we (and other animals) have an integrated perceptual system. It is not that the presence of a unified perceptual system explains what individual senses can do; rather the presence of a unified system is grounded in what the special senses do24. Aristotle’s suggestion has important consequences. I shall conclude by noting two. In recent years, it has become fashionable to present Aristotle as suggesting, no doubt correctly, that we possess a reliable route to gain, on the basis of perception, knowledge of kinds and of universals. Some suggest that one of our natural, human abilities: if all goes well, is to proceed from perception to knowledge of kinds. There is, no doubt, some truth in this way of interpreting Aristotle’s account. However, unless we achieve a clearer grip on what perception and its distinctive contribution to such a process are, we will fail to grasp what is distinctive about the type of reliable process which Aristotle advocates. Both Kantian and empiricist can, no doubt, also talk of “a reliable process” from perception to knowledge of cross-modal mindindependent objects. Without a clear understanding of the distinct stages in Aristotle’s account of knowledge acquisition, our discussion of his view must remain at best incomplete, at worst superficial. Let us return finally to the general problem which began. How does perception or sensory experience ground our knowledge or concept of mind-independent objects? Bishop Berkeley, as I noted, famously (and not implausibly) thought that perception or sensory experience, if of 24 For a contrasting view, see P. Gregoric (2007). Pavel Gregoric sees Aristotle as taking as his starting point the existence of a unified holistic perceptual system directed at, for example, cold bitter green (cross-modal) objects (such as a chili) and presents vision and touch as abstractions from it. However, Aristotle, as I understand him, aims to arrive at his account of a unified perceptual system (if this is what is briefly and elusively introduced in De Anima III 7, 431a21 sq.) on the basis of his account in De Anima II 6 and III 1 of the perception of common sensibles (not vice versa). Further, in III 1, 425b5 sq., he seems to accept that – even if there were only visual perception – there could still be perception of common sensibles. Thomas Johansen’s suggestion (in T. Johansen (2012)) that each individual sense has two capacities, one specific to (for example) colour, the other a general one (targeted at the common sensibles) is closer to the view developed in this essay. However, my concern is with a further question: how do differing senses combine to yield a perception of one cross-modal object as yellow and bitter?

38

DAVID CHARLES

modal-specific, mind-dependent, objects could not give rise to knowledge or concepts of mind- independent objects. Others, inspired by Kant, have suggested that such knowledge is not based solely on sensory experience: there are a priori conditions for experience which are also part of the basis of such knowledge. However, if I am right, in the De Anima Aristotle’s account has a fundamentally different shape. It takes as its starting point perceptual (e.g. visual) experience of cross-modal objects, located in space, with size and bulk. These cross-modal objects are, in his view, the causes of our perceptual experience (see talk of acting on etc. in De Anima 416b33 sq., 429a13 sq.) and prior to them in definition (De Anima 415a20 sq.). If they are – as they seem to be on these grounds – mind-independent, Aristotle’s starting point will not be one – adopted by Kantians and British empiricists alike – in which perceivers initially have perceptual access only to modal-specific, mind-dependent objects from which they (or we as philosophers) seek (with great difficulty) to generate knowledge and understanding of mind-independent cross-modal objects. Instead, from the outset, our visual and tactile perceptual experience will be of mindindependent, cross-modal objects. While Aristotle does not attempt to establish in De Anima III 1 that the cross-modal objects of vision or the other senses are mind-independent, this chapter marks, or so I conjecture, an important step in his development and defence of his distinctive, and challenging, epistemological viewpoint25.

25 An earlier version of this paper was given as a talk at NYU in May 2015 and Lille in March 2016. I am indebted to those present for their comments and questions, and especially to Klaus Corcilius, Robbie Howton, Jean-Louis Labarrière, Marko Malink, Jessica Moss, Scott O’Connor and Jennifer Whiting. It is a pleasure to contribute this essay to a volume in honour of Michel Crubellier, from whose acute readings of Aristotelian texts so many of us have profited greatly.

LA CONSCIENCE DE SENTIR N’EST PAS UNE SENSATION : DE ANIMA III 2 Annick STEVENS

Il est bien connu que ce chapitre entame l’examen de la conscience de sentir, mais ne le mène pas jusqu’à une formulation claire de ce qu’est cet acte de conscience ni de la faculté qui le réalise, que l’on trouvera finalement dans les Parva Naturalia. Je voudrais mettre en évidence la structure argumentative de ce chapitre et tout ce qui y est déjà acquis, en particulier les hypothèses éliminées et les éléments nécessaires à la solution. Cette structure demande une reconstitution, parce qu’il y a peu de connexions explicites entre les différentes informations, dont certaines semblent à première vue des digressions ou des considérations extérieures à la question. Comme on le sait, il n’y a pas de verbe en grec pour exprimer la conscience (au sens de « se rendre compte », « s’aviser de quelque chose »1), de sorte qu’on utilise ou bien la négation de λανθάνειν (« il n’échappe pas ») ou bien le verbe αἰσθάνεσθαι. C’est précisément l’ambiguïté de ce verbe qui commande la formulation de la question en termes de « sensation de la sensation ». Le chapitre s’ouvre sur la disjonction entre deux manières de comprendre une telle expression. Nous savons, grâce aux textes plus aboutis d’Aristote sur la question (qui seront évoqués 1 Il ne s’agit, dans tout ce chapitre, que de la conscience qui, selon Aristote, accompagne nécessairement toute sensation, excluant la possibilité d’une sensation subliminale ou inconsciente. Cette conscience appartient donc à tous les animaux dès lors qu’ils sentent. Elle ne doit pas être confondue avec la conscience réflexive, par laquelle le sujet se prend lui-même pour objet, mais elle implique un certain rapport à soi, car l’animal se sent concerné par les sensations qui lui arrivent puisqu’il réagit en conséquence. Cette distinction est loin d’épuiser toutes les significations de « être conscient d’une donnée sensible » — et je remercie très vivement David Charles de m’avoir suggéré une série de précisions dans la discrimination de l’objet, voire du sujet — mais elle suffit pour comprendre le propos d’Aristote. Pour plus de détails sur cette question, je me permets de renvoyer à A. Stevens (2009).

40

ANNICK STEVENS

au cours de l’exposé), que les deux branches de la disjonction seront finalement éliminées, mais ni les étapes ni l’aboutissement de leur examen ne sont évidents dans le présent chapitre. Je vais proposer que la deuxième branche soit éliminée dès le premier paragraphe, et la première par un raisonnement qui s’étend tout au long du chapitre et dont l’argument central sera celui de la relativité tout à fait particulière du sensible et du sens. 1. Structure de la question et première élimination Ἐπεὶ δ’ αἰσθανόμεθα ὅτι ὁρῶμεν καὶ ἀκούομεν, ἀνάγκη ἢ τῇ ὄψει αἰσθάνεσθαι ὅτι ὁρᾷ, ἢ ἑτέρᾳ. ἀλλ’ ἡ αὐτὴ ἔσται τῆς ὄψεως καὶ τοῦ ὑποκειμένου χρώματος, ὥστε ἢ δύο τοῦ αὐτοῦ ἔσονται ἢ αὐτὴ αὑτῆς. ἔτι δ’ εἰ καὶ ἑτέρα εἴη ἡ τῆς ὄψεως αἴσθησις, ἢ εἰς ἄπειρον εἶσιν ἢ αὐτή τις ἔσται αὑτῆς· ὥστ’ ἐπὶ τῆς πρώτης τοῦτο ποιητέον. Puisque nous sentons que nous voyons et entendons, il est nécessaire qu’on sente qu’on voit soit par la vue soit par une autre sensation. Mais (A) la même sensation sera celle de la vue et celle de la couleur qui en est la base2, de sorte que, ou bien (A1) il y aura deux pour la même chose ou bien (A2) la même sera d’elle-même. En outre, même si (B) la sensation de la vue était une sensation d’un autre type, ou bien (B1) on irait à l’infini, ou bien (B2) l’une d’entre elles serait sensation d’elle-même, de sorte qu’il vaut mieux concevoir cela pour la première (A2). (DA III 2, 425b12-17)

2 J’ai choisi cette traduction pour ὑποκείμενον parce qu’il ne s’agit pas du « sujet » au sens aristotélicien habituel (la matière pour la forme ou la substance composée pour ses attributs) et parce que les traductions par « qui lui est soumis » ou « assujetti » donnent l’impression d’une action du sens sur le sensible, alors que c’est l’inverse. Le terme ἀντικείμενα en I 1, 402b12-16 et II 4, 415a16-22 est moins précis car il signifie seulement une opposition réciproque.

DE ANIMA III 2

41

(B1) est éliminé implicitement par la remontée à l’infini (cette sensation devant à son tour être sentie par une autre sensation, et ainsi de suite), car celle-ci n’est jamais une solution ; (B2) est ramené à (A2) Il reste à mettre à l’épreuve (A1) et (A2), ce qui est fait explicitement au paragraphe suivant. Cependant, avant d’aller plus loin, il faut signaler que la plupart des traducteurs comprennent d’une autre façon les lignes 425b13-14 (cf. Rodier (1900), Ross (1961), Barbotin (1966), Tricot (1988) et Bodéüs (1993)3), à savoir : Mais (B) la même sensation sera celle de la vue et celle de la couleur qui en est la base, de sorte que, ou bien (B1) il y aura deux pour la même chose ou bien (B2) la même sera d’elle-même. (425b13-14)

Cette reconstitution me semble moins bonne, étant donné que: selon (B) et (B1) l’autre sensation doit saisir directement la couleur, en plus de saisir la vue ; or la couleur est le sensible propre de la vue, donc aucune autre sensation ne peut la saisir ; il est douteux qu’Aristote formule une telle hypothèse sans en souligner l’impossibilité. Quant à (B2), on ne voit pas comment (B) pourrait impliquer que l’autre sensation se saisisse elle-même – à moins que cette sensation soit une autre vue, comme le proposent Ross, Tricot et Bodéüs ; mais alors il s’agit d’une reprise de l’hypothèse (A) suivant un schéma logique très bizarre : si B, alors ou B1 ou A ! Il vaut beaucoup mieux respecter la construction syntaxique en ὥστε ἢ... ἢ..., qui exprime clairement un choix entre deux conséquences possibles de la même hypothèse. Par conséquent, les lignes 13-14 ne peuvent développer que (A). On peut supposer que la traduction la plus fréquente est motivée par le ἔτι δὲ de la ligne 15, qui semble indiquer qu’on poursuit l’examen de 3 Thillet (2005) fait en sorte de ne pas se prononcer : « Mais il y aura alors la même sensation et de la vue et de la couleur qu’elle supporte, de telle sorte qu’il y aura ou deux sensations du même objet, ou la sensation d’elle-même. »

42

ANNICK STEVENS

la même hypothèse, désormais explicitement la (B) ; cependant, comme le ἔτι est suivi de εἰ καὶ, il peut indiquer que le même résultat est atteint si on retient la deuxième hypothèse : même si on considère B, on revient à A2 (mais le καὶ n’est évidemment pas traduit par ces traducteurs). Les deux interprétations aboutissent à l’élimination de (B) et à un plus ample examen de (A), mais l’élimination de (B) par la traduction majoritaire me semble moins bien construite, aussi bien grammaticalement que logiquement. 2. Les raisons d’admettre une « vue de la vue » Voyons maintenant l’examen de (A). Ἔχει δ’ ἀπορίαν· εἰ γὰρ τὸ τῇ ὄψει αἰσθάνεσθαί ἐστιν ὁρᾶν, ὁρᾶται δὲ χρῶμα ἢ τὸ ἔχον, εἰ ὄψεταί τις τὸ ὁρῶν, καὶ χρῶμα ἕξει τὸ ὁρῶν πρῶτον. φανερὸν τοίνυν ὅτι οὐχ ἓν τὸ τῇ ὄψει αἰσθάνεσθαι· καὶ γὰρ ὅταν μὴ ὁρῶμεν, τῇ ὄψει κρίνομεν καὶ τὸ σκότος καὶ τὸ φῶς, ἀλλ’ οὐχ ὡσαύτως. ἔτι δὲ καὶ τὸ ὁρῶν ἔστιν ὡς κεχρωμάτισται· τὸ γὰρ αἰσθητήριον δεκτικὸν τοῦ αἰσθητοῦ ἄνευ τῆς ὕλης ἕκαστον· διὸ καὶ ἀπελθόντων τῶν αἰσθητῶν ἔνεισιν αἰσθήσεις καὶ φαντασίαι ἐν τοῖς αἰσθητηρίοις. Or, il y a là une difficulté, car, si sentir par la vue c’est voir, et que l’on voie la couleur ou ce qui la possède, alors, si l’on voit ce qui voit, ce qui voit en premier aura aussi de la couleur [ceci est valable pour A1 comme pour A2]. Il est donc manifeste que sentir par la vue n’est pas une seule chose car, même lorsque nous ne voyons pas, c’est par la vue que nous distinguons l’obscurité et la lumière, mais pas de la même façon. En outre, d’une certaine manière ce qui voit est aussi coloré, car chaque organe sensoriel est réceptacle du sensible sans la matière ; c’est pourquoi, même quand les sensibles ne sont plus là, des sensations et des impressions4 persistent dans les organes sensoriels. (425b17-25)

L’hypothèse A (on voit qu’on voit par la vue, et de même pour chacune des autres sensations) présente la difficulté que la vue de l’objet doit elle-même être un objet visible, c’est-à-dire un objet coloré, si du moins « sentir par la vue » n’a pas d’autre signification que « voir ».

4 Sur la distinction entre la rémanence des αἰσθήματα et celle des φαντάσματα, voir De Insomniis, 459b3-22. La rémanence se fait en général sous la forme de φαντάσματα, mais dans certains cas particuliers une sensation forte peut recouvrir temporairement les suivantes, par exemple la vision d’une couleur très lumineuse se surimprime pendant un certain temps sur les nouvelles visions.

DE ANIMA III 2

43

Viennent alors deux manières de répondre à la difficulté et donc de confirmer la possibilité d’une vue de la vue. La première manière (425b2122) consiste à accorder précisément deux significations à « sentir par la vue », à savoir, d’une part, voir au sens propre, c’est-à-dire voir des couleurs, et d’autre part voir de l’obscurité ou de la lumière (quand on est ébloui) sans rien voir d’autre ; dans ce cas, c’est par la vue que nous jugeons que nous ne voyons pas (et Aristote marque la différence dans les termes en utilisant ὁράω pour la vue au sens propre et τῇ ὄψει κρίνειν pour cette vue du fait de ne pas voir). Il y a ici une extension du domaine de la vue au-delà des couleurs, comme déjà en II 10, 422a20-22 : « La vue porte sur le visible et l’invisible, car l’obscurité est invisible et c’est la vue qui en juge, de même que sur le trop lumineux, car il est aussi invisible » ; mais on peut déjà préciser que cette extension de la vue n’est pas la conscience qui accompagne l’acte de voir, puisqu’elle a lieu seulement lorsqu’il n’y a pas d’acte de voir au sens propre. Cette extension de la signification de voir ne correspond donc pas à ce que serait « une vue de la vue ». La deuxième manière d’envisager une vue de la vue (425b22-25) consiste à reconnaître que, d’une certaine manière, ce qui voit est coloré, donc est objet de la vue ; et cette « certaine manière » est que l’organe sensoriel (τὸ αἰσθητήριον) « est réceptacle du sensible sans la matière », c’est-à-dire reçoit la forme de la couleur. Cette réception de la forme, qui constitue proprement la sensation, ne doit pas être confondue avec une affection matérielle telle que le reflet de l’objet dans l’œil, comme le précise De Sensu 2, 438a5-9. Par ailleurs, la forme ne s’identifie pas au réceptacle, celui-ci étant une grandeur corporelle tandis que « l’être du sensoriel » est une puissance et un logos (II 12, 424a17-b18). Il faut donc se demander si recevoir la forme de la couleur est vraiment suffisant pour être coloré et visible. 3. L’argument central contre une « vue de la vue » C’est le rôle du paragraphe suivant, qui formule pour la première fois la théorie selon laquelle l’acte et la forme sont une seule et même chose. En vertu de cette identité, on pourrait dire, conformément à (A1), que la deuxième vue voit la première vue parce que celle-ci se confond avec la couleur qu’elle saisit, ou, conformément à (A2), que la vue en tant qu’elle voit la couleur se voit elle-même. Or, la suite montre qu’il n’en est rien, et que ces deux conclusions résultent d’une mauvaise compréhension de

44

ANNICK STEVENS

la théorie de l’identité, qui est longuement exposée afin précisément d’éviter une telle erreur : Ἡ δὲ τοῦ αἰσθητοῦ ἐνέργεια καὶ τῆς αἰσθήσεως ἡ αὐτὴ μέν ἐστι καὶ μία, τὸ δ’ εἶναι οὐ τὸ αὐτὸ αὐταῖς· λέγω δ’ οἷον ὁ ψόφος ὁ κατ’ ἐνέργειαν καὶ ἡ ἀκοὴ ἡ κατ’ ἐνέργειαν· ἔστι γὰρ ἀκοὴν ἔχοντα μὴ ἀκούειν, καὶ τὸ ἔχον ψόφον οὐκ ἀεὶ ψοφεῖ, ὅταν δ’ ἐνεργῇ τὸ δυνάμενον ἀκούειν καὶ ψοφῇ τὸ δυνάμενον ψοφεῖν, τότε ἡ κατ’ ἐνέργειαν ἀκοὴ ἅμα γίνεται καὶ ὁ κατ’ ἐνέργειαν ψόφος, ὧν εἴπειεν ἄν τις τὸ μὲν εἶναι ἄκουσιν τὸ δὲ ψόφησιν. εἰ δή ἐστιν ἡ κίνησις καὶ ἡ ποίησις [καὶ τὸ πάθος] ἐν τῷ ποιουμένῳ, ἀνάγκη καὶ τὸν ψόφον καὶ τὴν ἀκοὴν τὴν κατ’ ἐνέργειαν ἐν τῷ κατὰ δύναμιν εἶναι· ἡ γὰρ τοῦ ποιητικοῦ καὶ κινητικοῦ ἐνέργεια ἐν τῷ πάσχοντι ἐγγίνεται· διὸ οὐκ ἀνάγκη τὸ κινοῦν κινεῖσθαι. ἡ μὲν οὖν τοῦ ψοφητικοῦ ἐνέργειά ἐστι ψόφος ἢ ψόφησις, ἡ δὲ τοῦ ἀκουστικοῦ ἀκοὴ ἢ ἄκουσις· διττὸν γὰρ ἡ ἀκοή, καὶ διττὸν ὁ ψόφος. ὁ δ’ αὐτὸς λόγος καὶ ἐπὶ τῶν ἄλλων αἰσθήσεων καὶ αἰσθητῶν. ὥσπερ γὰρ καὶ ἡ ποίησις καὶ ἡ πάθησις ἐν τῷ πάσχοντι ἀλλ’ οὐκ ἐν τῷ ποιοῦντι, οὕτω καὶ ἡ τοῦ αἰσθητοῦ ἐνέργεια καὶ ἡ τοῦ αἰσθητικοῦ ἐν τῷ αἰσθητικῷ. ἀλλ’ ἐπ’ ἐνίων μὲν ὠνόμασται, οἷον ἡ ψόφησις καὶ ἡ ἄκουσις, ἐπ’ ἐνίων δ’ ἀνώνυμον θάτερον· ὅρασις γὰρ λέγεται ἡ τῆς ὄψεως ἐνέργεια, ἡ δὲ τοῦ χρώματος ἀνώνυμος, καὶ γεῦσις ἡ τοῦ γευστικοῦ, ἡ δὲ τοῦ χυμοῦ ἀνώνυμος. ἐπεὶ δὲ μία μέν ἐστιν ἐνέργεια ἡ τοῦ αἰσθητοῦ καὶ τοῦ αἰσθητικοῦ, τὸ δ’ εἶναι ἕτερον, ἀνάγκη ἅμα φθείρεσθαι καὶ σώζεσθαι τὴν οὕτω λεγομένην ἀκοὴν καὶ ψόφον, καὶ χυμὸν δὴ καὶ γεῦσιν, καὶ τὰ ἄλλα ὁμοίως· τὰ δὲ κατὰ δύναμιν λεγόμενα οὐκ ἀνάγκη· ἀλλ’ οἱ πρότερον φυσιολόγοι τοῦτο οὐ καλῶς ἔλεγον, οὐθὲν οἰόμενοι οὔτε λευκὸν οὔτε μέλαν εἶναι ἄνευ ὄψεως, οὐδὲ χυμὸν ἄνευ γεύσεως. τῇ μὲν γὰρ ἔλεγον ὀρθῶς, τῇ δ’ οὐκ ὀρθῶς· διχῶς γὰρ λεγομένης τῆς αἰσθήσεως καὶ τοῦ αἰσθητοῦ, τῶν μὲν κατὰ δύναμιν τῶν δὲ κατ’ ἐνέργειαν, ἐπὶ τούτων μὲν συμβαίνει τὸ λεχθέν, ἐπὶ δὲ τῶν ἑτέρων οὐ συμβαίνει. ἀλλ’ ἐκεῖνοι ἁπλῶς ἔλεγον περὶ τῶν λεγομένων οὐχ ἁπλῶς. Par ailleurs, l’acte du sensible et celui du sens sont un seul et même, quoique leur être ne soit pas le même – je veux dire, par exemple, le son en acte et l’ouïe en acte, car il est possible, pour qui a l’ouïe, de ne pas entendre, et ce qui peut émettre un son ne l’émet pas toujours ; mais lorsqu’est en acte ce qui a la puissance d’entendre et qu’émet un son ce qui a la puissance de résonner, alors l’ouïe en acte se produit en même temps que le son en acte, ce qu’on pourrait appeler, respectivement, l’audition et la résonance. Si le mouvement et la production se trouvent dans ce qui subit5, il est nécessaire que le son et l’ouïe en acte se trouvent dans l’ouïe en puissance, car l’acte de ce qui peut produire et mouvoir se trouve dans ce qui est affecté ; c’est 5 La théorie générale de la localisation de l’action dans le patient se trouve en Phys. III 3, 202a13-b22, ainsi que l’unicité de l’acte de l’agent et du patient. Je suis pour cette phrase l’édition de Ross de 1956.

DE ANIMA III 2

45

pourquoi il n’est pas nécessaire que le moteur soit mû. Ainsi donc, l’acte de ce qui peut émettre un son est le son ou la résonance, et l’acte de ce qui peut entendre est l’ouïe ou l’audition6, car double est l’ouïe et double le son. Et le même raisonnement vaut pour les autres sens et sensibles. En effet, de même que la production et l’être-affecté se trouvent dans ce qui est affecté et non dans ce qui produit, de même l’acte du sensible et celui du sentant se trouvent dans le sentant. Dans certains cas, on dispose de noms pour les dire, comme la résonance et l’audition, dans d’autres l’un des deux reste sans nom : on appelle « vision » l’acte de la vue, mais celui de la couleur est sans nom, et on appelle « gustation » l’acte de ce qui peut goûter, mais celui de la saveur est sans nom. Et puisque l’acte du sensible et du sentant est le même, quoique leur être soit différent, nécessairement l’ouïe et le son en acte s’arrêtent et demeurent ensemble, de même que la saveur et la gustation, et de même pour les autres ; en revanche, pour ceux qui sont en puissance ce n’est pas nécessaire, mais les anciens physiologues se sont trompés à ce propos en pensant qu’il n’y avait ni blanc ni noir sans la vue, ni saveur sans la gustation. En un sens c’est correct, mais en un autre ce ne l’est pas, car, puisque le sens et le sensible se disent de deux façons, de l’une selon la puissance et de l’autre selon l’acte, pour ces dernières ce qu’ils disent est vrai, mais ce ne l’est pas pour les premières. Mais ils parlaient de façon univoque de ce qui ne se dit pas de façon univoque. (425b26-426a26)

Par cet exposé, Aristote refuse implicitement que la vue se saisisse ellemême en saisissant son objet, même si en acte elle est identique à la forme de l’objet. En effet, le visible, en tant que puissance, est la couleur portée par un corps, mais le vu en acte n’est plus dans ce corps, il est la forme qui se produit dans l’organe sensoriel correspondant, sous la stimulation du sensible extérieur. Ainsi, l’acte de voir se confond avec la forme vue mais pas avec le visible. Il n’est pas lui-même du visible et n’a pas la puissance de susciter un acte de voir. Il ne peut pas s’auto-susciter, à la différence de la pensée dont les objets sont intériorisés (selon la distinction établie en II 5, 417b20-26). De la même manière, une forme ne devient pas voyante du seul fait qu’elle est vue en acte (c’est le coup de force de Plotin de s’être autorisé d’Aristote pour donner aux intelligibles platoniciens l’acte d’intelligence). C’est pourquoi Aristote dit que leur être n’est pas le même : il reste une différence ontologique, définitionnelle, entre le même acte comme voyant et comme vu, parce qu’ils se fondent sur des potentiels différents. 6 Pour éviter que l’acte désigne à la fois la puissance et l’acte, Ross propose de lire implicitement : « le son ou la résonance, l’ouïe ou l’audition ».

46

ANNICK STEVENS

Le fait que le voyant soit ontologiquement autre que le visible, et donc non visible en tant que voyant, est confirmé dans les passages de la Métaphysique où Aristote réfute le relativisme attribué à Protagoras, selon lequel esse est percipi : D’une manière générale, s’il est vrai que seul est ce qui est sensible, rien ne pourrait exister sans les êtres animés, car il n’y aurait pas de sensation. Que donc n’existent ni les sensibles ni les sensations, c’est peut-être vrai (car c’est là une affection de l’être sentant), mais que n’existent pas les sujets7 qui produisent la sensation, y compris sans la sensation, c’est impossible. En effet, la sensation n’est certes pas sensation d’elle-même mais il y a quelque chose d’autre, en dehors de la sensation, qui est nécessairement antérieur à la sensation, car ce qui meut est par nature antérieur à ce qui est mû, et ne l’est pas moins même si elles sont dites l’une par rapport à l’autre. (Met. Γ 5, 1010b30-1011a2)

Certes, le sentant et le sensible sont des relatifs, mais même parmi les relatifs l’un peut être ontologiquement antérieur à l’autre. Cette particularité de certains relatifs est étudiée au chapitre 15 du livre Δ, où Aristote signale que, dans le cas de la sensation, de la pensée et de la connaissance en général, la dépendance ontologique n’est pas réciproque : Le mesurable, le connaissable et le pensable sont dits relatifs du fait qu’une autre chose est dite par rapport à eux ; en effet, le pensable signifie qu’il y en a une pensée mais la pensée n’est pas dite par rapport à ce dont elle est pensée (car on dirait deux fois la même chose), et de même la vue est vue de quelque chose, non de ce dont elle est la vue (même s’il est vrai de le dire) mais par rapport à la couleur ou à autre chose du même genre – car de l’autre manière on dira deux fois la même chose, à savoir que la vue est vue de ce dont elle est la vue. (Met. Δ 15, 1021a29-b3)

Il est remarquable qu’au livre I de la Métaphysique (6, 1056b34), Aristote affirme que les relatifs tels que la science ne sont pas « par soi » des relatifs (καθ’ αὑτά), non qu’ils le soient par accident, mais parce qu’ils le sont indirectement, du fait que leurs corrélatifs sont définis par eux (le sensible est défini comme tel par la sensation, il est donc un relatif au sens strict) tandis que ces actes ne sont pas définis par leurs corrélatifs en tant que tels mais en tant qu’ils sont des qualités possédant leur être propre (la vue est définie par la couleur, c’est-à-dire par une certaine qualité des surfaces des corps se révélant à la lumière). 7 Les sujets (ὑποκείμενα) qui produisent la sensation ne sont pas les corps eux-mêmes mais les qualités sensibles de ces corps, selon le même usage qu’en De anima 425b14.

DE ANIMA III 2

47

On dirait actuellement que la couleur existe par soi en tant qu’elle est une certaine longueur d’onde, et qu’elle dépend de nos organes sensoriels seulement en tant qu’elle devient par eux une impression visuelle (tandis que d’autres longueurs d’ondes sont transformées en impressions sonores ou tactiles) ; elle est donc relative à nous en tant que visible, mais non en tant que longueur d’onde. C’est ce que Démocrite cherchait à exprimer par l’opposition entre les figures des atomes et la manière dont elles apparaissent à nos organes ; dans cette transformation seulement réside la « convention » (fr. B125 D.-K.). C’est pourquoi, je ne pense pas qu’il soit visé parmi les « anciens physiologues » à qui Aristote reproche de ne pas faire cette distinction, à moins que ce soit seulement parce qu’il ne l’a pas identifiée à une distinction entre puissance et acte8. 4. Confirmation de l’argument par la notion de proportion Pour en revenir à la suite du texte du De anima, le paragraphe suivant confirme l’immatérialité de l’acte-forme en rappelant la différence entre la sensation et certaines altérations que des qualités sensibles peuvent aussi provoquer, mais pas en tant qu’elles sont senties. La particularité de la sensation est qu’elle est une proportion (λόγος). Le raisonnement qui le montre commence par un syllogisme plutôt sophistique, avant d’être plus persuasif par le recours à l’expérience. Εἰ δ’ ἡ συμφωνία φωνή τις ἐστιν, ἡ δὲ φωνὴ καὶ ἡ ἀκοὴ ἔστιν ὡς ἕν ἐστι (καὶ ἔστιν ὡς οὐχ ἓν τὸ αὐτό), λόγος δ’ ἡ συμφωνία, ἀνάγκη καὶ τὴν ἀκοὴν λόγον τινὰ εἶναι. Puisque l’accord est un certain son9 ; puisque le son et l’ouïe sont d’une certaine manière une seule et même chose (et d’une certaine manière non)10 ; 8 Pas davantage en De Sensu 4, 442a29-b23, lorsqu’Aristote reproche à la théorie de Démocrite de réduire tous les sens au toucher, il ne suppose que cette théorie nie l’existence des qualités sensibles en puissance. 9 L’usage de φωνή au lieu de ψόφος ne semble pas constituer une réduction au cas particulier du son vocal (défini en II 8, 420b5-6 et 29-33 comme un son signifiant accompagné d’une représentation, que peuvent émettre plusieurs animaux par l’intermédiaire de l’air qu’ils respirent), car la conclusion concerne l’audition en général, donc elle doit être valide pour n’importe quel son. Il me semble qu’il y a plutôt ici une influence du Timée (67bc) où Platon explique l’ouïe et le son en général en utilisant le mot φωνή dans son sens général et non limité. Que l’on choisisse l’un ou l’autre, de toute façon cela ne change rien à la difficulté principale du raisonnement. 10 Il n’y a pas de raison de supprimer cette parenthèse, qui se trouve dans tous les manuscrits et distingue simplement l’acte et la puissance (même si c’est peut-être une glose).

48

ANNICK STEVENS

puisque l’accord est une proportion ; il est nécessaire que l’ouïe aussi soit une proportion. (DA III 2, 426a27-30)

Si l’on maintient, comme je l’ai fait, le texte unanime des manuscrits (ἡ συμφωνία φωνή τις ἐστιν, comme le font Barbotin et Tricot, suivant St Thomas), le syllogisme est faux car il faudrait que tous les sons soient des accords pour pouvoir conclure que toutes les auditions sont des proportions ; or la première prémisse dit seulement que certains sons sont des accords. Si l’on accepte la modification des manuscrits proposée par Trendelenburg et éditée par Ross (suivi par Bodéüs et Thillet), le syllogisme devient correct mais la première prémisse est fausse : « tout son est un accord ; tout son est identique à l’audition ; donc toute audition est un accord » ; il n’est pas vrai que tout son est un accord, car l’accord suppose toujours un mélange de plusieurs sons, or il y a des sons simples (ce sera confirmé par 426b3-6). Je pense qu’il ne s’agit pas d’un problème de corruption du texte mais d’un sophisme volontaire de la part d’Aristote, qui cherche à introduire la notion de proportion par son usage familier à propos d’un mélange. Or, s’il y a bien des mélanges proportionnés de sensations, ce n’est pas cet usage qui peut éclairer la question présente, où le λόγος définit toute sensation, même celle d’une donnée non composée, et ne désigne pas une certaine mesure entre les sensibles mais la sensation elle-même comme relation mesurée entre le sensible et la faculté sensorielle. C’est ce qu’on finit par comprendre grâce au recours à l’expérience : Καὶ διὰ τοῦτο καὶ φθείρει ἕκαστον ὑπερβάλλον, καὶ τὸ ὀξὺ καὶ τὸ βαρύ, τὴν ἀκοήν· ὁμοίως δὲ καὶ ἐν χυμοῖς τὴν γεῦσιν, καὶ ἐν χρώμασι τὴν ὄψιν τὸ σφόδρα λαμπρὸν ἢ ζοφερόν, καὶ ἐν ὀσφρήσει ἡ ἰσχυρὰ ὀσμή, καὶ γλυκεῖα καὶ πικρά, ὡς λόγου τινὸς ὄντος τῆς αἰσθήσεως. διὸ καὶ ἡδέα μέν, ὅταν εἰλικρινῆ καὶ ἄμικτα ὄντα ἄγηται εἰς τὸν λόγον, οἷον τὸ ὀξὺ ἢ γλυκὺ ἢ ἁλμυρόν, ἡδέα γὰρ τότε· ὅλως δὲ μᾶλλον τὸ μικτόν, συμφωνία, ἢ τὸ ὀξὺ ἢ βαρύ, ἁφῇ δὲ τὸ θερμαντὸν ἢ ψυκτόν· ἡ δ’ αἴσθησις ὁ λόγος· ὑπερβάλλοντα δὲ λυπεῖ ἢ φθείρει. ἑκάστη μὲν οὖν αἴσθησις τοῦ ὑποκειμένου αἰσθητοῦ ἐστίν, ὑπάρχουσα ἐν τῷ αἰσθητηρίῳ ᾗ αἰσθητήριον, καὶ κρίνει τὰς τοῦ ὑποκειμένου αἰσθητοῦ διαφοράς, οἷον λευκὸν μὲν καὶ μέλαν ὄψις, γλυκὺ δὲ καὶ πικρὸν γεῦσις· ὁμοίως δ’ ἔχει τοῦτο καὶ ἐπὶ τῶν ἄλλων. Pour cette raison aussi, un son trop fort, que ce soit dans l’aigu ou dans le grave, détruit l’ouïe, et de même, dans les saveurs, une saveur trop forte détruit le goût, et dans les couleurs, une couleur trop brillante ou obscure détruit la vue, ainsi que, dans l’olfaction l’odeur trop forte, qu’elle soit douce ou piquante, en raison du fait que la sensation est une certaine

DE ANIMA III 2

49

proportion. Et c’est pourquoi les sensibles sont agréables, lorsque, purs et sans mélange, ils sont menés jusqu’à la proportion, par exemple l’acide ou le doux ou l’amer, car c’est alors qu’ils sont agréables, tandis que, d’une manière générale, l’agréable est plutôt le mélangé et l’accord que l’aigu ou le grave (et pour le contact, ce qui peut être réchauffé ou refroidi). Mais c’est la sensation qui est la proportion, et des sensibles excessifs la rendent désagréable ou la détruisent. Ainsi donc, toute sensation porte sur le sensible qui en est la base, en se trouvant dans l’organe sensoriel en tant qu’organe sensoriel, et elle distingue les différences du sensible qui en est la base, par exemple le blanc et le noir pour la vue, le doux et l’acide pour le goût ; et il en va de même pour les autres. (426a30-b12)

L’information importante est que la sensation en tant que telle est une proportion, et non l’éventuel mélange de données sensibles. Cela signifie que, pour qu’il y ait sensation, il faut que le sensible soit proportionnel à la capacité réceptrice du sentant, c’est-à-dire qu’il n’excède pas le champ sensoriel propre à chaque espèce animale. Un son, par exemple, qu’il soit une note simple ou une combinaison de notes, ne sera pas entendu s’il est audelà de nos limites auditives (ultrason ou infrason). De même, dans notre passage, la couleur trop lumineuse ou trop obscure tend vers l’invisible et empêche qu’il y ait vision. Cette distinction entre ce qui est sensible parce que proportionné et ce qui n’est pas sensible parce que non proportionné à notre champ perceptif ne doit pas être confondue avec la distinction, à l’intérieur de ce champ, entre les données qui se rapprochent le plus de la moyenne et, de ce fait, sont les plus agréables et celles qui, se rapprochant des extrémités du champ, sont plus désagréables (ceci étant indépendant du fait qu’elles soient simples ou mélangées)11. C’est pourquoi, Ross a raison de considérer que la destruction évoquée en 426b7 est celle de la sensation et non celle de l’organe ou de la faculté sensorielle, comme c’est le cas dans d’autres passages. En effet, si les organes sensoriels peuvent être endommagés par une lumière, un son ou une chaleur, ce n’est pas en tant que ceux-ci sont sentis, mais en tant qu’ils peuvent affecter une matière : par exemple, une plante peut être chauffée, l’air peut devenir odorant, mais aucun des deux ne sent 11 Cette dernière distinction se trouve également en II 11, où la sensation est également dite une μεσότης parce qu’elle juge du degré de la qualité sensible entre les deux extrêmes (424a5-10). Dans le cas particulier du toucher, le fait que le corps ne sente pas le degré de chaleur qu’il possède lui-même (424a2-4) concerne plutôt la chair que le sens du toucher proprement dit, car aucune faculté sensorielle n’a par elle-même une qualité sensible (voir aussi la note 12).

50

ANNICK STEVENS

ces qualités car ils sont affectés matériellement et non formellement (II 12, 424a25-b18). De même, si une chaleur endommage notre chair, ce n’est pas en tant qu’elle est sentie, car elle l’endommagerait aussi bien sous anesthésie. Ross paraphrase dès lors φθείρει par «  make impossible », et à la ligne 426b7 il choisit la correction de λυπεῖ en λύει (« is dissolved »), s’appuyant sur l’expression parallèle de 424a30-31 : λύεται ὁ λόγος. On peut à la rigueur conserver la leçon des manuscrits en considérant que l’excès qui provoque un désagrément (λυπεῖ) concerne la proportion qui produit l’agréable (le désagréable étant un certain degré d’écart par rapport au milieu), tandis que l’excès qui provoque une destruction (φθείρει) concerne la proportion comme condition de toute sensation (la suppression de la sensation résultant d’un degré d’écart tel qu’on sort du champ perceptif)12. Par la distinction entre la sensation et une altération de la matière, l’argument répond à une objection en faveur d’une « vue de la vue » qui serait : puisqu’une altération est sensible, pourquoi une sensation ne serait-elle pas sensible ? Et la réponse principale à cette question est rappelée : c’est que la sensation consiste en la distinction d’une certaine information qui parvient à l’organe sensoriel à partir du sensible extérieur ; mais elle-même n’est une telle information pour aucun organe. 5. Conclusion sur la conscience de sentir Dès lors qu’aucune manière n’a été trouvée pour sauver une « vue de la vue », la voie est désormais ouverte pour une autre expression de la conscience de sentir. Même si la proposition d’Aristote ne sera formulée tout à fait explicitement que dans les Parva Naturalia (en particulier dans le fameux passage du De Somno 2, 455a16-26), on en trouve certaines indications dans le paragraphe suivant, qui semble se tourner sans transition vers la solution d’un autre problème, celui de la distinction entre plusieurs sensibles de genre différent. Ἐπεὶ δὲ καὶ τὸ λευκὸν καὶ τὸ γλυκὺ καὶ ἕκαστον τῶν αἰσθητῶν πρὸς ἕκαστον κρίνομεν, τινὶ καὶ αἰσθανόμεθα ὅτι διαφέρει. ἀνάγκη δὴ αἰσθήσει· αἰσθητὰ γάρ ἐστιν. ᾗ καὶ δῆλον ὅτι ἡ σὰρξ οὐκ ἔστι τὸ 12 La distinction entre l’agrément ou le désagrément d’une sensation et la nocivité de la qualité sensible correspondante se trouve également en De Sensu 5, 444b28-445a4.

DE ANIMA III 2

51

ἔσχατον αἰσθητήριον· ἀνάγκη γὰρ ἂν ἦν ἁπτόμενον αὐτὸ κρίνειν τὸ κρῖνον. οὔτε δὴ κεχωρισμένοις ἐνδέχεται κρίνειν ὅτι ἕτερον τὸ γλυκὺ τοῦ λευκοῦ, ἀλλὰ δεῖ ἑνί τινι ἄμφω δῆλα εἶναι – οὕτω μὲν γὰρ κἂν εἰ τοῦ μὲν ἐγὼ τοῦ δὲ σὺ αἴσθοιο, δῆλον ἂν εἴη ὅτι ἕτερα ἀλλήλων, δεῖ δὲ τὸ ἓν λέγειν ὅτι ἕτερον· ἕτερον γὰρ τὸ γλυκὺ τοῦ λευκοῦ· λέγει ἄρα τὸ αὐτό·ὥστε ὡς λέγει, οὕτω καὶ νοεῖ καὶ αἰσθάνεται – ὅτι μὲν οὖν οὐχ οἷόν τε κεχωρισμένοις κρίνειν τὰ κεχωρισμένα, δῆλον· Par ailleurs, puisque nous distinguons entre eux le blanc, le doux et chacun des sensibles, par quoi sentons-nous qu’ils diffèrent ? Ce doit être par un sens, puisqu’ils sont des sensibles. (Par là il est clair aussi que la chair n’est pas l’organe sensoriel final, car ce qui distingue devrait nécessairement distinguer en étant lui-même en contact13). Il n’est donc pas possible que ce soit par des instances séparées14 que l’on distingue que le doux est autre chose que le blanc, mais il faut que les deux soient clairs pour une seule, sinon, même si moi je sentais l’un et toi l’autre, il serait clair qu’ils diffèrent entre eux, mais il faut qu’une seule instance dise qu’ils diffèrent ; car le doux est différent du blanc, donc la même instance l’affirme et, comme elle l’affirme, elle le pense et le sent. Que donc il n’est pas possible de distinguer des sensibles séparés par des instances séparées, c’est clair. (426b12-23)

La distinction entre des sensibles de genres différents est attribuée à une instance unique qui réunit et distingue les différents sens, mais cette solution n’est pas explicitement étendue à la question de la conscience, alors que tous les éléments sont désormais réunis pour le faire. Un dernier indice se trouve dans le νοεῖ de la ligne 426b22. La mention d’une pensée semble étrange puisqu’il est établi dès le départ que, s’agissant des sensibles, seule peut intervenir la faculté de sensation (et ce n’est pas une erreur de manuscrit car on trouve également νόησις jointe à αἴσθησις en 427a1 et a9). St Thomas suppose que le verbe νοεῖν n’est pas encore explicitement spécialisé et inclut la sensation ; Ross traduit par « know  » mais ne l’explique pas ; Bodéüs l’explique 13 Aristote a déjà montré au chap. II 11 que la chair est un milieu intermédiaire. On en trouve ici une confirmation par le fait que, comme la chair est en contact avec le corps sensible, si le toucher se faisait par elle, alors la discrimination entre le toucher et un autre sens devrait se faire aussi par contact, or ce n’est pas ainsi que les autres sensibles sont perçus. L’expression τὸ ἔσχατον αἰσθητήριον signifie l’organe final du toucher, et non l’organe centralisant toutes les sensations, puisqu’elle n’est qu’un rappel du πρῶτον αἰσθητήριον dont Aristote disait qu’il devait se trouver à l’intérieur de la chair (II 11, 422b22 et 423b30-31). 14 Le choix du terme « instance » pour compléter en français les adjectifs au neutre pluriel est commandé par la fonction de κρίνειν qui est centrale dans ce passage.

52

ANNICK STEVENS

comme une sorte d’équivalent de κρίνει  ; Thillet estime que la pensée est nécessaire dès lors qu’il y a langage (dire). Je propose d’y voir plutôt une signification ancienne du verbe νοεῖν, la signification large de « avoir à l’esprit » (que l’on trouve d’Homère à Xénophon), qui peut évoquer ce qu’Aristote cherche à exprimer depuis le début du chapitre. Ceci indiquerait qu’il a déjà identifié l’instance sensorielle commune comme responsable de la conscience de sensation, même s’il n’y fait qu’une allusion très vague. La fin du chapitre (426b23-427a16) apporte une précision sur l’indivisibilité de la sensation dans le temps, question complexe et controversée que je me suis permis de laisser de côté, car elle demanderait beaucoup de développements sans apporter davantage d’information à la question principale du chapitre.

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3 Dorothea FREDE

Introductory remarks Among the many difficult chapters of De Anima, Chapter 3 presents special challenges to the editor’s and translator’s skills, as well as to the reader’s ingenuity. For the chapter’s organization is difficult to penetrate and its train of thought is hard to follow. Even after a careful study of the text it is not easy to give a succinct summary of the chapter’s content, its progress, and its results. It contains some repetitions and also some contradictory statements. The chapter therefore looks more like a draft, awaiting further organization and clarification, than like a well-worked-out essay on the conception of φαντασία. And that is not the fault of the text’s editor in the Renaissance who made the division into chapters. It would indeed be difficult to make a better division of the text. For even the fact that φαντασία is going to be the chapter’s central topic is not recognizable at once. Instead, the chapter starts with a critique of the alleged failure of the philosophers of old to distinguish between sense-perception and thought (427a17-b16). It is only at the tail-end of the critique of his predecessor’s mistaken views that Aristotle first mentions φαντασία. Even from a grammatical point of view φαντασία’s entrance is strange, for it appears, so to speak, in mid-sentence. It is appealed to as an extra indication that sense-perception and thought cannot be one and the same thing: Φαντασία γὰρ ἕτερον καὶ αἰσθήσεως καὶ διανοίας, αὕτη τε οὐ γίγνεται ἄνευ αἰσθήσεως, καὶ ἄνευ ταύτης οὐκ ἔστιν ὑπόληψις. For φαντασία is different from both perception and thought; and it does not come about without perception and without it there is no supposition. (427b14-16)1 1

The translation follows with modification, that of C. Shields (2016).

54

DOROTHEA FREDE

Of the many uses of γάρ that puzzle Aristotle’s readers, this is one of most puzzling. How could the claim that φαντασία is different from both perception and thought show that perception and thought are different things? Prima facie it sounds like the negation of the Euclidean axiom concerning equals: “Things which are equal to the same thing are equal to one another.” But the opposite claim that things which are unequal to the same thing are unequal to one another, would be nonsensical, and cannot be what Aristotle wants to say. He, rather, seems to be concerned to show that φαντασία is a factor that both separates but also somehow links together perception and thought. Nevertheless, this appeal to φαντασία as a criterion comes as a surprise, because that capacity has not yet been properly introduced and defined. To be sure, φαντασία has been mentioned several times before as a concept the reader is expected to be familiar with, but its nature has not been explained2. The oblique introduction of φαντασία is not the only problem concerning that concept. There is also the problem of its translation. That problem is due to the fact that Aristotle in De Anima III 3 applies it to two different kinds of “appearances”, without giving more than a hint that there is a shift in his usage. For φαντασία is discussed in two different parts, parts that are of quite unequal length and importance with respect to Aristotle’s own conception of φαντασία. In the first, quite short part, φαντασία is used in the “active” sense of imagination (427b14-26). The second, long part, leads to a quite different conception of φαντασία: it turns out to be the after-effect of sense-perception. Already the way that Aristotle proceeds to establish the latter type shows that this sense of φαντασία represents a novelty (427b26-429a9). For Aristotle reaches its result in two stages. In the first stage Aristotle explains at length what φαντασία is not – namely that it is neither senseperception nor thought of any kind nor a combination of the two (427b27-428b9). It is only in the discussion’s last stage that Aristotle finally informs his readers of the nature he attributes to φαντασίαι in the intended sense: that they are the after-effects of sense-perceptions that outlast their causes (428b10-429a9). 2 That φαντασία has an important cognitive function is indicated right at the beginning (I 1, 402b22-403a2; 403a8-10). II 3, 414b16 sq. indicates the need for further elucidation: “But regarding φαντασία things are not clear. One must inquire into that later.”

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

55

But there is a further problem concerning the meaning of φαντασία that is unique to our chapter and that makes it inappropriate to use a unitary translation of φαντασία throughout this chapter. For it starts out with an active sense of φαντασία as “imagination”, but then switches to the passive sense that applies to the after-images of sense-perception. Because the active sense concurs with our present-day use of creative imagination, while the passive sense does not, it is misleading to use “imagination” as the uniform translation. Some translators therefore treat φαντασία as untranslatable and stick to the transliteration; other translators use “imagination” throughout3. But it should become clear that while “imagination”, in the sense that is common in ordinary language nowadays, fits the active use of φαντασία, it does not really fit the sense intended in the chapter’s second part4. For the capacity to receive, to retain, and to reproduce sensory impressions is not what we mean by “imagination”, nor do we apply to it the verb “to imagine”. Readers expecting an elucidation of creative imagination are therefore quite surprised by the discussion in De Anima III 3. An explanation of why Aristotle regards his own conception of φαντασία as “sensory impression” as the more important psychic phenomenon is the aim of this article. 1. Φαντασία as imagination (427b14-26) As indicated before, φαντασία is introduced as a topic in a somewhat oblique fashion, namely as a supplementary argument to show that the philosophers of old were wrong if they identified sense-perception and thought. The chapter’s first part points out why this identification is mistaken, quite generally, without resorting to φαντασία (427a17-b14). Whether Aristotle’s imputation of that confusion to his predecessors is justified need not concern us here. What concerns us, however, is the way φαντασία is used as a proof that perception and thought must be different because φαντασία is not identical with either of them. The argument does, of course, not rest on the inversion of Euclid’s axiom, as 3

On this issue cf. T. Johansen (2012), p. 199 sq. On “imagination” in contemporary philosophy cf. P.F. Strawson (1970), p. 31: “The uses, and applications, of the terms “image”, “imagine”, “imagination”, and so forth make up a very diverse and scattered family. Even this image of a family seems too definite. It would be a matter of more than difficulty to identify and list the family’s members, let alone their relations of parenthood and cousinhood.” 4

56

DOROTHEA FREDE

facetiously suggested earlier, but on the further claim that (i) φαντασία does not come about without sense-perception, and (ii) that supposition does not come about without φαντασία, but that φαντασία is not the same as either of them. Aristotle clearly wants to argue, then, that although φαντασία is necessarily connected with both sense-perception and thought, the fact that it differs from both is a further proof that the two are different from each other. The use of “supposition” (ὑπόληψις) here, a term not used in De Anima before and largely confined to this chapter, seems to be due to Aristotle’s wish to use a neutral term designating rational thought of any kind. The argument itself makes quite clear that φαντασία is here to be understood in the sense of what we would call “free ranging imagination”: Ὅτι δ’ οὐκ ἔστιν ἡ αὐτὴ φαντασία5 καὶ ὑπόληψις, φανερόν. τοῦτο μὲν γὰρ τὸ πάθος ἐφ’ ἡμῖν ἐστιν, ὅταν βουλώμεθα (πρὸ ὀμμάτων γὰρ ἔστι τι ποιήσασθαι, ὥσπερ οἱ ἐν τοῖς μνημονικοῖς τιθέμενοι καὶ εἰδωλοποιοῦντες), δοξάζειν δ’ οὐκ ἐφ’ ἡμῖν· ἀνάγκη γὰρ ἢ ψεύδεσθαι ἢ ἀληθεύειν. ἔτι δὲ ὅταν μὲν δοξάσωμεν δεινόν τι ἢ φοβερόν, εὐθὺς συμπάσχομεν, ὁμοίως δὲ κἂν θαρραλέον· κατὰ δὲ τὴν φαντασίαν ὡσαύτως ἔχομεν ὥσπερ ἂν εἰ θεώμενοι ἐν γραφῇ τὰ δεινὰ ἢ θαρραλέα. It is evident that φαντασία and supposition are not the same. For this kind of state is up to us whenever we wish. For it is possible to produce something before one’s eyes, as those do who set out something in mnemonic systems and form images of them, whereas believing is not up to us, since it is necessary that it be true or false. Moreover, when we come to believe something is terrible or frightful, we correspondingly get affected right away, and similarly with something encouraging. But in the case of φαντασία we are just as if we had seen terrible or encouraging things in a picture. (427b17-24)

Φαντασία is here clearly depicted as the active capacity to create images at will. So it is what is called “imagination” in ordinary parlance. For the ability to “set up images before our eyes” is what we mean when we say that people either have or lack imagination when we refer to the senses. And because such imaginings are up to us, they are not connected with any suppositions. When we imagine a lion jumping at us, we do not assume that there actually is a lion there – and that is why we are not frightened by

5 Following C. Shields (2016): « reading φαντασία with C2U2S1 and secluding νόησις with Madvig » (p. 77, n. 44).

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

57

the image we have created6. We are, as Aristotle says, in the same frame of mind as we would be if we looked at the lion in a picture. We might regard the lion in that picture as fierce or friendly, as the case would be. But we would not find the picture or the imagination to be a reason for either fear or confidence, as we would do if we believed that there was such a lion facing us. Thus, φαντασία here designates the active capacity to create such sights in our mind at will. This fact is clearly indicated by the vocabulary: such states are “up to us”, for we put them in front of our eyes, on account of our ability of “image-making” (εἰδολοποιεῖν). Εἰδολοποιεῖν is a hapax legomenon in Aristotle. Its use is probably an allusion to Plato who first introduces the term in his critique of poetry in Rep. X 606c and later employs it again in order to “catch the sophist” in the Sophist (235b-239e; 260d et pass.): the sophist turns out to be the quintessential image-maker by words (εἰδωλοποιός). Aristotle, instead, uses the expression in the sense of intentional image-making in one’s own soul, without the illusion that the images are real. The active aspect of that capacity is further confirmed by a reference to mnemotechnic. For mnemotechnic works with self-created images: in order to memorize their speeches speakers envisage, for instance, a house with several rooms and assign the different parts of their speech to the different rooms. When delivering a speech, the speaker imagines moving through the house’s different rooms, so that he can present the different parts of his speech in the right order7. Because these images are our own creations, we do not get frightened when we imagine something frightening. For such images are fictions, even though their content may consist of sensory images that are drawn from past experiences. But as figments of our own making they are neither true nor false. Beliefs, by contrast, assume that their subject-matter is the case, they are true or false and for that reason they are not “up to us”; for we cannot at will believe that something is or is not the case. This is all Aristotle has to say about “imagination” in the sense that we usually connect with that term8. For he concludes its discussion with 6 My make-believe image of a lion that is jumping at me is neither true nor false, pace R. Polansky (2012), p. 414, n. 21. 7 On “mnemotechnic” cf. Top. VIII 14, 164b29 sq. 8 The only place where Aristotle uses that concept is at Poet. 17, 1455a22-34; he recommends to the poet to put the actual scenes as far as possible before his eyes (ὅτι μάλιστα πρὸ ὀμμάτων τιθέμενον) in order to avoid inconsistencies and to get the emotions right.

58

DOROTHEA FREDE

a short statement concerning suppositions: there are different kinds, they comprise knowledge (ἐπιστήμη), belief (δόξα), practical reason (φρόνησις) and their opposites, but their difference should be discussed elsewhere (427b24-26). This statement is somewhat disconcerting, because apart from its redundancy (it repeats, in essence, what has already been said at 427b8-11) it suggests that no more need be said about the different types of “suppositions” and their difference from φαντασία, so that the topic seems closed. But in fact, knowledge, reason, and δόξα are taken up again in the second part of the discussion, where Aristotle points out in quite an elaborate way that they are not identical with φαντασία in the sense of “sensory impressions”. The concluding statement at this point may, therefore, be a reader’s gloss asserting that ὑπόληψις covers all intellectual types of understanding, because that term has not been used before in De Anima. However that may be, the next section ignores the active sense of φαντασία and its concluding remarks and starts the discussion of the difference between thinking, perceiving, and φαντασία all over again. 2. What φαντασία is not (428a1-b9) Εἰ δή ἐστιν ἡ φαντασία καθ’ ἣν λέγομεν φάντασμά τι ἡμῖν γίγνεσθαι καὶ μὴ εἴ τι κατὰ μεταφορὰν λέγομεν, μία τις ἔστι τούτων δύναμις ἢ ἕξις καθ’ ἅς κρίνομεν καὶ ἀληθεύομεν ἢ ψευδόμεθα; τοιαῦται δ’ εἰσὶν αἴσθησις, δόξα, ἐπιστήμη, νοῦς. So, if φαντασία is that in virtue of which we say that a particular image comes to us and we do not speak about it as if in a metaphorical sense, will it be one of these: a capacity or disposition in virtue of which we discriminate and issue truths or falsehoods? But such are perception, belief, knowledge, and reason. (428a1-5)

That this is a fresh start is indicated by the fact that in this case the images in question “come to us” and are not “made by us”. What Aristotle is now concerned with is a “passive” type of sensory impressions. That he regards this as a major distinction concerning φαντασία is signaled by the fact that he states that he is now not speaking of φαντασία in a metaphorical sense. The metaphorical sense must, I submit, refer to the active sense of “imagination” discussed in the previous section. For this “active sense” of creating images at will is not picked up again in what follows – nor is it referred to again later on in De Anima or in the

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

59

discussion of memory and images in dreams in the Parva naturalia or anywhere else in Aristotle. The claim that “imagination” in the discussion’s first part is thereby downgraded as φαντασία in a metaphorical sense only has not found many adherents9. It is certainly an unusual procedure for Aristotle to first introduce an explanation of a phenomenon as if it represented the standard case, only to put it aside with the explanation that it is such a case only in a metaphorical sense. But there are two points that deserve attention concerning this question. (i) Aristotle often uses the term “metaphor” in cases where there is quite a close relation between the two items. Thus he calls the types of courage that are not concerned with facing danger in battle “courage κατὰ μεταφοράν”, i.e. as courage in an “extended sense” (EN III 6, 1115a14-24). Similarly, the types of ἀκρασία that do not concern pleasures of the body but love of honor, family or victory, are called ἀκρασία κατὰ μεταφοράν (EN VII 5, 1149a23). A “metaphor” for Aristotle does, then, not necessarily refer to a figurative way of speaking but to an extended sense. (ii) Commentators tend to overlook the importance of the difference between the active and the passive sense of φαντασία: the one type we make, the other type “happens” to us; the one type does not discriminate objects and issue truths or falsehoods, while the other type does both10. But if “creative imagination” is φαντασία in an extended sense only, why does Aristotle give it first place in his discussion? If this was not a sheer oversight, a kind of blunder that he forgot to delete later, he must have decided to start out his elucidation of φαντασία with a type of “appearance” that people will be most familiar with. For, his own conception of φαντασία as “sensory impression” as a kind of after-effect is not a familiar one, but requires careful introduction11. Thus he can use 9 Cf. D. Frede (1992), p. 280, n. 3. Most commentators just ignore the active aspect of εἰδολοποιεῖν, as e.g. M. Schofield (1992); T. Johansen (2012), p. 208; K. Corcilius (2013), p. 75; J. Dow (2014), p. 206 sq. C. Rapp (2001), p. 92, n. 90 suspects that that metaphorical use refers to “homonymous uses of φαίνεται” that are typically not covered by φαντασία (similarly C. Shields (2016), p. 281). D.W. Hamlyn (1993), p. 131 simply has it the other way round: he translates φαντασία by imagination and claims that “sensory impression”, as defined afterwards, is φαντασία in a metaphorical sense. 10 Both types are concerned with sensory images, albeit in quite a different way. “Imagetheory” used to be the butt of heavy criticism by contemporary philosophers of mind; but the opposition to images of sense-perception seems to have died down in recent years. 11 C. Shields (2016), p. 275 explains the difference in approach to φαντασία as due to the specialty of the subject-matter: Aristotle is interested in isolating it by differentiating

60

DOROTHEA FREDE

the “imaginative” type of φαντασία to show, in a kind of preparatory way, that φαντασία need be neither identical with sense-perception nor with “supposition”. For although such imaginary objects are related to sense-perception, they are not sense-perceptions12; and they are also not suppositions either, because we do not assume that the scenario in our imagination represents matters of fact. If Aristotle afterwards dismisses this type as a kind of φαντασία in an extended sense, it is because it differs from his own conception of φαντασία that is concerned with appearances that are not created at will, that are either true or false and that are, in a way yet to be explained, related to suppositions. Hence the circuitous way that his further procedure first explains what φαντασία is not. It consists of a review of the different capacities and dispositions that make us get things either right or wrong in order to show why these capacities and dispositions are not identical with φαντασία in the sense he is going to explain in the chapter’s last part. For they either lack a common characteristic or are not coextensive13. (1) φαντασία vs. αἴσθησις (428a5-13): That φαντασία and senseperception are not the same is justified by several arguments that are formulated in quite a compressed way and that seem at times hard to reconcile with each other. (i) Perception is either a potentiality or an activity – like the potentiality to see (ὄψις) or the actuality of seeing (ὅρασις). This is not so with visual φαντασίαι in dreams, for in dreams there is neither the potentiality for sight, nor the actuality. (ii) Perception is always there, φαντασία is not. This claim must exclude sleep – and it presupposes that while we can always make use of sense-perception when we are awake, given the fulfillment of certain conditions, “images” do not always “come to us”. (iii) If perception and φαντασία were the same, all animals would have both. But some animals like ants, bees, and grubs have it from other more robust faculties because he does not presuppose that his readers know what phenomenon he is talking about: “On the contrary, he thinks he owes his readers some proof of our even having such a capacity or faculty, and he thinks that such a proof intimately involves distinguishing imagination from some other capacities to which one might be inclined to reduce it.” 12 In order to actively phantasize we for the most part use sensory impressions we are familiar with. Artists sometimes have to strain their imagination when it comes to envisage objects of a kind they have never seen, such as lions or elephants in medieval art. 13 For a detailed critical evaluation of these arguments see C. Shields (2016), p. 282-289.

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

61

perception but not images14; a claim that is going to be modified later in the chapter (428a21 sq.) and this modification is later on explained: even the lowliest animals are at least capable of indeterminate φαντασίαι, i.e. of indeterminate sensory after-impressions (11, 434a4 sq.). (iv) Perceptions are always true, φαντασίαι are for the most part false. This claim is also subsequently modified in the chapter’s last part (428b18-25): only sense-perception of their specific objects (like color for sight) are usually true, in all other cases there is the possibility of falsehood. (v) We speak of φαντασίαι when we do not have clear perceptions and are not sure of the identity of the object. (vi) It is possible to have “visions” (ὁράματα) with our eyes closed. This example is the counterpart of visions in dreams. All these discrepancies are supposed to show that though there are close relations between perceptions and φαντασίαι, they are not identical. The arguments also provide some information of what Aristotle means by φαντασία in the intended sense: they are sensory impressions that are closely related to perceptions but that can be detached from actual perception, as in the case of visions in dreams or when our eyes are closed. And these impressions can be unclear and “false” if they do not represent anything real, in contradistinction to clear perceptions that – at least at this point − Aristotle treats as always true. When Aristotle refers to unclear perceptions as cases of φαντασία – as in the case of the unclear perception we take to be a man (428a13 sq.) – one may wonder whether he is, somewhat inadvertently, falling back here on the general use of “seeming” (φαίνεται). Although this assumption cannot be excluded, Aristotle may rather anticipate his own theory of φαντασία and assume that unclear sense-perceptions evoke φαντάσματα that are stored in our memory, so that an unclearly perceived object invokes the association with the image of a man. For if the impression concerns an object that we have never seen before, no such association would occur to us. An analogous explanation may explain the (incorrigible) impression of the size of the sun – i.e. that it is a foot wide, referred to below (428b3 sq.). Aristotle may not just assume that the sun appears a foot wide when we look at it but that we have an impression of such a size that we associate with the sun’s when we see it. Whether this is the correct 14 Given that Aristotle regards bees and ants as intelligent animals elsewhere (HA 488a7-10 et al.), A. Torstrik has suggested an emendation of the text to “φαντασία belongs to the bees… but not to the grub”.

62

DOROTHEA FREDE

explanation of the nature of unclear or misleading sense-impressions must remain speculative, but the possibility is worth considering. In all of these cases it is misleading to translate φαντασία by “imagination”. For though they all concern sensory images that may or may not be real or unclear, it is wrong to call the capacity to receive such images “imagination”15. For “imagination” in modern languages means the capacity to envisage objects or states of affairs at will. Visions in dreams are images, but not the objects of imagination of that kind, nor are visions that we see when our eyes are closed; and it is highly controversial whether animals should be attributed imagination in the creative sense at all. (2) φαντασία vs. knowledge and reason (428a16-18): The relation of φαντασία to knowledge (ἐπιστήμη) and reason (νοῦς) is given short shrift: they cannot be the same, because reason and knowledge are always true while φαντασίαι can be false. This criterion excludes their identity, but it does not rule out some connection between them, as will emerge in subsequent chapters. (3) φαντασία vs. belief (428a18-24): That φαντασία is the same as belief is ruled out by a somewhat unexpected move (428a5-16): belief supposedly involves conviction (πίστις), a thing that animals don’t have, although many of them have φαντασίαι. Now, the question is whether πίστις adds anything to δόξα that δόξα does not have by itself. So why not just deny that animals have δόξα, given that they are not capable of reasoning16? Aristotle seems rather concerned to emphasize a commitment to the facts that the belief is about, a commitment that weak beliefs may not convey17. His appeal to πίστις may therefore be due to two reasons: (a) He can avoid, at this point, an explanation of why animals do not have beliefs – as the interpolated text affirms. For that would make superfluous any further discussion of whether φαντασία is a combination 15 M. Schofield (1992), p. 251 et al. It is significant to note that no German translation of the De Anima uses either “Einbildungskraft” or “Phantasie” as translations of φαντασία. If they don’t leave the term untranslated, they use “Vorstellung” – a term wide and vague enough to accommodate φαντασία both in the active and in the passive sense. 16 Aristotle sometimes speaks as if certain higher animals are capable of reasoning of some kind (cf. EN VI 7, 1141a26-28), but this must be taken with a grain of salt. 17 That belief contains commitment to truth and fact is also mentioned in the distinction between belief and προαίρεσις in EN III 2, 1112a5-7; it is repeated in the distinction between εὐβουλία and δόξα in VI 9, 1142b11 sq.

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

63

of belief and perception. (b) Conviction, πίστις, or rather πιστεύειν, is going to play an important role in the case of incorrigible senseimpressions like that of the apparent size of the sun. (4) φαντασία vs. a combination of belief and perception (428a24-b9): It is puzzling that Aristotle does not regard the reasons given against the identification of φαντασία with perception and belief as a sufficient reason to rule out that it is a combination of both of them. For if φαντασία is not the same as either of them it should a fortiori not be a combination of those two capacities. Some interpreters think, therefore, that Aristotle’s real concern is a critique of Plato’s definition of φαντασία as a combination of sense-perception and belief in the Sophist. But it is unlikely that Aristotle means to target Plato here18. He has reasons of his own to discuss the combination, as indicated by his insistence that the perception and the belief in question must be concerned with the same object. His concern here is not so much with the difference between φαντασία and senseperception-cum-belief as with the possibility of conflicting impressions and beliefs concerning one and the same object at the same time. That conflict is illustrated by the famous example of the apparent size of the sun: the sun appears to be just a foot wide, but we are convinced (πιστεύομεν) that it is much larger than the earth. The example explains, at the same time, why πίστις has been added to δόξα: our conviction is unshakeable even by direct sensory evidence to the contrary. For as Aristotle adds, if φαντασία and belief refer to the same thing, and the phenomenon has not changed, then we must (a) either give up or forget about the belief, or (b) the same belief will be both true and false. This dilemma shows, therefore, that φαντασία is neither a combination of these capacities or dispositions, nor is it a product of them. We shall not take up here the discussion of what is problematic in Aristotle’s specification of the apparent size of the sun, most of all why he claims that it appears to be a foot wide19. It may well have been part 18 Πίστις is not mentioned in the Sophist’s discussion of φαντασία, and nothing further is said there about φαντασία as a combination of sense-perception and belief. As K. Corcilius (2013), p. 75 notes, Plato represents φαντασία not as a mental capacity but as an occurrence in the mind. 19 The real size of the sun, the moon, and the stars must have been a matter of controversy among scientists. Aristotle repeats the claim that the sun looks a foot wide twice in Insomn. 458b29; 460b18.

64

DOROTHEA FREDE

of the arsenal of skeptical arguments, a particularly intriguing one, because unlike other cases of trompe l’œil, it is incorrigible, for there is no way to get a different view of the sun, as there is in the case of the stick that looks bent when inserted in water or the octagonal tower that looks round at a distance. 3. What φαντασία is (428b10-429a9) The explanation of the nature of φαντασία is introduced in a rather unobtrusive way, without mention we are finally presented with the solution to the problem: Ἀλλ’ ἐπειδὴ ἔστι κινηθέντος τουδὶ κινεῖσθαι ἕτερον ὑπὸ τούτου, ἡ δὲ φαντασία κίνησίς τις δοκεῖ εἶναι καὶ οὐκ ἄνευ αἰσθήσεως γίνεσθαι ἀλλ’ αἰσθανομένοις καὶ ὧν αἴσθησις ἔστιν, ἔστι δὲ γίνεσθαι κίνησιν ὑπὸ τῆς ἐνεργείας τῆς αἰσθήσεως, καὶ ταύτην ὁμοίαν ἀνάγκη εἶναι τῇ αἰσθήσει, εἴη ἂν αὕτη ἡ κίνησις οὔτε ἄνευ αἰσθήσεως ἐνδεχομένη οὔτε μὴ αἰσθανομένοις ὑπάρχειν, καὶ πολλὰ κατ’ αὐτὴν καὶ ποιεῖν καὶ πάσχειν τὸ ἔχον, καὶ εἶναι καὶ ἀληθῆ καὶ ψευδῆ. Since it is possible when something is set in motion for something else to be moved by it, and since φαντασία seems to be a sort of motion and not to occur without perception, but to occur in things which are perceiving and to be of those things of which there is perception; and since it is possible for this motion to occur as the result of the activity of perception and must be like that perception, therefore this motion would not be possible without perception nor belong to things that do not perceive; and it is possible for what has φαντασία both to affect and to be affected by many things in accordance with it, and to be either true or false. (428b10-17)

As has been remarked by others, this oblique introduction of φαντασία does not present us with a regular definition of the Aristotelian type. Nor does Aristotle return to the question raised in 428a3 sq. whether φαντασία is a capacity (δύναμις) or disposition (ἕξις)20. But at least he 20 In Insomn. 1, 458b30-32; 459a15-22 Aristotle speaks of a “phantastic” capacity – that is the same as that of perception but not the same “in being” – and repeats the definition that φαντασία is a motion in the soul caused by sense-perception. So he treats the φανταστική there as a capacity. Despite its somewhat anomalous state there is indeed no reason to deny that there is a δύναμις to experience φαντασία (pace T. Johansen (2012), p. 204 sq.; p. 220). For there are passive as well as active δυνάμεις (cf. Met. Θ 1, 1046a11-29). The difference is that perceptions are caused by the objects outside, while φαντασίαι are caused by perception. But what certain species by nature do not have – or have only in an indeterminate way − must certainly be a capacity.

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

65

does tell us the genus which φαντασία belongs to: it is a κίνησις, a movement. And in a round-about way he also informs us about what is specific to that movement: it is caused by sense-perception and it is like that sense-perception. Thus it turns out that φαντασία, although it is not sense-perception, is directly related to sense-perception. Φαντασία, then, neither seems to have a genus of its own, nor is it a species of senseperception, it is a secondary, dependent, kind of passive motion. This is not the occasion to enter into the controversial question whether perception is a regular kind of κίνησις or not. That Aristotle assumes so in book III is undeniable, for that is the gist of his general comments on perception at 2, 425b26-426a19 where he states, without reservations, that perceptions are passive motions in the recipient. The sound that strikes the ear, the ψόφος, is the active motion, while the perception itself is the passive motion of being struck, the ψόφησις. Aristotle especially coins the expression πάθησις as the common name of such passive activities (426a10; cf. Phys. III 3, 202a21-202b5). So in III 3 he can count on his reader’s awareness of the fact that perceptions are motions of that kind. Clearly, it is that discussion that Aristotle refers to here when he resorts to “motion” as the genus of φαντασία. It is somewhat disconcerting that Aristotle, having defined φαντασία as the after-effect of sense-perception, does not provide any further information about its nature and function in the rest of the chapter, but confines himself to the assurance that there are φαντασίαι corresponding to all three types of perception, i.e. to specific perception such as hearing and seeing, to common perception such as an object’s size, motion or number, and to accidental perception of its subject and its properties (428b17-30)21. His prime interest at that point seems to lie in a comparison of the likeliness of truth and falsity of the three types of perception and of the corresponding φαντασίαι. As he states, perceptions of specific objects are always true (or almost so); perception of accidentals can be false – you can be mistaken about the identity of that 21 The example of accidental perception in DA II 6, 418a20-23: that the white thing is the son of Diares, might suggest that it has propositional content. But that would deprive the animals of this type of perception. Accidental perception must rather consist in the associations of properties that are perceived as somehow connected, like in the case of perceiving something as both yellow and bitter (III 1, 425a30-b4). Animals can recognize “ensembles” that constitute an object – the dog recognizes that pale thing as the son of Diares but not qua “son of Diares”.

66

DOROTHEA FREDE

white object you see, and the same is true of “common perception”, for it is possible to be wrong about an object’s size, motion or number. Of the corresponding φαντασίαι Aristotle has little more to say than that φαντασία concerning perception proper is true while the object is present, but that the other two can be false whether or not the object is present, and most of all when the object is at a distance. There are several points about this evaluation of φαντασία that give interpreters pause. First it seems strange that Aristotle assumes that perceptions and φαντασίαι should occur simultaneously, given that the φαντασίαι are the after-effects of perception; second, that he extends that simultaneity to all three types of perception and φαντασία22. Third, it seems strange that Aristotle spends so much time on the conception of φαντασία only to note that most of them are likely to be false and comes to the conclusion that φαντασία is operative in animals and in humans when reason is not: Καὶ διὰ τὸ ἐμμένειν καὶ ὁμοίας εἶναι ταῖς αἰσθήσεσι, πολλὰ κατ’ αὐτὰς πράττει τὰ ζῷα, τὰ μὲν διὰ τὸ μὴ ἔχειν νοῦν, οἷον τὰ θηρία, τὰ δὲ διὰ τὸ ἐπικαλύπτεσθαι τὸν νοῦν ἐνίοτε πάθει ἢ νόσῳ ἢ ὕπνῳ, οἷον οἱ ἄνθρωποι. Because instances of φαντασία persist and are similar to perception, animals do many things in accordance with them, some because they lack reason, e.g. beasts, and others because their reason is sometimes shrouded by passion, or sickness, or sleep, as in the case of humans. (429a4-9)

These questions deserve careful consideration. But in order to answer them a further problem needs to be addressed, namely what general use Aristotle assigns to the φαντασίαι. 4. The use of φαντασία There is just one indication in our passage that provides an answer concerning the question of the use and value that Aristotle assigns to φαντασία. That indication is very brief: it consists of the indication that the φαντασίαι persist (429a4: ἐμμένειν). Why is that important? It is so because sense-perceptions are short-lived, especially in the Aristotelian conception. For, there is sight only as long as the object’s impact on the 22 Cf. the critique in M. Schofield (1992), p. 264: “Aristotle has here been overwhelmed by the scholasticism of his attempt to distinguish three sorts of φαντασία corresponding to his three kinds of sense-perception, which strikes most readers as a baroque extravagance.”

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

67

eye lasts, and the same applies to all the other senses. As soon as one averts or closes one’s eyes, or the object disappears, there is no sight any longer, and the same is true in the case of hearing, tasting, smelling and touching. Hence all our lasting impressions of what we have seen, heard, touched etc. are φαντασίαι, once the perceptions themselves have passed. The short-livedness of perception at the same time provides an answer to the first question, why Aristotle assumes that there is simultaneity of perception and φαντασία: if the impressions are to be like their originals they must be caused by the sense-impressions while they are in operation, so that they produce impressions that closely resemble their originals. And the longer perception and φαντασία exist simultaneously the better is the chance that the imprints are reliable. Contiguity of the two types of impressions would not do. This fact at the same time provides an answer to the second question and to the “baroque extravagance” of assuming φαντασίαι of all three types of sense-perceptions: unless I retain an impression of the object’s color, I will not remember it when I try to recall it. And the same applies to its size, motion, number and identity, for instance that the white object I saw was a book. Otherwise it would be impossible to recollect anything of that sort, whether its color or size. If I have no such lasting impressions, I will not be able to find the book again. I may remember that I had that book, i.e. the fact. I may remember its title and that I have used it yesterday, but without the impression of its color and size I will not be able to retrieve it on the shelf by sight23. Although Aristotle does not comment on these phenomena, the need for such imprints is even stronger in the case of the recollection of a succession of sounds in melodies or in language. For every individual sound is gone as soon as it stops; the ability to retain a succession of sounds in a melody or in speech requires φαντασία. It is necessary therefore that we take a good look, listen attentively, pay attention to that particular taste, to that color, to each sound etc. if we are to recollect them properly later. Later recollections in a way are “up to us”, but not in the way certain interpreters seem to think24. For while it is up to me to try to 23 At I 1, 402b21-403a2 Aristotle has assigned to φαντασία an important role in the cognition of an object’s incidental properties because without that the demonstration must remain “empty dialectic”. According to DA III 2, 425b24 sq. there are perceptions and φαντασίαι after the perceptibles are no longer there. 24 T. Johansen (2012), p. 209 treats such retrievals as similar to mnemotechnic.

68

DOROTHEA FREDE

“conjure up” such an appearance, I can actually do no more than try to recollect it, and it may or may not “come to me”, such as the pattern of the dress someone was wearing yesterday. So “making use” of φαντάσματα is quite different from making up images in one’s mind. For retrieving an image of what has been perceived in the past is not the same thing as entertaining some imagination about it. Φαντασία, then, turns out to be the repository of sense-perceptions that constitute our sensory experience; they concern what modern psychology calls our “sensory memory”. Every particular recollection can, of course, be false, not only because our initial vision and other kinds of perceptions have been too perfunctory, but also because some of us are better at such retention and recollection than others. Inattentive witnesses and witnesses with bad sensory memory are notorious as a bane to the police. What is it that a witness has seen, heard, or smelled − or not seen, heard, or smelled? Often enough the police get the sheepish reply: “Had I known that anything might depend on it, I would have been more attentive”. It should be mentioned here, as an aside, that such sensory repositories should not be confused with “decaying sense-perceptions”. That Aristotle was well aware of the difference between them can be concluded from his description of such phenomena elsewhere. He describes in detail what happens to the visual imprint of the sun when we avert our eyes: at first it turns crimson, then purple, then black, until it finally fades away25. And the same applies to the reverberations of loud noises and strong after-tastes or smells – they all eventually fade away, even the most troublesome kinds. But these vanishing sense-impressions are not the same as the more or less clear φαντασίαι that we retain in our memory and that we can, at least often, retrieve when we try to recall what we have experienced before, as we can in case of the apparent size of the sun even when it is not visible. Because sense-perceptions are short-lived their impressions left in us, the φαντάσματα, also play an important role in both thought and action. I will be brief on both issues because they are addressed in some of the subsequent chapters (chapters 7-11). That φαντασία is an important precondition of knowledge has been mentioned at the beginning of De Anima (I 1, 403a7-10) and repeated in our chapter earlier (427b16): “without φαντασία there is no ὑπόληψις”. In his later discussion of the 25

Insomn. 459b13-16.

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

69

conditions of thought Aristotle is even more emphatic about the need of φαντασία. For he repeatedly asserts that φαντάσματα must accompany all our thoughts and that it is not possible even to “theorize” without them (7, 431a14; 8, 432a3-10). Moreover, he makes them co-responsible for conceiving of future goods and evils to be obtained or to be avoided in action (7, 431b2-10). Now we may think that in the case of actions such foresight of what is good or bad for us is a matter of imagination in the active sense. But envisaging a future good or evil, requires calculation and anticipation, rather than creative imagination. For it is not “up to us” what future good or evil we are to expect. Instead, it is the result of calculations that start with the present. And we regard such anticipations at least as likely to be true or false, and for that reason we pursue or avoid them: if I expect that a certain lion is going to pounce on me, I will try to get out of its reach. It therefore comes as no surprise that in his discussion of the need for φαντάσματα in thought and action Aristotle never repeats the claim that most of them are false. For in order to make choices we often need not only to know what object we want but we need to have reliable impressions of what it will be like, so that we either desire or avoid it. And because animals cannot calculate or reason about future goods or evils they are entirely dependent on the respective φαντάσματα. The richer a projected φαντασία of the future good or evil is, the more we will desire to obtain or to avoid it. It is φαντασία that allows us to retain such perceptual information beyond the “given” and to project it into the future26. But if φαντασίαι in the sense of “sensory impressions” are important repositories and constitute our sensory experience tout court, why do they receive such negative grades at the end of our chapter (429a4-9)? For humans are said there to act in accordance with φαντασία only when their reason is obstructed by emotions, illness or sleep. Some interpreters have assumed, therefore, that the main purpose of the determination of φαντασία in De Anima III 3 is to account for unclear and misleading sense impressions27. I take it that Aristotle’s concluding remarks in our chapter are concerned with people who, like animals, rely on φαντασίαι only and do not engage in deliberation and choice. In opposition to most sense-perceptions, such φαντασίαι are bound to mislead. But given that 26 On the fact that φαντάσματα of the future are necessary for animal motion cf. H. Lorenz (2006), p. 141 sq. 27 V. Caston (1996). W.D. Ross (1961), p. 39 even calls it “not a valuable faculty but a disability”.

70

DOROTHEA FREDE

φαντασία is a necessary prerequisite of thought (427b16) that verdict is obviously not Aristotle’s last word on φαντασία. If we wonder that he does not even hint at the importance of φαντασία at this point, we have to acknowledge that this is not the only occasion where he singles out one particular feature in an evaluation to the neglect of others and omits a reference to what is yet to come. There are, in addition, good reasons for Aristotle’s cageyness in III 3 concerning the way φαντασία works. For, there are difficult questions concerning the nature of those after-motions caused by sense-perception. To name just two: Where does this “phantastic” motion actually take place? Where and how is that motion preserved, once the original senseperceptions themselves are no longer there? In De Anima Aristotle confined himself to the statement that both sense-perception and φαντασία are contained in the sense-organs (III 2, 425b24 sq.: ἐν τοῖς αἰσθητηρίοις), by which he must mean the central or inner sense. For the rest he remains rather silent on the question of how and where these impressions are stored, and how and by what mechanism they are retrieved when they are needed. 5. Further information on the use of φαντασία in Parva Naturalia Φαντασία and φαντάσματα receive quite some attention in De Memoria and De Insomniis because they represent important elements in memory and in dreams. In both works Aristotle refers back to his determination of φαντασία in De Anima. Concerning the nature of φαντασία itself Aristotle gives the following information: If asked of which among the parts of the soul memory is a function, we reply: manifestly of that part to which φαντασία also appertains; and all objects of which there is φαντασία are in themselves objects of memory. (Mem. 1, 450a21-23)

In De Memoria Aristotle provides some further information of how the impressions come about by resorting to the comparison with the imprint of a seal on wax, a comparison that is familiar from the explanation of sense-perception in DA II 12, 424a17-25: The process of movement stamps, as it were, a sort of impression (τύπος) of the percept (αἴσθημα), just as persons do who make an impression with a seal. This explains why, in those who are strongly moved owing to passion, or time of life, no memory is formed; just as no impression would be

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

71

formed if the movement of the seal were to impinge on running water; while there are others in whom, owing to the receiving surface being frayed, as happens to old walls, or owing to the hardness of the receiving surface, the requisite impression is not implanted at all. (Mem. 450a31-b6)

That there is a difference in quality among those imprints is due not only to the material that retains the impressions, but also due to the difference in temperament: the memory of young persons is bad because they are too quick, of very old people because they are too slow28. Concerning the receptacle of such stamping, its quality and its place, Aristotle is, again, quite silent, except that he assigns it to “the primary faculty of perception” (πρῶτον αἰσθητικόν) and calls it an “affection” (πάθος) of the common sense. In Insomn. he adds that φαντασίαι take place either at the surface or deeper down near the “source” (ἀρχή) or “ruling and discerning part” (Insomn. 461b23 sq.: τὸ κύριον καὶ τὸ ἐπικρῖνον). In addition, φαντασίαι are called “residues of perceptions” (ὑπολείμματα). Just like in DA III 2 these residues are said to remain in the sense-organs (Insomn. 461b25-462a15 – ἐν τοῖς αἰσθητηρίοις). It is clear that the reference is to the inner sense around the heart. The φαντασίαι are supposedly preserved in the blood (Insomn. 2, 459b1-6; 3, 461a24-30), a claim that is repeated in DMA (7, 701b16-32). How this is supposed to work is difficult to envisage. But in view of the multiple functions that Aristotle attributes to the area around the heart, his silence about physiological questions should come as no surprise. It must be a very busy spot; and if it is fluid in consistency then it must be at the same time an impressionable and a retentive fluid, more like Plato’s wax than something in flux. Aristotle mentions that these impressions can get distorted by other impressions, especially under the influence of the emotions. But we learn nothing about the mechanisms by which the φαντάσματα are preserved, especially over a long time, or about the processes by which they are recalled. So that aspect of Aristotle’s theory remains quite mysterious. But De Memoria does provide some valuable information about the φαντάσματα themselves. For, Aristotle introduces an important distinction 28 In other works, Aristotle often uses αἰσθήσεις or αἰσθήματα to designate what must sensu stricto be φαντάσματα, because he has lasting sensory experience in mind and not actual perceptions. A witness to this fact is A.Po. II 19, 99b36-100a3 where Aristotle speaks of animals that are capable of the retention (μονή) of αἰσθήματα and thereby are capable of acquiring experience.

72

DOROTHEA FREDE

there with respect to sensory memories. He distinguishes between plain sensory memory and memory that intentionally focuses on what it is the impression of. As his comments on the “pictures” in our memory show, Aristotle has given quite some thought to the phenomenology of φαντάσματα (1, 450a29-451a18). He distinguishes between φαντάσματα that are pictures (ζωγράφημα) and are regarded as likenesses of a certain object and the pictures that are regarded only as pictures (450b11-451a18): Just in the same way we have to consider the impression (φάντασμα) in us both as something in itself and as relative to something else. In so far as it is regarded in itself, it is only an object of thought, but when considered as relative to something else, e.g. as its likeness, it is also a reminder. (Mem. 450b24-27)

And Aristotle adds that sometimes we are unsure whether an impression is a memory of something that has actually occurred so that we may be deluding ourselves in that respect: certain people are prone to assume that their impressions are likenesses of something else where there is, in fact, no such counterpart (Mem. 451a8-12)29. Thus, the φαντάσματα not only need to be stored, they also need to be “filed” and “labeled”. That is one of the points where things can go wrong. Sometimes when such a “movement” comes up again we are uncertain whether we actually did experience the corresponding perception or not. That is why we sometimes confuse mere images with past experiences. Sensory after-impressions can indeed not only get misfiled and mislabeled; they can also get distorted, overlaid in all sorts of ways; they can also get dissociated from the original experience and associated with other experiences, and they can remain in us as free-floating impressions that are no longer tied to any particular experience. And such experiences sometimes come to mind quite unbidden, without any conscious effort. And that possibility applies to all three types of φαντασία that correspond to the three types of sense-perception. But we speak of “imagining” only if we regard our impressions as mistaken and “imagination” in that case is used as a synonym for “illusion”. But we do not speak of “imagining” or “imagination” when our retrievals of past sense-perceptions are right. For, to get them right is not a matter of imagination nor can we procure them at will, although we can try to do so. The attempt to recollect such an impression will remain unsuccessful 29

See H. Weidemann (2001).

THE TWO MEANINGS OF PHANTASIA IN DE ANIMA III 3

73

if the images do not “come to us”, and often enough they don’t come, no matter how hard we try. I can try to recall a certain taste, a shade of colour, a melody, or a person’s voice. But in all such cases I am at the mercy of my sensory memory, of the φαντασίαι that are somehow stored. While I can actively imagine a purple dog at will, at any time, I may not be able to recall the exact shade of the color of my neighbor’s dog or its exact size, despite the fact that I see that animal every day. And we should all be familiar with agonizing attempts to recall a melody – or any other visual, audible, tactile etc. memory. But these are clearly all cases of “recalling” certain sensory impressions, not of “making” them in our imagination, for they clearly are not cases of εἰδωλοποιεῖν. Why is this difference important for the comprehension of Aristotle’s conception of φαντασία? Apart from the fact that φαντασίαι are the necessary components of thought, as Aristotle is going to explain in the subsequent chapters in De Anima, φαντασίαι are important, because the repository of sensory impressions is what allows both man and beast to find their way around in everyday life. For without sensory memory we would not find our way home, because we would not recognize any of the objects around us; imagination, by contrast, will be of no use to us. In everyday life we do not notice our dependence on these sensible images; this happens only if we are ill, suffer from some delusion or loss of memory. The following story will illustrate the distinction between impressions in memory and imagination. Many years ago in the investigation of a murder in the Metropolitan Opera House the prima ballerina had to be hypnotized, because she was dimly aware of having seen something connected with the victim in an elevator that she could not recall. Under hypnosis she managed to “re-visualize” the crucial scene. The murderer was thereby identified, caught, tried, and sentenced. It is clear that in that investigation the investigators would not have been interested in the prima ballerina’s imagination. Had she started garbling about what she imagined had happened, the police would have instantly dismissed her, one should hope. Mere imaginations are not φαντάσματα in the sense that Aristotle is concerned with in DA III 3. For he is not concerned with fictions or figments of any kind but with the types of sensory experience that we are capable of recollecting when we need them, either because we are trying to identify something in the present or in the past or we are trying to

74

DOROTHEA FREDE

envisage something we have to face in the future, be it good or bad. For our thoughts would be purely abstract if they were not accompanied by “images” of what we have in mind, and our memories would be rather poor if all that we could recollect were the bare facts that something or other has happened. Similarly, our calculations of future goods and ills would be rather poor without sensory impressions that stir up our desire or aversion. How, then, should we translate φαντασία, if “imagination” is misleading and in certain cases quite inappropriate? Should we resign ourselves to the use of φαντασία, as I did throughout this article, and admit that there are some terms in Aristotle that defy translation? My suggestion is that “sensory impression” is the right term. It sounds rather pale in comparison with the more fanciful “imagination”, but it is at least not misleading about its nature and object. Aristotle has discovered an important phenomenon that did not have a proper name at his time and that has remained elusive throughout most of history: our sensory memory and its contribution to both thought and action30.

30 The importance of φαντασία in that respect was noted and commented on extensively by Thomas Aquinas, cf. D. Frede (2001). He called it a “treasury of sensible forms” that ties thinking to the sensible objects.

DE ANIMA III 4, 429a10-430a91 Annick JAULIN

Le chapitre DA III 4 ouvre l’examen thématique « de la partie de l’âme par laquelle l’âme connaît (γινώσκει) et comprend (φρονεῖ) ». Cet examen s’inscrit dans une étude plus large de « la pensée et de l’intelligence », ouverte en III 3 (427a17-19)2, en des termes analogues à ceux qui figurent à la fin de l’examen, au début de III 9 (432a15-19)3, avant l’étude du mouvement selon le lieu. L’âme est définie en III 3 « surtout par deux différences » : le mouvement local d’une part et penser et comprendre d’autre part (III 3, 427a18-20) qui sont nommées en III 9, 432a15-17 « deux capacités (ou facultés) » : capacité critique d’une part, « œuvre de la réflexion (διανοίας) et de la sensation (αἰσθήσεως) », et mouvement local d’autre part. La communauté du thème, exposé au début et à la fin de l’examen de la faculté critique, n’a pas seulement l’avantage d’enrichir le débat traditionnel entre facultés (ou capacités : δυνάμεις) et parties (μορία / μερή) de l’âme par l’apport d’une variante : différence (διαφορά), il montre aussi que le jugement de R. Polansky, selon lequel cette partie est plus « fermement » organisée qu’il n’y paraît, est fondé4. 1 Le texte utilisé est celui établi par A. Jannone et traduit par E. Barbotin (1966) ; il est sur plusieurs points différent de celui édité par W.D. Ross (1956). Les traductions sont les nôtres, sauf cas contraires qui seront indiqués. 2 Ἐπεὶ δὲ δύο διαφοραῖς ὁρίζονται μάλιστα τὴν ψυχήν, κινήσει τε τῇ κατὰ τόπον καὶ τῷ νοεῖν καὶ φρονεῖν : « puisqu’on définit principalement l’âme par deux différences : le mouvement selon le lieu, la pensée et l’intelligence » ; les éditions anglaises ont, en outre, αἰσθάνεσθαι donné par ELCXy. 3 Ἐπεὶ δὲ ἡ ψυχὴ κατὰ δύο ὥρισται δυνάμεις ἡ τῶν ζῴων, τῷ τε κριτικῷ, ὃ διανοίας ἔργον ἐστὶ καὶ αἰσθήσεως, καὶ ἔτι τῷ κινεῖν τὴν κατὰ τόπον κίνησιν : « puisque l’âme est définie selon deux puissances – l’âme des animaux – par la fonction critique, œuvre de la réflexion et de la sensation et aussi par le fait de produire un mouvement selon le lieu ». 4 R. Polansky (2007), p. 434.

76

ANNICK JAULIN

On pourrait, en effet, en comparant les deux passages parallèles de III 3 et III 9 faire l’hypothèse que ce qui est posé comme une δόξα au début de III 3 : « on est d’avis (δοκεῖ δὲ) que penser et comprendre sont une sorte de sentir » (427a19-20)5 donne lieu à une thèse positive au début de III 9 : la capacité critique est donnée comme fonction ou œuvre (ἔργον) de la réflexion (διάνοια)6 et de la sensation (αἴσθησις). La thèse proposée en III 9 serait ainsi le résultat de l’examen critique et argumenté de l’opinion présentée en III 3. On pourrait donc établir une progression entre les chapitres DA III 3 à III 9 dans l’analyse de la capacité critique de l’âme ; dans cet ensemble, l’objet de III 4 serait précisément d’établir la différence de la partie par laquelle « l’âme connaît et comprend ». Or le chapitre III 4, à son tour, va confronter la position aristotélicienne avec les opinions antérieures relatives au νοῦς. En effet, au départ en III 3, la question du rapport νοεῖν / αἰσθάνεσθαι se situe dans un cadre plus large que celui de la seule doctrine aristotélicienne du rapport entre percevoir et penser, et prend aussi en considération une opinion qui affirme que « penser et intelliger est comme une sorte de sentir / percevoir » (III 3, 427a19-22). Même si cette opinion n’équivaut pas strictement (puisqu’elle est modalisée par ὥσπερ) à celle des anciens, selon laquelle « intelliger et sentir sont la même chose » (427a21-22), qui sera, elle, explicitement et rapidement refusée : « que ne soient pas la même chose sentir et intelliger, c’est évident » (427b67), on ne peut poser que l’opinion qui affirme que « penser et intelliger est comme une sorte de sentir / percevoir » exprime sans réserve une thèse aristotélicienne. Surtout si l’on suit les conséquences que cette opinion entraîne, à savoir une opinion proche de celle de Protagoras : « tout ce qui apparaît est vrai » (427b3)7. Or la formule réduite de cette δόξα de III 3 : δοκεῖ δὲ καὶ τὸ νοεῖν καὶ τὸ φρονεῖν ὥσπερ αἰσθάνεσθαί τι εἶναι (427a19-20), fournit le contenu de la protase conditionnelle ouvrant l’examen en III 4 : εἰ δή ἐστι τὸ νοεῖν ὥσπερ τὸ αἰσθάνεσθαι (429a13-14), dont dépend la longue suite des conséquences introduites par ἄρα jusqu’en 429a29.

Δοκεῖ δὲ καὶ τὸ νοεῖν καὶ τὸ φρονεῖν ὥσπερ αἰσθάνεσθαί τι εἶναι. Il n’y a pas lieu de faire de distinction entre διάνοια et νοῦς qui lui est substitué deux lignes plus bas. 7 Voir notamment le passage parallèle de Met. Γ 5. Tel était déjà le thème du Théétète de Platon (151e-160e). 5 6

DE ANIMA III 4, 429a10-430a9

77

On ne peut donc exclure la possibilité que cette première partie du texte de III 4 soit une reconstruction dialectique de la logique interne d’opinions qui ne seraient pas celles d’Aristote, ou du moins pas toujours ou pas seulement8. L’usage de la particule ἄρα qui introduit, à partir de 429a15, les conséquences de la protase conditionnelle va en ce sens. Elle est donnée par Liddell-Scott comme un opérateur de conclusion « more subjective than οὖν » ; elle peut être utilisée dans des conclusions « pseudo-syllogistiques ». J.D. Denniston va plus loin dans la même direction, lorsqu’il décrit ainsi le sens de ἄρα : « In reported speech, and after verbs of thinking and seeming, ἄρα denotes the apprehension of an idea not before envisaged. Usually ἄρα conveys either, at the most, actual scepticism, or, at least, the disclaiming of responsibility for the accuracy of the statement. But sometimes the context implies acceptance of the idea, and ἄρα merely denotes that its truth has not before been realized »9. Il signale sous ce sens que « sometimes the particle is repeated  »10, avec plusieurs exemples extraits de Platon. Le sens et la fonction de la particule ἄρα, omniprésente dans le début de III 4, peuvent ainsi indiquer une suite de conséquences de la protase conditionnelle qui ne sont pas toutes acceptées comme des thèses propres par Aristote. D’où l’hypothèse qui guidera ma lecture : le chapitre III 4 est construit selon une structure dialectique et organisé en trois parties (et non en deux, selon le découpage suivi par la plupart des commentateurs) qui seraient les suivantes : – Après une phrase d’introduction générale (429a10-13), une première partie (429a13-29) énumère une série logique de conséquences dépendantes d’une δόξα exprimée par la protase conditionnelle : εἰ δή ἐστι τὸ νοεῖν ὥσπερ τὸ αἰσθάνεσθαι. Certaines des conséquences issues de cette protase, comme cela a été souligné par de nombreux interprètes11, suscitent de réelles difficultés d’interprétation, si on les considère comme des thèses aristotéliciennes. Or, ce n’est sans doute pas nécessaire de les considérer ainsi. On comprend sans peine que le début du traitement d’une nouvelle question, ce qui 8 Parmi les commentateurs récents, Malcolm Lowe est l’un des rares à remettre en cause l’identité entre pensée et sensation, voir M. Lowe (1993) p. 110. 9 J.D. Denniston (1954), p. 38. 10 J.D. Denniston (1954), p. 39. 11 Voir C. Cohoe (2013).

78

ANNICK JAULIN

est le cas de la question traitée en III 4, suscite un exercice dialectique doxo-aporétique ; tel est le cas du début de l’examen du νοῦς. Cet exercice dialectique produit le renversement de la thèse anaxagoréenne, à partir de 429a21 et introduit la conception aristotélicienne. – Une seconde partie, introduite par ὅτι δ’ οὐχ ὁμοία ἡ ἀπάθεια τοῦ αἰσθητικοῦ καὶ τοῦ νοητικοῦ, φανερὸν ἐπὶ τῶν αἰσθητηρίων καὶ τῆς αἰσθήσεως (429a29-b22), expose la réception aristotélicienne des analyses précédentes ou ce qu’il en conserve comme conforme à ses propres analyses. Il montre alors précisément la raison de la différence entre l’ἀπάθεια du sens et de l’intellect ; il ne peut donc avoir accepté auparavant n’importe quelle forme d’analogie entre sentir et penser. Il précise alors de quelle manière le νοῦς est séparable. – Une troisième partie (429b22-430a9) expose les difficultés qui pourraient surgir si l’on interprétait la théorie aristotélicienne du νοῦς sur le modèle de la théorie anaxagoréenne du νοῦς, ainsi que la manière dont on peut éviter les difficultés en question par une certaine conception de la séparation des intelligibles. La logique générale de III 4 semble donc être la suivante : une première partie propose certaines conséquences logiques de la δόξα rapportée dans la protase conditionnelle. Ces conséquences permettent cependant à Aristote de présenter, dans les parties 2 et 3, sa propre conception du νοῦς, après avoir fait un usage particulièrement renversant de la thèse anaxagoréenne (élément doxographique évident). Dans une deuxième partie, il reprend et précise une conception du νοεῖν qui lui est propre, fondée sur l’identité du νοεῖν et des νοητά. Il la conforte enfin en montrant, dans une troisième partie, comment cette théorie peut éviter les difficultés que la théorie anaxagoréenne a rencontrées. 1. Introduction générale : 429a10-13 Le thème général de l’étude (429a10-13) concerne la différence « de la partie de l’âme par laquelle l’âme connaît (γινώσκει) et comprend (φρονεῖ) » et « comment se produit le penser », que cette « partie soit séparable selon la grandeur ou non séparable selon la grandeur, mais selon l’énoncé de définition »12. Les commentateurs anciens (Thémistius, 12 Les traductions sont variables : « notion », E. Barbotin (1966), P. Thillet (2005) ; « raison », R. Bodéüs (1993) ; « en définition », D.W. Hamlyn (1968).

DE ANIMA III 4, 429a10-430a9

79

172, 23 ; Philopon, 520, 25 ; ‘Simplicius’, 222, 5), suivis par G. Rodier, P. Siwek et H. Seidl ont vu dans la double activité de l’âme une référence à « la double fonction (théorique / γινώσκει et pratique / φρονεῖ) de l’intellect » ; R.D. Hicks, suivi par G. Movia, fait, au contraire, de γινώσκει le terme renvoyant à la connaissance en général et de φρονεῖ l’indication de la pensée intellectuelle ; ils sont suivis par R. Polansky13. Il est sans doute plus important de souligner que, conformément au privilège des opérations sur les capacités ou facultés, on s’occupe des opérations communes aux deux activités intellectuelles, présentées en 427a2021 dans une comparaison avec la sensation, comme des opérations de jugement (κρίνει) et de connaissance (γνωρίζει), à savoir des activités critiques dans le domaine de la connaissance14. L’examen de l’activité de νοεῖν sera mené quelle que soit la réponse donnée à la nature de la séparation de cette partie, séparation spatiale ou seulement logique, la séparation elle-même ne faisant pas de doute. On peut remarquer que la question, évoquée depuis le début du traité par Aristote15, est considérée ici, c’est-à-dire précisément au moment où l’examen de la partie noétique de l’âme commence et où tout le monde attend qu’elle soit tranchée, comme secondaire. Les deux manières de penser la séparation, séparation physique ou logique, sont traditionnelles chez Aristote et sont exprimées dans les mêmes termes pour la séparation « logique » et selon des différences lexicales mineures en ce qui concerne la séparation physique : ici et plus loin, 432a20 et 433b23-25, par la distinction μέγεθος / λόγος ; en 413b11-16, par l’opposition λόγῳ / τόπῳ, dans un passage où la question se pose pour toutes les capacités de l’âme ; en Met. H 1, 1042a26-31, la séparabilité suit la division λόγῳ / ἁπλῶς. La question de la séparation semble évoquée au début de chaque nouvel examen d’une « faculté » de l’âme, voir 432a20 pour la faculté motrice.

13 Alexandre d’Aphrodise, De l’âme 81, 5-9, distingue, comme les autres commentateurs anciens, deux sortes d’ὑποκείμενα de la δύναμις λογική : les πρακτικά d’un côté et les ἀίδια καὶ ἀναγκαῖα de l’autre. La référence de R. Polansky (2007), p. 435, n. 1, à 89, 4-18 concerne le commentaire d’Alexandre à un autre passage, nommément celui du ποιητικὸς νοῦς. 14 Cependant III 3, 427b10 qui énumère comme parties du νοεῖν ὀρθῶς la φρόνησις, dans une distinction avec l’ἐπιστήμη, semble poser une différence entre les deux et renvoyer la φρόνησις à l’exercice critique dans le domaine pratique. 15 403a5-16 ; 413b11-16, 24-27.

80

ANNICK JAULIN

Sur le point de savoir de quoi le νοῦς est séparable, les commentateurs ont également des réponses divergentes : s’agit-il des autres parties de l’âme ou du corps ? G. Rodier16 cite Thémistius qu’il suit en pensant qu’il ne s’agit pas de la séparabilité d’avec le corps (Simplicius admet même qu’il pourrait s’agir d’une différence interne à la pensée), mais plutôt d’avec les autres facultés de l’âme ; il fait référence aussi à DA I 1, 403a8-10 et II 2, 413b13, alors que R. Bodéüs17 pense que la séparabilité s’entend par rapport au corps. On peut remarquer que la question de la séparation n’a toujours pas de réponse ferme, quand elle est posée de nouveau en III 7, 431b19, lorsqu’on se demande si le νοῦς, qui n’est pas séparé des grandeurs spatiales, peut penser ou non des choses séparées de ces mêmes grandeurs (431b17-19). 2. Première partie : 429a10-29 Il s’agit donc d’une série logique de conséquences dépendantes d’une protase conditionnelle : εἰ δή ἐστι τὸ νοεῖν ὥσπερ τὸ αἰσθάνεσθαι, dont le contenu est identique à celui d’une δόξα exprimée en III 3, 427a19-20. Πάσχειν ἀπαθές La δόξα pourrait fort bien être celle d’Aristote qui pratique l’analogie entre intelliger et sentir / percevoir (II 5, 417b18-20). Même le statut paradoxal d’un pâtir (πάσχειν) impassible (ἀπαθές) pourrait être aristotélicien. Il pourrait correspondre à l’un des deux sens de pâtir en DA II 5, 417b2-16 où l’altération de la puissance vers son accomplissement était présentée comme un « autre genre d’altération » qui, loin d’être une destruction « sous l’action du contraire », était une « conservation / σωτηρία de l’être en puissance par l’être en entéléchie ». En effet, en ce cas, il y a passion sous l’action du semblable et la passion sous l’action du semblable n’est pas identique à la passion sous l’action du dissemblable

16

G. Rodier (1900), p. 435. R. Bodéüs (1993), p. 222, n. 1. Selon G. Movia (1991), p. 372-373, R.D. Hicks pense que la séparabilité d’avec les autres facultés implique la séparabilité d’avec le corps et H. Seidl se réfère à l’immatérialité (429b5 ; II 1, 413a3-7) et à l’immortalité du νοῦς (II 2 413b24-7 ; III 5, 430a22-23). 17

DE ANIMA III 4, 429a10-430a9

81

(417a18-21)18. Le même thème sera repris en III 7, 431a4-7, où il est alors question « d’une autre espèce (εἶδος) de mouvement ». Il était souligné en II 5, 417b29-418a3, que ce double sens de « pâtir », corrélatif du double sens « d’être en puissance », du fait d’une différence sans nom, contraignait d’user des noms impropres comme s’ils étaient propres19. En réalité, dans cet usage impropre de πάσχειν, πάσχειν équivaut à ἐνεργεῖν (417a15-16). Une analogie entre la science et la sensation (417b18) était alors posée, cependant que la différence entre percevoir et penser était également soulignée (417b24-26) ; elle tenait à la situation des agents producteurs : les singuliers sensibles, producteurs de la sensation, étaient extérieurs, tandis que les universels, objets de la science, étaient décrits comme étant « en quelque sorte dans l’âme elle-même » (417a22-24). Cette première conséquence, celle du πάσχειν ἀπαθές pourrait donc décrire une thèse aristotélicienne, mais la déduction n’est pas absolument nécessaire, car elle vaut aussi dans le cas de théories non aristotéliciennes. En effet, le ὥσπερ en 429a14 qui exprime le rapport entre sentir et intelliger ne suffit cependant pas à déterminer la nature ni le contenu de l’opération ainsi décrite, car il peut valoir aussi pour des conceptions somatiques du sentir et du penser, comme le chapitre précédent du livre III l’a montré, en présentant les conceptions qui aboutissent à la vérité de tout ce qui apparaît : « en effet, tous ceux-là conçoivent que le fait de penser est corporel, comme le fait de sentir »20. Les conceptions somatiques du penser comportent, elles aussi, la thèse d’un πάσχειν ἀπαθές : l’impassibilité du sentir est ensuite transférée au penser ; la thèse est déclarée « absurde » en I 5, 410a23-2621, alors qu’elle ne semble pas formellement éloignée de celles d’Aristote. Le πάσχειν ἀπαθές peut donc renvoyer à une opinion aristotélicienne comme à une opinion qui ne l’est pas. Dans les deux cas différents, la déduction de l’ἀπάθεια de l’opération est bonne. 18 διὸ ἔστι μὲν ὡς ὑπὸ τοῦ ὁμοίου πάσχει, ἔστι δὲ ὡς ὑπὸ τοῦ ἀνομοίου, καθάπερ εἴπομεν· πάσχει μὲν γὰρ τὸ ἀνόμοιον, πεπονθὸς δ’ ὅμοιόν ἐστιν. 19 νῦν δὲ διωρίσθω τοσοῦτον, ὅτι οὐχ ἁπλοῦ ὄντος τοῦ δυνάμει λεγομένου, ἀλλὰ τοῦ μὲν ὥσπερ ἂν εἴποιμεν τὸν παῖδα δύνασθαι στρατηγεῖν, τοῦ δὲ ὡς τὸν ἐν ἡλικίᾳ ὄντα, οὕτως ἔχει τὸ αἰσθητικόν. ἐπεὶ δ’ ἀνώνυμος αὐτῶν ἡ διαφορά, διώρισται δὲ περὶ αὐτῶν ὅτι ἕτερα καὶ πῶς ἕτερα, χρῆσθαι ἀναγκαῖον τῷ πάσχειν καὶ ἀλλοιοῦσθαι ὡς κυρίοις ὀνόμασιν. 20 πάντες γὰρ οὗτοι τὸ νοεῖν σωματικὸν ὥσπερ τὸ αἰσθάνεσθαι ὑπολαμβάνουσιν. (DA III 3, 427a26-27). 21 ἄτοπον δὲ καὶ τὸ φάναι μὲν ἀπαθὲς εἶναι τὸ ὅμοιον ὑπὸ τοῦ ὁμοίου, αἰσθάνεσθαι δὲ τὸ ὅμοιον τοῦ ὁμοίου καὶ γινώσκειν τῷ ὁμοίῳ τὸ ὅμοιον· τὸ δ’ αἰσθάνεσθαι πάσχειν τι καὶ κινεῖσθαι τιθέασιν· ὁμοίως δὲ καὶ τὸ νοεῖν τε καὶ γινώσκειν.

82

ANNICK JAULIN

En réalité, la règle de l’ἀπάθεια concerne le rapport du semblable au semblable (par opposition au rapport du contraire au contraire), elle vaut dans les théories matérialistes des physiologues comme dans celle d’Aristote et ne permet pas de caractériser précisément le contenu des opérations décrites, ni par conséquent la théorie dont il s’agit. On pourrait même penser que lorsque Aristote parle de « pâtir ou d’autre chose de tel du fait de l’intelligible »22 il fait allusion à des doctrines imprécises ou générales qui ne sont pas nécessairement les siennes. Ἀπαθές n’est donc pas un critère suffisant pour caractériser une théorie du sentir et du penser, pour la déclarer aristotélicienne ou non. On dira que cette imprécision n’est plus le cas, lorsque l’impassibilité est définie dans les termes d’une certaine sorte de « réception » de l’εἶδος, laquelle serait un marqueur différentiel suffisant de la teneur aristotélicienne de la thèse. De fait, il n’y a aucune difficulté à établir l’identité des termes de l’analyse de III 4, 429a15-18 avec ceux de la conclusion de l’analyse de la sensation en II 12, 424a17-2823. Pourtant, ce qui vaut de la sensation ne vaut peut-être pas exactement de la même manière pour l’intelligence, car Aristote remettra précisément en question le ὁμοίως ἔχειν, posé en 429a16, de l’ἀπάθεια entre l’αἰσθητικόν et le νοῦς, quelques lignes plus bas dans ce même texte en 429a29-30, lorsqu’il affirmera comme une évidence (φανερόν) que l’ἀπάθεια du sens et celle de l’intellect ne sont pas semblables24. L’hypothèse que la description de 429a16 ne renvoie pas à une thèse aristotélicienne, malgré une similitude terminologique certaine avec des thèses d’Aristote, éviterait donc une franche contradiction à quelques lignes d’intervalle. De fait, le vocabulaire de l’εἶδος ainsi que la différence entre τοῦτο et τοιοῦτον25 ne sont pas des inventions spécifiquement aristotéliciennes ; on trouve déjà ce lexique, et notamment cette dernière différence dans un passage du Timée de Platon (48e-50c)26, où apparaît le thème de l’ἐκμαγεῖον27. ἢ πάσχειν τι ἂν εἴη ὑπὸ τοῦ νοητοῦ ἤ τι τοιοῦτον ἕτερον (DA III 4, 429 a14-15). ἡ μὲν αἴσθησίς ἐστι τὸ δεκτικὸν τῶν αἰσθητῶν εἰδῶν. 24 ὅτι δ’ οὐχ ὁμοία ἡ ἀπάθεια τοῦ αἰσθητικοῦ καὶ τοῦ νοητικοῦ. 25 Qui se trouve aussi en II 12, 424a23-24 sous la forme de l’opposition entre ᾗ ἕκαστον et ᾗ τοιονδί. 26 ἀεὶ ὃ καθορῶμεν ἄλλοτε ἄλλῃ γιγνόμενον, ὡς πῦρ, μὴ τοῦτο ἀλλὰ τὸ τοιοῦτον ἑκάστοτε προσαγορεύειν πῦρ, μηδὲ ὕδωρ τοῦτο ἀλλὰ τὸ τοιοῦτον ἀεί (49d4-7). 27 La différence concerne notamment les figures (σχήματα) et leur porte-empreinte (ἐκμαγεῖον) en 50a-c. L’ἐκμαγεῖον est dans le Théétète (191c-e) une cire sur laquelle viennent se graver des empreintes et des signes en lesquels consiste la mémoire. 22 23

DE ANIMA III 4, 429a10-430a9

83

Ἀμιγής La référence à Anaxagore est ici explicite. Il est clair que la présentation de la doctrine ne vaut pas pour une adhésion, puisque les thèses anaxagoréennes fourniront le thème des difficultés de la fin de III 4 (429b22-430a9). En outre, l’absence de mélange du νοῦς anaxagoréen, qui semble se déduire de l’ἀπάθεια précédente, pourrait être une objection à la ressemblance entre intelliger et percevoir, car l’absence de mélange du νοῦς anaxagoréen est aussi le signe de son absence de similitude avec toute sensation. Le « sans mélange » du νοῦς est un marqueur anaxagoréen : ἀμιγής n’est utilisé par Aristote que lorsqu’il s’agit d’Anaxagore28. De fait, selon Anaxagore, le νοῦς est seul à être « simple, sans mélange et pur »29 (I 2, 405a16-17), puisqu’il affirme que « tout est à l’état de mélange sauf l’intelligence et qu’elle seule est sans mélange et pure » (Met. A 8, 989b15-16) ; en outre, Anaxagore, seul parmi les physiologues, déclare que « l’intellect est impassible (ἀπαθῆ) et n’a rien de commun avec aucun des autres êtres »30 (I 2, 405b19-21), bien que, selon Aristote, il identifie parfois l’intellect et l’âme (404b1-3). La déduction de ἀπαθής à ἀμιγής est le déroulement d’une logique anaxagoréenne, non nécessairement fondée, d’où l’usage du ἄρα. C’est « parce qu’il pense tout » que le νοῦς doit être « sans mélange », car toute participation serait une limitation de son κράτος. Dans son commentaire au DA, R. Polansky31, après avoir posé la question « comment Aristote connaît-il que le νοῦς pense toute chose », fait référence au fragment d’Anaxagore DK 59B12 pour expliquer l’assertion aristotélicienne. Il est vrai que, dans ce fragment32, l’absence de mélange du νοῦς 28 Ce caractère est présent dans toutes les occurrences relatives à Anaxagore : DA 404a25-b7 ; 405a13-17 ; 405b19-21 ; Met A 8, 989b15 ; Phys. 256b25. Le même trait est relevé par Platon dans le Cratyle 413c6. 29 Μόνον γοῦν φησὶν αὐτὸν τῶν ὄντων ἁπλοῦν εἶναι καὶ ἁμιγῆ τε καὶ καθαρόν. 30 Ἀναξαγόρας δὲ μόνος ἀπαθῆ φησιν εἶναι τὸν νοῦν, καὶ κοινὸν οὐθὲν οὐθενὶ τῶν ἄλλων ἔχειν. 31 R. Polansky (2007), p. 437. 32 τὰ μὲν ἄλλα παντὸς μοῖραν μετέχει, νοῦς δέ ἐστιν ἄπειρον καὶ αὐτοκρατὲς καὶ μέμεικται οὐδενὶ χρήματι, ἀλλὰ μόνος αὐτὸς ἐπ’ ἐωυτοῦ ἐστιν. εἰ μὴ γὰρ ἐφ’ ἑαυτοῦ ἦν, ἀλλά τεῳ ἐμέμεικτο ἄλλῳ, μετεῖχεν ἂν ἁπάντων χρημάτων, εἰ ἐμέμεικτό τεῳ· ἐν παντὶ γὰρ παντὸς μοῖρα ἔνεστιν, ὥσπερ ἐν τοῖς πρόσθεν μοι λέλεκται· καὶ ἂν ἐκώλυεν αὐτὸν τὰ συμμεμειγμένα, ὥστε μηδενὸς χρήματος κρατεῖν ὁμοίως ὡς καὶ μόνον ἐόντα ἐφ’ ἑαυτοῦ. ἔστι γὰρ λεπτότατόν τε πάντων χρημάτων καὶ καθαρώτατον, καὶ γνώμην γε περὶ παντὸς πᾶσαν ἴσχει καὶ ἰσχύει μέγιστον· καὶ ὅσα γε ψυχὴν ἔχει καὶ τὰ μείζω καὶ τὰ ἐλάσσω, πάντων νοῦς κρατεῖ.

84

ANNICK JAULIN

est mise en rapport avec son aspect illimité (ἄπειρον) et sa toute-puissance, associée à la pensée (l’équivalence indiquée par τοῦτό ἐστι est, elle, une interprétation aristotélicienne) ; tout mélange lui ferait donc obstacle (ἐκώλυεν). En même temps, dans le passage de 429a20 : παρεμφαινόμενον γὰρ κωλύει τὸ ἀλλότριον καὶ ἀντιφράξει, le seul terme qui rappelle le texte d’Anaxagore est κωλύει, ce qui est peu. G. Rodier33 note que G. Teichmüller34 rapproche ce passage de Timée 50d : παρεμφαινόμενον en 429a20 serait un écho du παρεμφαῖνον de 50d5-e535. L’absence de forme du réceptacle, décrit comme ἄμορφον, est la condition de la réception de « toutes » les formes. La réception du fragment 12 d’Anaxagore sur le νοῦς serait donc ici médiatisée par la description platonicienne du réceptacle. Ainsi R.D. Hicks36 suggère qu’Aristote, en plus d’une référence à Anaxagore, suit la description platonicienne de la manière dont le réceptacle dans le Timée doit lui-même manquer de toute forme afin de les recevoir toutes. La remarque est décisive ; car sans que l’on puisse décider si Platon a construit le réceptacle du Timée comme représentation parodique du νοῦς anaxagoréen, ou si c’est Aristote qui établit un rapport « renversant » entre le νοῦς d’Anaxagore et le réceptacle du Timée, il est clair que le νοῦς est passé d’un statut de toute-puissance au statut de simple puissance. Aristote interprète en effet la χώρα du Timée comme matière, puisqu’elle est sans forme, et la matière n’a d’autre fonction que celle de puissance. Le renversement pourrait expliquer les hésitations entre deux choix de construction et de traduction de 429a2037 : si l’on suit le ἂν ἐκώλυεν αὐτὸν τὰ συμμεμειγμένα du fragment B12 d’Anaxagore, on construira la phrase avec τὸ ἀλλότριον en position de sujet : « le corps étranger paraissant en même temps fait obstacle et s’interpose » ; le vocabulaire est physique : ἀντιφράττειν est le terme utilisé pour l’éclipse (A.Po. I 31, 33

G. Rodier (1900), p. 438. G. Teichmüller (1874), p. 333. 35 νοῆσαί τε ὡς οὐκ ἂν ἄλλως, ἐκτυπώματος ἔσεσθαι μέλλοντος ἰδεῖν ποικίλου πάσας ποικιλίας, τοῦτ’ αὐτὸ ἐν ᾧ ἐκτυπούμενον ἐνίσταται γένοιτ’ ἂν παρεσκευασμένον εὖ, πλὴν ἄμορφον ὂν ἐκείνων ἁπασῶν τῶν ἰδεῶν ὅσας μέλλοι δέχεσθαί ποθεν. ὅμοιον γὰρ ὂν τῶν ἐπεισιόντων τινὶ τὰ τῆς ἐναντίας τά τε τῆς τὸ παράπαν ἄλλης φύσεως ὁπότ’ ἔλθοι δεχόμενον κακῶς ἂν ἀφομοιοῖ, τὴν αὑτοῦ παρεμφαῖνον ὄψιν. 36 R.D. Hicks (1907), p. 476-477. Cité par F.A. Lewis (2003), p. 101-102. 37 Les traducteurs hésitent entre deux choix de traduction et d’interprétation selon que τὸ ἀλλότριον est considéré comme sujet (Ross, Burnyeat) ou comme objet (Politis, Dillon, Cohoe). 34

DE ANIMA III 4, 429a10-430a9

85

87b40 ; II 2, 90a15 ; Cael. II 13, 293b25) ; si l’on suit la logique du réceptacle du Timée, on comprendra comme Alexandre d’Aphrodise : « car sa propre forme, en se manifestant, empêcherait la réception d’une forme étrangère »38 (84, 15), avec τὸ ἀλλότριον en position d’objet. La manière dont Alexandre comprend la phrase conserve la logique de la déduction opérée par le texte, de sorte que s’accomplit bien un renversement de la toute-puissance du νοῦς anaxagoréen. Qu’il s’agisse d’une interprétation contextuelle et ad hoc (c’est-à-dire dialectique) du νοῦς anaxagoréen, on peut le supposer, puisque Aristote donne une interprétation quasiment opposée du νοῦς anaxagoréen, comme le représentant de l’un par opposition à l’autre qui est « indéfini / ἀόριστον » dans le texte de Met. A 8, 989b16-2139. Il est possible que le jeu interprétatif d’Aristote soit permis par les caractères du νοῦς anaxagoréen, jugés par lui incompatibles, à la fois ἄπειρον et αὐτοκρατές ; on sait que l’ἄπειρον est « cause comme matière et son être propre est la privation » (Phys. III 7, 207b35-208a1). Δυνατόν Ce caractère δυνατόν du νοῦς est introduit par ὥστε comme une conséquence logique du rapprochement qui vient d’être effectué entre les caractères du νοῦς anaxagoréen et ceux de la χώρα du Timée. Le renversement de la conception anaxagoréenne du νοῦς importe à Aristote dans la mesure où il permet de récuser toute théorie naturaliste du νοῦς  : le νοῦς n’a aucune nature, sauf d’être un « possible » (δυνατόν)40. Ce caractère « possible » du νοῦς rencontre la conception de « ce qu’on appelle le νοῦς de l’âme », à savoir « celui par lequel l’âme réfléchit et conçoit » et qui « n’est en acte aucun des êtres avant de penser » (429a24). En d’autres mots, le problème est maintenant inscrit dans le lexique de l’Académie, à l’intérieur duquel Aristote manifestera sa problématique personnelle quelques lignes plus bas (429a27-29). παρεμφαινόμενον γὰρ τὸ οἰκεῖον εἶδος κωλύει τὴν τοῦ ἀλλοτρίου λῆψιν. φησὶ δ’ εἶναι μεμιγμένα πάντα πλὴν τοῦ νοῦ, τοῦτον δὲ ἀμιγῆ μόνον καὶ καθαρόν. ἐκ δὴ τούτων συμβαίνει λέγειν αὐτῷ τὰς ἀρχὰς τό τε ἕν (τοῦτο γὰρ ἁπλοῦν καὶ ἀμιγές) καὶ θάτερον, οἷον τίθεμεν τὸ ἀόριστον πρὶν ὁρισθῆναι καὶ μετασχεῖν εἴδους τινός, ὥστε λέγει μὲν οὔτ’ ὀρθῶς οὔτε σαφῶς, βούλεται μέντοι τι παραπλήσιον τοῖς τε ὕστερον λέγουσι καὶ τοῖς νῦν φαινομένοις μᾶλλον. Pour une étude de ce texte, voir C. Louguet (2013), p. 133-136. 40 ὥστε μηδ’ αὐτοῦ εἶναι φύσιν μηδεμίαν ἀλλ’ ἢ ταύτην, ὅτι δυνατόν (429a21-22). 38 39

86

ANNICK JAULIN

Le plus important est que ce caractère « possible » du νοῦς pose et définit la manière dont l’analogie aristotélicienne entre sentir et penser se distingue de la similitude mise en place par la δόξα antérieure : il permet ainsi d’expliciter comment le ὥσπερ aristotélicien se distingue du ὥσπερ de la δόξα antérieure qui posait une similitude entre sentir et penser. En effet, le caractère δυνατός du νοῦς est développé selon la différence de trope du δυνατός, décrite en II 5, 417a26-29, qui met en perspective deux manières d’être savant en puissance (417a30-b2). Ces deux manières différentes d’être savant en puissance mettent en évidence comment d’abord par l’enseignement (διδασκαλία), puis par la pratique de la θεωρία, il y a « changement vers les possessions et la nature » (417b16), cette nature dont l’absence, au départ, signe précisément le statut δυνατόν du νοῦς. L’intérêt du passage de II 5 pour la question qui nous intéresse est qu’il compare ces deux modes de la puissance du savoir, lequel implique le double changement de l’acquisition et l’exercice, avec les modes du sentir qui ne comprennent pas ce double changement : « dans le cas de la faculté de sentir, le premier changement se produit du fait du générateur, et quand l’être a été engendré, il possède désormais le sentir à la manière d’une science »41 (417b16-19), autrement dit, à la naissance, l’être engendré possède le sentir comme une ἕξις. Le sentir n’est donc pas un pur δυνατόν, il a déjà quelque φύσις, lorsque naît l’animal. Telle est la raison pour laquelle les A.Po. II 19, 99b35 présentent la sensation (αἴσθησις) comme une « puissance critique naturelle » (δύναμιν σύμφυτον κριτικήν). C’est donc parce que le νοῦς n’est pas σύμφυτον, mais δυνατόν et que, pour cette raison, il n’est rien avant d’apprendre et de penser qu’il est « raisonnable qu’il ne soit pas mêlé au corps » (429a24-25). Le νοῦς est sans qualité et n’est pas prédéterminé par un organe : le νοῦς vient, au sens propre, « du dehors » (θύραθεν)42. Sans entrer dans une étude approfondie de cette question qui serait d’ailleurs hors du sujet, on peut au moins déduire de cette différence entre αἴσθησις et νοῦς ce qu’en disent les lignes des A.Po. II 19, à savoir qu’il n’y a pas en nous des « états noétiques dont nous ignorerions l’existence » (99b25-26),

41 τοῦ δ’ αἰσθητικοῦ ἡ μὲν πρώτη μεταβολὴ γίνεται ὑπὸ τοῦ γεννῶντος, ὅταν δὲ γεννηθῇ, ἔχει ἤδη, ὥσπερ ἐπιστήμην, καὶ τὸ αἰσθάνεσθαι. 42 Selon l’expression que l’on trouve en GA II 3, 736b19-29.

DE ANIMA III 4, 429a10-430a9

87

autrement dit l’âme n’est pas le « lieu de formes »43 en acte, dont elle ignorerait l’existence. Ce qui vaut pour une critique des thèses platoniciennes sur l’intelligible et éclaire la distinction aristotélicienne entre l’âme et le νοῦς. L’âme noétique, non l’âme entière, est « le lieu » des formes en puissance, comme l’affirment les dernières lignes (429a2829) de cette première partie. L’assertion est cohérente avec celle du traité des Parties des animaux (I 1, 641a32-b4) qui affirme que, si la science naturelle s’occupait de toute sorte d’âme, la philosophie serait privée d’objet, car la physique traiterait aussi des intelligibles, puisque la même science porte, en chaque domaine, sur les termes qui sont relatifs l’un à l’autre. Le texte confirme ainsi que l’analogie entre sensation et intellection tient à la forme de la relation entre une sorte d’âme et son objet, mais non à son contenu, puisque le contenu est précisément ce qui distingue l’âme, objet de la physique et l’âme, objet de la philosophie. Les âmes diffèrent dans le même rapport que le sensible et l’intelligible se distinguent. L’usage de l’expression « âme noétique » (429a27-28) comme équivalent de « ce qu’on appelle intellect de l’âme » (429a22) met en évidence que la distinction de l’âme et du νοῦς ne saurait valoir pour une séparation substantielle. Le résultat de l’analyse de cette première partie est que par le statut δυνατόν du νοῦς, déduit d’une théorie anaxagoréenne interprétée ad hoc dans une analogie avec la théorie de la χώρα du Timée, la théorie aristotélicienne du νοῦς est située par rapport aux théories antérieures. Il y a bien entre elles une ressemblance quant au rapport entre le sens et le sensible, l’intellect et l’intelligible, mais cette ressemblance n’induit pas une similitude, car le statut du sens et celui de l’intellect diffèrent pour Aristote : le sens est une puissance critique naturelle, ce qui n’est pas le cas du νοῦς qui n’est pas une ἕξις, inscrite dans la structure organique. Il y a ainsi trois états du νοῦς : (1) celui de pure puissance au départ, (2) un état de puissance seconde dans l’apprentissage et l’acquisition et enfin (3) un état d’ἐνέργεια dans l’exercice, alors qu’il y a deux états seulement de l’αἴσθησις : ἕξις au départ et ἐνέργεια ou exercice ensuite. Le sens du νοῦς δυνατός dans ce contexte est donc que le νοῦς n’est pas une ἕξις naturelle ; l’on ne peut ainsi concevoir le rapport entre 43 Si l’on ne trouve pas chez Platon l’expression τόπος εἰδῶν, on trouve bien l’expression τόπος νοητός en Rep. VI, 508c2 et 509d2.

88

ANNICK JAULIN

l’intellect et l’intelligible sur un mode semblable à celui qui existe entre le sens et le sensible. Cette critique vaut aussi contre la théorie platonicienne de la réminiscence. L’objet de cette première partie est donc de défaire la δόξα de tous les auteurs qui « conçoivent la pensée, à l’instar de la sensation, comme une réalité corporelle et que le semblable sent et pense par le semblable, comme nous l’avons exposé au début du traité »44. Il suit de là que le νοῦς δυνατός ne peut être identifié au νοῦς παθητός dont il sera question en 430a2445 : le νοῦς δυνατός a une fonction critique par rapport à la doxa antérieure qui développait une conception naturaliste du νοῦς ; le νοῦς παθητός remplit une fonction interne à l’analyse aristotélicienne du νοῦς. Le νοῦς παθητός concerne le νοῦς comme ἕξις et prend sens dans les limites de son exercice. 3. Seconde partie : 429a29-b22 Que la finalité de la démarche précédente était bien de défaire une similitude erronée entre sentir et penser, le montre l’affirmation forte de leur dissemblance introduite par ὅτι (429a29), à la fin de l’examen critico-doxographique, destiné à établir ce point. Les thèses se développent maintenant à partir de l’analyse des fonctions clairement distinctes que sont sentir et penser, du point de vue de leur exercice. À ce titre, sont évoquées des différences intensives et fonctionnelles qui explicitent la différence des relations entre sens et sensible d’une part, intellect et intelligible d’autre part. Trois points sont notés : a) L’absence de similitude entre l’ἀπάθεια de la faculté sensitive et celle de la faculté intellective ; cette différence tient au fait que la faculté sensitive n’est pas sans le corps, tandis que l’intellect est séparable (429a29-b5). b) Une différence dans l’être en puissance (2)46 du νοῦς, par rapport au δυνατόν (1) évoqué précédemment, dans le passage de la possession 44 πάντες γὰρ οὗτοι τὸ νοεῖν σωματικὸν ὥσπερ τὸ αἰσθάνεσθαι ὑπολαμβάνουσιν, καὶ αἰσθάνεσθαί τε καὶ φρονεῖν τῷ ὁμοίῳ τὸ ὅμοιον, ὥσπερ καὶ ἐν τοῖς κατ’ ἀρχὰς λόγοις διωρίσαμεν (427a26-29). 45 Alexandre d’Aphrodise qui insiste sur l’aspect potentiel du νοῦς (De l’âme, 81, 13-15) les identifie, au contraire. Il est vrai que son interprétation du νοῦς δυνατός implique déjà une certaine détermination, puisqu’il y trouve déjà une distinction entre l’intellect pratique et l’intellect théorique (81, 9-12). 46 Δυνάμει en 429b8.

DE ANIMA III 4, 429a10-430a9

89

de l’intellect à son exercice ; quand l’intellect peut, par lui-même, passer à l’acte, il peut alors se penser lui-même (429b5-10). c) Le rapport entre les composés et la forme qui est une autre manière de poser le problème du rapport entre les facultés sensibles et intelligibles (429b10-22), puisque la dernière partie du chapitre pose que le νοῦς est intelligible comme le sont les νοητά (430a2). La dissemblance de l’ἀπάθεια entre les facultés sensible et intelligible Elle est affirmée, alors qu’est maintenue l’analogie du rapport entre sens et sensible, intelligence et intelligible. Cette différence d’impassibilité expose la signification de la séparabilité de l’intellect d’avec le corps et sa différence d’avec le sens, quant à lui somatiquement déterminé. La différence est intensive : une passion excessive détruit l’organe sensoriel (III 2, 426b7-8)47, quand il « pâtit » sous l’effet des sensibles, d’où l’importance de la question de la moyenne et des excès dans l’analyse du pâtir propre au sens (II 11, 424a1-6)48. La capacité intelligible, dans son exercice, ne connaît pas la condition d’une moyenne du fait qu’elle est un exercice interne. Au contraire du sens, dont l’exercice intensif rend impossible la perception des sensations moins fortes, l’exercice intensif de l’intellect (la pensée du plus intelligible) n’empêche pas de penser des intelligibles de rang inférieur : penser les formes n’interdit pas de penser les formes dans la matière. Le νοῦς est donc « séparable » du corps49. Il est à remarquer que cet aspect séparable concerne le νοῦς en général ; on suivra Théophraste sur ce point50. Le fait que le νοῦς soit séparable du corps n’implique pas que l’exercice de la pensée soit sans condition ; on le sait, l’exercice de la pensée peut être affecté par l’affaiblissement des organes : même si le νοῦς est ἀπαθής, l’exercice de la pensée, comme les autres formes d’exercice, est le fait du sujet (I 4, 408b18-30). ἡ δ’ αἴσθησις ὁ λόγος· ὑπερβάλλοντα δὲ λύει ἢ φθείρει. τὸ γὰρ αἰσθάνεσθαι πάσχειν τι ἐστίν· ὥστε τὸ ποιοῦν, οἷον αὐτὸ ἐνεργείᾳ, τοιοῦτον ἐκεῖνο ποιεῖ, δυνάμει ὄν. διὸ τοῦ ὁμοίως θερμοῦ καὶ ψυχροῦ, ἢ σκληροῦ καὶ μαλακοῦ, οὐκ αἰσθανόμεθα, ἀλλὰ τῶν ὑπερβολῶν, ὡς τῆς αἰσθήσεως οἷον μεσότητός τινος οὔσης τῆς ἐν τοῖς αἰσθητοῖς ἐναντιώσεως. 49 Ce qui rappelle la reconnaissance par Théétète que « les communs n’ont pas comme les sensibles d’organe propre » (185d). 50 Voir F. Brentano dans M. C. Nussbaum et A.O. Rorty (1992), p. 313. 47 48

90

ANNICK JAULIN

On peut remarquer à ce propos que, depuis le début de III 4, Aristote s’est exprimé en disant « l’âme connaît et pense » (429a10-11), il dit maintenant « la sensation ne peut sentir » (429a31), « l’intellect quand il a pensé » (429b3), en semblant avoir oublié la règle qu’il a posée en I 4, 408b14-18, selon laquelle il ne faut pas dire que l’âme « apprend ou pense », mais « l’homme par son âme ». Ces expressions, comme le montrent leurs variations, doivent être considérées comme des facilités d’expression. Les deux niveaux en puissance du νοῦς Il s’agit de la distinction faite par Aristote entre l’état en puissance du νοῦς quand il existe comme une ἕξις et son état en puissance « avant d’avoir appris ou trouvé » (429b8-9). Ces deux niveaux distincts de la puissance s’inscrivent dans la droite ligne de la « double chasse » aux colombes du Théétète51 : les deux niveaux de puissance, antérieurs à l’exercice de la pensée, équivalent aux deux niveaux de la chasse du Théétète ; la première chasse est l’acquisition qui vise la possession (κτῆσις), elle est l’apprentissage ; la seconde est la possession qui vise la prise en mains (ἕξις) de ce que l’on possède. L’exercice de la pensée, l’ἐνέργεια du νοῦς, est analogue à la prise en mains de ce que l’on possède, puisque la possession n’est elle-même encore qu’une puissance (δύναμις). Il y a bien également ici deux niveaux distincts de la puissance (apprentissage et possession), avant la « prise en main », malgré le décalage terminologique entre le Théétète et le DA, puisque Aristote nomme ἐνέργεια (429b7) ce qui est désigné comme ἕξις par Socrate, après qu’il a distingué entre « avoir sous la main » ou posséder, et « prendre en main » (197c). Ainsi, il y a un état du νοῦς en puissance qui est différent du νοῦς δυνατός dont il était question dans la première partie ; c’est seulement dans l’exercice de la pensée, ou quand il est en ἐνεργείᾳ, que le νοῦς peut se penser lui-même52. 51 198d : « Donc nous reviendrons à l’image de la possession et de la chasse des colombes et nous dirons qu’il y avait là double chasse : l’une avant acquisition et visant la possession ; l’autre, par qui possède, mais désire prendre et avoir en mains ce que depuis longtemps il possède ». Il a été signalé (197e) que la « cage est vide dans l’enfance ». 52 Selon la formule de Met. Λ 7, 1072b19-20 : « l’intellect a l’intellection de luimême par une saisie de l’intelligible ».

DE ANIMA III 4, 429a10-430a9

91

La chose et son essence C’est bien parce que le τὸ τί ἦν εἶναι est la forme ou la substance au sens premier53, et ainsi l’expression même de l’intelligible, que la suite du texte va évoquer la différence entre la substance composée et la substance formelle ou forme, afin de continuer l’examen de la manière dont le νοῦς est séparable. La conclusion de l’examen sera, en effet, que « de la même manière que les choses sont séparables de la matière, ainsi en est-il pour ce qui concerne l’intellect » (429b21-22). Cette conclusion a pourtant suscité de multiples interprétations ; G. Rodier54 constate qu’il y a autant d’interprétations de 429b22 que de commentateurs. La perplexité demeure sur le sens de ce passage55, où l’on réactualise parfois la question de savoir quel est le sujet du κρίνει (429b13, 15, 17, 21) : le νοῦς lui-même ou celui qui juge, qui avait déjà été l’objet d’un débat chez les commentateurs anciens56. La remarque faite plus haut sur les « facilités d’expression » aristotéliciennes rend le débat peu significatif. On ne peut considérer la proposition introduite par ἐπεὶ en 429b10, au début de cet examen, comme un acquis suffisant pour régler la question, car elle introduit une alternative : on juge de l’essence de la chose et de la chose « soit par une autre [faculté] soit par une [faculté] autrement disposée » (429b12-13). Le développement de cette alternative est donné quelques lignes plus bas, quand on dit que l’on juge de l’essence ou de la forme par une faculté « ou séparée, ou comme est la ligne brisée par rapport à elle-même quand elle a été redressée » (429b16-17). Or il faudrait avoir choisi l’un des deux membres de l’alternative pour avoir une thèse positive sur le point considéré. On suggérera donc de considérer ce passage comme l’énoncé d’un problème qui reste à résoudre, une fois que l’on a distingué entre l’être d’une chose et cette chose : comment penser cette altérité, une fois qu’elle a été posée ? La répétition de ἄλλο sous des formes diverses (429b10, 13, 16) montre que le problème porte sur cette altérité et il est de savoir s’il y a une double instance de jugement ou simplement une même instance, disposée autrement, comme l’est une ligne brisée par rapport à une ligne brisée redressée. L’image suscite l’interrogation. Dans un contexte où la question posée est celle 53 54 55 56

Voir par exemple Met. Z 7, 1032b1-2 ou DA II 1, 412b9-16. G. Rodier (1900), p. 453. R. Polansky (2007), p. 445-451. M. Lowe (1993), p. 114-119.

92

ANNICK JAULIN

du type de rapport que peut avoir la saisie de l’essence avec une capacité de perception, il semble que Timée 43e puisse expliquer le choix de l’image. Le passage expose en effet comment il est possible de redresser les lignes brisées et les proportions dérangées par l’arrivée de la sensation dans l’âme ; un mouvement trop violent, qui ne peut cependant les défaire entièrement, brise à ce moment-là et dérange les proportions de l’âme (il est là aussi question de μεσότης). Le second membre de l’alternative opposerait donc à l’altérité pure et simple entre la chose composée et son essence la possibilité d’un « redressement » de la substance composée dans la saisie de son essence. Les exemples choisis tendent à renforcer l’hypothèse du développement d’un problème. Le premier ensemble de cas (429b10-18) rassemble des objets physiques : grandeur, eau, chair qui sont des composés, comparables au camus. On peut vraisemblablement considérer que ces objets sont des composés de matière et de forme dont le sens peut rendre compte. La question devient plus délicate lorsque le même traitement est appliqué à des objets mathématiques, comme dans le deuxième exemple (429b1820) qui porte sur un objet « d’abstraction » tel que la ligne droite ; or elle est également comparée au camus, ce qui signifie qu’elle est, elle aussi, considérée comme composée de matière et de forme : l’essence de la ligne droite et la ligne droite sont différentes, si l’on suppose que l’essence de la ligne droite serait la dyade. La question des objets mathématiques suscite une interrogation, car il est difficile de rendre compte, par l’activité du sens, de la composition des objets qui résultent d’une abstraction. Autrement dit, il y aurait aussi des composés dans ce qui peut avoir un statut intelligible ; on notera, en outre, que le ἐπεὶ du début s’est transformé en un εἰ : « si est différente l’essence de la ligne droite et la ligne droite » (429b19-20). Enfin, ce même exemple de la ligne droite et de la dyade est donné, en Met. Z 11, 1036b7-20, comme l’illustration d’une mauvaise séparation entre les parties de la forme et les parties du composé ; cette séparation, pratiquée sous des formes différentes par les platoniciens, les conduirait ainsi à poser « une seule forme de plusieurs choses dont la forme est manifestement différente ». De manière plus générale, la question de savoir si la chose composée et sa forme sont identiques ou différentes est longuement développée dans le livre Z des traités métaphysiques, notamment dans les chapitres Z4-6. Or quand Z6 examine si « l’être ce que c’est et chaque chose sont les mêmes ou différents » (1031a15-16), la conclusion de l’examen est que « à partir

DE ANIMA III 4, 429a10-430a9

93

de ces arguments donc, chaque chose et l’être ce que c’est de chaque chose sont un et identiques » (1031b18-20). Le livre Z présente les universels comme des composés de forme et de matière et tend à attribuer à la forme la même fonction dans les composés individuels et dans les composés universels (1033b19-26), alors que, pourtant, les éléments de la définition sont seulement les éléments de la forme, y compris lorsqu’il s’agit du composé (Z 11, 1037a21-b5). On sait enfin qu’Aristote introduit, en Z 11, la notion de matière intelligible (1037a1-5). On peut donc supposer que, plutôt qu’une altérité de l’essence par rapport à la chose composée, c’est l’autre branche de l’alternative que choisirait Aristote, de sorte que ce serait une même faculté, autrement disposée, qui jugerait de la chose et de son essence. Si donc ce qu’il en est de l’intellect suit la manière dont l’essence de la chose est séparable de la matière, selon l’affirmation de 429b21-22, une altérité pure et simple entre l’essence et la chose composée ne pourrait que conduire à une aporie sur la manière dont le νοῦς connaît. Or telle est l’aporie soulevée dans la dernière partie du texte qui examine les difficultés auxquelles on se trouve confronté, si l’on conçoit le νοῦς à la manière d’Anaxagore comme « simple et impassible » (429b22-24). 4. Troisième partie : 429b22-430a9 La troisième partie pose successivement deux questions, issues cependant d’une même théorie de l’intellect, celle d’Anaxagore dont il a déjà été question de manière critique dans la première partie du texte57. Dans la mesure où l’intellect est donné par Anaxagore comme « simple » et « impassible », autrement dit sans rien de commun avec ce qui existe, comment peut-il penser, s’il est vrai que « penser est subir une certaine affection », et s’il faut un élément « commun » entre ce qui agit et ce qui est affecté (429b22-26) ? Autre question issue des mêmes caractéristiques du νοῦς : peut-il se penser lui-même (429b26) ? Comment sera caractérisé le rapport avec ses objets ? Par une identité spécifique ? L’intelligible serait alors une espèce, partagée par différentes instances. 57 En effet, on ne partage pas l’avis selon lequel Aristote reprendrait à son compte les thèses d’Anaxagore qui ont été récusées, selon nos analyses, dans la première partie du texte. Les thèses d’Anaxagore paraissaient faire déjà difficulté sur ce même point en I 2, 405b19-23.

94

ANNICK JAULIN

Ou bien y aura-t-il entre l’intellect et ses objets un mélange identique ? (429b26-29). Pour répondre à ces difficultés, Aristote reprend la position qui fut la sienne (429a22-24) après avoir renversé la thèse d’Anaxagore. Il la précise même en reprenant la thèse des deux états en puissance du νοῦς et montre le contenu de l’usage positif qu’il en fait : le νοῦς n’est rien avant que de penser, ce que l’image de la tablette (allusion à l’ἐκμαγεῖον platonicien dont on a parlé plus haut58) illustre clairement. Si donc le νοῦς n’est rien d’autre que les νοητά (conséquence de la thèse du νοῦς δυνατός), le rapport entre le νοῦς et les νοητά a trouvé sa solution : le νοῦς sera intelligible à la manière dont ses objets sont intelligibles. Dans les composés de matière et de forme, il sera en puissance chacun des intelligibles, mais identique aux intelligibles dans les intelligibles sans matière (430a2-9). Cette réponse à la seconde difficulté repose sur la distinction qui a été faite dans la seconde partie du texte entre une chose et son essence (429b10-22). La solution de l’aporie semble alors parallèle à celle qui est proposée en Met. Λ 9, 1074b38-1075a5 ; elle repose sur l’identité de la science et de son objet dans le cas de ce qui n’a pas de matière, à savoir le τὸ τί ἦν εἶναι. Cette solution semble se distinguer et d’une identité spécifique entre le νοῦς et l’intelligible et d’une identité du mélange entre l’intellect et ses objets, les deux solutions évoquées précédemment en 429b28. Ainsi le νοῦς peut-il se penser lui-même (429b9) seulement quand il est dans le statut d’une ἕξις constituée et en acte sous la forme d’une science. Dans la mesure où sentir que l’on sent (III 2, 425b11-15) et penser que l’on pense (429b9-10) sont des phénomènes secondaires et dépendants de l’actualisation des sens par les sensibles et de l’intellect par les intelligibles, on pourrait dire qu’Aristote a répondu à l’un des problèmes qu’il posait au début du traité (I 1, 402b14-16) : si les actes (ἔργα) sont antérieurs aux parties (μόρια), doit-on étudier les corrélats (ἀντικειμένα) de ces actes avant ces actes mêmes, notamment « l’intelligible avant la faculté noétique ». Il semble bien que l’on puisse affirmer, à partir de là, que la faculté noétique n’a d’autre substance que celle des intelligibles, quand ils sont sans matière.

58

Voir ci-dessus, « Les deux niveaux en puissance du νοῦς ».

FROM DE ANIMA III 4 TO DE ANIMA III 51 Stephen MENN

1. I made a promise a quarter of a century ago, in my paper “Aristotle and Plato on God as Noûs and as the Good”2, that the way of thinking about Aristotle on νοῦς that I was developing in the context of Metaphysics Λ would also shed light on De Anima III 5. Conversations and written exchanges since then, especially with Victor Caston, Michel Crubellier, Klaus Corcilius and Robert Roreitner, have stimulated me to try to deliver on this promise. In this paper I will mainly be working through De Anima III 5, in the context of the problems from De Anima III 4 that I think it is immediately addressing. But I will briefly recall some issues about νοῦς that arise in thinking about Λ, which will help to motivate an investigation of DA III 5. It looks as if Aristotle intends the account of νοῦς in Λ to rest on foundations laid in DA III and especially 1 This paper is a longer version of a talk I gave at the workshop on De Anima III in honor of Michel Crubellier at Lille in March 2016; I also gave related talks at conferences in Berlin in May 2011 and Prague in June 2018, and I thank the participants at all of those meetings for their comments. The paper comes out of a much earlier talk that I gave at a conference organized by Victor Caston at the University of California at Davis in October 2002 on “Mind and Nature”, and its immediate occasion was Victor’s 1999 paper. One important change is that in 2002 I took for granted a construal of De Anima III 4, 430a6-7 which turns out to be highly controversial: I discuss the issue at length below. I remember discussing this issue with Michel when I was in Lille in March-April 2005 on a “European Research Fellowship”, an approved time somewhere in the EU outside Germany as a side-trip from an Alexander-von-Humboldt fellowship – officially I was working on Aristotle’s Metaphysics with André Laks, but I discovered Michel’s dissertation in the library of the UMR, and that showed me what a commentary on the Metaphysics could be. Michel was the first person who thought I might be on to something about De Anima 430a6-7. I am delighted to be able to dedicate this paper to him. 2 S. Menn (1992).

96

STEPHEN MENN

DA III 5. So DA III 5 might help us understand Λ, if only we could understand DA III 5; and seeing how DA III 5 supports Λ 7 and Λ 9 might help us understand what DA III 5 is trying to accomplish. Metaphysics Λ 6 argues that the first moving and acting principle is essentially ἐνεργείᾳ: that it is essentially acting rather than merely possessing the capacity for action, or, more generally, that by its essence it is actually everything that it can potentially be. More sharply, Aristotle says that the οὐσία of the principle is an ἐνέργεια (Λ 6, 1071b20)3, whereas on his opponents’ view the οὐσία of the principle would be a δύναμις, since its action would be merely an accident of some underlying substance whose essence involves only the capacity for action. In the following chapters Aristotle draws consequences for what the principle is and does. If the principle is never potentially anything other than what it is actually, it must be eternally acting, and eternally acting in the same way, eternally moving and ordering the world – against, notably, the Timaeus, which represents the demiurge as originally quiescent and not intervening in the primordial chaos. The mover must also move whatever it moves (at least the outermost heaven) without either needing to be moved itself in order to move something else (as my body or some parts of it must be moved in order to press keys on a keyboard) or being moved as a result of moving something else. This threatens to eliminate most or all of the ways that we might imagine the moving principle as acting on the world: Aristotle seems to allow only that it moves by being desired, and it is desired because it is intellectually grasped [νοητόν], and, presumably, grasped as good. But Aristotle investigates not only how this first principle acts on the world, but also what it is and does in itself. By sometime in Λ 7, by an inferential step that is not made quite explicit, Aristotle is describing the first principle as νοῦς (agreeing with Anaxagoras and Plato4, but with a new justification), and in Λ 9 he investigates what kind of νοῦς it is, what it νοεῖ and how. Because the first principle is not just a power of νοῦς, but a νόησις, an act of thinking, it must be a νόησις of something in particular. For Plato νοῦς performs its act of ordering the world by νοεῖν the forms in the intelligible paradigm, the animal-itself, 3 There is MS disagreement between ἐνέργεια and ἐνεργείᾳ at 1071b22, but at 1071b20 all manuscripts reported by S. Alexandru (2014) have ἐνέργεια. 4 For the identification of the demiurge of the Timaeus with the “νοῦς which has ordered all things” of the Laws and the νοῦς which is “king of heaven and earth” in the Philebus, see S. Menn (1995a), (2002).

FROM DE ANIMA III 4 TO DE ANIMA III 5

97

but Aristotle argues in Λ 9 (briefly and cryptically) that what the first principle νοεῖ must be the first principle itself. Aristotle gives (or suggests) several arguments for this conclusion, but one thing he says is that “the νοούμενον and the νοῦς are not different [in] whatever things do not have matter [i.e. in cases where the νοούμενον has no matter]”, so that “they will be the same, and the νόησις will be one with the νοούμενον” (1075a3-5). Aristotle certainly does not mean that each individual soul is identical with each νοούμενον that it νοεῖ – which would amount to saying that each soul νοεῖ only itself. Rather, his argument is that every science (ἐπιστήμη, 1075a1) or every νόησις (1075a3) is identical with its νοούμενον if that νοούμενον is non-material: so when he says that in such cases the νοῦς is not different from the νοούμενον, νοῦς here must mean a disposition of knowledge or an act of knowing, not a distinct soul or soul-part or soul-power underlying the disposition or act. And the claim that every ἐπιστήμη or νόησις (1075a3) is identical with its νοούμενον if that νοούμενον is non-material is apparently inferred in turn from the more general claim that the ἐπιστήμη or νόησις of X is X itself existing without its matter: if X had no matter to begin with, then this ἐπιστήμη or νόησις is simply X (1074b381075a5). Since the soul as subject of knowledge is not identical with the knowledge it possesses, the knowing subject is also not identical with the object it knows; but the first principle is not a soul possessing knowledge but is pure νόησις by its essence, so that in this case the knowing subject is identical with the known object. Indeed, it is likely that Aristotle is already drawing on this claim about knowledge in Λ 7’s mysterious inference from the premise that the first mover of the heavens is νοητόν (1072a26) to the conclusion that it is νοῦς (by 1072b22-24 at the latest): because this thing, call it X, must be νοούμενον by the heavens in order to impel them to move, and because X is without matter, X would be identical with the νοῦς of X, i.e. with the knowledge whose object is X (Aristotle does say in Λ 7 ὥστε ταὐτὸν νοῦς καὶ νοητόν, 1072b21, although it is not obvious what the argument-context is). Since X is a substance, the knowledge whose object is X is not a mere accident of some other thinking substance, but is rather a substance in its own right5. Λ offers nothing to support the idea that this substance is or has a soul.

5

On the issues in the argument of Λ see S. Menn (2012).

98

STEPHEN MENN

These passages of Metaphysics Λ have close echoes in De Anima III, and especially in De Anima III 5. In III 5 we find a higher kind of νοῦς, which is distinct from the παθητικὸς νοῦς (430a24) and νοεῖ in a different way from it. This higher νοῦς is essentially ἐνέργεια (430a18)6, “it is not the case that at one time it νοεῖ and at another time it does not” (430a22), and it is identical with its object (430a19-20). It is separate and impassible (430a17-18), and it alone is immortal and eternal (430a23). The phrase “it alone” here contrasts it with the παθητικὸς νοῦς (430a24): this παθητικὸς νοῦς is evidently a part or power of the soul, since “the part of the soul by which the soul knows and understands [φρονεῖ]” (III 4, 429a10-11) has its νοεῖν as “either being acted on [πάσχειν τι] by the intelligible, or something else of this kind” (III 4, 429a14-15). But the impassible νοῦς, described in III 5 as an activity of knowledge which is identical with its object, cannot be a soul, or an accident inhering in a soul, any more than the νοῦς of Λ 9 can, unless this soul knows nothing other than itself. (I will come back below to the passage III 5, 430a13-14, which is often held to imply that this νοῦς is a soul or something in a soul.) It is natural to hope that what Aristotle is doing in Λ 7 and Λ 9 will be illuminated by these parallels in De Anima III 5. Indeed, it is natural to hope that they are more than just parallels, that Λ is building on De Anima III and applying it to the heavens and to the problem of understanding the first intelligible principle and its causality, just as Λ also draws on Physics VIII, on the Ethics for pleasure and the divine activity of contemplation, and probably on other treatises too: Λ would be pulling the “high points” or conclusions of different treatises together to give an account of the first principle, and inviting the reader to turn to those treatises to fill in the details. Unfortunately, DA III 5 will help us understand Λ only if we can understand DA III 5. More: while it is reasonable to hope that Aristotle’s analysis of intellection in De Anima III will be applied in his account of 6 Reading τῇ οὐσίᾳ ὤν ἐνέργεια with Förster and Corcilius (I cite Förster from Klaus Corcilius (2017) – Corcilius notes his rare divergences from Förster’s text) and Ross, following two manuscripts (and a correction in a third) and the majority of citations in the Greek commentators; the majority of the manuscripts and a minority of the commentators have τῇ οὐσίᾳ ὤν ἐνεργείᾳ. (See a note above on a similar issue at Metaphysics Λ 6, 1071b22.) But the implications will be the same either way. If S is, by its essence, actually F, then F-ness will not be an accident inhering in S, but rather will be really identical with S. In particular, if S is, by its essence, actually νοοῦν, then the ἐνέργεια of νοεῖν will not be an accident of S but will be really identical with S.

FROM DE ANIMA III 4 TO DE ANIMA III 5

99

the heavenly bodies and their νοητά, and while III 5 seems like the closest point of contact, III 5 can help us understand how the De Anima’s analysis of intellection supports Λ only if III 5 is itself a logical development of Aristotle’s theory of the soul, and not a sudden theological intrusion into an otherwise more autonomous (perhaps “naturalistic”) psychology. If in III 5 Aristotle is simply slipping in a bit of the fifty-drachma course, giving his readers an anticipation of Λ, then it will not help. And that is what it looks like to many readers. Victor Caston’s article “Aristotle’s Two Intellects: a Modest Proposal”7 states the problem sharply. Nothing before III 5 has led us to expect that there will be two intellects (I will for the moment follow Caston’s translation of νοῦς as “intellect”), and yet in III 5 they are certainly two, not just two aspects of the same thing, since “this alone is immortal and eternal” and “this is impassible, whereas the passive νοῦς is corruptible” (430a23-25). More strongly: it is not just that the ποιητικὸς νοῦς (to use the traditional name for what is νοῦς τῷ πάντα ποιεῖν at 430a14-15) cannot be identical with the παθητικὸς νοῦς; it also cannot be a part or power of the human soul. There is nothing absurd in saying that our souls are immortal and ungenerated, or that they have an immortal and ungenerated core. But the ποιητικὸς νοῦς is essentially actual (whether we read τῇ οὐσίᾳ ὢν ἐνέργεια or τῇ οὐσίᾳ ὢν ἐνεργείᾳ at 430a18), with no unactualized potentialities, so that at every moment it is actually doing everything that it ever does or ever can do; “it is not the case that at one time it νοεῖ and at another time it does not” (DA III 5, 430a22). If it were a part of our souls, we would be eternally knowing and contemplating all the intelligible truths that we are ever capable of knowing, which is absurd. Aristotle says elsewhere that “if we have [already the ἕξεις of knowledge of the first principles], it is absurd: for it would follow that we have knowledges more precise than demonstration without noticing it” (Posterior Analytics II 19, 99b2627)8; but if the ποιητικὸς νοῦς were part of our souls, we would even more absurdly have to have the ἐνέργεια of knowledge, and not just the ἕξις, at every instant without noticing it. And since this ἐνέργεια of knowledge is identical with its object, if the ποιητικὸς νοῦς were a part or power of our soul, it would be a knowledge only of that same part or 7 8

V. Caston (1999). Similarly, Metaphysics A 9, 992b33-993a2.

100

STEPHEN MENN

power of our soul. De Anima III 4 says that “what is called the νοῦς of the soul [ὁ καλούμενος τῆς ψυχῆς νοῦς] – I am calling νοῦς that by which the soul reasons and affirms – is none of the beings in ἐνέργεια until it thinks / knows them” (429a22-24), and that “it has no nature except this, that it is δυνατόν” (429a21-22)9; this is clearly the παθητικὸς νοῦς, not the ποιητικὸς νοῦς which is essentially ἐνέργεια, and which is not the νοῦς of the soul but a cause which acts on the νοῦς of the soul. Caston’s view, then, is that the ποιητικὸς νοῦς is not a human intellect but a divine intellect – the intellect of a divine soul, or itself a divine soul that is pure intellect, without nutritive or sensitive powers. And, even though Aristotle had said nothing in III 4 (or anywhere earlier in the De Anima) about such a divine intellect, there would be nothing absurd in his introducing such an intellect in III 5, if the analysis of the human intellect in III 4 leads to the discovery of some phenomenon which would justify an inference to a divine intellect as its cause. But, as Caston argues, there is no such explanatory gap: “the tasks which commentators have invented for the Agent Intellect to fill – such as abstraction, selective attention, or free choice – are factitious. They are not problems Aristotle even acknowledges; a fortiori, they cannot be the reasons he appeals to for the existence of a second intellect” (Caston p. 200). Caston asks whether, if III 5 had somehow dropped out of the manuscript tradition of the De Anima, we would notice that anything was missing, and his answer, apparently, is no. He thinks that III 5 is doing something important in setting the De Anima’s analysis of the human intellect in a cosmological and theological comparative context (“cosmological” because he takes the divine intellect of III 5 to be the mover of the outermost sphere), showing how far we attain and how far we fall short of the divine ideal of knowledge, but this is an interlude where Aristotle steps outside the natural-philosophical project of the De Anima to comment on the project from the standpoint of first philosophy, “before turning back again to his main work”: “in the chapters both before and after De Anima 3.5, there is virtually no trace of the second intellect. It simply plays no role in the details of Aristotle’s psychology” (Caston p. 216). 9 For the MSS reading δυνατόν, followed by Förster but not Corcilius (who has δυνατός like Ross), see below; it makes no difference for the present issue.

FROM DE ANIMA III 4 TO DE ANIMA III 5

101

I am not convinced by Caston’s “minimalist” reading of III 5, and can take issue with him on many details (perhaps most importantly, I don’t agree that the ποιητικὸς νοῦς is introduced as a final rather than efficient cause), but that does not really take away the force of his challenge. And if he is even broadly right, then we would not expect De Anima III 5 to give much help in understanding Metaphysics Λ: it would be an application of Λ, rather than Λ being an application of it. I agree that the ποιητικὸς νοῦς cannot be a human intellect or a part or power of the human soul, and I also agree that there is no clear explanatory gap in human psychology that Aristotle cites it as filling. I do not agree that, if III 5 were missing from our manuscripts, the rest of Book III would make enough sense without it that we would not suspect anything was missing. Most obviously, III 6, 430b24-26, in a discussion of indivisibles and how we cognize them – in some cases we know them as privations through their positive contraries, but “if there is some cause which has no contrary, it knows itself and is in actuality and separate” – would be bizarre and unintelligible without III 5 preceding10. But that does not help us on the question: what is Aristotle’s justification for positing the ποιητικὸς νοῦς when he does? Is there something in the argument of III 4 that supports it? I want to say that Aristotle introduces the ποιητικὸς νοῦς in III 5, not as a cause explaining some phenomenon in human psychology which he has discovered in III 4 (he does think it’s an efficient cause, but that’s not his justification for positing it), but rather as helping solve an aporia that arises in the course of III 4. But to see how it can help, we need to challenge the assumption, made by Caston but also by a broad range of other scholars, that if the ποιητικὸς νοῦς is not a part or power of the human soul, then it must be a part or power of a divine soul, or itself a divine soul 10 DA III 6, 430b24-26: Εἰ δέ τινι μηδὲν ἔστιν ἐναντίον τῶν αἰτίων, αὐτὸ ἑαυτὸ γινώσκει καὶ ἐνεργείᾳ ἐστὶ καὶ χωριστόν. This is not to say that the passage is easy even under current conditions. I would think it is clear enough that the subject of “knows itself and is in actuality and separate” is the indivisible cause without a contrary: the harder questions are what Aristotle’s argument for this conclusion is, and why he finds it helpful to say it here. But Corcilius, with among others E. Berti (1978, 1996), takes the subject to be the cognizer (“wenn einer der Ursachen jedoch nichts entgegengesetzt ist, erkennt (das Erkennende) sich selbst und ist wirklich und getrennt”). The disagreement at least shows that the passage is difficult, even with III 5 in our manuscripts and editions. I discuss this passage and its place in Aristotle’s larger argument in a talk, “The Program of De Anima III”, that I gave at a conference organized by Robert Roreitner in Prague in June 2018. I hope to return to these issues.

102

STEPHEN MENN

which has no other parts or powers. Aristotle says nothing here to suggest that this νοῦς is a soul, or something that exists only as an accident of a soul. As we have seen, Aristotle says of “what is called the νοῦς of the soul” that “it has no nature except this, that it is δυνατόν” (III 4, 429a2124), and he says nothing to restrict this to human souls11. There is no reason to think that the ποιητικὸς νοῦς is the νοῦς of any soul at all, just as there is no reason to think that the νοῦς which Anaxagoras and Plato (in the Philebus and Laws) posit as the ordering cause of the cosmos is a soul or an accident of soul, although for Plato of course souls participate in it. As a matter of Greek, νοῦς can very easily be a name for an intellectual virtue, akin to (perhaps sometimes synonymous with) ἐπιστήμη or φρόνησις: this usage is common in Plato, and it is in Aristotle notably in Nicomachean Ethics VI, and Posterior Analytics II 19. I argued in Plato on God as Noûs that in Plato and also in Anaxagoras the world-ordering νοῦς is such an intellectual virtue, existing as a substance in its own right and participated in by other things, and I argued in “Aristotle and Plato on God as Noûs and as the Good” that the νοῦς of Metaphysics Λ is also such a self-subsisting virtue. It is of course completely unsurprising, if νοῦς is an intellectual virtue, that Plato would think that it subsists by itself apart from souls, although perhaps it is a bit more surprising that he would think of such as self-subsisting virtue as an efficient cause of cosmic order. It is more surprising that Aristotle would accept such a selfsubsisting virtue, but, again as I argued in “Aristotle and Plato on God as Noûs and as the Good,” he has no in-principle reason why he would not. He has reasons for thinking that the moral virtues do not exist apart from souls – indeed, not apart from irrational souls and mortal bodies – for an intellectual virtue, the knowledge of X, it will depends on what X is. If X is said like the snub, then the knowledge of X cannot exist apart from souls – indeed, not apart from φαντασία, which cannot exist apart from the senses, which cannot exist apart from their bodily organs. But if X is a substance existing separately from matter, then Aristotle is committed to the claim that the knowledge of X is simply X, and so it is a self-subsisting substance, not an accident depending for its existence on whatever souls 11 For the phrase “the νοῦς of the soul” compare Theophrastus Fragment 307B Fortenbaugh-Sharples-Huby-Gutas (Priscianus, Metaphrasis 26, 5-7), contrasting ὁ ψυχικὸς νοῦς with ὁ ἐνεργείᾳ νοῦς, τουτέστι ὁ χωριστός (but it is not clear how much of this may be Priscian's gloss). At Aristotle’s Metaphysics α 1, 993b10-11, it is less clear what the contrast is with.

FROM DE ANIMA III 4 TO DE ANIMA III 5

103

may possess this knowledge. But this does not tell us the justification for positing such an immaterial substance in the first place. 2. As I have said, I think Aristotle introduces the ποιητικὸς νοῦς, not as a cause posited to explain some effect in the soul, but to help solve aporiai that arise in the course of III 4. So we have to go back to III 4 to see how the aporiai arise, and then work through III 5 to see how it helps solve them. I think that the text can and should be read continuously – the chapter-division is of course not part of Aristotle’s text, and in my view would be better placed at 429b22, with the aporiai beginning a new chapter – and that if we do this correctly, III 5 will not seem like a surprise or an intrusion. If III 5 were not there, the solution to the aporiai, begun in III 4, would be radically incomplete. De Anima III 4 divides into two main parts. In 429a10-b22, Aristotle gives a sketch of the soul’s power of νοῦς and activity of νοεῖν, based chiefly on an extended comparison with sensation; in 429b22-430a9, he raises two aporiai about the relation between νοῦς and the νοητόν, and begins to solve these aporiai, but the solution, in particular the solution to the second of these aporiai, is not finished until the end of III 5 at 430a25. My main concern here is with Aristotle’s posing and solving of the aporiai in 429b22-430a25, but the aporiai make sense only against the background of Aristotle’s basic principles about cognition. He has first laid out these principles in developing his theory of sensation in Book II, and he takes them up again in the first part of III 4, 429a10-b22, in developing a theory of νοῦς by drawing on what he has said about sensation especially in II 5 and II 12, and bringing out the similarities and the differences between sensation and νοῦς. Aristotle states his basic theses about νοῦς, by analogy with what he has said about sensation, in two sentences near the beginning of III 4: Eἰ δή ἐστι τὸ νοεῖν ὥσπερ τὸ αἰσθάνεσθαι, ἢ πάσχειν τι ἂν εἴη ὑπὸ τοῦ νοητοῦ ἤ τι τοιοῦτον ἕτερον. ἀπαθὲς ἄρα δεῖ εἶναι, δεκτικὸν δὲ τοῦ εἴδους καὶ δυνάμει τοιοῦτον ἀλλὰ μὴ τοῦτο, καὶ ὁμοίως ἔχειν, ὥσπερ τὸ αἰσθητικὸν πρὸς τὰ αἰσθητά, οὕτω τὸν νοῦν πρὸς τὰ νοητά. If νοεῖν is like sensing, it would either be being affected by the intelligible object, or something else similar to this. So [νοῦς] must be unaffectable but receptive of the form, and potentially such [as the object] – not potentially

104

STEPHEN MENN

this [object]; and νοῦς must be to the intelligible objects as the sensitive [power] is to the sensible objects. (429a13-18)

He is here drawing on the thesis of II 12 that “sense is what is receptive of sensible forms without the matter, as the wax receives the sign of the signet ring without the iron or the gold” (424a17-20). He also uses the corollary that this kind of receptivity implies that the recipient is neutral with respect to the contrary qualities of the objects: plants, although they have souls and although they are affected by heat and cold, do not sense heat and cold, and “the reason is that they do not have a mean and a principle such as to receive the forms of sensible things, but only such as to be affected by them together with the matter” (II 12, 414b1-3; so too vision requires that the medium [De Anima II 7] and the organ [the pupil, De Sensu 2] be transparent and thus neutral). Now, in III 4, Aristotle makes more explicit than he had in Book II why the recipient must be neutral: the soul’s νοῦς, “since it νοεῖ all things, must be unmixed, as Anaxagoras says, in order to dominate, that is, in order to know – παρεμφαινόμενον γὰρ κωλύει τὸ ἀλλότριον καὶ ἀντιφράττει – so that it has no nature except this, that it is δυνατόν” (III 4, 429a18-22). I have left the phrase explaining the reason for the neutrality in Greek because its meaning and construal are disputed. Ross renders the phrase as “for the intrusion of anything foreign to it interferes with it”: in other words, he takes τὸ ἀλλότριον, modified by the participle παρεμφαινόμενον, to be the subject of the verbs κωλύει and ἀντιφράττει12. But all ancient and medieval commentators that I have checked take παρεμφαινόμενον as the subject and τὸ ἀλλότριον as the object of the verbs, and a closer look at the meaning of παρεμφαίνειν and the parallel contexts of its use shows that they must be right13. In the Timaeus, the receptacle must be “unshaped by all those forms which it is going to receive from anywhere: for if it were similar to any of the things that enter into it, then when things of a contrary or entirely different nature come to it, it would not receive their likenesses well, since it would display its own 12 This translation is from Ross’ analysis p. 290 of his editio maior, W.D. Ross (1961). Ross’ construal of the syntax is explicit in his commentary, p. 292, and is shared by Corcilius. 13 So at least Alexander De Anima 84, 14-18, Themistius In de Anima paraphrasis 94, 20-24, Averroes Commentarium magnum in Aristotelis de Anima libros (F. Stuart Crawford (1953)), Book III, text 4, lines 22-5 (p. 384), Thomas In de Anima #680. I am less sure that everyone agrees that the subject implicitly modified by παρεμφαινόμενον is νοῦς (see below for the gender issue); but there is really nothing else it could be.

FROM DE ANIMA III 4 TO DE ANIMA III 5

105

appearance alongside them [τὴν αὑτοῦ παρεμφαῖνον ὄψιν]” (50d7-e4); less metaphysically, in the Aristotelian Problemata, “water is more transparent [or more reflective] than olive oil: for olive oil has color, whereas water, being displayed without color alongside [ἄχροον παρεμφαινόμενον], makes the image clearer” (XXIII 9, 932b22-4)14. In both of these texts, an object is being displayed or imaged in some medium, and the medium is also displaying some quality of its own alongside the object – this is what παρεμφαίνειν means – and the more the medium displays a quality of its own, the worse it will display the object (in the Timaeus, this happens particularly if the quality of the object is opposed to the quality of the medium). This is also what Aristotle is saying in the De Anima passage, and indeed he must be deliberately echoing the Timaeus passage. Thus, in the De Anima passage, παρεμφαινόμενον γὰρ κωλύει τὸ ἀλλότριον καὶ ἀντιφράττει must mean that the soul’s νοῦς must not itself have any determinate nature, “because if it were itself displayed alongside its objects, it would hinder and block what is of a different character”15. Aristotle is here making some allusions to earlier philosophers, interpreting Anaxagoras (“in order to dominate, that is, in order to know”) as referring to a cognitive power in the soul rather than to an ordering principle of 14 “Image” here is ἔμφασις: if the medium itself is visible, it παρεμφαίνει something παρὰ the ἔμφασις of the object. 15 Ross greatly weakens the force of παρεμφαίνειν. Ross says, editio maior p. 292, that τὸ ἀλλότριον is “plainly” the subject of the verbs. Perhaps he does not have the parallels in mind (although Hicks cites them) and does not consider the possibility that νοῦς is itself the antecedent of παρεμφαινόμενον. But he may be consciously rejecting this possibility on the ground that the gender would have shifted from the masculine accusative ἀμιγῆ at 429a18 (picking up τὸν νοῦν at a17) to the neuter presupposed by παρεμφαινόμενον at a20. But the gender has indeed shifted: δυνατόν at 429a22 (transmitted by all manuscripts) certainly refers back to νοῦς (and is naturally also taken as picking up παρεμφαινόμενον), and we should not correct it to δυνατός with Ross (now followed by Corcilius, in one of his rare departures from Förster). Hicks prints δυνατόν and construes as I do; see his extended note, R.D. Hicks (1907), p. 478-479. (Robert Roreitner’s Humboldt-Universität Berlin dissertation, The Unmoved Causes of Receptivity: Sense-Perception and Thinking as Passive Activities in Aristotle’s De Anima (R. Roreitner (2018)), has convinced me that I was too hasty in my interpretation of this passage. I still think the syntax has to be what I said it was. But it is possible that τὸ ἀλλότριον means not what is alien in the sense of “contrary to the nature of the subject” but simply what is alien in the sense of “belonging to the object rather than to the subject.” So perhaps, if the subject were F, that would prevent it from being acted on by, and thus from perceiving, the F-ness that belongs to the object. In any case Aristotle’s considered opinion is that the form through which νοῦς cognizes F is the same form through which it cognizes the contrary of F, if F has a contrary.)

106

STEPHEN MENN

the cosmos16. At the same time, he assimilates Anaxagoras’ νοῦς, not to the demiurge of the Timaeus, but to the receptacle. This assimilation is eased by the fact that, a few lines further down (50e8-51a1), the Timaeus compares the receptacle to a smooth surface for impressing shapes, which would suggest a human cognitive power, the wax tablet of the Theaetetus (Timaeus 50c2 calls the receptacle an ἐκμαγεῖον, the same word used in the Theaetetus for the wax tablet): Aristotle must be thinking of the Theaetetus when further down in De Anima III 4 he compares the soul’s νοῦς, prior to any act of νοεῖν, to “a tablet in which nothing is present written in actuality” (430a1-2). But beyond any literary resonances between Anaxagoras and the Timaeus and Theaetetus, the deeper point is that in the physical and epistemological cases alike we must posit a principle, the receptacle / matter or what receives forms in the soul, which must be distinct from ordinary objects, having no features in common with them and no distinctive features of its own, in order to receive their forms. And this idea of neutrality helps to explain Aristotle’s saying that νοῦς “must be unaffectable [ἀπαθές] but receptive of the form” (429a15-16). This sounds strange, since receiving the form of an object sounds like a way of being affected by the object, and it sounds especially strange as an inference from the previous sentence, that “if νοεῖν is like sensing, it would either be being affected [πάσχειν τι] by the intelligible object, or something else similar to this” (429a13-15). But presumably the point is that, like Anaxagoras’ νοῦς (cited as ἀπαθές at DA III 4, 429b22-23) and like the receptacle, the soul’s νοῦς, and equally the sensory powers, must undergo no alteration or change of intrinsic quality in interacting with their objects, so that they can remain receptive to all objects equally17. So when the form of the object is received in a cognitive power, it must not be present in it in the normal way that a form is present in matter: perhaps something like the way that the color of an object is 16 He cites the same tag of Anaxagoras – νοῦς must be unaffectable and unmixed, in order to dominate – as a point about unmoved movers, Physics VIII 5, 256b24-27. But perhaps Aristotle is right that Anaxagoras’ implicit argument in B12 is that if νοῦς had intermixed with it a share of the ingredients of the cosmos, that would prevent it from perceiving the low concentrations of these ingredients in the precosmic mixture (as the intermixture of gold in our bodies prevents us from perceiving the low concentrations of gold in our surroundings), and so would prevent it from dominating the mixture in such a way as to separate these ingredients from it. 17 See DA II 5, where exercising the sensory powers, like exercising an art, is “either not alteration ... or a different kind of alteration” (417b6-7): surely the artisan’s quality, i.e. his art, has not changed simply by being exercised. But a sensory power need not be neutral to all qualities, only to those qualities it can sense.

FROM DE ANIMA III 4 TO DE ANIMA III 5

107

present in a mirror, or in an actually transparent medium. This does not mean that the bodily organ of a sensitive power is not qualitatively changed in sensing – at least sometimes it certainly is, as when we are dazzled by a bright light (DA III 4, 429a29-b5) – but this is not what it is to sense the object, and if the organ is altered too much it will interfere with sensation (the sensory power is not itself qualitatively affected, although its exercise may be blocked when the organ is affected, DA I 4, 418b18-24). Aristotle starts in DA III 4 from the similarities between νοεῖν and sensation, but he also points to two important differences. First, “the ἀπάθεια of the sensitive and of the intellectual [powers] is not alike” (429a29-30), just because the sensory powers have organs which can be affected, whereas νοῦς has no bodily organ – “the sensitive [power] is not without a body, but [νοῦς] is separable” (429b4-5) – and so its exercise cannot be blocked by an alteration of the body. Second, when the intellectual power has been exercised in such a way that we acquire a ἕξις of knowledge [ἐπιστήμη], we can then exercise that ἕξις in contemplating [θεωρεῖν] without any need of an external object to exercise it on, whereas to exercise sensation an appropriate external object and medium are necessary and sufficient. Thus the actualization of νοῦς takes two steps, from the power to the ἕξις and from the ἕξις to the activity of contemplation, while the actualization of sensation takes only one step, from the power to the activity, which depends on the presence of the object (for all this compare DA II 5, 417b2-27 with III 4, 429b5-9)18. Both of these points about the differences between νοεῖν and sensation will need qualification. Some of the qualification comes in Aristotle’s paragraph on the relation between cognizing flesh and cognizing the essence of flesh, DA III 4, 429b10-22, which I will pass over for now; the paragraph is important, and I will come back to it later, but Aristotle’s formulations of the aporiai about νοῦς and the νοητόν at 429b22-29 do not depend on it. 3. Aristotle raises two aporiai. First, if νοῦς is, as we have said following Anaxagoras, “simple and unaffectable and has nothing in common” with anything else, how will it νοεῖν, if νοεῖν is something like being 18 This holds whether or not we accept the Bywater-Ross emendation of δὲ αὑτόν at 429b9 to δι᾿ αὑτοῦ as in b7, not accepted by Förster or Corcilius.

108

STEPHEN MENN

affected – as we have also said, and as seems to follow if it is receiving the form of the object (429b22-25)? And there seems a special difficulty in νοῦς’ being affected by the νοητόν, since it seems that X can be affected by Y only if X and Y share something in common (most obviously by belonging to the same genus), and we have said that νοῦς has nothing in common with any of its νοητά (thus 429b25-26). Aristotle seems to have the resources to handle this aporia, but it leads to a second and deeper aporia. We have said that νοῦς is simple and has nothing in common with any of its νοητά. But it seems that νοῦς can νοεῖν itself, since we must cognize it somehow if we can talk about it, and we are certainly not grasping it by sensation; De Anima III 2 allows even sight to see that it sees yellow, since the sight’s act of seeing the yellow and the yellow’s act of moving the sight are a single act of the agent and patient and so are cognized simultaneously, and surely the same argument should show that νοῦς cognizes itself cognizing the νοητόν. But if νοῦς can νοεῖν itself, then νοῦς is itself something νοητόν. But then either its being νοητόν is an additional attribute, distinct from its own nature, which it shares with the other νοητά (or, to put it the other way around, its being νοῦς is an additional attribute, distinct from the common nature which it shares with the other νοητά), in which case νοῦς is “mixed” or composite, contrary to what we have assumed; or else being νοῦς and being νοητόν are the same thing, so that “νοῦς will belong to the other [νοητά] as well, and the νοητόν will be one in species” (429b26-29, this quote b27-28), although it seems absurd that everything that is thinkable and knowable should be itself thinking and knowing. In answer to the first aporia, which asked how νοῦς will νοεῖν without being affected by its object, or how it will be affected without having anything in common with its object, Aristotle answers briefly that in a certain sense it is affected, and in a certain sense it does have something in common with its objects, since “νοῦς is in a way potentially the νοητά, but actually none of them [or: actually nothing] before it νοεῖ” (429b30-31, cf. 429a21-24); “the way it is potentially is as in a writingtablet in which nothing actually written is present, which is what happens in the case of νοῦς” (429b31-430a2), like the wax tablet of the Theaetetus, and like the material principle of bodies. So the action of the νοητόν on the soul’s νοῦς does not depend on their having any common predicate in the same way: rather, the νοῦς has potentially the same predicate that the νοητόν has actually. However, although this is all that

FROM DE ANIMA III 4 TO DE ANIMA III 5

109

Aristotle says here, it is not precise enough: for it is also true that when fire acts on some matter which currently has the form of earth, and turns it into more fire, the patient (at the beginning) has only potentially the predicate which the agent has actually; and yet if the νοῦς and the νοητόν were related as the earth and the fire, there would not be a sufficient answer to the aporia. Although earth is only potentially hot and fire is actually hot, earth and fire belong to the same genus (“simple corruptible body”, or more broadly “corruptible body”): it is because they both belong to this genus that they are susceptible to the contrariety hot / cold, everything in this genus being actually cold and potentially hot or vice versa, and the different things in the genus can interact by heating and cooling each other. If νοῦς and the νοητόν were related in this way, then, contrary to Aristotle’s assumptions, they would have something in common in the most straightforward sense, and νοῦς would be affected in the most straightforward sense; also, the νοητόν would not be an unmoved mover, but could be reciprocally affected by acting on the νοῦς, as a hot body can be cooled in heating a colder body. So when Aristotle says that “νοῦς is in a way potentially the νοητά” (my stress), it must be in a different way from the way that earth is potentially fire, or is potentially hot. One model for how νοῦς and the νοητόν could have the same predicate in different ways, or could be the same thing present in different modes, is given by the case of an art and the matter that the art acts on: this is elsewhere Aristotle’s standard model for an agent and patient not sharing a genus, or not sharing (the same kind of) matter, and therefore for action without reciprocal action. This is certainly one model he has in mind here too, but it cannot be fully adequate. The locus classicus is from On Generation and Corruption I 7: The same account holds for acting and being acted on as for being moved and moving. For “mover” is said in two ways: that in which the principle of motion exists is said to be the mover, and so is the last thing, the thing proximate to the thing moved and the coming-to-be. So likewise with “agent” [or “maker”]: for we say both that the doctor is what heals and that the wine is. So nothing prevents the first mover in a motion from being unmoved (and in some cases this is even necessary), whereas the last [mover] always moves by being itself moved; and so too in action the first [agent] is unaffected, but the last is itself affected. For those [agents] which do not have the same matter [as their patients] act without being affected (like the art of medicine, which in producing health is in no way affected

110

STEPHEN MENN

by the person who is being healed); but the food is also an agent [of health] and it is affected (for it is heated or cooled or affected in some other way at the same time that it acts). Here the art of medicine is [the agent] as the principle, and the food as the last [agent] and as the thing in contact [with the patient]. So those agents which do not have their form in matter are unaffected, whereas those which are in matter [scil. the same kind of matter as the patient] are subject to affection. (GC I 7, 324a24-b6)

Here the first agent in healing a person is not the doctor, but the art of medicine which is present in the doctor: the doctor belongs to the same genus and has the same kind of matter as the sick person, and he must be moved at least incidentally in order to heal (he must move his limbs or at least his mouth, etc.), but the art of medicine, which does not belong to the same genus and does not have matter, remains unaffected. Furthermore, while we can say that the art of medicine, or the doctor as its bearer, possess something actually that the sick person possesses only potentially, namely the form of health, they do not possess it in the same way that the sick person (once healed) will possess it; this is why the art, or the doctor qua bearer of the art, do not risk losing the form of health in the act of healing the sick person, as the fire risks losing its heat in the act of heating the earth. As will become clear a few lines further down in the De Anima (III 5, 430a10-14, to be discussed below), Aristotle has the art model in mind in thinking about the relation between the soul’s νοῦς (analogous to the person being healed) and its νοητόν (analogous to the art of medicine), and this avoids many difficulties of the earthand-fire model. But this model cannot be exact, since the person being healed is in no way ἀπαθές and receives the form of health in the full straightforward sense, whereas νοῦς is somehow ἀπαθές and receives the form of its object in some less straightforward way. One feature that distinguishes νοῦς both from the earth and from the sick person is that νοῦς, like a blank writing-tablet, is “actually none [of the νοητά] before it νοεῖ”: when it passes from not-knowing to knowing X, in the prior state it is neither X nor the contrary ¬X19. So the transition to knowing X is not a transition from ¬X to X; indeed, since the knowledge of contraries is the same, the end-state of the transition is no more X than it is ¬X. Now of course not every νοητόν has a contrary: for instance, one necessary condition for X to have a contrary is that X 19

I will use “¬” as a sign for the contrary rather than for the contradictory.

FROM DE ANIMA III 4 TO DE ANIMA III 5

111

should exist in some ὑποκείμενον which is capable of receiving both X and its contrary. But in cases where the νοητόν X does have a contrary, then when the νοητόν comes to be present in the νοῦς, it does not come to be present in it in the same way that it is present in its ὑποκείμενον, since the νοητόν and its contrary cannot both at the same time be present in something in the way that they are present in their ὑποκείμενον. Nonetheless, Aristotle insists that S’s knowing X is X’s being in some way present in S. This seems to hold for ascriptions of knowledge both in the ἕξις-sense and in the ἐνέργεια-sense. S’s ἐνέργεια of intellectually knowing X, like S’s ἐνέργεια of sensing X, is also X’s ἐνέργεια in S, since the passive ἐνέργεια of the patient (here the knower S) is also the active ἐνέργεια of the agent (here the object X) in the patient; and for S to have ἕξις-knowledge of X is for X to be present in S in such a way as to enable it to operate there. But in the present passage Aristotle says nothing further to make clear the distinctive way in which νοῦς is potentially the νοητά, or the way in which they come to be in νοῦς. More clarity may emerge in his treatment of the second aporia. Having briefly handled the first aporia (429b29-430a2), Aristotle turns to the second aporia: can νοῦς νοεῖν itself, so that it is itself something νοητόν? If so, how do we avoid the conclusions that, if being νοῦς and being νοητόν are different, then νοῦς is composite, and that, if they are the same, then everything that is νοητόν also νοεῖ? Aristotle gives at least some of the answer in a compressed text, 430a2-9. I give the Greek and then how I would translate it: there are no serious issues in the constitution of the text (as there will be in III 5), but there is at least one serious dispute about the construal, as well as a dispute about the meaning of the fundamental distinction into two cases that Aristotle is drawing in the passage. καὶ αὐτὸς δὲ νοητός ἐστιν ὥσπερ τὰ νοητά. ἐπὶ μὲν γὰρ τῶν ἄνευ ὕλης τὸ αὐτό ἐστι τὸ νοοῦν καὶ τὸ νοούμενον, ἡ γὰρ ἐπιστήμη ἡ θεωρητικὴ καὶ τὸ οὕτως ἐπιστητὸν τὸ αὐτό ἐστιν (τοῦ δὲ μὴ ἀεὶ νοεῖν τὸ αἴτιον ἐπισκεπτέον)· ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν.20 ὥστ’ ἐκείνοις μὲν οὐχ ὑπάρξει νοῦς (ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων), ἐκείνῳ δὲ τὸ νοητὸν ὑπάρξει. 20 I am punctuating slightly differently from the editions I have checked, keeping Ross’ parentheses around τοῦ δὲ μὴ ἀεὶ νοεῖν τὸ αἴτιον ἐπισκεπτέον and his colon after, to try to preserve the balance between the clauses ἐπὶ μὲν γὰρ τῶν ἄνευ ὕλης and ἐν δὲ τοῖς ἔχουσιν ὕλην. I don’t think there is any substantive disagreement here.

112

STEPHEN MENN

And [νοῦς] itself is νοητός, as the νοητά are. For in [ἐπί + gen.] things that are without matter the νοοῦν and the νοούμενον are the same thing, for theoretical knowledge and what is known in this way are the same (but we must investigate the cause why we do not [it does not? there is not?] always νοεῖν); but in [ἐν] things that have matter, [the νοῦς] is potentially each of the νοητά, so that νοῦς will not belong to them [= the νοητά] (for the νοῦς of such things is a δύναμις without matter), but the νοητόν [= being νοητόν] will belong to it [= νοῦς]. (430a2-9)

Aristotle here answers the question of the second aporia in the affirmative, by saying that νοῦς is itself something νοητόν, and tries to show how to avoid the dilemma of 429b27-29. He does not want to say that νοῦς is “mixed”, of a substrate by which it is νοῦς and an attribute by which it is νοητόν, or of a genus by which it is νοητόν and a differentia by which it is νοῦς, but he also wants to avoid the absurd conclusion that νοῦς belongs to all the νοητά, i.e. that all νοητά are themselves intelligent. It is, or should be, clear that he proceeds in showing how to avoid this absurd conclusion by distinguishing two cases, depending on whether the νοητά “are without matter” or “have matter”: ἐπὶ μὲν γὰρ τῶν ἄνευ ὕλης at 430a3 introduces the first case, and the answering ἐν δὲ τοῖς ἔχουσιν ὕλην at 430a6 takes up the second case, and ἐπί and ἐν come to the same thing (as noted by Hicks on 430a3), meaning “in the case of”21. Where Aristotle puts the dividing line between νοητά without matter and νοητά that have matter has been more controversial. I say “it is, or should be, clear” because, although Ross sees what should be the obvious structure of the passage in his paraphrase in his editio maior (W.D. Ross (1961), p. 291) and in his commentary (p. 295), he apparently does not see it in his introduction, p. 41, where he paraphrases: “To the second question he answers that reason is knowable just as its objects are: theoretical knowledge and its objects are one. Its objects do not possess reason (since reason is a faculty unaccompanied by matter, while they have matter in them); but it possesses knowability”. Ross apparently intends all of this to apply to all kinds of νοητά, but Aristotle says “theoretical knowledge and its objects are one” only in the case of objects without matter, and “[reason’s] objects do not possess reason” only in the case of objects that have matter – there are no objects of 21 And Ross editio maior p. 295 says “in the case of” both times, although this causes trouble both for his construal of 430a6-7 and for his interpretation of III 5, 430a13-14; see discussion below.

FROM DE ANIMA III 4 TO DE ANIMA III 5

113

which Aristotle is making both assertions. Indeed, in the case of objects without matter, where Aristotle says “the νοοῦν and the νοούμενον are the same thing, for theoretical knowledge and what is known in this way are the same” (430a3-5), he cannot possibly say that the νοητά do not themselves have νοῦς – he has just said the contradictory. Aristotle deals with the argument of the aporia that, if νοῦς is νοητός and is unmixed, the νοητά will themselves have νοῦς, by distinguishing two cases. In the case of νοητά that are without matter, he simply admits the conclusion, but will presumably try to show that in this case it is not absurd; in the case of νοητά that have matter, he will deny the conclusion, “so that νοῦς will not belong to them” (430a7). The main controversial issues will be (i) where Aristotle puts the dividing line between the two cases (that is, what counts as “being without matter” or “having matter”), and (ii) what he is saying at 430a6-7 ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν (the syntax is disputed) and how it is supposed to justify the conclusion at 430a7-8 that these things do not have νοῦς. There will also be issues about how much of the solution of the aporiai he is deferring here, and where and how he comes back to it. Alexander, paraphrasing our passage in his own De Anima, says “the essences and forms of composite things νοῦς makes intelligible to itself in separating them from the things with which they have their existence. But if there are some forms, like the things existing by themselves [ὡς τὰ καθ’ αὑτά], separate from matter and from any substratum, then these are intelligible in the strict sense [κυρίως], since they have being-such [i.e. being intelligible] in their own nature rather than acquiring it by the help of what thinks them” (87, 24-28)22. It is clear from comparing 22 I’ll give his full paraphrase of 430a2-9: Τὰ μὲν οὖν τῶν συνθέτων τὸ τί ἦν εἶναί τε καὶ τὰ τούτων εἴδη ὁ νοῦς αὑτῷ νοητὰ ποιεῖ χωρίζων αὐτὰ τῶν σὺν οἷς αὐτοῖς τὸ εἶναι. εἰ δέ τινά ἐστιν εἴδη, ὡς τὰ καθ’ αὑτά, χωρὶς ὕλης τε καὶ ὑποκειμένου τινός, ταῦτα κυρίως ἐστὶ νοητά, ἐν τῇ οἰκείᾳ φύσει τὸ εἶναι τοιαῦτα ἔχοντα, ἀλλ’ οὐ παρὰ τῆς τοῦ νοοῦντος αὐτὰ βοηθείας λαμβάνοντα. τὰ δὲ τῇ αὑτῶν φύσει νοητὰ κατ’ ἐνέργειαν νοητά, δυνάμει γὰρ νοητὰ τὰ ἔνυλα. ἀλλὰ μὴν τὸ κατ’ ἐνέργειαν νοητὸν ταὐτὸν τῷ κατ’ ἐνέργειαν νῷ, εἴ γε ταὐτὸν τὸ νοούμενον τῷ νοοῦντι. τὸ ἄρα ἄυλον εἶδος νοῦς ὁ κυρίως τε καὶ κατ’ ἐνέργειαν. καὶ ὁ νοῶν ἄρα τοῦτο νοῦν νοεῖ οὐ γινόμενον νοῦν ὅτε νοεῖται, ὡς ἐπὶ τῶν ἐνύλων εἰδῶν ἔχει, ἀλλὰ ὄντα νοῦν καὶ χωρὶς τοῦ ὑπὸ τοῦ νοῦ νοεῖσθαι. εἰ δὴ ὁ νοῶν νοῦς ἐν τῷ νοεῖν ὃ νοεῖ γίνεται, καὶ ὁ ἐν ἡμῖν νοῦς, οὗτος δέ ἐστιν ὁ καθ’ ἕξιν, ὅταν ταῦτα τὰ εἴδη νοῇ, ὁ αὐτὸς ἐκείνοις τότε γίνεται· ὁποῖα γάρ ἐστι κατὰ τὴν οἰκείαν ὑπόστασιν καὶ χωρὶς τοῦ νοεῖσθαι (ἁπλᾶ γὰρ τοιαῦτα), καὶ ἐν τῷ νοοῦντι αὐτὰ γίνεται. ὥσθ’ ὁ ταῦτα νοῶν νοῦς ὁ αὐτὸς αὐτοῖς,

114

STEPHEN MENN

Alexander’s whole paraphrase (87, 24-88, 16) with its original 430a2-9 that Alexander intends these forms “separate from matter” to be the νοητά “without matter” that Aristotle discusses in 430a3-6, and the “essences and forms of composite things” to be νοητά “that have matter” that Aristotle discusses in 430a6-9. And it is clear that Alexander takes the forms “without matter” to be eternally unchanging substances really existing apart from matter, not forms of composite things abstracted from their matter by the human mind. They are eternally νοητά “in the strict sense” because they are eternally known by themselves, whether anyone else is thinking about them or not: as Alexander says a bit further on in the passage, “the immaterial form is νοῦς in the strict sense and in actuality” (88, 2-3). Where Aristotle speaks of νοητά or νοούμενα, Alexander substitutes “forms”. Alexander assumes that we cannot νοεῖν a matter-form composite but only the form that exists in the matter; and, here as elsewhere, he calls the separate immaterial substances “forms”, which Aristotle never does in talking about substances that he accepts, only when talking about Platonic forms. I think Alexander is right that the νοητά without matter of 430a3-6, which are themselves intelligent, are the separate immaterial substances. But this does not seem to be the interpretation of most modern authors, going back to Zabarella’s commentary on the De Anima (published posthumously in 1605)23. Zabarella gives a surprising-sounding objection to taking the νοητά without matter here to be separate immaterial substances, namely that the νοητά here must be the same as the νοητά in the formulation of the aporia (at 429b26-29), and that there Aristotle must have meant “not ideas, or separate substances, but material intelligibles”, since “the doubt [= aporia] did not arise for abstract [= really separated] substances, since no one doubts that they are intellects, but ὅτε νοεῖ, γίνεται. νοῦς δὴ ἐπὶ τούτων τό τε νοούμενον καὶ τὸ νοοῦν, καὶ ἄμφω τότε ταὐτό. ὅταν δέ γε τῶν ἐνύλων τι εἰδῶν νοῇ καὶ τὸ τί ἦν εἶναι ἐνύλου τινὸς λαμβάνῃ, οὐκέθ’ ὁ αὐτὸς πάντῃ γίνεται τῷ νοουμένῳ πράγματι, ὅτι τὸ μὲν νοούμενον ὑπ’ αὐτοῦ ἐπὶ ὕλῃ τινὶ τὸ εἶναι ἔχει, ὁ δ’ αὐτὸ ὡς κεχωρισμένον ὕλης λαμβάνει. διὸ τὸ τοιοῦτον εἶδος νοούμενον μὲν νοῦς ἐστι καὶ αὐτό, ἔξω δὲ τοῦ νοεῖσθαι γενόμενον οὐκέτι. ἔτι ὁ μὲν νοῦς χωρὶς ὕλης εἶδός τι, τὸ δὲ ἐν ὕλῃ τὸ εἶναι ἔχει. (87, 24-88, 16) 23 It is also not the interpretation of Thomas Aquinas (In Aristotelis librum de Anima commentarium, ed. Pirotta, Marietti, 1925, #724), who takes the “things without matter” here to be the intelligible species of an object as they exist in our intellect, through which our intellect knows itself as well as the object. (He takes “what is known in this way” to mean what is known in actuality.)

FROM DE ANIMA III 4 TO DE ANIMA III 5

115

only for material things, which undoubtedly are not intellects: for a stone is not an intellect”24. As Zabarella goes on to say, when Aristotle says at 430a3-4 “in things that are without matter the νοοῦν and the νοούμενον are the same thing”, this would certainly be true of separate νοητά too, but that is not what Aristotle is saying here – because it is too obvious to need saying! It seems to me that Zabarella’s view – that a separate immaterial νοητόν is obviously itself a νοῦς, and that the knowledge of the object is identical with the object – is exactly what Aristotle is trying to bring about. But Zabarella is anachronistically assuming that this is already obvious at the present stage of Aristotle’s argument. Aristotle has not said thus far in the De Anima that there are no Platonic forms, much less argued that there are no Platonic forms or shown what is wrong with the arguments for Platonic forms, and he cannot be expecting that his readers have already mastered the Metaphysics. Some of his readers or interlocutors may very well believe that there are separate Platonic forms or numbers, and that we can νοεῖν them, and they may think it would be a serious objection against Aristotle’s theory if it implies that the horse-itself, or the three-itself, also νοεῖ. What Aristotle says here – that immaterial νοητά are separately existing knowledges, and that “the νοοῦν and the νοούμενον are the same thing” (430a3-4) – is not only not obvious, but indeed makes the problem more acute for the Plato-sympathetic reader. Zabarella’s interpretation is also in part motivated by Aristotle’s supporting argument that “theoretical knowledge and what is known in this way are the same” (430a4-5). Metaphysics E 1 describes physics as a kind of theoretical knowledge alongside mathematics and theology, and indeed Zabarella says that properly speaking we do not have knowledge [ἐπιστήμη, scientia], but rather non-discursive intellection, of separate immaterial substances; so when Aristotle talks about the objects of theoretical knowledge here, Zabarella thinks, he must mean the forms of material things. But if the “things that are without matter” of 430a3-6 are the forms of material things, what are the “things that have matter” of a6-9? If form-matter composites, why are these a case of νοητά at all? Zabarella’s answer is that the “things that are without matter” and the “things that have matter” are not two disjoint classes of things, but 24 In the 1605 Venice edition, this is In de Anima, Liber III (which has its own pagination), 56r, column 2, toward the end of Textus XV.

116

STEPHEN MENN

rather two modes of being that the same thing can have: “intelligibles have two modes of being: in one way they have being in matter outside the intellect, and in another way they have being without matter in the intellect; so if they are taken as they are in the intellect, abstracted from matter, then they are actually intelligible, and the argument that they are also intellects is sound, since the intellect in thinking becomes what it thinks, so also conversely what is actually thought becomes the intellect. Thus it is true that a stone is intelligent, namely the stone which is in the soul, abstracted from matter and actually intelligible; but if stone is taken the other way, as having real being outside the soul in matter, then it is not actually intelligible, and therefore it is not intelligent, for it is only potentially intelligible, and hence it is also potentially intellect” (55v, columns 1-2, near the beginning of Textus XV). And indeed Zabarella construes a6-9 as saying that a stone is potentially intellect or potentially intelligent. We’ll examine the construal issue below, but if on any reading a stone is potentially intelligent, then something has gone wrong and we need to reflect on how we have gotten there. The crucial issue is what Aristotle means by saying “in things that are without matter the νοοῦν and the νοούμενον are the same thing, for theoretical knowledge [ἐπιστήμη] and what is known in this way are the same” (430a3-5). For this to be an argument, the νοοῦν in this case must be the theoretical knowledge: the νοοῦν is not the individual human rational soul (which is certainly not the same thing as its object, or you and I would be the same as each other), but rather the science or theoretical knowledge which we possess. “Theoretical knowledge” is apparently the same as knowledge of “things that are without matter”, and contrasts with the “things that have matter” of 430a6-9, in which the knowledge is not identical with its object. But where is the dividing line between the two cases? Zabarella and many more recent interpreters put the line between the forms of material things when universalized by being mentally abstracted from their matter and the forms of material things as existing in individual matter, so that Aristotelian physics, which is a universal knowledge about the forms of material things, would count as a theoretical knowledge and as being identical with its object. And, as noted, Metaphysics E1 says that physics is θεωρητική τις, although it says nothing about physics being identical with its object. However, De Partibus Animalium I 1, 639b30-640a9 contrasts physics with the “theoretical knowledges”, on the ground that these argue from “what is” to

FROM DE ANIMA III 4 TO DE ANIMA III 5

117

what follows from that, whereas physics argues from “what will be” to what must be in order for that to come about. This is a characteristic that physics shares with the arts, and the implication is apparently that physics is a productive knowledge, as if we were sharing nature’s own deliberations about how to produce things, or more strictly, knowledge of how it would deliberate if it deliberated: this remains a productive knowledge even though we cannot use it to produce anything, just as the art of architecture remains a productive knowledge even when we are using it only to understand how someone else build a house and not to build anything ourselves25. Indeed, when Metaphysics E1 argues that physics “is neither practical nor productive” (1025b21, argument through 1025b24) and therefore that “if all thought is either practical or productive or theoretical, physics would be θεωρητική τις, but [ἀλλά] theoretical about that sort of being which is capable of being moved, and about an οὐσία-inthe-sense-of-λόγος for the most part only as inseparable” (1025b25-28), Aristotle is either weakening the sense of θεωρητική or at least controversially widening its extension. It is thus not shocking that in De Partibus Animalium I 1 Aristotle should assume that the θεωρητικαὶ ἐπιστῆμαι are only the sciences of unchanging things26. And I think this is what he 25

Physics is certainly not practical knowledge, and that is the only other kind of knowledge that Aristotle sometimes mentions. Physics II 8, 199a12-15 says that if what is in fact a natural thing were constructed by an art, the art would proceed by the same steps by which nature actually proceeds, following the same means-end connections (and likewise if a house were built by nature, nature would proceed by the same steps as the art). James Lennox (2002) records a range of interpretations, partly following G.E.R. Lloyd (1996), p. 29-30. Lloyd offers three options, the obvious one on which Aristotle is contrasting physics with the theoretical sciences, and two alternatives on which he would be contrasting both physics and the theoretical sciences jointly with something else. But these two alternatives are both pretty obviously untenable, and Lloyd is candid enough about their difficulties; his only argument against the interpretation on which Aristotle is contrasting physics with the theoretical sciences is that it would contradict Metaphysics E 1, which indeed it does. Lennox mentions Lloyd’s three readings and goes through their difficulties, but, as he says, “the context nevertheless favours this reading”, i.e. the reading on which Aristotle is contrasting physics with the theoretical sciences. This is also the interpretation given by David Balme in his older Clarendon Aristotle De Partibus Animalium I and De Generatione Animalium I (D. Balme (1972)), and indeed there is no choice. 26 De Partibus Animalium I 1 itself speaks liberally of θεωρία and θεωρεῖν, and says that physics or the physicist is θεωρητικός of this or that, but not just θεωρητικός without a dependent genitive (641a10-12, mentioned by Lennox as a difficulty for the reading that contrasts physics with theoretical knowledge, says merely that “physics is not θεωρητική of any of the things that exist by abstraction [i.e. of mathematical objects], since nature does everything for the sake of something [and apparently mathematical objects are not for the sake of anything, cf. Metaphysics B 2, 996a21-b1]”). But when it

118

STEPHEN MENN

must be doing in De Anima III 4 as well. It would be very strange for him to describe physics as knowledge of “things that are without matter”: it is either about matter-form composites or about their forms (surely it is about both, in different ways), but those forms are “neither without matter nor according to matter” (Physics II 2, 194a14-15, cf. Metaphysics E 1, 1026a4-6). And the forms that physics knows are not any kind of knowledge. By contrast, when I know some object X that really does exist without matter, then the knowledge of X that I possess (the ἐπιστήμη of X, or the νοῦς in the sense of non-demonstrative knowledge) is simply X: it is not X-minus-its-matter, since X has no matter to begin with, and it is not a “representation” of X, but simply X, and X is therefore a kind of knowledge. Such an X is a separately existing knowledge that I and others can come to have, so that it can become my knowledge and your knowledge, without its being an accident ontologically dependent on my soul or your soul, and without its being divided into a part in me and a part in you. To say that S has [ἔχει] X (whether X is a knowledge or something else) is equivalent to saying that X is in S, with different senses of having, corresponding to different senses of being-in (so Metaphysics Δ 23, 1023a23-25). When X is a knowledge, this will happen when S is capable of ἐνεργεῖν according to X, i.e., if X is a theoretical knowledge, when S is contemplating what X is a knowledge of. In the case where X is a knowledge of something without matter, the knowledge and its object are the same, so this just means contemplating X itself. As noted above, S’s ἐνέργεια of contemplating X is both the passive ἐνέργεια of the patient S and the active ἐνέργεια of the agent X acting on S: for X to be in S in the relevant sense is just for it to be capable of acting on S27. gives an official classification of the sciences, physics is not among the θεωρητικαί. In Metaphysics E 1, to make an important point (some forms cannot exist apart from matter and cannot even be studied scientifically apart from matter), Aristotle classifies differently. The Metaphysics E view is presumably Aristotle’s considered decision, but the view of the De Partibus Animalium (and of De Anima III 4 as I read it, and I think also of Metaphysics Λ 9, 1074b 38-1075a5) is a natural default position for him when he is not putting any special effort into revising the standard classification. The Stoics apparently think that physics, like ethics, is both theoretical and practical (where they, like Aristotle sometimes, seem not to distinguish practical from productive knowledge): see S. Menn (1995b). On Metaphysics Λ 9, 1074b38-1075a5, which I think supports the same theoretical/productive distinction, see the following note. 27 Metaphysics Λ 9, 1074b38-1075a5: ἢ ἐπ’ ἐνίων ἡ ἐπιστήμη τὸ πρᾶγμα, ἐπὶ μὲν τῶν ποιητικῶν ἄνευ ὕλης ἡ οὐσία καὶ τὸ τί ἦν εἶναι, ἐπὶ δὲ τῶν θεωρητικῶν ὁ λόγος

FROM DE ANIMA III 4 TO DE ANIMA III 5

119

De Anima III 4, 430a6-9, ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν. ὥστ’ ἐκείνοις μὲν οὐχ ὑπάρξει νοῦς (ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων), ἐκείνῳ δὲ τὸ νοητὸν ὑπάρξει, is supposed to deal with the case that contrasts with theoretical knowledge of things without matter. Where in the case of theoretical knowledge Aristotle simply accepts the apparently absurd conclusion that the νοητόν is itself νοῦς, in the case of things that have matter he says that the νοῦς or knowledge of such things is also νοητός (apparently without the objectionable consequence that this νοῦς will be somehow “mixed” or composite out of a νοητόν-element and a distinctively νοῦς-element), but that these νοητά, the “things that have matter”, will not themselves have νοῦς. How 430a6-9 is supposed to support that conclusion is disputed: the dispute turns in part on a dispute about the syntax of ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν, and also in part on how the δυνάμει of that clause connects with ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων. What objects 430a6-9 is supporting the conclusion for is also disputed: whatever objects are not the “things without matter” that are the objects of theoretical knowledge, which are both νοητά and νοῦς, but as we have seen it is disputed which objects those are. I have translated ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν as “but in things that have matter, [the νοῦς] is potentially each of the νοητά, so that νοῦς will not belong to them [= the νοητά]”. Ἐν δὲ τοῖς ἔχουσιν ὕλην means “in the case of things that have matter”, as ἐπὶ μὲν γὰρ τῶν ἄνευ ὕλης meant “in the case of things that are without matter”. Νοῦς (or τὸ νοοῦν or ἐπιστήμη, all of which mean the same τὸ πρᾶγμα καὶ ἡ νόησις; οὐχ ἑτέρου οὖν ὄντος τοῦ νοουμένου καὶ τοῦ νοῦ, ὅσα μὴ ὕλην ἔχει, τὸ αὐτὸ ἔσται, καὶ ἡ νόησις τῷ νοουμένῳ μία, is probably itself too controversial to be usefully cited as evidence about the meaning of De Anima III 4, 430a3-5. But I think that when Aristotle says in Λ 9 that in some cases the knowledge is its object, he means that this holds not in the productive knowledges but only in the theoretical ones (where, as he says, the same thing is the object and the νόησις); when he says that in ὅσα μὴ ὕλην ἔχει (and only in these) the knowledge and its object are the same, he means the kinds of knowledge which he has described as theoretical in the previous sentence; that the first case, where [the knowledge, I take it] is the essence of the thing without the matter, describes all knowledge of enmattered things and their forms; and that since this is being contradistinguished from the theoretical case, Aristotle must be restricting the theoretical case to knowledge of separate eternal things, and counting physics as productive knowledge. So it would agree exactly with what I think he is saying in De Anima III 4, and would put the theoretical/productive distinction where De Partibus Animalium I 1 puts it, not where Metaphysics E 1 puts it.

120

STEPHEN MENN

thing), carrying over from the previous clauses, is the subject of δυνάμει ἕκαστον ἔστι τῶν νοητῶν: in the immaterial case, the νοῦς simply is its νοητόν, while the νοῦς of a material thing is δυνάμει its νοητόν, and while this is perhaps initially a puzzling assertion, Aristotle explains and justifies it when he says ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων. In the theoretical sciences, the knowledge simply is its object; the paradigm case of what is contrasted with theoretical knowledge is a productive art such as the art of housebuilding, which is not a house, but is rather a δύναμις for houses, i.e. a δύναμις for producing houses, existing without matter in the soul of the housebuilder, and so it can be said to be δυνάμει a house, not because it can become a house, but because it can produce a house. But Alexander, and indeed almost the whole interpretive tradition from Alexander down to Ross and beyond, construe the text differently, and take it as saying that enmattered forms are only potentially νοητά28. This is important, because it gives the ποιητικὸς νοῦς the task of making the potentially νοητά actually νοητά; if Aristotle is not committed to enmattered forms’ being only potentially νοητά, then there is no such task and no need to search for something to perform it. Aristotle does, of course, think that enmattered forms are not always by the necessity of their nature actually νοούμενα, actually being thought, whereas a separate essentially actual νοῦς as described in De Anima III 5 and Metaphysics Λ is necessarily always being thought, since it it is necessarily always thinking itself. But that is not sufficient for what Alexander and the tradition following him are saying. If that were all, then the enmattered form could simply act on the νοῦς of our soul, making the form actually thought and our νοῦς actually thinking, without needing a prior ποιητικὸς νοῦς to turn the potentially νοητόν 28 In fact, the only translator or commentator I know who construes the sentence my way, as saying that, in the case of things that have matter, the νοῦς is potentially each of the νοητά (because, in the paradigmatic case, it is able to produce them), rather than as saying that the νοητά are potential or are potentially in the things that have matter, is Christopher Shields, who translates “in the case of those things which have matter it [scil. reason] is each of the objects of reason in potentiality” in C. Shields (2016). Oddly, given that he is going against the overwhelming majority of previous interpreters in his construal of his sentence, he seems to say nothing in his commentary to explicate or defend his translation, or to say how in these cases the νοῦς is potentially each of the νοητά, or how it contributes to solving the aporia. Perhaps Shields simply thought, as I did at the Davis conference in 2002, that this construal is just obviously how the syntax of the sentence (in its context) has to work, and did not realize that it was controversial.

FROM DE ANIMA III 4 TO DE ANIMA III 5

121

form into an actually νοητόν form capable of interacting with the νοῦς of our soul. There seem to be two possible construals on which 430a6-7 ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν would be saying what Alexander needs. On the construal apparently accepted by most modern interpreters, notably Hicks and Ross, ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν means “in things containing matter each of the objects of thought is present potentially” (Hicks’ translation)29. On this construal ἔστι, modified by the adverbial δυνάμει, is not simply predicative but locative: its subject is not (as I have taken it) the νοῦς or ἐπιστήμη carried over from the previous clauses but ἕκαστον τῶν νοητῶν, and its complement is ἐν τοῖς ἔχουσιν ὕλην. This breaks the apparent parallel between ἐπὶ μὲν γὰρ τῶν ἄνευ ὕλης and ἐν δὲ τοῖς ἔχουσιν ὕλην, which both seem to mean “in the case of”. It makes sense to say that each of the intelligibles is potentially present in matter, although this is not true without restriction: this piece of matter potentially contains most of the intelligibles, but sublunar matter does not contain the forms of celestial things even potentially (or vice versa), and no matter contains God even potentially. But it makes less sense to each that each of the intelligibles is potentially present in the things containing matter: the form of a table is actually present in this composite table, and if the form of an elephant is potentially present in this composite table it is by a very remote potentiality. The answer of Hicks and Ross, so far as I understand it, is that while the form of a table is actually present in the composite table, it is only potentially intelligible, and so the intelligible is only potentially present in the composite table. It is possible that Alexander is also construing ἐν δὲ τοῖς ἔχουσιν ὕλην as locative, but he may instead be construing the phrase, as I do, as “in the case of the things that have matter”; but then, like Hicks and Ross, he takes δυνάμει ἕκαστον ἔστι τῶν νοητῶν to mean “each of the intelligibles is potentially”, which must ultimately mean “is potentially intelligible”30. How would this 29 Ross “in things that possess matter each of the objects of reason is potentially present” (editio maior p. 291). Ross’ comment, “meaning presumably that these objects are there, ready to be picked out and recognized by reason” (p. 295), does not clarify things for me. 30 Thanks to Sean Kelsey for getting me to realize that Alexander may be taking the second option. While almost all interpreters have taken ἕκαστον τῶν νοητῶν as the subject, it is often hard to say whether they are taking ἐν δὲ τοῖς ἔχουσιν ὕλην as meaning “present in” or “in the case of”. Zabarella says in habentibus materiam non est intelligibile

122

STEPHEN MENN

connect to the following sentence, ὥστ’ ἐκείνοις μὲν οὐχ ὑπάρξει νοῦς (ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων), ἐκείνῳ δὲ τὸ νοητὸν ὑπάρξει, which it is supposed to support, and how would it contribute to solving the aporia? The thought-connection must be: “if the things that have matter, or something present in them, were actually intelligible, then they would be actually intelligent; but since they are only potentially intelligible, they are only potentially intelligent, and the problem is solved”. To say, as Zabarella explicitly does, that a stone or its form is potentially intelligent, is a high price to pay for solving the aporia31. But, even if we accept this, how would the conclusion that a stone does not have νοῦς actually (but only potentially) be supported by the γάρ clause, ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων? If Aristotle is trying to make the point that something must be actual to count as νοῦς, then it would be very strange to support this by saying that νοῦς is a δύναμις. And it is hard to believe that Aristotle does not intend the δύναμις here to be somehow connected with the δυνάμει of the previous sentence ἐν δὲ τοῖς ἔχουσιν ὕλην δυνάμει ἕκαστον ἔστι τῶν νοητῶν. Zabarella, the only commentator I know of who has tried to explain the connection, paraphrases ἄνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων by saying “the meaning [ratio] of intellect which is attributed to a stone and such-like material things is a power, i.e. an ability [virtus], without matter: for we call a stone an intellect when it is received in the mind without matter and nisi potestate (57r, column 1), but it is not clear to me how he takes the in. Thomas says in rebus habentibus materiam, species non est intelligibilis secundum actum, sed secundum potentiam tantum (In de Anima #727): the sensible species in the material things becomes an intelligible species, and so becomes actually intelligible, only when it has been stripped of matter by the agent intellect. I think Thomas must mean “in the material things”, not merely “in the case of material things”, since there is indeed an actually intelligible species in the case of material things, not in the material things but in the intellect. But it seems to me extremely unlikely that Aristotle thinks that, when the intellect actually knows something, what it knows is something present only in the intellect itself and not in external things – that its cognitions can only be directed to itself and its accidents, and not to anything outside itself. 31 Alexander says not only that an enmattered form becomes νοητόν, but also that it becomes νοῦς: τὰ δὲ τῇ αὑτῶν φύσει νοητὰ κατ’ ἐνέργειαν νοητά, δυνάμει γὰρ νοητὰ τὰ ἔνυλα. ἀλλὰ μὴν τὸ κατ’ ἐνέργειαν νοητὸν ταὐτὸν τῷ κατ’ ἐνέργειαν νῷ, εἴ γε ταὐτὸν τὸ νοούμενον τῷ νοοῦντι. τὸ ἄρα ἄυλον εἶδος νοῦς ὁ κυρίως τε καὶ κατ’ ἐνέργειαν. καὶ ὁ νοῶν ἄρα τοῦτο νοῦν νοεῖ οὐ γινόμενον νοῦν ὅτε νοεῖται, ὡς ἐπὶ τῶν ἐνύλων εἰδῶν ἔχει, ἀλλὰ ὄντα νοῦν καὶ χωρὶς τοῦ ὑπὸ τοῦ νοῦ νοεῖσθαι (87, 28-88, 5). So he must think, like Zabarella, that an enmattered form is potentially νοῦς. Similarly Alexander (DA 90, 19-22), the “other νοήματα” (forms of composites) are ἐν τῷ νοεῖσθαι γενόμενα νοῦς.

FROM DE ANIMA III 4 TO DE ANIMA III 5

123

is actually intelligible” (57r, column 1, halfway through Textus XVI). So he takes ὁ νοῦς τῶν τοιούτων as “the sense in which these things are [or have] νοῦς”; and apparently he takes “power without matter” to mean the ability to be dematerialized and thus to be received in the mind32. So Aristotle would be saying not that νοῦς is a δύναμις, but that the things that have matter, because they can be dematerialized, are potentially νοητά and potentially νοῦς. I think it is much better to read “the νοῦς of such things” as an objective genitive: the knowledge of such things, i.e. of the objects of productive knowledge, is a δύναμις without matter like the art of housebuilding. This is why, in the case of things that have matter, the νοῦς is only potentially each of its νοητά, e.g. each individual house. Because it is not identical with them, as it would be if they were without matter, it can be νοητόν like them without their being νοῦς like it. And this gives a resolution of the aporia: it can be both νοητόν like its objects, and νοῦς unlike its objects, without its being “mixed” out of one constituent that makes it νοητόν and another that makes it νοῦς, because what makes it νοῦς is not something it has that the other νοητά lack, but rather something it lacks that they have, because it is a δύναμις without matter. The νοῦς is not composite but simple; its material νοητά are composite, and that is what explains their non-identity. We do not have to say that an individual house or enmattered house-form is not actually νοητόν, or that it is even potentially νοῦς. An enmattered form is indeed νοητόν, but it is not νοῦς and cannot become νοῦς by being somehow stripped of its matter: only the δύναμις without matter for such a thing (in the paradigm cases, the power for producing it) is νοῦς. This νοῦς is not identical with its object, but such identity holds only in “theoretical knowledge”, where the object is “without matter”, and that means not the forms of form-matter composites, but separately existing immaterial substances.

32 Zabarella says a bit further on (57r, column 2) that when Aristotle says “the intellect of such things is a power without matter”, “we have clearly shown that Aristotle asserts these words not about the intellect but about the material intelligibles, to which also the preceding words [i.e. ‘in things that have matter each of the intelligibles is potentially’] pertained”. Zabarella denies that the relevant intellect (the human potential intellect) is a power without matter, since it is a form inseparable from the human body as its matter. Rather, a material intelligible is a power without matter in that it is potentially without matter, or its being intelligible is its power for being without matter, inasmuch as it can be dematerialized by the intellect.

124

STEPHEN MENN

However, in solving the aporia, Aristotle has left some difficulties remaining, or perhaps he has created new ones. These difficulties seem to arise in the case of objects without matter, not in the case of enmattered forms. Aristotle explicitly calls attention to one of them: “we must investigate the cause why we do not [it does not? there is not?] always νοεῖν”. This is mentioned at the end of the discussion of objects without matter, and the difficulty seems to arise only for them. In this case Aristotle has conceded that the object is itself νοῦς, indeed that it is identical with our knowledge of it; but since it is without matter, it is eternal and unchanging. So it seems to follow that it always νοεῖ; indeed, since it is our knowledge and it is eternal and unchanging, it seems to follow that we always νοοῦμεν, not just thinking about something-or-other but knowing this particular object, not just within our human lifetimes but from eternity and to eternity. Perhaps these conclusions are true, but they are certainly surprising, and we need to learn either how to avoid them or how to live with them. More generally, even if the object νοεῖ, it is unclear how its νοεῖν is related to ours. The νοῦς that is identical with its object, and is itself νοητόν, is the knowledge, the ἕξις, not a bare first δύναμις: the first δύναμις is not νοητόν, except that we can infer its existence, and get some grasp on its nature, by analogy, like matter – it is nothing in actuality before it νοεῖ, and there is nothing to grasp except its relation to actual knowledge. But what is that relation? The nature of the object too is mysterious. We have been told that it is without matter and that it is νοῦς. But for all we know at this stage in the argument there are Platonic forms, and it is surprising if a Platonic form (other than a form of some knowledge or virtue) itself νοεῖ. Even if it does, it must be very different from us: it seems to be one thing for it to νοεῖν, and another for it to have the intelligible content that makes it a form, and specifically the form of X or Y. But Aristotle has denied that νοῦς can have any such composition. 4. These kinds of questions are why De Anima III 5 is necessary. The answer to the aporiai raised at De Anima III 4, 429b22-29 is not simply De Anima III 4, 429b29-430a9, but extends to the end of De Anima III 5 at 430a25: so that, as I said above, the chapter-division between III 4 and III 5 could be made at least as reasonably at 429b22, putting the aporiai

FROM DE ANIMA III 4 TO DE ANIMA III 5

125

and their answers in a single chapter. It is thus wrong to say with Caston that III 5 could simply be removed without damage to the overall argument: if it were removed, we would have no answers to the questions about immaterial νοητά which are naturally raised by the end of III 4. The interpretation of the sixteen lines of De Anima III 5 is of course extremely controversial. I will proceed through the text point by point, but without explicitly addressing every controversy, so as not to lose sight of the argument that I think the text as a whole is making, in the context of the aporiai of De Anima III 4. I will quote the text (in the form in which I accept it, with notes on the more important textual disputes), and offer a provisional translation, in two installments. Ἐπεὶ δ᾿33 ἐν ἁπάσῃ34 τῇ φύσει ἐστὶ35 τὸ μὲν ὕλη ἑκάστῳ γένει (τοῦτο δὲ ὃ πάντα δυνάμει ἐκεῖνα), ἕτερον δὲ τὸ αἴτιον καὶ ποιητικόν, τῷ ποιεῖν πάντα, οἷον ἡ τέχνη πρὸς τὴν ὕλην πέπονθεν, ἀνάγκη καὶ ἐν τῇ ψυχῇ ὑπάρχειν ταύτας τὰς διαφοράς· καὶ ἔστιν ὁ μὲν τοιοῦτος νοῦς τῷ πάντα γίνεσθαι, ὁ δὲ τῷ πάντα ποιεῖν, ὡς ἕξις τις, οἷον τὸ φῶς· τρόπον γάρ τινα καὶ τὸ φῶς ποιεῖ τὰ δυνάμει ὄντα χρώματα ἐνεργείᾳ χρώματα. Since in every nature there is one thing which is matter for each genus (this is what is potentially all those things), and another which is the cause and agent / maker through making [them] all, as the art is related to the matter, necessarily these distinctions must exist also in the case of the soul, and what is like this [= what plays the role of matter] is νοῦς through becoming all things, and the latter [is νοῦς] through making them all, as a kind of ἕξις, like light: for in a way light too makes what are potentially colors actually colors. (430a10-17) 33 Deleting the MSS’ ὥσπερ (after δ᾿) with Ross (in both editions), against Förster and Corcilius. In 430a13 ἀνάγκη καὶ can pick up either ἐπεὶ or ὥσπερ, but I don’t see any way that both words can be picked up. If we keep the transmitted text, we have to assume that Aristotle lost control of his syntax. Another option might be to delete καὶ before ἔστιν in a14, yielding a structure “since, as in every nature … so too in the case of soul …, therefore one thing is νοῦς through becoming all things, and other νοῦς through making all things.” 34 The variant πάσῃ for ἁπάσῃ in two MSS may be right. It doesn’t make much difference: as LSJ say s.v. ἅπας, “the use of ἅπας for πᾶς is chiefly for the sake of euphony after consonants”. Either way, the overall sense of the sentence seems to require the meaning “in every nature”, not “in the totality of nature”: ἐν ἁπάσῃ τῇ φύσει is parallel to ἑκάστῳ γένει. (The article before φύσει does not rule this out: ἅπαν τὸ X can mean “every X”, as at On Sophistical Refutations 22, 178b37-9, τὸ ἄνθρωπος καὶ ἅπαν τὸ κοινὸν οὐ τόδε τι ... σημαίνει, and Metaphysics Θ 8, 1050a7-8: ἅπαν ἐπ᾿ ἀρχὴν βαδίζει τὸ γιγνόμενον καὶ τέλος). 35 Deleting the MSS’ τι (after ἐστὶ) with Ross (in both editions), against Förster and Corcilius. If the τι can be saved, it will be in a way that does not change the sense.

126

STEPHEN MENN

From where we were at the end of III 4, it was surprising to learn that every immaterial νοητόν X – perhaps something like a Platonic form, since the De Anima has given no arguments against the existence of such forms – is itself a νοῦς, apparently exercising eternal νοεῖν: certainly there are many questions about the manner of such νοεῖν. Aristotle responds to these concerns by distinguishing two types of νοῦς, one (the παθητικὸς νοῦς, as he calls it further down at 430a24-25) which plays the role of matter and is potentially each thing, that is, potentially each thing which it is able to νοεῖν, and one which plays the role of ποιητικόν (thus traditionally called the ποιητικὸς νοῦς) and makes the potential νοῦς to be actually each of the things which it is potentially, that is, each of the things which it is able to νοεῖν, as the art makes the matter to be each of the things which it is potentially. The παθητικὸς νοῦς is the kind of νοῦς that has been described in III 4 – “what is called the νοῦς of the soul [ὁ καλούμενος τῆς ψυχῆς νοῦς] – I am calling νοῦς that by which the soul reasons and affirms – is none of the beings in ἐνέργεια until it thinks / knows them” (429a22-24), and “it has no nature except this, that it is δυνατόν” (a21-2) – while the ποιητικὸς νοῦς is something new, not mentioned in III 4. Victor Caston asks, “why on earth should Aristotle have thought there were two intellects?” (p. 202). But this is precisely why it is so important not to translate νοῦς by “intellect”, and to recognize that νοῦς has several senses in Greek philosophical discourse, not all of which refer to rational souls or rational parts or powers of souls. Νοῦς can be an act, and νοῦς can also be a ἕξις, a virtue, which may be conceived, as by Plato, as existing separately, as a Reason-in-general in which souls participate in order to think and act rationally, or as a separately existing knowledge of some particular content. And something like this sense of νοῦς is very useful at this juncture in Aristotle’s argument, in responding to the aporia about whether every immaterial νοητόν is itself a νοῦς. It would be very surprising if every immaterial νοητόν X were itself an individual disembodied mind; but, Aristotle now points out, there is a different and higher sense of νοῦς, or a different and higher way of being νοῦς, and if the argument has established only that every immaterial νοητόν is a νοῦς in that higher way, then while the result may still be news to the Platonist, it is not absurd, and we can accept it and explore the consequences. Doing this will not mean accepting uncritically Plato’s concept of νοῦς in the higher sense, but rather refining it, by means especially of the ἐνέργεια / δύναμις

FROM DE ANIMA III 4 TO DE ANIMA III 5

127

distinction, and drawing some un-Platonic consequences from some starting-points that Aristotle and Plato share. Aristotle says now that the higher νοῦς is to the lower as an art to the raw material for that art. This was his model from On Generation and Corruption of an agent and patient which are not in the same genus and do not have the same (kind of) matter, and where therefore the patient does not act reciprocally on the agent, so that the agent is an unmoved mover: I cited this above as a possible model for how the νοητόν can act on νοῦς (on νοῦς in the δύναμις-sense) without being reciprocally affected. Aristotle stresses both in De Anima III 10 and in Metaphysics Λ7 that the νοητόν and ὀρεκτόν are unmoved movers, acting first on the soul, and then through the soul on other things; so it is reasonable to say, following the model from On Generation and Corruption, that the νοητόν is to the soul’s νοῦς as the art is to the matter, and that the soul’s νοῦς receives forms from the νοητόν as the matter receives forms from the art. The difference is that in the present text, Aristotle is saying, not that the νοητόν is to the soul’s νοῦς as art to matter, but that a higher νοῦς is to the soul’s νοῦς as art to matter. But this substitution makes sense here, since Aristotle is here interested specifically in the case of an immaterial νοητόν: at the end of De Anima III 4, we knew that such a νοητόν was itself a νοῦς, but we wanted to know what kind of νοῦς it is, how it νοεῖ, how it is related to the soul’s νοῦς when the soul νοεῖ it, and De Anima III 5 is intended to clear up these questions. Aristotle is now saying that the immaterial νοητόν is not a νοῦς in the same way that the soul’s νοῦς is, but rather is a kind of νοῦς that is to the soul’s νοῦς as the art to the matter. The immaterial νοητόν is not, of course, a part of the human soul. As we have noted, it is eternally exercising νοεῖν, which no part of the human soul is doing. Also, independently of Aristotle’s claim that an immaterial νοητόν is itself a νοῦς, you and I can know the same immaterial νοητόν, and it is no more a part of your soul than it is of mine. Nonetheless, a strong tradition, going back to Themistius and the neo-Platonic commentators on the De Anima and endorsed by Thomas Aquinas and more recently by Brentano and Ross, holds that the ποιητικὸς νοῦς of De Anima III 5 is indeed a part or faculty of the human soul. (The chief motivation for the ancient writers and many of their successors is to save Aristotle for the doctrine of the immortality of [at least a part of] the human soul: this can be done only by making the ποιητικὸς νοῦς part of the human soul,

128

STEPHEN MENN

since Aristotle says that “this alone is immortal and eternal”.) All of these writers cite, as proof for their interpretation, Aristotle’s saying here that “these distinctions must exist ἐν τῇ ψυχῇ”36. Ross argues that “ἐν τῇ ψυχῇ can hardly mean only ‘in the case of the soul’”37, but to find ἐν meaning “in the case of”, we need only look seven lines further up in the De Anima, where ἐν δὲ τοῖς ἔχουσιν ὕλην at 430a6 meant “in the case of things that have matter”, parallel to ἐπὶ μὲν τῶν ἄνευ ὕλης, “in the case of things that are without matter”, at 430a338. And just now Aristotle has said that the agent is to the patient as the art to the matter, where the art is an agent external to the matter, not a part or aspect of the matter (or a part or aspect of the same substance that the matter is a part or aspect of). Caston proposes that ἐν τῇ ψυχῇ means: within the genus “soul”, so that the παθητικὸς νοῦς would be a human soul (or the rational part or faculty of a human soul) and the ποιητικὸς νοῦς would be a divine soul. But there is no reason to think that the ποιητικὸς νοῦς is a soul (or a part or faculty of a soul) at all: certainly Plato’s νοῦς-itself is not a soul but rather what souls participate in (e.g. Laws X, 897b1-4); we have seen Aristotle qualifying the potential νοῦς as ὁ […] τῆς ψυχῆς νοῦς (DA III 4, 429a22), and Theophrastus contrasting ὁ ψυχικὸς νοῦς with ὁ ἐνεργείᾳ νοῦς, τουτέστι ὁ χωριστός (Fr. 307B FHS&G); there is no hint whatever that the God of Metaphysics Λ is or has a soul. And in the parallel case of “the matter for each genus” (430a10-11), the art is not itself a member of that genus, but an agent possessing the same form in a higher way39: so too in the case of soul, the ποιητικὸς νοῦς, which acts on the soul, has or is νοῦς in a higher way than the soul does. Aristotle compares the ποιητικὸς νοῦς to light, “for in a way light too makes what are potentially colors actually colors”. In the immediate context, this is most easily taken to mean that the ποιητικὸς νοῦς makes the 36 So F. Brentano (1867), p. 168 (citing Themistius and Thomas among others), and Ross’ editio maior p. 45. 37 W.D. Ross (1959), p. 304, n. 85. Ross had been less dogmatic about the issue in the first edition (1923), p. 148-149 and p. 149, n. 1 (corresponding to the cited footnote in the 1959 edition). 38 Ross, of course, denied the parallelism between ἐν δὲ τοῖς ἔχουσιν ὕλην at 430a6 and ἐπὶ μὲν τῶν ἄνευ ὕλης at 430a3, and interpreted ἐν δὲ τοῖς ἔχουσιν ὕλην at 430a6 as meaning “present in the things that have matter”: see discussion above. As also noted above, Ross himself recognizes the parallelism in his editio maior p. 295. 39 And thus in ἐν ἁπάσῃ φύσει too, if it means (as I think it must) “in every nature” rather than “in the totality of nature”, ἐν must mean “in the case of” rather than “inside”, against Ross.

FROM DE ANIMA III 4 TO DE ANIMA III 5

129

potential νοῦς to be actually each of the things which it is potentially (that is, each of the things which it is able to νοεῖν) as light makes the potential colors to be actual colors. Aristotle may also be thinking that, as light makes the potential colors to be actual colors and so to be actually seen, the ποιητικὸς νοῦς makes the potential νοητά to be actual νοητά, or actual νοούμενα. Presumably this comes to much the same thing, since to make the potential νοῦς actually νοῶν and to make the potential νοητά actually νοούμενα would be the same act viewed from two different sides. Either way, as has often been observed, Aristotle is recalling the Sun passage of Republic VI, where the light of the sun “makes our sight to see, and the visibles to be seen, in the best way” (508a5-6); without the presence of light, “sight will see nothing, and the colors will be invisible / unseen” (507e2). Plato here is interested in sight and the visible only as an analogy for νοῦς and the νοητόν: “what this [scil. the good] is in the intelligible domain in relation to νοῦς and the νοούμενα, that [scil. the sun] is in the visible domain in relation to sight and the things seen” (508b13-c2). Plato keeps up a systematic analogy, soul:νοῦς:good:truth: intelligibles :: eye:sight:sun:light:visibles, where the good is the cause of νοῦς to soul and of truth and thus intelligibility to the intelligibles, as the sun is the cause of sight to the eyes and of light and thus visibility to the visibles (“νοῦς” in this passage is not used for a being superior to souls, but always for the δύναμις in the soul analogous to sight in the eye)40. 40

Plato systematically distinguishes here between light, corresponding to truth or intelligibility, and the sun, corresponding to the good, which is the best (but not necessarily the only) source of light. Note that although the sight, and the eye in which it exists, are not a sun, the eye is “sunlike”, and has its δύναμις as an overflow from the sun (508a11-b8), and that “the sun is not sight, but, being the cause of it, is seen by it” (508b9-10), all of which carry over well for Aristotle to the relation between the soul’s νοῦς (sight or the eye) and the ποιητικὸς νοῦς (the sun or its light). Plato probably intends that light is to the sun as sight [ὄψις] is to the eye: while we most naturally think of sight as a power residing in the eye, Plato is likely to be thinking of the kind of optical theory later presented in Euclid’s Optics, where ὄψεις radiate out from the eye in straight lines – the most obvious way for the eye to be sunlike is that ὄψεις radiate out from it as light radiates out from the sun. (The sun is often described as the eye of a god, presumably seeing by means of its rays, and αὐγαί can be used of rays either from the sun or from the eyes, which are both propagated in straight lines, reflected by mirrors, and so on.) Vision occurs only when both ὄψις, coming from the eye, and light, coming from the sun or from some other light-source, fall upon the same object. (Plato describes ὄψις as coming out of the eye to meet either the object or something emerging from the object both in the Theaetetus and in the Timaeus; the details of the accounts differ, and only the Timaeus gives a role to light.)

130

STEPHEN MENN

Aristotle simplifies this picture. His theory of vision gives no special role to the sun, but only to light, which links the visibles with the eye by actualizing the potentially transparent medium so that the visibles can act on the medium and thus on the eye. So here in his account of intellection he does not distinguish between an analogue of light and an analogue of the sun: he mentions just one cause, which is the cause of actual νοεῖν to the soul’s power of νοῦς and of actual νοεῖσθαι to the intelligibles. A further difference from Plato is that Aristotle here describes this cause, not as the good, but as νοῦς, νοῦς in a higher sense than that in which the soul’s δύναμις is called νοῦς. As I’ve noted, Plato uses the word only in the lower sense in this passage, but it is far from clear that he would identify the good even with νοῦς-itself, the Reason in which souls participate. That νοῦς for Plato is the demiurgic principle responsible for imposing form on matter in an orderly way, but the good-itself seems to be a higher principle prior to, and somehow giving rise to, the immaterial forms themselves41. By contrast, while Aristotle agrees with Plato that there is a separate gooditself (Metaphysics Λ 10, 1075a11-15), he denies that it is anything beyond νοῦς (indeed, since the good-itself is an immaterial νοητόν, and since every immaterial νοητόν X is identical with the knowledge of X, the good-itself must be identical with knowledge of the good, that is, with the highest kind of νοῦς). So the causal role that Plato ascribes to the gooditself, Aristotle gives to a νοῦς that is purely ἐνέργεια. And indeed, it seems reasonable that such a νοῦς should be sufficient, by its acting on the soul (or by the soul’s coming to “participate” in it), to actualize the soul’s δύναμις of νοῦς, with no need to invoke any cause superior to νοῦς. (The hard question is whether an external νοῦς is necessary for cognition, not whether it is sufficient.) However, the most important difference between De Anima III 5 and Plato’s Sun passage is that, while for Plato the source of νοεῖν and νοεῖσθαι is the single first principle above the many νοητά, Aristotle’s argument as we have traced it implies that every immaterial νοητόν must be a ποιητικὸς νοῦς. And this seems to lead to a tension. Aristotle 41 There is no sign that the demiurge of the Timaeus creates the Forms that he looks to as his model. Both Republic VI and the Philebus argue, in an ethical context, against identifying the good with νοῦς or φρόνησις or ἐπιστήμη. The later Platonist tradition is of course divided on the relation between νοῦς and the good, with Alcinous identifying νοῦς with the good (and the Forms with its ἐνέργειαι = νοήσεις), while Plotinus and most later Platonists make the good a first principle superior to νοῦς and to the Forms.

FROM DE ANIMA III 4 TO DE ANIMA III 5

131

says that the ποιητικὸς νοῦς makes the potential νοῦς to be actually each of the things which it is potentially, that is, each of the things which it is able to νοεῖν (and thus presumably also makes the potential νοητά to be actually νοούμενα), as light “makes what are potentially colors actually colors”. This suggests the Platonic picture on which there is a single cause of νοεῖν and νοεῖσθαι, itself νοητόν but also a cause of νοεῖσθαι to many inferior νοητά; whereas for Aristotle, as we have just seen, at least every immaterial νοητόν is itself a ποιητικὸς νοῦς and thus presumably sufficient to cause the soul to νοεῖν it without help from further above. I will discuss different possible ways of resolving this tension after I have gone through the second half of III 5. καὶ οὗτος ὁ νοῦς χωριστὸς καὶ ἀπαθὴς καὶ ἀμιγής, τῇ οὐσίᾳ ὢν ἐνέργεια42· ἀεὶ γὰρ τιμιώτερον τὸ ποιοῦν τοῦ πάσχοντος καὶ ἡ ἀρχὴ τῆς ὕλης. τὸ δ᾿ αὐτό ἐστιν ἡ κατ᾿ ἐνέργειαν ἐπιστήμη τῷ πράγματι· ἡ δὲ κατὰ δύναμιν χρόνῳ προτέρα ἐν τῷ ἑνί, ὅλως δὲ οὐδὲ χρόνῳ, ἀλλ᾿ οὐχ ὁτὲ μὲν νοεῖ ὁτὲ δὲ οὐ νοεῖ43. χωρισθεὶς δ᾿ ἐστὶ μόνον τοῦθ᾿ ὅπερ ἐστί, καὶ τοῦτο μόνον ἀθάνατον καὶ ἀίδιον· οὐ μνημονεύομεν δέ, ὅτι τοῦτο μὲν ἀπαθές, ὁ δὲ παθητικὸς νοῦς φθαρτός44. καὶ ἄνευ τούτου οὐθὲν νοεῖ. And this νοῦς is separate and impassible and unmixed, being essentially actuality[/activity]: for the agent is always superior to the patient and the principle to the matter. Knowledge in actuality is the same as the object; knowledge in potentiality is temporally prior [to knowledge in actuality] in the individual, but universally it is not prior even temporally. Rather, Reading τῇ οὐσίᾳ ὢν ἐνέργεια with Förster and Ross and Corcilius, rather than τῇ οὐσίᾳ ὢν ἐνεργείᾳ: see a note above for discussion. 43 The οὐχ in 430a22 is missing in two MSS and some indirect sources (see Förster’s apparatus). If we delete the οὐχ, which I don’t think any modern editor or interpreter proposes to do, the subject would have to be ἡ δὲ κατὰ δύναμιν [ἐπιστήμη]. Ross in his editio maior (not in the OCT) brackets the whole passage a19 τὸ δ᾿ … a22 οὐ νοεῖ, an irresponsible move rightly rejected by Corcilius; see discussion below. (K. Corcilius (2017), p. 238 says that Ross brackets only through the end of a21 οὐδὲ χρόνῳ, but in fact Ross brackets ἀλλ᾿ οὐχ ὁτὲ μὲν νοεῖ ὁτὲ δὲ οὐ νοεῖ as well.) 44 Ross puts οὐ μνημονεύομεν δέ, ὅτι τοῦτο μὲν ἀπαθές, ὁ δὲ παθητικὸς νοῦς φθαρτός in parentheses, i.e. he takes καὶ ἄνευ τούτου οὐθὲν νοεῖ as picking up from before this phrase. That may well be right, but I am printing καὶ ἄνευ τούτου οὐθὲν νοεῖ as a separate sentence so as not to prejudge whether it is picking up from before, or whether it is a continuation of ὁ δὲ παθητικὸς νοῦς φθαρτός, with ὁ παθητικὸς νοῦς as the subject of νοεῖ and οὐθὲν as the object. Corcilius, following Förster, prints a full stop before οὐ μνημονεύομεν δέ, ὅτι τοῦτο μὲν ἀπαθές, ὁ δὲ παθητικὸς νοῦς φθαρτός and then a colon before καὶ ἄνευ τούτου οὐθὲν νοεῖ, but his translation has dashes which may have the same effect as Ross’ parentheses around οὐ μνημονεύομεν δέ, ὅτι τοῦτο μὲν ἀπαθές, ὁ δὲ παθητικὸς νοῦς φθαρτός. 42

132

STEPHEN MENN

[knowledge or νοῦς in the actuality-sense] does not sometimes think / know and sometimes not think / know: and when it has been separated it is just what it is [i.e. it is just knowledge / νοῦς and nothing else], and this alone is immortal and eternal: but we do not remember, because this is impassible, whereas the passive νοῦς is corruptible. And without this [scil. the impassible νοῦς] nothing thinks / knows [or “and without this it (scil. the passive νοῦς) thinks / knows nothing”]. (430a17-25)

The first sentence picks up the Anaxagorean predicates that Aristotle had applied in De Anima III 4 to the potential νοῦς in the soul, and argues that they apply in a stronger way to the ποιητικὸς νοῦς. When the agent and the patient νοῦς encounter each other, the patient has the knowledge of X in potentiality and the agent has the knowledge of X in actuality: this is because the agent simply is the knowledge of X existing separately, which is because the agent simply is X itself existing separately, and the knowledge of X is identical to X. Of course, this argument applies only to an immaterial νοητόν X, and not to a form existing in matter. In the case of an enmattered form X, Aristotle has given no argument why the rational soul’s potential knowledge of X must be actualized by a separately existing νοῦς or knowledge of X, rather than simply by a concrete material instance of X. And, rather than reconstructing an argument for this conclusion on his behalf, I think we should question whether he is really committed to the conclusion; I will argue below that he does not in fact believe it. In any case, in De Anima III 5 he is speaking only about a case where “knowledge in actuality is the same as the object”, and, as we saw in De Anima III 4, this holds only for a νοητόν existing separately from matter. At any rate, this is true if 430a19-22, τὸ δ᾿ αὐτό ἐστιν ἡ κατ᾿ ἐνέργειαν ἐπιστήμη τῷ πράγματι· ἡ δὲ κατὰ δύναμιν χρόνῳ προτέρα ἐν τῷ ἑνί, ὅλως δὲ οὐδὲ χρόνῳ, ἀλλ᾿ οὐχ ὁτὲ μὲν νοεῖ ὁτὲ δὲ οὐ νοεῖ, are part of De Anima III 5. Ross in his editio maior of 1961 proposes to delete the passage, on the ground that “these words, all except the final words ἀλλ᾿ οὐχ ὁτὲ μὲν νοεῖ ὁτὲ δὲ οὐ νοεῖ, recur in chapter 7, 431a13. They cannot have been meant to stand in both places; one early editor must have placed them in chapter 5 while another placed them in chapter 7, and a third included them in both places, They are harmless in chapter 7, which is in any case a collection of scraps; here they seriously interfere with the course of the thought, which without them would be continuous” (W.D. Ross (1961), p. 296). But even if the duplication

FROM DE ANIMA III 4 TO DE ANIMA III 5

133

between III 5, 430a19-22 and III 7, 431a1-3 were perfect, and even if the same passage could not stand in both places, the question is whether it contributes in III 5, or whether it disrupts an argument that would be better without it45. Contra Ross, it does contribute in III 5, since the κατ᾿ ἐνέργειαν ἐπιστήμη which is the same as its object – that is, the knowledge of an immaterial νοητόν – is precisely the ποιητικὸς νοῦς that Aristotle has been discussing46. He has just said that this νοῦς is the ἀρχή, i.e. that it is prior (in one or more senses) to the παθητικὸς νοῦς: he has said that it is τιμιώτερον, i.e. that it has a priority of honor, but we might want a demonstration that it is prior in some other sense. The issue here is a special case of the issue of priority between δύναμις and ἐνέργεια as discussed in Metaphysics Θ 8 (which Aristotle echoes especially closely in the De Anima III 7 variant of our passage). Here as in Metaphysics Θ 8, the sense of priority to which δύναμις seems to have the strongest claim is priority in time, and in both texts Aristotle concedes that in one way δύναμις is prior in time, but argues that in another way δύναμις is not even temporally prior (οὐδὲ χρόνῳ in our passage). Knowledge κατὰ δύναμιν – that is, νοῦς in the sense of the bare δύναμις in the individual soul – is temporally prior to actual knowledge in the history of the individual soul, but, Aristotle claims, it is not temporally prior in the history of the universe, since there was never a time when there was no actual knowledge. And this is true not just because there has always been some human being or other who possesses actual 45 But the two passages are not identical: the first two lines’ words are identical, or almost identical, depending on which manuscripts we follow, but then III 5 continues ἀλλ᾿ οὐχ ὁτὲ μὲν νοεῖ ὁτὲ δὲ οὐ νοεῖ, and III 7 continues ἔστι γὰρ ἐξ ἐντελεχείᾳ ὄντος πάντα τὰ γιγνόμενα. Both make sense, and each is adapted to its context, in III 5 picking up from 430a17 “this νοῦς”, which eternally exists and is eternally thinking, in III 7 pointing out something common to intellectual and sensory cognition and making a transition to an account of sensory cognition and its cause. Ross says nothing about how these two different last lines would have arisen on his account. If Ross were right in calling III 7 a mere collection of scraps, so that the shared passage would make no contribution there, then we should delete the shared passage from III 7, not from III 5. But I agree with Klaus Corcilius, in his contribution to the present volume, that III 7 is not a mere collection of scraps, and that Aristotle does build on this passage there, as well as in III 5, which he is deliberately recalling in III 7 to bring out analogies between intellectual and sensory cognition. 46 When Ross decides that 430a19-22 makes no contribution to the argument of III 5, he may not have considered the possibility that νοῦς and ἐπιστήμη here are the same thing, perhaps due to a preconception that νοῦς means something like “intellect”, a power of reasoning in general, rather than a rational cognition of some determinate object.

134

STEPHEN MENN

knowledge (and, even granting the eternity of the human species, it is not obvious that, for any given immaterial νοητόν X, there has always been some human being who has knowledge of X), but because οὐχ ὁτὲ μὲν νοεῖ ὁτὲ δὲ οὐ νοεῖ – that is, because the actual knowledge itself is identical with the immaterial νοητόν and therefore exists separately and eternally, and is itself eternally knowing. As Aristotle then says, “when this has been separated, it is just what it is”. The subject of this assertion is ἐπιστήμη or νοῦς: grammatically, since χωρισθείς is masculine, its antecedent is νοῦς, but Aristotle is drawing no distinctions here between νοῦς and ἐπιστήμη. In speaking of this thing’s being separated, or of its being just what it is, Aristotle is calling on one of his basic technical distinctions: if a thing A exists, it exists either “separately and καθ᾿ αὑτό” (as far as I can tell there is no difference in meaning between these two terms) or not separately and καθ᾿ αὑτό. A exists καθ᾿ αὑτό if it is not predicated of some other underlying nature, not καθ᾿ αὑτό if it is so predicated. In Aristotle’s official phrase, A exists καθ᾿ αὑτό if “it is not, being something else, what it is”. Or, putting it the other way around: A exists not καθ᾿ αὑτό if “being something else, it is what it is”, as “the walking [thing], being something else [e.g. man or Socrates], is walking”47. In other words, A exists not καθ᾿ αὑτό if the thing which is A has some other underlying nature B, of which A is predicated, so that A exists only because B exists and is A; whereas if A exists καθ᾿ αὑτό, then A exists because there is something whose nature is just to be A. Cutting more finely than Aristotle usually does, we can distinguish two ways that A can exist not καθ᾿ αὑτό: A exists concretely not καθ᾿ αὑτό if A exists because B exists and is A (A = white, B = Socrates); A exists abstractly not καθ᾿ αὑτό if A exists because B exists and is called by some name paronymous from A (A = whiteness, B = Socrates, who is not whiteness but white)48. 47 That exists καθ᾿ αὑτό which “is not said of some other underlying thing [ὃ μὴ καθ᾿ ὑποκειμένου λέγεται ἄλλου τινός]: for example, the walking [thing], being something else, is walking [τὸ βαδίζον ἕτερόν τι ὂν βαδίζον ἐστί], and likewise the white, but substance, and whatever signifies a this, are not, being something else, what they are [οὐχ ἕτερόν τι ὄντα ἐστὶν ὅπερ ἐστίν]. So the things that are not [said] of some underlying thing [καθ᾿ ὑποκειμένου], I call καθ᾿ αὑτά, and the things that are [said] of some underlying thing I call accidents” (Posterior Analytics I 4, 73b5-10). 48 Compare Aristotle’s distinction between the modes of existence of τὸ βαδίζον (in my terms, existing concretely not καθ᾿ αὑτό) and of τὸ βαδίζειν (existing abstractly not καθ᾿ αὑτό), Metaphysics Z 1, 1028a20-29.

FROM DE ANIMA III 4 TO DE ANIMA III 5

135

Whenever whiteness exists, it exists abstractly not καθ᾿ αὑτό, and this is because whiteness cannot exist in separation from a body or surface which is white. We might also think that whenever knowledge exists, it exists abstractly not καθ᾿ αὑτό, because knowledge cannot exist in separation from a person or soul which is knowing. However, Aristotle’s view is, instead, that some kinds of knowledge can exist in separation from a knowing person or soul, and that some cannot. When X is an enmattered form, the knowledge of X cannot exist in separation from a soul (indeed, it cannot exist in separation from a body, see discussion below), but when X is a separate immaterial νοητόν, then the knowledge of X can exist in separation from a soul and from any other underlying nature, since the knowledge of X is just X: and, as we have seen, its being the knowledge or νοῦς of X is not some further attribute additional to its being the νοητόν X. When some knowledge is capable of existing separately, and when it has been separated – in particular, separated from any soul whose knowledge it has been – then it is just what it is. It is not some other underlying nature which is knowing, or some other underlying nature’s knowledge, but simply knowledge49, 50. 49 I intend this to be neutral as to whether the knowledge was at one time an attribute of something else and at another time exists separately, or whether it has merely at one time been considered as an attribute of something else and at another time as existing separately – it does in fact exist separately, even if it turns out at each time there are one or more souls that possess it. Aristotle does sometimes use χωρίζειν for a mental act of separating (so Metaphysics Z11 1036b7, and in a number of places, collected in Bonitz’ Index Aristotelicus, where Plato or Platonists are the subject – this usage is already in Plato). Compare Caston’s discussion (1999), especially p. 208. I think he goes too far: an aorist participle, unlike a perfect participle, does signify prior action, but this need not entail that the thing has really at one time existed separately and at another time not. In the example Caston cites from DA 403a14-15, I would say that we have a temporal sequence in a thought-experiment. 50 I thus disagree with Caston’s discussion of what it means for νοῦς to be separate or separable, see especially his p. 210. Caston talks about separating the rational “capacity” from other psychic capacities, but Aristotle is talking about separating an ἐνέργεια, not a δύναμις. It is presumably true that the rational power of a soul can be instantiated without the non-rational powers (in the souls of the heavenly bodies), but why would that make it a ποιητικὸς νοῦς? Caston’s use of the word “God” here is dangerous – perhaps every rational soul without non-rational powers can be called a god (or, if it has a body, the soul-body composite can be called a god), but there are many gods, and it cannot be said of gods in general that they are essentially ἐνέργεια or are identical with their νοητά, and there is no Aristotelian support for saying that the God or gods of whom these things are true are souls or have souls. The parallel Caston cites from DA II 2, 413b24-7 is not genuinely parallel: this is explicitly about νοῦς as a θεωρητικὴ δύναμις, which is described as a type of soul. I also think that, in context (and with a reference back to the

136

STEPHEN MENN

It needs stressing that Aristotle’s view is not just that the immaterial νοητόν X is some νοῦς or ἐπιστήμη, but that it is the very νοῦς or ἐπιστήμη that the soul has of X. When someone ἔχει the ἕξις of ἐπιστήμη of X, the ἕξις which he ἔχει is just X itself. Numerically one and the same thing can be your ἕξις, and also my ἕξις, and also a separately existing substance. This thing is νοῦς κατ᾿ ἐνέργειαν; and the rational power of the soul is called νοῦς κατὰ δύναμιν, not because it is able to become this thing, but because it is able to ἔχειν this thing and so to be called by a name paronymous from it, νοῶν or ἐπιστήμων51. Now “when this has been separated it is just what it is, and this alone it is immortal and eternal”. In other words, the knowledge that the soul possesses (of an immaterial νοητόν) is eternal and is capable of existing without the soul, both before the soul came to be and after the soul passes away. The fact that the soul possesses something eternal does not end of DA II 1), this passage is talking about the immortality of a part of the soul of a rational animal, and not just about “taxonomic” separation, i.e. about whether the rational power is sometimes instantiated without the irrational powers or without a body. The view I take Aristotle to be suggesting in the II 2 passage, that the rational soul is immortal, is contradicted by the view I take him to be asserting in III 5, that only the ποιητικὸς νοῦς is immortal and that the rational soul is not. But the several references in DA I-II to the possible immortality of the rational soul are highly tentative, and defer the problem for a further scientific investigation; when that investigation arrives in DA III 4-5, and the necessary distinctions are made, it turns out that the arguments for the immortality of νοῦς apply only to the ποιητικὸς νοῦς, and not to the νοῦς which is part of the soul. 51 Describing this knowledge as a ἕξις raises the question: when Aristotle calls it an ἐνέργεια, does he mean a first ἐνέργεια, i.e. a habit of knowledge, or does he mean a second ἐνέργεια, i.e. an activity of contemplating? His descriptions of νοῦς in the δύναμις-sense clearly refer to a first δύναμις, so we would expect that the contrasting ἐνέργεια-sense would be a first ἐνέργεια. On the other hand, the parallel with Metaphysics Λ9, where the separate νοῦς is described more precisely as νόησις, i.e. as a second ἐνέργεια, suggests that the separate νοῦς of DA III 5 too is a νόησις, i.e. that it is always contemplating and that its contemplation is essential to it and not a superadded attribute. This is connected with the question of what Aristotle means by saying that, in the “theoretical”/immaterial case, ἐπιστήμη κατ᾿ ἐνέργειαν is identical with its object: does he mean the first ἐνέργεια, the ἕξις (which is what he normally calls ἐπιστήμη) or does he mean the second ἐνέργεια, the θεωρεῖν? I think the answer must be that the ἐπιστήμηἕξις is the object itself, which is present in the soul in the sense that it is capable of acting in and on the soul (in the material case it is the λόγος of the object present in the soul), and that the θεωρεῖν is not precisely the object, but is simultaneously an active ἐνέργεια of the object and a passive ἐνέργεια of the soul. It remains true that (in the immaterial case) the object is essentially ἐνέργεια, i.e. is essentially νόησις, but for us to have the object, or for the object to be present in us (or present to us), is not the same as for us to ἐνεργεῖν, or for it to ἐνεργεῖν in us or on us: ἔχειν, here as everywhere else, is merely a potentiality for ἐνεργεῖν.

FROM DE ANIMA III 4 TO DE ANIMA III 5

137

imply that the soul is eternal. If Alexander possesses an incorruptible diamond, this does not imply that Alexander is incorruptible: when Alexander passes away, Alexander’s diamond will not pass away, but will simply cease to be Alexander’s diamond. When I pass away, my knowledge (of an immaterial νοητόν) will not pass away, but will simply cease to be my knowledge – and it was never my knowledge alone, but was my knowledge in so far as it was present in me, and your knowledge in so far as it was present in you. At any rate, it will cease to be mine, and I will cease to be, if “I” means my rational soul (or the whole soul, or the soul-body composite) rather than meaning the knowledge present that is currently present in my soul. That seems like the more natural way to use the pronoun “I”. But what Aristotle means by “I” and “we” is contentious, as becomes clear from what follows. I do not claim certainty about what Aristotle means by 430a23-25: οὐ μνημονεύομεν δέ, ὅτι τοῦτο μὲν ἀπαθές, ὁ δὲ παθητικὸς νοῦς φθαρτός. καὶ ἄνευ τούτου οὐθὲν νοεῖ.

But I will offer what I think are the three most plausible options, and say which I prefer and why. Recall that Aristotle has just said that only the ποιητικὸς νοῦς (that is, that knowledge of an immaterial νοητόν which we possess) is immortal and eternal. He may now be saying (Option 1) that after our potential νοῦς has ceased to exist, and after the ποιητικὸς νοῦς has therefore been separated from it, “we” (whatever was in us and survives) will not remember. “We” will certainly νοεῖν, but νοεῖν is different from remembering (and even from διανοεῖσθαι, reasoning), because remembering (and reasoning) depend on a potentiality and a process of actualizing that potentiality: they cannot happen when there is only the ποιητικὸς νοῦς, which is always actually knowing everything that it is capable of knowing. As Aristotle says here, “this is without πάθος”, whereas memory depends on a πάθος (or on two successive πάθη, first of forgetting and then of remembering): the only “νοῦς” that is capable of these πάθη is corruptible, and once it has been corrupted, there is no longer anything capable of remembering. This is important for the issue of immortality which Aristotle has just mentioned, because it means that “we” will not remember anything from this life, but will simply continue to know the same eternal truths that “we” knew from eternity before this life.

138

STEPHEN MENN

Option 1 seems to be supported by the quasi-parallel in De Anima I 4, 408b18-30. As Aristotle says there, νοεῖν or θεωρεῖν is ἀπαθές, although it “is quenched” [μαραίνεται] when something else within us is corrupted: “reasoning [διανοεῖσθαι] and loving or hating are not its πάθη, but the πάθη of what possesses this, inasmuch as it possesses it: whence when this [= the possessor] perishes, it [= the νοῦς] does not remember or love, since these [πάθη] belonged not to it but to the compound [κοινόν] which has perished: but νοῦς is perhaps [ἴσως] something more divine and ἀπαθές” (408b25-30). Particularly the distinction here between διανοεῖσθαι, a πάθος belonging to something composite, and νοεῖν, which is something higher and simpler and ἀπαθές, suggests the DA III 5 distinction between the παθητικὸς νοῦς (introduced in DA III 4 as “that by which the soul διανοεῖται and affirms”, 429a23) and the ποιητικός. The DA I 4 passage would then be saying that the activity of νοεῖν is “quenched” in us only in the sense that conditions may prevent that ποιητικὸς νοῦς from being present in, or from acting in and on, the παθητικός; and DA I 4 and III 5 would both be saying that what remembers is not the ποιητικὸς νοῦς but rather the παθητικὸς νοῦς when the two are conjoined, and thus that we will not remember after death. However, the DA I 4 passage is very tentative, like all the passages in DA I-II talking about separation or immortality, and when it suggests that νοῦς is ἀπαθής it may be saying this of all νοῦς, not yet distinguishing ποιητικός from παθητικός. So DA I 4’s contrast between possessor/ composite and possessed may be not between παθητικός and ποιητικός, but between the body-soul composite and νοῦς considered as a separable part of the soul; in which case it will not be directly parallel to III 5, 430a23-2552. But one thing that clearly emerges from the DA I 4 passage is that νοεῖν and μνημονεύειν are different: although οὐ μνημονεύομεν may well mean (and on Option 1 does mean) that there is no remembering without the potential νοῦς, ἄνευ τούτου οὐθὲν νοεῖ cannot mean that there is no νοεῖν without the potential νοῦς (this is clearly false on Aristotle’s view), but only that there is no νοεῖν without the agent νοῦς. 52 This is how Caston reads it (V. Caston (1999), p. 213-214, n. 19), and this is what is suggested by the immediately preceding comparison with what happens to the senses in old age. But I am not sure this interpretation is fully determined by the text of DA I 4, and the echoes with DA III 4-5 are surprisingly close if this is all that is going on.

FROM DE ANIMA III 4 TO DE ANIMA III 5

139

One difficulty with Option 1 is that it is hard to see what motivates the last sentence ἄνευ τούτου οὐθὲν νοεῖ. It does not seem to connect with the point about memory, but seems to be just a general comment, “without this [i.e. without the ποιητικὸς νοῦς], nothing νοεῖ”, picking up from before οὐ μνημονεύομεν – the reading implied by Ross’ parentheses around οὐ μνημονεύομεν δέ, ὅτι τοῦτο μὲν ἀπαθές, ὁ δὲ παθητικὸς νοῦς φθαρτός. Perhaps the best thing to say is that ἄνευ τούτου οὐθὲν νοεῖ picks up the praise-attributes of the ποιητικὸς νοῦς from before the parentheses: this alone is immortal and eternal, and all νοεῖν depends on it. Another difficulty is that, since Aristotle would be talking about memory in the proper sense as distinguished from intellectual cognition (which is about atemporal things, not about the past, and so cannot be memory), it is unclear why he would suddenly introduce this topic into III 5, when the discussion has been entirely about νοεῖν. The usual answer has been: Aristotle has just argued that at least something in us (but not our whole soul) is immortal, and so we want to know whether we will preserve our memories post mortem: the answer is no, which is disappointing but worth knowing, because it tells us that if only this is immortal it will not be a personal immortality. But I think this is a suspiciously “modern” reading: I do not see much sign that, for Aristotle or his interlocutors, a main concern in assessing some promised immortality is whether it would be personal, still less that the criterion for its being personal immortality is that we will remember our present lives. For these reasons it seems more promising to say that Aristotle is talking, not about whether “we” in a future state will remember things from this life, but about whether we in our present life remember the knowledge which has existed from eternity. There seem to be two options here. He could (Option 2) be picking up the question, deferred at the end of DA III 4, about the reason why we do not always νοεῖν. Then οὐ μνημονεύομεν would mean, not that we never remember, but that we do not always remember: the explanation would then follow, namely that although the ποιητικὸς νοῦς is ἀπαθής, the soul’s νοῦς is not ἀπαθής, but is subject to πάθη on account of which it sometimes forgets, and sometimes remembers, knowledge which is available to it. On this construal, saying that the παθητικὸς νοῦς is corruptible would not seem to contribute much to the argument; perhaps it is just an emphatic way of making the point that it cannot be expected to stay in the same state. On the other hand, there might be a point to saying ἄνευ τούτου οὐθὲν νοεῖ,

140

STEPHEN MENN

i.e. “without this [scil. the ποιητικὸς νοῦς], it [the παθητικὸς νοῦς] does not νοεῖν anything”: the soul’s νοῦς is not self-sufficient for knowledge but depends for knowing on participating in something extrinsic, so is not surprising that it is not always knowing. But Aristotle cannot want to say that we are not always contemplating because the ποιητικὸς νοῦς is not always acting – being eternal unmixed ἐνέργεια, it is always acting (as the sun is always shining), even if it is not always acting on us. And the reason why at some times it is not acting on us cannot come from any change in the ποιητικὸς νοῦς itself: rather, at some times it is not acting on me, either because I do not exist then, or because of some change in my παθητικὸς νοῦς which obstructs it from receiving the action of the ποιητικὸς νοῦς. So saying ἄνευ τούτου οὐθὲν νοεῖ, or mentioning the ποιητικὸς νοῦς in general, does not seem to help explain why we do not always νοεῖν: it should be enough to mention that the παθητικὸς νοῦς is sometimes in better and sometimes in worse states (and that, since it is corruptible, of course it does not think when it does not exist). So this is one difficulty for Option 2. Another difficulty is that on this interpretation (unlike Option 1, or Option 3 below), Aristotle would have to be using μνημονεύειν very broadly, to mean the same as νοεῖν or perhaps “continue to νοεῖν”. It would be mysterious why he has suddenly switched to speaking of μνημονεύειν: this passage and the quasi-parallel from DA I 4, cited above in discussing Option 1, are the only uses of μνημονεύειν in the De Anima. For these reasons, I prefer a variant on Option 2, on which, as on Option 2, 430a23-25 would be talking not about whether “we” in a future state will remember things from this life, but about whether we in our present life remember the knowledge which has existed from eternity. Aristotle may (Option 3) be taking for granted the explanation of why we do not always νοεῖν, and making instead the point that even when, in this life, we participate in the eternal knowledge and make it our knowledge, we are still not remembering it from before this life, as Plato says we are. As in Option 1, the ποιητικὸς νοῦς, being ἀπαθής, does not remember, so it could only be the παθητικὸς νοῦς that remembers. But “the παθητικὸς νοῦς is corruptible”, and so (we would have to add) it is also generated: so it cannot remember the knowledge that it had before this life, because it did not have the knowledge before this life, because it did not exist before this life, although the knowledge did. On this interpretation, we can translate οὐ μνημονεύομεν as “the theory

FROM DE ANIMA III 4 TO DE ANIMA III 5

141

of recollection is false”. This is worth saying, because Aristotle has come close to the theory of recollection in saying that our knowledge of immaterial νοητά has existed from eternity, and that we can come to possess and to exercise this knowledge because it is already there and available to us, and we are already in potentiality to it. But, Aristotle would now be saying, Plato is wrong to conclude that our soul preexisted and possessed this knowledge before our present life. This alone, the essentially actual knowledge, is immortal, existing from eternity, and we, the παθητικὸς νοῦς or rational soul, are not. While our knowledge preexisted, it was not at that time our knowledge, because our soul did not yet exist to possess it, and so, when it comes to be our knowledge, we are not recollecting it, but simply being acted on by it. On this interpretation, ἄνευ τούτου οὐθὲν νοεῖ would be saying: although we don’t remember this preeternal knowledge, it acts on us, and we cannot think without its acting on us. This starts a theme that Aristotle will pursue in later chapters of DA III, of the need of an already actual external actualizing cause for our episodes of cognition (of simple objects, which we can then compound in various ways), both of intellectual cognition as here, and of sensory cognition, which he will compare in III 753. 5. This reading of De Anima III 5 leaves a number of issues open. In this concluding section, I want to show how I would respond to Victor Caston’s challenge in “Aristotle’s Two Intellects: a Modest Proposal”, and where my solution differs from his; in the process, I will have to deal with another interpretive issue from DA III 4, from before the aporetic section 429b22-430a9. Caston says that his own interpretation identifies the ποιητικὸς νοῦς with God, and he lists me and Michael Frede as the only modern interpreters who agree with him. I did in fact make this identification in my 1992 paper54, but I would now want to be more careful. Strictly speaking, the question “what is the ποιητικὸς νοῦς?” is ill-posed, since we have no reason to think there is only one of it. Any separate immaterial 53

See Klaus Corcilius’ paper in the present volume. This identification explicitly in S. Menn (1992), p. 562, n. 26. On the other hand, I also said something which, if thought through, should imply the same qualification I would want to make now, p. 566, n. 29. 54

142

STEPHEN MENN

νοητόν is a ποιητικὸς νοῦς, and any separate immaterial νοητόν that my soul can νοεῖν is a ποιητικὸς νοῦς that can act on my soul. When Aristotle says “in every nature there is one thing which is matter for each genus (this is what is potentially all those things), and another which is the cause and agent/maker”, he does not mean that in each nature there is numerically only one agent and one patient (obviously ὁ παθητικὸς νοῦς signifies a type, not a unique individual), and there is no good reason to think that he means even that the agent νοῦς is numerically one. There are as many of them as there are separate immaterial νοητά. There might be one, or ten, or 47, or 55, or infinitely many: Aristotle is not concerned with this question in the De Anima, and the methods of the De Anima would not be able to resolve it55. Another problem – connected, as we will see, with the problem of whether the ποιητικὸς νοῦς should be identified with the one first God – is the problem of how it is a cause to our soul, and of what exactly it causes in our soul. Caston raises this problem in acute form, noting that the account of the soul’s intellectual activity in De Anima III 4 seems self-sufficient without the help of the ποιητικὸς νοῦς. Caston’s own solution is to say that the ποιητικὸς νοῦς – that is, on his account, God – is a final cause, or perhaps more precisely an exemplar: God “constitutes [by which I think Caston just means ‘exemplifies’] the complete actualization towards which all of our intellectual striving is directed, in emulation of his perfect state” (ibid.). Caston thinks that Aristotle introduces God in DA III 4-5, not because there is something that happens in the soul that could not be explained without God, but because we understand the soul better by putting it in cosmological and theological context, by comparing it with the divine exemplar. 55 Caston argues (V. Caston (1999), p. 212) that “there can be only one such intellect [scil. meeting the description of the divine intellect given in Λ 7-9], just because it is actuality”, citing Λ 8, 1074a35-7: “the first essence does not have matter – for it is actuality. Therefore the first mover is one both in account and in number, since it cannot be moved” (Caston’s translation). This shows that it cannot be numerically multiplied within its species, but Λ 8 also says that there are many separate immaterial substances, thus presumably many species of separate immaterial substances, each with one instance. Why shouldn’t each of these be a νοῦς (and a ποιητικὸς νοῦς)? What distinguishes them? Well, they are sciences, and so they are distinct because their objects are distinct. Of course, they are their objects, so if we don’t already know them, we won’t be able to understand how they differ. But that may be our problem rather than theirs. Perhaps they are distinguished by relations of essential subordination, by being said per prius et posterius. But Aristotle doesn’t give us much to go on. Certainly only one of them is the good-itself.

FROM DE ANIMA III 4 TO DE ANIMA III 5

143

I do not see any basis in the De Anima for saying that the ποιητικὸς νοῦς is a final cause to the soul. We have here a pair of something ποιητικόν and something παθητικόν, and Aristotle certainly thinks of the former as an efficient cause to the latter56. Of course, Metaphysics Λ 7 says that God moves the heaven as final cause and as ὀρεκτόν and νοητόν; and Eudemian Ethics VIII 3 also says that God is a final cause (as “to possess which” rather than “to benefit whom”, same distinction Λ 7, 1072b1-3) of human actions57. But for God to cause the heaven to move, he has to cause it to desire him, and to do this he has to cause it to know him: We desire because it appears [good or beautiful], rather than its appearing so because we desire it: for the starting-point [ἀρχή] is νόησις; and νοῦς [i.e. the παθητικὸς νοῦς of the heaven] is moved by the νοητόν. (Λ 7, 1072a29-30)

So while God is a final cause to the heaven of its moving, he is an efficient cause to the heaven’s παθητικὸς νοῦς of its knowing him, as a color is an efficient cause of its being seen58. So here as in the De Anima, the ποιητικὸς νοῦς seems to be simply an efficient cause to the παθητικὸς νοῦς, not a final one, and we cannot use its being a final cause to explain the peculiar way in which it is an efficient cause (the mere fact that it is an unmoved mover does not imply that it is a final cause, since the color is not a final cause of its being seen)59. In the 56

Caston allows that Aristotle would describe this final causality, like God’s causality on the heavens, as a special kind of efficient causality, but he says that it is not “what we would call a ‘causal’ relation” (V. Caston (1999), p. 224, cf. p. 200; Caston’s emphasis). I am not comfortable talking about “what we would call causality”. Modern science, unlike Aristotelian science, does not use the notion of cause (and without the scientific anchor, modern philosophers can, and do, use the word however they want to). If Aristotle says that arts are efficient causes, I think we should adapt to his usage, rather than saying that he does not mean what “we” mean by efficient cause. 57 On the meaning of the distinction between the two senses of οὗ ἕνεκα (which remains controversial), and on the text of Λ 7, 1072b1-3 (where there is a complicated manuscript situation), see S. Menn (2012). 58 And of course, at Λ 6, 1071b12 the mover of the heaven is κινητικὸν ἢ ποιητικόν; there is no reason why this conclusion should be subverted when we learn that it is an object of thought and desire. Likewise in DA III 10, where the ὀρεκτόν is an unmoved mover τῷ νοηθῆναι ἢ φαντασθῆναι, it seems to be an efficient cause of νοηθῆναι or φαντασθῆναι. 59 I disagree with Caston’s claim (V. Caston (1999), p. 219), that GC I 6 says that being productive κυρίως requires mutual contact. “Things which cannot touch each other cannot ποιεῖν and πάσχειν κυρίως” (322b22-4) is not making a point about mutuality,

144

STEPHEN MENN

Eudemian Ethics passage, God is a final cause of our actions because we do them for the sake of God, i.e. in order to possess God, i.e. to maximize the quantity and quality of our contemplation of God (to do this we need leisure, a soul undisturbed by passions, etc.), but there is no suggestion that God is the final cause of our contemplating God, and while we can say that we act in order to contemplate God, we cannot say that we contemplate God in order to contemplate God (we might say this negatively, meaning that the contemplation of God is not for the sake of anything else, but it would not be positively causa sui). Rather, God is the efficient cause of our contemplating him (once we are properly prepared and all obstacles are removed), and both God and our contemplation of God can in different ways be called the final cause of our other actions. I may not be disagreeing very radically with Caston here. So far, I have claimed only that God, or any ποιητικὸς νοῦς X, is the efficient cause of our contemplating X. I am not sure that Caston would disagree with this claim. But, he might say, if this is all that the ποιητικὸς νοῦς is cause of, then Aristotle is not positing it in De Anima III 5 to explain anything in the soul: if I have not already been contemplating the ποιητικὸς νοῦς, then I am not aware of any psychological phenomenon that needs explaining, and if I have already been contemplating the ποιητικὸς νοῦς, then I do not need reasons for positing it. The ποιητικὸς νοῦς is only doing explanatory work if it is the cause of our cognizing something other than itself. Aristotle does seem to imply that the ποιητικὸς νοῦς is the cause of our cognizing something other than itself when he compares it to light, which “makes what are potentially colors actually colors”. As I noted above, Aristotle is taking this comparison from Republic VI, but, as I also noted, he is modifying Plato’s account in ways that threaten to undermine the comparison. For Plato, a single first form of the Good is the cause of our νοεῖν each of the other νοητά. For Aristotle, if X is a nor is b26-29 (καὶ τούτοις ὡσαύτως needn’t mean that the contact must be mutual in this case too). In any case, 323a25 sq. makes clear that contact is not always mutual, and that movers are not always moved, although these do hold in the majority of cases and in the most familiar cases; there is no warrant for saying that only such mutual cases are cases of κινεῖν or ποιεῖν κυρίως. I agree with Caston (citing GC I 7, 324b13-15) that the final cause, qua final cause, is not a mover κυρίως, but being a final cause is not the only way of being an unmoved mover.

FROM DE ANIMA III 4 TO DE ANIMA III 5

145

separate immaterial νοητόν (like a Platonic form if there were any, and the De Anima hasn’t argued that there aren’t), then the argument of De Anima III 4-5 shows that X is itself a ποιητικὸς νοῦς, a pure ἐνέργεια with no potentialities needing to be actualized by something else: X is itself a cause of our νοεῖν X, and there seems to be no room for anything higher than X (like God, if X is not itself the first God) to be a cause of our νοεῖν X. So how is any ποιητικὸς νοῦς the cause of our νοεῖν anything other than itself? Let me mention one possible answer which I think can be rejected quickly. Aristotle might think that a ποιητικὸς νοῦς can be the cause of our knowing a plurality of intelligible contents because it itself knows those contents, that is, because it is itself the separately existing knowledge of those intelligible contents, so that it must itself be those intelligible contents: where for the one thing X to be the many things Y, Z, etc., they must be something like different parts of X or different aspects of X. I think this is a consistent and reasonable position. It was Plotinus’ position: for Plotinus, the many Platonic forms, or many sciences, are inseparable parts within νοῦς as a whole, like many theorems within a single universal science60. But Aristotle rejects all this, and more generally he denies that a separate immaterial substance can consist of parts (so Metaphysics N 2), because he denies the possibility of inseparable parts: for Aristotle as for Plato, to say that something is a whole is to say that it is both one and many, and while Plato in the second part of the Parmenides is apparently willing to tolerate such a compresence of contraries in the forms, for Aristotle such a compresence of unity and multiplicity is intolerable unless the whole is actually one and potentially many, because it can be divided into many parts and so become actually many. But in separate immaterial things there are no unactualized potentialities, and so a whole cannot be potentially many without being actually many: in which case it is not actually one, and thus not a whole. This does not force Aristotle to deny plurality in separate immaterial things: Y and Z can be two separate immaterial things, but then they cannot also be a single whole. So Y and Z can each be a ποιητικὸς νοῦς, Y being a knowledge of Y and Z being a knowledge of Z, but X cannot be a single ποιητικὸς νοῦς which is a knowledge of both Y and

60

See S. Menn (2001).

146

STEPHEN MENN

Z. Or, as Aristotle puts it with drastic compression in Metaphysics Λ 9, after asserting the identity of a separate νόησις with its object: There remains an aporia, whether the νοούμενον is composite: [if it were, the νόησις] would change among the parts of the whole. Perhaps everything which does not have matter is indivisible. (1075a5-7)

where “perhaps” [ἤ] is Aristotle’s way of introducing his solution to an aporia, and does not express any doubt. The only obvious remaining way that a ποιητικὸς νοῦς could be the cause of our νοεῖν something other than itself is if it is the cause of our νοεῖν the forms of material things: this, of course, has been the view of most commentators. But, once we reject Plotinus’ option of positing complexity within the separate νοῦς, it is mysterious how a single simple νοῦς can be the cause of our knowing a plurality of contents. Caston rightly emphasizes that the De Anima III 4 account of how we know the forms of material things makes no mention at all of such a higher cause; and, as I hope to have shown above, De Anima III 5 introduces the ποιητικὸς νοῦς in order to solve a problem from III 4 about our cognition of separate immaterial things, not about our cognition of enmattered forms. In fact, I think that things Aristotle says in De Anima III 4 imply, and are intended to imply, that a separately existing νοῦς has no role at all in our cognition of enmattered forms. God cannot give us knowledge of the form of a horse, because God himself does not know the form of a horse; speaking more precisely, the knowledge of the form of a material thing cannot be a separately existing substance, because it cannot exist apart from matter. Recall that in De Anima III 4, after describing νοῦς on the analogy of sensation, Aristotle had noted various differences between νοῦς and sensation. While sensation, like νοῦς, is in some way ἀπαθές, the exercise of sensation depends on an organ which is subject to πάθη: “the sensitive [power] is not without a body, but [νοῦς] is separable” (429b4-5). The exercise of sensation depends on external conditions (the appropriate object must be present and acting on the organ), and it may be obstructed if the organ is damaged, whereas νοῦς is immune to these limitations: once we have acquired an ἐπιστήμη, we can exercise it in contemplation independently of any bodily organ or any external body.

FROM DE ANIMA III 4 TO DE ANIMA III 5

147

Or so Aristotle says. But then, in a passage I skipped over before, he adds some qualifications. ἐπεὶ δ’ ἄλλο ἐστὶ τὸ μέγεθος καὶ τὸ μεγέθει εἶναι, καὶ ὕδωρ καὶ ὕδατι εἶναι (οὕτω δὲ καὶ ἐφ’ ἑτέρων πολλῶν, ἀλλ’ οὐκ ἐπὶ πάντων· ἐπ’ ἐνίων γὰρ ταὐτόν ἐστι), τὸ σαρκὶ εἶναι καὶ σάρκα ἢ ἄλλῳ ἢ ἄλλως ἔχοντι κρίνει· ἡ γὰρ σὰρξ οὐκ ἄνευ τῆς ὕλης, ἀλλ’ ὥσπερ τὸ σιμόν, τόδε ἐν τῷδε. τῷ μὲν οὖν αἰσθητικῷ τὸ θερμὸν καὶ τὸ ψυχρὸν κρίνει, καὶ ὧν λόγος τις ἡ σάρξ· ἄλλῳ δέ, ἤτοι χωριστῷ ἢ ὡς ἡ κεκλασμένη ἔχει πρὸς αὑτὴν ὅταν ἐκταθῇ, τὸ σαρκὶ εἶναι κρίνει. πάλιν δ’ ἐπὶ τῶν ἐν ἀφαιρέσει ὄντων τὸ εὐθὺ ὡς τὸ σιμόν· μετὰ συνεχοῦς γάρ· τὸ δὲ τί ἦν εἶναι, εἰ ἔστιν ἕτερον τὸ εὐθεῖ εἶναι καὶ τὸ εὐθύ, ἄλλο· ἔστω γὰρ δυάς. ἑτέρῳ ἄρα ἢ ἑτέρως ἔχοντι κρίνει. ὅλως ἄρα ὡς χωριστὰ τὰ πράγματα τῆς ὕλης, οὕτω καὶ τὰ περὶ τὸν νοῦν. Since magnitude and being-magnitude, water and being-water, are different (and so in many other cases, but not in all: in some cases they are the same), [the soul, or the person] judges being-flesh and flesh either by different [powers] or by [the same power] differently disposed. For flesh is not without matter, but is like the snub, this-in-this. So by the sensitive [power] [the soul, or the person] judges hot and cold, i.e. the things of which flesh is a ratio, but it judges being-flesh by a different [power], either by a separate [power] or [by a power which is to the sensitive power] as a bent line is to the same line when it is stretched out. Again, even in the case of things which are by abstraction, the straight is like the snub, for it is together with [μετά; i.e. cannot exist without] the continuous, whereas the essence, if being-straight is different from the straight, is something else, let it be the dyad. So it judges it either by a different [power] or by [the same power] differently disposed. So, in general, as the objects are separable from matter, so too will what is concerned with νοῦς [be likewise separable from matter]. (DA III 4, 429b10-22)

This is appallingly condensed, but it is possible to tease out the points Aristotle is making. It is often possible to distinguish that which is X from what-it-is-to-be-X, the essence of X: this will be true whenever X is a composite, a-form-in-a-matter, like flesh, which consists (say) in a certain ratio among the elements or among their primary qualities. So, in these cases, the question arises whether the thing which is X (an instance of X) and the essence of X are discerned by the same or different cognitive powers. Everything that Aristotle has said up to this point in De Anima III 4 leads us to expect that, in a case like flesh, they will be different: the essence or form of flesh will be grasped by νοῦς (which is “receptive of the form”, 429a15-16), while this particular instance of flesh will be discerned by sensation; or, more precisely, sensation will

148

STEPHEN MENN

perceive the matter (the elements or qualities) in which this particular instance of flesh resides, and so bring it before the judgment of νοῦς, and then νοῦς will judge that it falls under the concept of flesh. However, Aristotle disrupts this expectation by suggesting that the two judgments, about this particular flesh and about the essence of flesh, may be made not by two different powers but by the same power in two different conditions – like the same line when straight and when bent. There seem to be at least three broad views of what Aristotle is suggesting here, and of whether he is endorsing it. (1) While Alexander’s paraphrase of this passage in his De Anima (87, 4-23) seems to take it for granted that sensation will know flesh and that νοῦς, a really distinct power, will know the essence or form of flesh – and so Alexander simply skips the suggestion that they might be known by the same power in two different conditions – Themistius suggests that in order to be able to make the judgment “flesh is not the same as the essence of flesh” we must somehow grasp both terms by the same power (as we must grasp both white and sweet by the common sense in order to judge whether the white thing and the sweet thing are the same). The power that grasps both will be νοῦς: νοῦς more obviously grasps the essence of flesh, but it must somehow also be able to grasp flesh, and in doing so it “becomes as it were composite when it νοεῖ the composite” (Themistius In de Anima paraphrasis 96, 25), and can therefore be compared to a line bent back on itself. So, for Themistius, Aristotle would be suggesting, and endorsing, that it is νοῦς in different conditions that knows the essence of flesh and also flesh; and something like this interpretation seems to have been followed by most ancient and medieval commentators. Thomas Aquinas, perhaps just making explicit what Themistius is thinking, says that Aristotle is endorsing both the suggestion that flesh and the essence of flesh are known by different cognitive powers and the suggestion that they are known by the same power in different conditions. Usually sensation grasps flesh and intellect grasps the essence of flesh, but when intellect compares flesh with the essence of flesh, then as Themistius says intellect must grasp them both: “[intellect] knows the nature of the species, or what something is, by ‘stretching itself out straight’, but it knows the singular itself by a certain ‘bending-back’, inasmuch as it goes back upon the phantasms from which intelligible species are abstracted” (Thomas, In de Anima #713).

FROM DE ANIMA III 4 TO DE ANIMA III 5

149

This view is attractive in its insistence that it is possible to think about sensible things, and not merely to sense them. But it is not compatible with the details of the text, “by the sensitive [power] [the soul, or the person] judges hot and cold, i.e. the things of which flesh is a ratio, but it judges being-flesh by a different [power], either by a separate [power] or as a bent line is to the same line when it is stretched out”. The first power mentioned is the sensitive power, and the second power is either a power separate from the sensitive power (that is, presumably, νοῦς), or else it is to the sensitive power as the bent line is to the straight line. The text does not open the possibility that both powers are νοῦς differently disposed: either the first is sensation and the second is νοῦς, or they are both sensation differently disposed. And, if they are both the same power differently disposed, then this power is compared to a straight line when it is directed toward flesh, and to a bent line when it is directed toward the essence of flesh, not vice versa as Themistius and Thomas insist61. (2) So we might say, as Hicks and Ross suggest, that Aristotle is asking not whether a particular act of cognition should be attributed to sensation or to νοῦς, but whether sensation and νοῦς are themselves two separate powers: “it seems more probable that Aristotle is merely saying in l. 14-17 that the faculty of sense-perception and that of reason are either separate faculties or one faculty operating on different objects, the one on sensible things, the other on essences” (Ross, ad loc.). Hicks thinks that Aristotle is in fact favorably inclined to the suggestion that intellect is “only … sense in a different relation”, on the ground of a general suspicion of positing distinct parts of the soul (Hicks, ad loc., p. 487), whereas Ross thinks that Aristotle would be raising the suggestion to immediately 61 Zabarella considers and rejects a view like this, p. 47r columns 1-2 (in the middle of his long and complicated commentary on Textus X, immediately before introducing his own view). Zabarella’s own view is that the subject of the sentence is not the soul, or the composite human, but rather the intellect. The intellect can grasp by the sensitive power in that the intellect acts in a certain way when moved by the sensitive power, or more precisely by the phantasm, whereas the intellect acts in another way when moved by the ποιητικὸς νοῦς (which Zabarella takes to be a substance existing outside us). Or, as he says, you can also say that the intellect grasps the two objects when moved by the same thing in two different conditions, namely when it is moved by a phantasm acting according to the phantasm’s own nature, or when it is moved by a phantasm illuminated by the ποιητικὸς νοῦς. This shows ingenuity, but also, I would say, desperation. Zabarella very much wants to avoid the suggestion that νοῦς is just sensation disposed in a particular way.

150

STEPHEN MENN

reject it (“there is no doubt that he thought of the two faculties as entirely different, except in the fact that both are forms of apprehension”). But, against both Hicks and Ross, it would be bizarre if Aristotle were suddenly expressing doubts about whether νοῦς and sensation are separate, ten lines after declaring, “the sensitive [power] is not without a body, but [νοῦς] is separable” (III 4, 429b4-5). (3) The key to the correct interpretation is to see that Aristotle is comparing sensation, not with νόησις as such, but specifically with the νόησις of enmattered forms such as the essence of flesh or the essence of snubness. Of course, νοῦς is separable, but Aristotle is suggesting that the νόησις of enmattered forms may be inseparable from sensation (and thus inseparable from matter, since sensation is not without a bodily organ): such νόησις would then be the act of a power which is numerically the same as the sensitive power, but differs from it in λόγος. More precisely, the power would be the sensitive power differently disposed, and essentially dependent or parasitic on the sensitive power, as bent line is on straight line, and as snub is on nose, or flesh on the elements and their qualities. So it would follow, as Aristotle says it does, that “in general, as the objects are separable from matter, so too will what is concerned with νοῦς [be likewise separable from matter]”: the knowledge of separate immaterial things is itself separable from matter, but the knowledge of inseparable enmattered forms is itself inseparable from matter. Aristotle certainly means to endorse this suggestion, and it fits closely with things he says elsewhere. In several places in the De Anima Aristotle raises the question whether νοεῖν can take place without imagination [φαντασία] or without an image [φάντασμα]. In De Anima I 1, in asking whether the soul or any of its activities can exist apart from bodies, Aristotle says that νοεῖν is the most plausible case, but notes that “if this too is a kind of imagination, or not without imagination, then this too would not be able to exist without a body” (403a8-10); in Book III the problem of the relation between νοῦς and imagination is a constant theme62. The “collection of scraps” DA III 7 says that “the soul never νοεῖ without an image” 62 So in III 3 it “seems” [δοκεῖ] that νοεῖν consists of two activities, imagination and judgment [ὑπόληψις], where imagination is a necessary precondition of judgment, 427b14-16 and b27-9.

FROM DE ANIMA III 4 TO DE ANIMA III 5

151

(431a16-17; cf. 431b2, “the νοητικὸν νοεῖ the forms in images”). Most remarkably, De Anima III 8, after recalling that the soul’s νοῦς is potentially the forms of intelligible things, as its sensory power is potentially the forms of sensible things, says: Ἐπεὶ δὲ οὐδὲ πρᾶγμα οὐθὲν ἔστι παρὰ τὰ μεγέθη, ὡς δοκεῖ, τὰ αἰσθητὰ κεχωρισμένον, ἐν τοῖς εἴδεσι τοῖς αἰσθητοῖς τὰ νοητά ἐστι, τά τε ἐν ἀφαιρέσει λεγόμενα καὶ ὅσα τῶν αἰσθητῶν ἕξεις καὶ πάθη. καὶ διὰ τοῦτο οὔτε μὴ αἰσθανόμενος μηθὲν οὐθὲν ἂν μάθοι οὐδὲ ξυνείη· ὅταν τε θεωρῇ, ἀνάγκη ἅμα φάντασμά τι θεωρεῖν· since nothing at all exists separated beyond sensible magnitudes, as it seems, the intelligibles are in sensible forms, both those [intelligibles] which are said by abstraction [i.e. mathematicals] and those which are states and affections of sensibles. And for this reason, if [νοῦς? the person?] did not sense anything, [it/he] would not learn or understand anything, and whenever [it/he] contemplates, [it/he] must always contemplate some image at the same time (432a3-9),

adding that “the first thoughts”, the simples as opposed to the compounds which are affirmed or denied, “are not images, but are not without images” (432a12-14). Now it would be very surprising if, so soon after De Anima III 5, Aristotle has decided that there are no intelligible substances separated from bodies. But I take it that the qualification “as it seems [ὡς δοκεῖ]” is crucial. While the Platonists believe that our knowledge of mathematics and of value-predicates plainly requires the existence of forms separate from matter, Aristotle thinks this is much more problematic. Our knowledge can arise without real separation, and in fact all of the obvious cases of our knowledge can be shown to depend on sensible things; in particular, our knowledge both of natural forms and of mathematical ones, although it is not simply an activity of the sensitive power in the way that imagination is, is essentially dependent on such an activity, so that our thinking of such forms must always be accompanied by an image. Aristotle himself believes that there is at least one ποιητικὸς νοῦς existing separately from matter, and that at least some human beings have knowledge of it, but he does not claim to have proved this in DA III 4-5: he has proved only that if there is an intelligible substance existing separately from matter and if some human beings have knowledge of it, then it is a ποιητικὸς νοῦς. All the knowledge that we can readily point to, and whose existence we can take for granted at this stage in the argument, is of forms that cannot exist apart from

152

STEPHEN MENN

matter, and the knowledge itself depends on sensation and thus on matter. And this is just the conclusion that Aristotle was drawing at DA III 4, 429b10-22, on the interpretation I am urging. The basis for this conclusion is that, as he says there, natural and even mathematical things are like the snub: that is, they each essentially presuppose some particular kind of matter, so that not only can they not exist without that matter, they cannot even be defined without it63. Aristotle likewise says that at least natural things are said like the snub in Metaphysics E 1 and Physics II 2, drawing the lesson is that the student of nature will study forms (and thus will use definitions and demonstrations), but that he will study forms in such a way that he studies their appropriate matter at the same time. But if we cannot know the forms of natural things without also knowing their matter, and if a pure νοῦς (a νοῦς that could exist and operate without any material substrate or organ) would necessarily grasp their forms alone without their matter, then it follows that they cannot be grasped by a pure νοῦς, but only by a νοῦς operating in dependence on sensation. So Aristotle has to qualify the claims he made just before our passage, that “the sensitive [power] is not without a body, but [νοῦς] is separable” (DA III 4, 429b4-5), and that once someone has ἐπιστήμη he can exercise it without dependence on external things. These claims are true of νοῦς and ἐπιστήμη as such, but not of the νοῦς and ἐπιστήμη 63 One Aristotelian way to explain why this is so comes from the Posterior Analytics II (and Metaphysics Z 17) analysis of the investigation τί ἐστι X: the question what X is is meaningful only on the assumption that X is, and what X is stands to that X is as cause to effect. And the only way to investigate the cause of X’s existence is to rewrite “X is” in the form “S is F” where S is the per se subject of X (e.g. “there is a lunar eclipse” = “the moon is darkened at opposition”), investigate the cause, to S, of its being F; and the definition of X, which expresses the essence of X, will include S and F and the cause of S’s being F. So we cannot grasp the essence of any natural thing X without knowing that X is instantiated, or at least representing X as instantiated (we can investigate the cause of lunar eclipses even though the moon is not eclipsed right now, but we will need an imaginative representation of an eclipse in memory, and even if the moon is now eclipsed, investigating its cause will also require memories of past eclipses), which means that we cannot know the essence without sensation or imagination. And the essence of X will contain a reference to the per se subject of X, which, if X is a natural substance, is the appropriate natural matter of X; so the essence cannot be known without the matter, and the matter cannot be known without sensation. In general, it is better to describe the progression from sensation of X to grasping the essence of X as a progression from knowledge that to knowledge why, rather than describing it as a gradual dematerialization to reveal the form of X stripped of its matter. We are trying to find the cause, not trying to escape from matter. There is helpful exploration of these themes (and others) in R. Roreitner (2018).

FROM DE ANIMA III 4 TO DE ANIMA III 5

153

of enmattered forms, which cannot be acquired or exercised without sensation and matter. So Aristotle concludes, qualifying his earlier claims, that: Ὅλως ἄρα ὡς χωριστὰ τὰ πράγματα τῆς ὕλης, οὕτω καὶ τὰ περὶ τὸν νοῦν. In general, as the objects are separable from matter, so too will what is concerned with νοῦς [be likewise separable from matter]. (429b21-22)

The conclusion of all this is that the separate νοῦς of De Anima III 5 does not exist in the case of our knowledge of enmattered forms, but only for our knowledge of separate immaterial things; or, more precisely, our knowledge of separate immaterial things is a separate νοῦς, whereas the knowledge of enmattered forms exists only in souls, indeed only in embodied souls. And consequently the ποιητικὸς νοῦς causes us to know it, and nothing other than it. By rejecting a good-itself superior to νοῦς, and by rejecting a single νοῦς-whole with many parts or aspects, and by denying that the forms of natural things can exist or even be cognized apart from matter, Aristotle has undermined his Platonic comparison with the light that “makes what are potentially colors actually colors” (DA III 5, 430a16-17) – that is, if this is taken to mean “making potentially intelligible objects outside the soul actually intelligible”, rather than “making the potentially intelligent soul actually intelligent”. Each ποιητικὸς νοῦς is a light that reveals itself, and only itself. So I agree with Caston that it does not explain anything that happens in the soul. It causes something that happens in the soul, by efficient causality, not final, but all it causes is that we know it. And if we do not already know it, what Aristotle says about it in De Anima III 4-5 gives us no reason to believe in its existence64. 64 Perhaps we can say, in trying to explicate a less restrictive interpretation of the ποιητικὸς νοῦς “making what are potentially colors actually colors”, that when it acts on us in such a way as to make us understand it, it also facilitates our understanding the things that are causally dependent on it – so perhaps grasping the movers of the heavens helps give us scientific knowledge of the heavens themselves, and of the things that the heavens in turn cause. Aristotle says in several passages that, in investigating principles, we start from a situation where the things most knowable to us are not the things most knowable in themselves, and seek to arrive at a situation where the things most knowable in themselves are also most knowable to us, and in Topics VI 4 he adds that “perhaps [ἴσως] too what is knowable without qualification is not what is knowable to everyone, but what is knowable to those whose thought is well disposed, just as what is healthy without qualification is what is healthy to those whose body is well disposed” (142a9-11; a similar comparison probably implied at Metaphysics Z3 1029b3-12, especially b5-8).

154

STEPHEN MENN

But I think this is perfectly all right with Aristotle. He is not trying to prove the existence of God, or prove the existence of separate immaterial νοητά, in De Anima III 5, any more than he is in the equally theological conclusion of the Eudemian Ethics. He does try to prove the existence of separate immaterial νοητά – each of which is a ποιητικὸς νοῦς, and the first of which is the first God and good-itself – in Metaphysics Λ. De Anima III 5 does not give any reason to believe in the existence of a ποιητικὸς νοῦς, or of any separate immaterial νοητόν, except in the sense that, if you already believe in the existence of a separate immaterial νοητόν, it gives you reason to think that this separate immaterial νοητόν is a ποιητικὸς νοῦς. De Anima III 5 is introduced to solve an aporia hanging over from the end of DA III 4, but this aporia does not arise for you unless you already believe in separate immaterial νοητά. De Anima III 5, like so many other texts in Aristotle, is arguing not against “atheists” or materialists, but against people with too “low” a conception of the divine things existing separately from matter: Although the Forms have manifold difficulties, what is most absurd is to say that there are natures besides those within the heaven [i.e. within the sensible world], but that these are the same as the sensibles except that these are eternal while those are corruptible. For they say that these are human-himself and horse-itself and health-itself, and nothing other [than human, horse, etc.], doing So the ποιητικὸς νοῦς, while intrinsically most knowable, would be initially very difficult for us to know (“as bats’ eyes are to the light of day, so is our souls’ νοῦς to the things that are by nature most manifest of all”, Metaphysics α 1, 993b9-11), and in coming to know it by a series of steps, we are being brought to cognitive health, coming to have our thought well disposed, so that the things more knowable in themselves will also be more knowable to us, i.e. so that we will know the effects through the causes and not the causes through the effects. (See S. Menn (2019), for a discussion of these texts and their parallels.) At De Anima III 4, 429a30-b5 Aristotle insists that, while our looking at very bright things makes us temporarily less able to see other things, “νοῦς, when it νοεῖ something intensely νοητόν, does not νοεῖν the inferior things less, but rather more” (b3-4): this is supposed to be diagnostic of νοῦς’ not having a bodily organ, but it looks as if it is intended to signal disagreement with Plato’s Sun comparison, where contemplating the Form of the Good makes us temporarily less able to discern the less bright things. So perhaps a ποιητικὸς νοῦς, besides making us actually knowing it, and making itself actually intelligible to us, will also make other things more intelligible to us to the extent that it causes them; I do not see how it would make anything other than its effects more intelligible to us. But none of these cognitive effects on us give us reason to believe in a ποιητικὸς νοῦς, unless we have recognized it as the cause of some other effect such as the motion of the heavens, i.e. by working through something like the argument of Physics VIII and Metaphysics Λ, not just De Anima III 4-5.

FROM DE ANIMA III 4 TO DE ANIMA III 5

155

much the same as those who say that the gods exist but are human-shaped: for neither were those [the poets] positing anything other than eternal humans, nor are these [the Platonists] making the forms anything other than eternal sensibles. (Metaphysics B 2, 997b5-12).

Against people who believe in separate mathematicals, separate virtues, and separate horses, and who fall into great difficulty in explaining how these things can interact with the soul so as to be known (Sophist 248a4e6), Aristotle argues that the only thing that exists separately from matter and is capable of acting on the soul is νοῦς, and νοῦς of a special kind: not a δύναμις of νοῦς but a pure ἐνέργεια of νόησις, and not a νόησις of natural things or of abstractions, but νόησις νοήσεως alone.

DE ANIMA III 61 Sylvain DELCOMMINETTE

1. Introduction Comme Michel Crubellier le faisait plaisamment remarquer lors du Workshop dont est issu le présent volume, on a parfois l’impression, à la lecture des commentaires du De Anima, qu’Aristote a tellement sollicité ses lecteurs dans les pages qui précèdent que ceux-ci semblent avoir épuisé leurs ressources à l’orée du chapitre 6 du livre III. De fait, ce chapitre ne semble pas avoir reçu toute l’attention qu’il mérite2. Je voudrais tenter de montrer qu’il est pourtant essentiel à la noétique d’Aristote et s’inscrit parfaitement dans le projet général du De Anima tel qu’il est développé depuis le livre II. Commençons par ce deuxième point. Pseudo-Simplicius (c’est-à-dire sans doute Priscien de Lydie) explique qu’alors que pour les autres facultés de l’âme (végétative et sensitive), Aristote partait de leur objet et considérait leur acte avant leur οὐσία, parce qu’il procédait du plus clair 1 C’est un grand honneur et un grand plaisir pour moi de dédier ce texte à Michel Crubellier, dont les lectures lentes et minutieuses du texte aristotélicien, que j’ai pu apprécier notamment lors des séminaires qu’il a co-organisés avec Annick Stevens sur Métaphysique Ζ et Η, ont beaucoup influencé ma propre approche. J’espère que les pages suivantes, consacrées à un texte que nous considérons tous deux comme important mais lisons de manière différente, contribueront à entretenir nos discussions. Je remercie l’ensemble des participants au Workshop pour leurs remarques et objections, en particulier David Charles, Klaus Corcilius, Michel Crubellier, Charlotte Murgier et Annick Stevens. Un grand merci également à Sean Kelsey qui m’a envoyé des commentaires écrits sur la dernière version de cet article. 2 Je tiens toutefois à souligner l’importance des deux contributions d’E. Berti (1978) et (1996) sur ce chapitre, même si je serai amené à être en désaccord avec les positions qu’il y défend sur plusieurs points fondamentaux. Voir à présent également Trentini (2016), article auquel je n’ai pu avoir accès qu’après la rédaction définitive de mon texte.

158

SYLVAIN DELCOMMINETTE

au moins clair pour nous, dans le cas de l’intelligence il procéderait de manière inverse, en commençant par déterminer le type de chose qu’elle est, qui serait mieux connu, avant de se tourner vers ses objets, ce qui dans les deux cas nous permettrait de saisir l’acte de l’intelligence dans son lien à l’intelligible3. Il poursuit en affirmant que telle est la démarche que nous avons suivie pour l’intelligence passive (en III 4), et que nous sommes à présent en train de reprendre à propos de l’intelligence productive, dont l’essence a été déterminée en III 5 et dont on va étudier les objets dans le présent chapitre (III 6)4. Même si l’on laisse de côté cette seconde division, qui semble traiter l’intelligence passive et l’intelligence productive comme deux entités indépendantes dont chacune aurait ses objets et ses actes propres, cette interprétation demeure problématique, dans la mesure où lorsque Aristote annonçait son plan en II 4, 415a14-22, consistant à commencer par étudier l’objet de chaque faculté, puis son acte, avant de se tourner vers son ce que c’est, il incluait explicitement la triade νοητόννοεῖν-νοητικόν parmi les notions devant être étudiées de cette manière. C’est pourquoi R. Polansky me semble avoir raison de considérer qu’Aristote suit en réalité au sujet de l’intelligence exactement le même plan qu’à propos de l’âme sensitive : de même que là, l’étude de chaque sens particulier, qui s’appuyait effectivement sur un examen du sensible correspondant, était précédée par un chapitre sur la nature de la sensation en général (II 5), de même ici, après avoir étudié différents problèmes généraux sur l’intelligence et sa nature dans les chapitres III 4 et 5, problèmes qui répondent d’ailleurs clairement et parfois explicitement à ceux étudiés en II 5 à propos de la sensation, il passe en III 6 à une étude plus précise des différentes formes d’intelligence qui s’appuie sur l’examen de ses objets5. Il n’y a donc aucune rupture entre l’étude de l’âme sensitive et celle de l’intelligence : le plan suivi est dans les deux cas le même, et correspond exactement à celui annoncé en II 4. Lu dans cette optique, le but du chapitre est donc d’éclairer les différents actes de l’intelligence, et partant la nature de l’intelligence elle-même 3 [Simplicius], In DA 248, 21-27. Même explication chez Philopon, In DA 542, 21-26. Cette interprétation est apparemment adoptée par G. Rodier (1900), p. 467-468. De manière analogue, mais plus simple, Thomas d’Aquin considère quant à lui qu’« après avoir déterminé ce qui concerne l’intellect, le Philosophe précise ici ce qui se rapporte à son activité » (Thomas d’Aquin (1999), p. 362). 4 [Simplicius], In DA 248, 27-33. 5 R. Polansky (2007), p. 473.

DE ANIMA III 6

159

comme faculté, en s’appuyant sur ses différents objets. C’est en ce sens qu’il revêt une importance capitale pour la noétique d’Aristote. Quant aux objets envisagés dans ce chapitre, ils sont de deux types : ceux dans lesquels il y a déjà une σύνθεσις τῶν νοημάτων (ou σύνθεσις νοημάτων, le τῶν de la ligne 430a28 étant omis dans certains manuscrits), que j’appellerai par commodité les composés, et les ἀδιαίρετα. La traduction de ce dernier terme est délicate, puisqu’il peut signifier aussi bien « indivisible », et donc ce qu’Aristote appelle ἀδιαίρετον en puissance, qu’« indivisé », correspondant à ce qu’Aristote appelle ἀδιαίρετον en acte. Si l’on souhaite conserver une traduction unique pour l’ensemble des occurrences6, ce qui paraît nécessaire afin de comprendre certains problèmes auxquels tente de répondre Aristote, sans doute vaut-il mieux suivre D.W. Hamlyn7 et opter pour « indivisé » (undivided), dans la mesure où « indivisible en acte » n’a pas beaucoup de sens, tandis qu’« indivisé en puissance » peut assez aisément se comprendre comme ce qui non seulement n’est pas divisé, mais ne peut pas l’être. Comme nous le verrons, Aristote distinguera trois espèces d’indivisés, ce qui portera à quatre les types et sous-types d’objets intelligibles étudiés dans le chapitre, dont chacun est censé correspondre à un type d’intellection différent. Tous ces objets partagent néanmoins un trait commun, sur lequel Aristote insistera tout au long du chapitre : ils se caractérisent par leur unité. Bien sûr, l’unité du composé n’est pas la même que celle d’un indivisé, mais il n’en reste pas moins que tout le chapitre semble destiné à montrer que quel que soit l’objet de l’intelligence, il présente toujours une certaine forme d’unité, qui garantit dès lors l’unité de l’acte qui s’y rapporte – tantôt parce que cet acte est ce qui produit l’unité de son objet, tantôt parce qu’il s’identifie purement et simplement à cet objet. Il s’agit certainement là d’un thème directeur du chapitre, qui me paraît beaucoup moins désordonné que ne le prétend Ross8 et dont le texte ne justifie pas les multiples corrections qu’il lui inflige. Je partirai dès lors du texte beaucoup plus conservateur de Jannone9, tout en discutant les variantes et corrections les plus significatives. Je suivrai également son découpage, qui me paraît plus cohérent que celui de Ross. 6 7 8 9

Contrairement à la solution retenue par C. Shields (2016), p. 331. Cf. D.W. Hamlyn (1993), p. 142. W.D. Ross (1961), p. 300. A. Jannone & E. Barbotin (1966).

160

SYLVAIN DELCOMMINETTE

2. 430a26-b6 Ἡ μὲν οὖν τῶν ἀδιαιρέτων νόησις ἐν τούτοις, περὶ ἃ οὐκ ἔστι τὸ ψεῦδος. ἐν οἷς δὲ καὶ τὸ ψεῦδος καὶ τὸ ἀληθές, σύνθεσίς τις ἤδη τῶν νοημάτων ὥσπερ ἓν ὄντων, καθάπερ Ἐμπεδοκλῆς ἔφη “ᾗ πολλῶν μὲν κόρσαι ἀναύχενες ἐβλάστησαν”, ἔπειτα συντίθεσθαι τῇ φιλίᾳ, οὕτω καὶ ταῦτα κεχωρισμένα συντίθεται, οἷον τὸ ἀσύμμετρον καὶ ἡ διάμετρος. ἄν δὲ γενομένων ἢ ἐσομένων, τὸν χρόνον προσεννοῶν καὶ συντιθείς. τὸ γὰρ ψεῦδος ἐν συνθέσει ἀεί· καὶ γὰρ ἂν τὸ λευκὸν μὴ λευκὸν, τὸ μὴ λευκὸν συνέθηκεν. ἐνδέχεται δὲ καὶ διαίρεσιν φάναι πάντα. ἀλλ’ οὖν ἔστι γε οὐ μόνον τὸ ψεῦδος ἢ ἀληθές, ὅτι λευκὸς Κλέων ἐστίν, ἀλλὰ καὶ ὅτι ἦν ἢ ἔσται. τὸ δὲ ἓν ποιοῦν, τοῦτο ὁ νοῦς ἕκαστον. L’intellection des indivisés a lieu dans les cas à propos desquels il n’y a pas d’erreur. Dans les cas où il y a aussi bien du faux que du vrai, en revanche, il y a déjà une certaine composition des10 notions comme étant11 une unité – comme le dit Empédocle : « là où de beaucoup poussèrent des têtes sans cou », elles entrèrent ensuite en composition grâce à l’amitié ; de même également, ces séparées entrent en composition, par exemple l’incommensurable et la diagonale. Et si elle porte sur ce qui a été ou ce qui sera, en pensant en plus le temps et en le faisant entrer en composition12. En effet, le faux se trouve toujours dans une composition ; car même si 13 le blanc non blanc, on a fait entrer le non-blanc en composition. Mais il est possible de dire que tout cela est aussi division. Mais en tout cas, le faux et le vrai, ce n’est pas seulement que Cléon est blanc, mais aussi qu’il l’était ou le sera. Et ce qui produit l’unité, c’est dans chaque cas l’intelligence.

Aristote commence donc par opposer les cas où il n’y a pas d’erreur aux cas où il y a aussi bien du faux que du vrai. Remarquons que sa formulation est ici très prudente. D’un côté, la structure du deuxième membre de l’opposition (καὶ τὸ ψεῦδος καὶ τὸ ἀληθές, 430a27) suggère que les cas en question sont non pas tous ceux auxquels s’applique soit la vérité soit l’erreur, mais seulement tous ceux qui se caractérisent Ou « de », si on ne lit pas le τῶν de 430a28. Malgré l’inélégance de la formule, je conserve le verbe « être » dans la traduction (plutôt que « constituer », « former », etc.), car comme nous allons le voir, c’est bien l’être qui est l’opérateur de la composition. 12 Je suis Jannone qui conserve le texte des manuscrits : τὸν χρόνον προσεννοῶν καὶ συντιθείς. Ross suit la proposition de Torstrik et corrige καὶ συντιθείς en συντίθησι. C’est sans doute plus élégant, mais cela ne me paraît pas nécessaire. 13 P. Thillet 2005, p. 388, n. 455, propose de lire ἂν τὸ μὴ λευκὸν λευκὸν en se fondant sur l’arabe, mais l’ajout ne paraît pas nécessaire. Cf. P. Siwek (1965), et A. Jannone & E. Barbotin (1966), qui supposent le même sens sans addition. 10 11

DE ANIMA III 6

161

par l’alternative entre le vrai et le faux. De fait, nous apprendrons à la fin du chapitre qu’un type de vérité est également possible là où il n’y a pas composition, mais que celui-ci exclut l’erreur (430b27-31)14. D’un autre côté, Aristote semble éviter à dessein de dire que là où il n’y a pas composition, il y a nécessairement vérité : il se contente de dire qu’il n’y a pas d’erreur. Ce n’est peut-être pas anodin, car cela rend ce texte compatible aussi bien avec le début du De Interpretatione, selon lequel les νοήματα sans composition ni division ne sont encore par eux-mêmes ni vrais ni faux (1, 16a9-18), qu’avec Métaphysique Θ 10, selon lequel la pensée des non-composés (ἀσύνθετα) est nécessairement vraie, mais d’un type de vérité différent de celle qui concerne les composés (1051b17-1052a4). Contrairement à ce que laisse entendre Philopon15, ces deux textes ne sont pas nécessairement incompatibles, à condition de considérer que le De Interpretatione également n’exclut que l’alternative du vrai et du faux, et que par ailleurs tous les νοήματα du De Interpretatione ne correspondent pas à des ἀσύνθετα au sens de Θ 10, ces derniers s’identifiant bien plutôt aux τί ἐστι (cf. 1051b26, 32), alors que les premiers incluent des notions telles que celle du bouc-cerf (τραγέλαφος, Int. 1, 16a16), qui n’a précisément pas de τί ἐστι au sens strict selon Aristote – la réponse à la question τί ἐστι présupposant une réponse affirmative à la question εἰ ἐστι (cf. Second analytiques II 1, 89b31-35). Autrement dit, seuls certains νοήματα au sens du De Interpretatione peuvent admettre la vérité (mais pas l’erreur, et donc pas l’alternative entre le vrai et le faux), à savoir ceux qui correspondent aux ἀσύνθετα de Met. Θ 10 – ou aussi bien aux ἀδιαίρετα du présent chapitre16. C’est important, car cela signifie, d’une part, qu’il peut y avoir composition de νοήματα qui ne sont pas, au sens strict, des « incomposés » ou des « indivisés » (les deux notions étant synonymes – j’y reviendrai), et, d’autre part, qu’il est erroné de considérer que la fin du présent chapitre (430b26-31) ne ferait que reprendre ce qui serait déjà annoncé dans son paragraphe introductif, puisque ce dernier n’a pas encore dit que l’intellection des indivisés était vraie – et il ne pouvait le faire, dans la mesure où la signification d’« indivisé » n’a pas encore été précisée à ce stade.

14 15 16

Sur ce point, voir J. Harvey (1978), p. 219-220. Philopon, In DA 544, 21-545.5. Sur ce point, comparer E. Berti (1996), p. 393-394.

162

SYLVAIN DELCOMMINETTE

Commençons toutefois par examiner le statut de la composition (σύνθεσις) dont il est ici question. De prime abord, la comparaison avec les « têtes sans cou » d’Empédocle peut surprendre. Selon Thémistius, elle se justifierait à trois niveaux : premièrement, pour insister sur le fait que la composition produite par l’intelligence est une unité résultant de la multiplicité et non un simple tas (σωρός) – qu’elle forme « un seul tout organique », comme l’écrit également Rodier17 – ; deuxièmement, parce qu’une telle composition peut être vraie ou fausse, tout comme la composition des membres chez Empédocle peut être viable ou non ; troisièmement, parce qu’elle peut être dissoute et retourner à ses éléments, comme chez Empédocle également18. Je risquerais, de manière très circonspecte, une explication supplémentaire : chez Empédocle, ce qui produit la composition est l’amitié ; or celle-ci semble être en rapport étroit avec le νόος, puisque Empédocle écrit que c’est avec ce dernier qu’il faut la regarder19 (fr. 17.21), ce qui, en vertu du principe selon lequel le semblable est connu par le semblable, semble impliquer une identité, ou à tout le moins une ressemblance, entre l’intelligence et l’amitié20. Quelle que soit la validité de cette conclusion, rien n’interdit de penser qu’elle a permis à Aristote de se sentir autorisé à trouver chez Empédocle un antécédent de son idée selon laquelle la production de l’unité à partir de la multiplicité est le fait de l’intelligence. Car telle me paraît être l’idée forte de ce passage : l’intelligence produit l’unité de composition. Sa qualification de ποιοῦν (430b6) ne peut évidemment passer inaperçue après le chapitre 5 (cf. 430a19) et semble justifier l’idée des commentateurs grecs selon laquelle l’intelligence dont il est question ici est à tout le moins en rapport étroit avec l’intelligence productive du chapitre précédent ; mais ce rapprochement aurait plutôt tendance à me faire penser que cette dernière n’est rien d’autre que l’acte de cette faculté qu’est l’intelligence « en puissance », bien que cet acte puisse également exister de manière absolument séparée dans le cas de dieu. Quoi qu’il en soit, d’après ce que nous venons de 17

G. Rodier (1900), p. 469. Thémistius, In DA 109, 8-18. 19 Cf. fr. 17.21 : τὴν [scil. Φιλότης, ligne 20] σὺ νόῳ δέρκευ μηδ’ ὄμμασι ἧσο τεθηπώς. 20 Ailleurs, Empédocle écrit également que nous percevons l’amour par l’amour (Στοργῇ δὲ Στοργήν, fr. 109.3), ce qui, rapproché du vers cité à la note précédente, conduit également à la même suggestion. 18

DE ANIMA III 6

163

voir, dire que l’intelligence produit l’unité de composition, c’est également dire qu’elle produit l’être – du moins l’être qui fonde la structure prédicative, c’est-à-dire la copule, celui dont le De Interpretatione nous dit qu’« en lui-même il n’est rien (αὐτό... οὐδέν ἐστιν ), mais [qu’]il ajoute la signification (προσσημαίνει) d’une certaine composition, qu’on ne peut penser sans les composants » (3, 16b23-25) et que Met. Θ 10 définit de la manière suivante : « L’être est la composition et l’être un, le non-être la non-composition et l’être multiple » (τὸ μέν... εἶναι ἐστι τὸ συγκεῖσθαι καὶ ἓν εἶναι, τὸ δὲ μὴ εἶναι τὸ μὴ συγκεῖσθαι ἀλλὰ πλείω εἶναι, 1051b11-13). L’objet propre de l’intelligence dans son rapport aux composés, ce ne sont pas les composants, même s’ils sont nécessairement présupposés, mais la composition elle-même dans son unité, c’est-à-dire l’être de cette composition, en tant qu’il est produit par l’intelligence elle-même. L’exemple d’Aristote est d’ailleurs intéressant à ce propos. En effet, si l’incommensurable et la diagonale peuvent être considérés comme deux notions séparées (κεχωρισμένα, 430a30) qu’il s’agit ensuite de composer, ce n’est évidemment pas au sens où l’incommensurable serait comme un accident qui viendrait s’ajouter à la diagonale pour en déterminer le sens de manière restrictive, puisque c’est bien par soi que la diagonale (diagonale d’un carré, s’entend) est incommensurable (au côté de ce carré) : il s’agit de ce qu’Aristote appelle parfois un « accident par soi », c’est-à-dire une détermination qui, si elle n’appartient pas à l’essence ellemême de la diagonale, en découle toutefois nécessairement. La composition en question est donc une composition nécessaire et non contingente, selon les deux modalités distinguées en Met. Θ 10 (1051b9-17), qui propose d’ailleurs le même exemple pour la première (1051b20-21). Ce type de composition est proprement l’objet de la science au sens strict, par opposition aussi bien à l’opinion qu’à ce qu’Aristote appelle νοῦς dans les Seconds analytiques (II 19) et qui correspond à l’un des modes de ce qu’il appelle ici, nous le verrons, intellection des indivisés. Il s’agit d’une composition syntaxique, correspondant à la structure prédicative d’une proposition ou de la pensée qu’elle exprime qui seule rend possible l’alternative entre la vérité et la fausseté de celles-ci. En d’autres termes, ce qui est susceptible d’être vrai ou faux, ce n’est pas la notion de « diagonale incommensurable », mais bien la proposition « la diagonale est incommensurable » (ou la pensée correspondante). En l’occurrence, cette proposition est nécessairement vraie, mais d’un type de vérité qui suppose la

164

SYLVAIN DELCOMMINETTE

composition, et donc la même structure syntaxique qu’une proposition fausse, y compris une proposition nécessairement fausse comme « le blanc est non blanc »21. Cela permet de comprendre en quoi toute composition peut également être appelée division (διαίρεσις, cf. 430b3-4). Contrairement à ce qui se passe dans d’autres contextes (notamment, bien que dans des termes un peu différents, dans le texte de Met. Θ 10, 1051b11-13 cité ci-dessus), la division n’est pas à comprendre ici comme correspondant à la négation, tandis que la composition correspondrait à l’affirmation, mais bien plutôt comme « l’envers » de la composition, qui lui est inséparablement lié – et ce précisément parce que « la composition ne peut être pensée sans les composants » : pour composer, l’intelligence doit en même temps diviser, c’est-à-dire considérer les composants comme étant d’une certaine manière « séparés ». Cette division ne semble toutefois pas avoir le même statut selon qu’il s’agit d’une prédication nécessaire ou d’une prédication contingente ou fausse : dans le premier cas, par exemple celui de la diagonale incommensurable, on peut considérer qu’elle s’opère à partir d’une unité préalable (qui n’est pas encore une unité de composition) que la pensée « déplie » pour lui donner une forme prédicative ; dans le second, il s’agit plutôt de la considération de deux notions séparées que la pensée compose au sein d’une proposition, sans qu’aucune unité préalable ne puisse être présupposée. Dans les deux cas, toutefois, composition et division s’opèrent d’un seul et même mouvement et s’impliquent mutuellement. Remarquons que cette coimplication de la σύνθεσις et de la διαίρεσις a pour conséquence que les termes ἀδιαίρετον et ἀσύνθετον peuvent eux-mêmes être considérés comme synonymes dans ce contexte, ce qui renforce encore la proximité entre ce chapitre et Met. Θ 10.

21 Aux lignes 430b2-3, je suis l’interprétation de D.W. Hamlyn (1993), p. 143, selon qui « [i]n the example involving white and not-white Aristotle is taking the most primitive kind of false statement – the assertion that one thing is its opposite ». Contra R. Bodéüs (1993), p. 231, n. 2, selon qui Aristote voudrait dire que « l’intelligence […] associe la couleur blanche perçue à l’idée de non-blanc et nie ainsi ce que le sens perçoit », de sorte que « l’erreur du jugement sur les sensibles propres est, pour Aristote, une erreur, non du sens, mais de l’intelligence, qui nie ce que le sens perçoit ». Il me semble au contraire qu’Aristote en reste ici au niveau de l’intelligence, plutôt qu’il ne passe à celui du rapport entre l’intelligence et la sensation, raison pour laquelle l’exemple d’erreur qu’il choisit est celui d’une proposition impossible.

DE ANIMA III 6

165

Le dernier point à souligner dans ce passage est la question du temps. Comme l’écrit Aristote, le temps peut être « pensé en plus » (cf. προσεννοῶν, 430b1) et donc entrer lui-même dans la composition produite par l’intelligence, rendant ainsi possible la vérité ou la fausseté d’assertions sur le passé ou l’avenir. Ce qu’il importe de remarquer ici, c’est que le temps est considéré comme une détermination de la composition elle-même  : en effet, il s’applique à la copule, que celle-ci soit explicite ou implicitement contenue dans le ῥῆμα (cf. Int. 3, 16b6-9 et 16-18, avec 12, 21b9-10 et Met. Δ 7, 1017a27-30). Ce point est important, car il signifie que la perception intellectuelle du temps, en tant que détermination de la liaison prédicative, s’opère au niveau de l’intellection des composés. Contrairement à ce qu’affirme Thémistius22, cela n’implique pas que le temps ne puisse être saisi par la sensation ou par la phantasia – Aristote affirme explicitement le contraire ailleurs, ce qui lui permet d’attribuer la perception du temps à d’autres animaux que l’homme (cf. Mem. 1, 449b28-30, 450a19, 451a16-17) – ; mais cela implique en revanche que la perception du temps est absente de l’intellection des indivisés, dont les objets doivent donc nécessairement être atemporels. Nous allons voir que cette atemporalité de l’indivisé, et partant de l’intellection qui le saisit, sera un thème essentiel de la suite du chapitre. 3. 430b6-14 Τὸ δ’ ἀδιαίρετον ἐπεὶ διχῶς, ἢ δυνάμει ἢ ἐνεργείᾳ, οὐθὲν κωλύει νοεῖν τὸ ἀδιαίρετον, ὅταν νοῇ τὸ μῆκος (ἀδιαίρετον γὰρ ἐνεργείᾳ) καὶ ἐν χρόνῳ ἀδιαιρέτῳ· ὁμοίως γὰρ ὁ χρόνος διαιρετὸς καὶ ἀδιαίρετος τῷ μήκει· οὔκουν ἔστιν εἰπεῖν ἐν τῷ ἡμίσει τί ἐννοεῖ ἑκατέρῳ· οὐ γὰρ ἔστιν, ἂν μὴ διαιρεθῇ, ἀλλ’ ἢ δυνάμει· χωρὶς δ’ ἑκάτερον νοῶν τῶν ἡμίσεων διαιρεῖ καὶ τὸν χρόνον ἅμα, τότε δ’ οἱονεὶ μήκει. Εἰ δ’ ὡς ἐξ ἀμφοῖν, καὶ ἐν τῷ χρόνῳ τῷ ἐπ’ ἀμφοῖν. Mais puisque l’indivisé se dit de deux manières, soit en puissance soit en acte, rien n’empêche de penser l’indivisé lorsqu’on pense la longueur23 (car elle est indivisée en acte), et dans un temps indivisé ; en effet, le temps est divisé et indivisé de manière semblable à la longueur. Il n’est donc pas possible de dire ce que l’on pense dans chaque moitié : en effet, n’est pas, si n’est pas divisé, sinon en puissance24 ; mais en pensant séparément chacune des moitiés, on divise simultanément le temps, et on le fait alors pour ainsi dire au moyen de la longueur25. Mais si comme constituée des deux ensemble, aussi dans le temps qui s’applique aux deux ensemble.

C’est ici qu’il est particulièrement important de traduire ἀδιαίρετον par « indivisé » plutôt que par « indivisible », car que pourrait bien signifier « indivisible en acte » ? L’ἀδιαίρετον en acte, c’est tout simplement ce qui n’est pas effectivement divisé, tandis que l’ἀδιαίρετον en puissance, c’est ce qui est proprement indivisible, bref ce qui ne peut pas être divisé26. L’ἀδιαίρετον en puissance est donc forcément aussi ἀδιαίρετον en acte, mais l’inverse n’est pas vrai ; et le présent paragraphe est précisément consacré à l’ἀδιαίρετον en acte qui est διαιρετόν en puissance – c’est-à-dire au continu, dont Aristote prend ici l’exemple de la longueur (τὸ μῆκος), et qu’il appellera dans la section suivante κατὰ ποσὸν ἀδιαίρετον (430b14). La longueur en question doit être non pas une longueur sensible, mais une longueur intelligible, puisqu’elle est considérée comme objet de l’intelligence. Il ne s’agit pas pour autant de l’εἶδος de la longueur – qui comme nous le verrons dans la section suivante est quant à lui indivisé non seulement en acte, mais aussi en puissance –, mais bien plutôt de la longueur mathématique, qui pour sa part est continue et est donc divisible, et ce parce qu’elle a une matière27 ; 24

Thillet (ainsi semble-t-il que Barbotin) comprend plutôt : chaque moitié de la longueur n’est pas si elle (= la longueur) n’est pas divisée. C’est vrai également, mais cette interprétation me semble plus difficile à tirer du grec que celle de Ross et de Hamlyn, que je suis ici. La traduction de Bodéüs est peut-être la meilleure, car la plus indéterminée : « puisqu’il n’est pas question de moitié sans division, si ce n’est potentiellement ». 25 À la ligne 430b10, je lis μήκει avec Jannone, qui me paraît être la lectio difficilior, contre μήκη retenu par Ross (« and the halves of the time are, as it were, separate lengths »), suivi notamment par Hamlyn (« and then it is as if they were lengths themselves ») et par Thillet (« c’est alors comme s’il y avait des longueurs »). 26 La note de R. Bodéüs relative à ce passage, selon laquelle Aristote distinguerait ici un indivisible potentiel, qui serait d’ordre sensible et serait effectivement divisible, d’un indivisible effectif, qui ne serait que potentiellement divisible (à l’infini) et correspondrait à l’intelligible continu des mathématiques (1993, p. 231, n. 5), est un bel exemple des confusions auxquelles conduit la traduction traditionnelle d’ἀδιαίρετον par « indivisible ». 27 La distinction entre la ligne mathématique et son essence est explicitement faite en DA III 4, 429b18-20 : πάλιν δ’ ἐπὶ τῶν ἐν ἀφαιρέσει ὄντων τὸ εὐθὺ ὡς τὸ σιμόν· μετὰ συνεχοῦς γάρ· τὸ δὲ τί ἦν εἶναι, εἰ ἔστιν ἕτερον τὸ εὐθεῖ εἶναι καὶ τὸ εὐθύ, ἄλλο· ἔστω γὰρ δυάς. Sur le fait que la matière intelligible des objets mathématiques ne

DE ANIMA III 6

167

cependant, cette matière est une matière intelligible, qui correspond aux différentes parties purement spatiales en lesquelles il est possible de la diviser, et non une matière sensible comme celle des corps physiques – sans quoi la longueur en question serait elle-même sensible et non plus intelligible28. Une telle longueur est proprement l’objet des démonstrations du mathématicien ; c’est elle, par exemple, qui peut être divisée en deux parties de longueur strictement égale, ce qui n’est le cas d’aucune longueur sensible (dont toute division dichotomique sera imparfaite) ni de l’εἶδος de la longueur (qui n’a pas d’étendue spatiale et est identique en toute longueur quelle qu’elle soit). Tel est donc le premier type d’indivisé que l’intelligence peut penser : ce qui n’est qu’actuellement indivisé, mais potentiellement divisible. La question sous-jacente d’Aristote à ce sujet, comme dans les cas qui suivront, peut être formulée de la manière suivante : en quoi la nature propre de cet objet de l’intelligence influe-t-elle sur celle de l’acte qui le saisit, et donc que nous apprend-elle au sujet de ce dernier, selon le principe méthodologique énoncé en II 4 ? Dans le cas des objets mathématiques, il ne peut sans doute y avoir une identité pure et simple entre la pensée et son objet, car cette identité ne prétendait valoir que pour ce qui est sans matière (cf. III 4, 430a3-4) ; et de fait, à la différence de la longueur mathématique, sa saisie par l’intelligence n’a pas d’étendue spatiale, même purement intelligible. En revanche, à l’indivision actuelle et à la divisibilité potentielle de la longueur correspond selon Aristote l’indivision actuelle et la divisibilité potentielle du temps en lequel on la saisit par l’intelligence. De fait, quand il écrit que l’intelligence saisit la longueur dans un temps indivisé, il ne peut aucunement vouloir dire, contrairement à ce qu’affirme Philopon, repris par certains commentateurs contemporains, qu’elle la saisit dans l’instant29, puisqu’il précise que ce temps est « divisé et indivisé de manière semblable à la longueur » – ce qui n’est aucunement le cas de correspond pas à l’extension pure, comme on le comprend généralement depuis le Pseudo-Alexandre, mais aux parties constitutives de telle ou telle figure (ou aux unités constitutives de tel ou tel nombre), je me permets de renvoyer à S. Delcomminette (2014), p. 106-113. 28 Contra R. Polansky (2007), p. 476, qui tout en considérant avec raison que la longueur ici en question n’est pas l’essence de la longueur mais une longueur étendue et donc dans la matière (enmattered), ajoute qu’elle pourrait aussi bien être celle d’un corps que l’étendue mathématique. 29 Philopon, In DA 549, 10-13 ; voir aussi T. De Koninck (1990), p. 219, et R. Bodéüs (1993), p. 231, n. 6. Voir les critiques d’E. Berti (1996), p. 395-396.

168

SYLVAIN DELCOMMINETTE

l’instant, qui correspond bien plutôt au point et est à la fois indivisé en acte et en puissance, comme nous le verrons. Il veut au contraire dire que la saisie de la longueur par l’intelligence prend un certain temps, fût-il imperceptible, mais un temps qui peut lui-même être considéré comme une unité actuellement indivisée, quoique potentiellement divisible. Qu’est-ce que cela signifie ? Non pas, comme le propose E. Berti, qu’il « faut un certain temps pour comprendre ce que c’est [qu’une longueur, par exemple] une diagonale, c’est-à-dire pour trouver sa définition exacte »30 : d’une part, parce qu’il n’est pas question ici du ce que c’est de la longueur, c’est-à-dire de son essence ou de son εἶδος, mais de la longueur elle-même, c’est-à-dire de cet εἶδος dans une matière (intelligible) ; et d’autre part, parce que cela revient à confondre le temps nécessaire à l’apprentissage ou à la découverte, c’est-à-dire à l’atteinte de l’intellection, et celui de l’intellection elle-même. Je pense qu’Aristote veut bien plutôt dire que c’est l’intellection elle-même qui prend un certain temps lorsqu’elle a pour objet une longueur mathématique, en ce sens que celleci, dans la mesure où elle a une extension spatiale (quoique purement intelligible), doit nécessairement être parcourue par la pensée pour être saisie, et que ce parcours prend du temps, même si celui-ci est trop infime pour être mesuré précisément31. Mais si l’on pense la longueur mathématique dans son unité, et donc comme indivisée en acte, le temps nécessaire à ce parcours doit lui-même être considéré comme unitaire et donc comme indivisé en acte, bien qu’il soit évidemment divisible en puissance, comme n’importe quel continu. L’intérêt de cette remarque n’apparaît qu’à l’étape suivante, où Aristote précise qu’il n’est pas possible de dire ce que l’on pense dans chaque moitié du temps en question, parce que chaque moitié n’est qu’en 30

E. Berti (1996), p. 395. Comme me l’a fait remarquer Charlotte Murgier, on pourrait objecter à ceci le passage célèbre de Met. Θ 6, 1048b18-35, selon lequel, notamment, la pensée étant un acte et non un mouvement, « on pense et on a pensé » simultanément (b24 et 34), tout acte étant complet à chaque moment de son occurrence. Je ne peux rentrer ici dans la discussion de ce passage très controversé ; qu’il me suffise de dire que je ne pense pas qu’il implique nécessairement que la pensée d’un objet intelligible quelconque soit toujours déjà complète, mais seulement que la pensée elle-même comme activité est toujours déjà complète dès qu’elle se produit ; bref, « on pense et on a pensé » n’implique pas nécessairement « on pense x et on a pensé x », pas plus que « on voit et on a vu » n’implique nécessairement « on voit x et on a vu x », la vision (et plus clairement encore l’audition) de certains objets pouvant réclamer un certain temps, pendant lequel on voit (ou on entend) néanmoins toujours déjà. 31

DE ANIMA III 6

169

puissance avant la division (430b10-11). Rodier considère qu’Aristote répondrait ici à un argument analogue à ceux de Zénon, qui aurait peutêtre été avancé par les Mégariques, consistant à nier la possibilité de penser une longueur indivisée en un temps par définition divisible à l’infini32. Mais si un tel argument a existé, Aristote y a en réalité déjà répondu dans les lignes qui précèdent, en soulignant la symétrie entre l’indivision en acte et la divisibilité en puissance de la longueur et du temps nécessaire à son appréhension. Il me semble qu’il est bien plutôt en train d’expliquer ici un fait psychologique, à savoir que s’il est vrai qu’il faut un certain temps à l’intelligence pour parcourir une longueur mathématique, en tant que celle-ci a une certaine étendue (intelligible), on ne peut assigner un ordre à ce parcours et dire qu’on a commencé par penser telle moitié avant l’autre. Aristote ne nie pas que si l’on pense la longueur AB dans le temps t, on doit nécessairement penser AB/2 dans le temps t/2, mais il nie que l’on puisse déterminer laquelle des deux moitiés AB/2 on pense dans chacun des deux temps t/2 compris dans le temps t, parce qu’aussi bien la division de la longueur que celle du temps correspondant sont seulement potentielles et non actuelles. Bien sûr, on peut aussi se demander en combien de temps on pense la longueur AB/2, et la réponse sera : dans le temps t/2 ; et on peut alors penser d’abord tel segment AB/2 (par exemple celui correspondant à la moitié gauche de AB) puis tel autre (par exemple celui correspondant à la moitié droite de AB), de telle sorte qu’à telle partie de la longueur corresponde telle partie du temps ; mais dans ce cas, la division du temps s’est faite au prix de la division de la longueur, et en quelque sorte « par son moyen », comme l’écrit Aristote, de sorte qu’on ne peut plus dire que l’on pense la longueur AB elle-même, mais seulement les deux segments AB/2 de manière successive (430b11-13). Enfin, troisième possibilité, on peut également penser AB comme constitué des deux moitiés, et dans ce cas on pensera nécessairement les deux segments dans un ordre déterminé (peu importe lequel) ; mais alors, on ne peut pas dire qu’on pense AB comme une unité indivisée (en acte), et c’est pourquoi il en va de même du temps nécessaire pour l’appréhender, qui lui aussi se caractérise par la succession de deux moitiés de temps (430b13-14)33. 32

G. Rodier (1900), p. 479 ; voir aussi P. Siwek (1965), p. 336-337. Thomas d’Aquin propose une autre explication de ce dernier cas : « S’il (l’intellect) comprend la ligne comme quelque chose d’un constitué de deux parties, il comprend également dans un temps non divisé, mais selon quelque chose qui est dans l’une et 33

170

SYLVAIN DELCOMMINETTE

L’intérêt de tout ce paragraphe est de montrer que lorsqu’elle porte sur un objet qui n’est indivisé qu’en acte, mais divisible en puissance, comme une longueur mathématique, l’intellection n’est elle-même indivisée qu’en acte, mais divisible en puissance. On voit donc en quoi la nature de l’objet influe sur l’acte qui le saisit. Mais ici, dans la mesure où l’objet en question n’est pas dépourvu de toute matière, il n’y a pas identité pure et simple entre l’acte et son objet ; en conséquence, la divisibilité de l’acte n’est pas exactement la même que celle de l’objet, bien qu’elle lui soit intimement corrélée : à la divisibilité de l’objet dans l’espace correspond la divisibilité de l’intellection dans le temps. Ce point est important, car il montre que le temps n’intervient pas au même niveau que dans le cas de l’intellection des composés : dans ce dernier cas, le temps était attribué à l’objet lui-même, c’est-à-dire à l’état de choses correspondant à la proposition exprimant la pensée composée, tandis qu’ici, il concerne seulement l’intellection, l’objet (la longueur mathématique, par exemple) étant quant à lui atemporel34. Il n’y a donc pas ici coïncidence parfaite entre la pensée et son objet, puisque la première est temporelle et le second atemporel. Cette absence de coïncidence introduit-elle la possibilité de l’erreur ? Aristote ne le dit pas, mais il est significatif que lorsqu’il opposera à la fin du chapitre l’intellection à l’œuvre dans la prédication, qui est susceptible d’être vraie ou fausse, à celle qui est toujours vraie, il dira que cette dernière porte sur « le ce que c’est selon le ce que c’est qu’être » (τοῦ τί ἐστι κατὰ τὸ τί ἦν εἶναι, 430b28) et « tout ce qui est sans matière » (ὅσα ἄνευ τῆς ὕλης, 430b30-31), deux caractérisations qui ne peuvent s’appliquer au continu mathématique, dont il n’est d’ailleurs pas fait mention en Met. Θ 10. L’activité « synthétique » à l’œuvre dans l’appréhension des objets mathématiques, qui suppose le parcours d’un divers et son rassemblement dans l’unité d’un acte, suffit-elle à rendre l’autre parties du temps, à savoir dans l’instant, et si la considération se poursuit pendant un temps, le temps n’est pas divisé de telle sorte qu’il saisisse autre chose dans une partie du temps et autre chose dans une autre, mais la même chose dans l’une et l’autre » (1999), p. 265. C’est ingénieux, mais outre que cela me paraît psychologiquement erroné, je ne pense pas du tout que cela corresponde à ce que dit Aristote, qui au contraire nie dans tout ce paragraphe que l’intellection de l’indivisé quantitatif s’effectue dans un temps indivisé en puissance, c’est-à-dire dans l’instant. Mon interprétation se rapproche davantage de celle de D. Hamlyn : « So, if a whole is compounded out of two halves, the time involved is similarly composite » (W.D. Hamlyn (1993), p. 144). 34 Sur ce point, comparer R. Polansky (2007), p. 475.

DE ANIMA III 6

171

une telle intellection susceptible d’être fausse, bien qu’elle soit d’une nature différente de la liaison syntaxique à l’œuvre dans la prédication ? Ou faut-il plutôt comprendre que l’intellection des objets mathématiques n’est ni vraie ni fausse ? La question reste ouverte, mais mériterait sans doute d’être explorée dans le cadre d’une investigation sur le statut des mathématiques chez Aristote35. 4. 430b14-20 Τὸ δὲ μὴ κατὰ ποσὸν ἀδιαίρετον ἀλλὰ τῷ εἴδει νοεῖ ἐν ἀδιαιρέτῳ χρόνῳ καὶ ἀδιαιρέτῳ τῆς ψυχῆς. κατὰ συμβεβηκὸς δέ, καὶ οὐχ ᾗ ἐκεῖνα, διαιρετὰ ᾧ νοεῖ καὶ ἐν ᾧ χρόνῳ, ἀλλ’ ᾗ ἀδιαίρετα· ἔνεστι γὰρ κἀν τούτοις τι ἀδιαίρετον, ἀλλ’ ἴσως οὐ χωριστόν, ὃ ποιεῖ ἕνα τὸν χρόνον καὶ τὸ μῆκος. καὶ τοῦθ’ ὁμοίως ἐν ἅπαντί ἐστι τῷ συνεχεῖ, καὶ χρόνῳ καὶ μήκει. Mais l’indivisé qui n’est pas selon la quantité mais selon l’εἶδος, on le pense dans un temps indivisé et par quelque chose d’indivisé de l’âme. Mais c’est par accident, et non à la manière de ceux-là, que sont divisés ce par quoi et le temps dans lequel on pense, mais à la manière d’indivisés ; en effet, même dans ceux-ci il y a quelque chose d’indivisé, mais qui n’est sans doute pas séparé, qui rend uns le temps et la longueur. Et ceci se trouve semblablement dans tout ce qui est continu, aussi bien le temps que la longueur.

J’ai traduit le texte de Jannone sans modification. Ross, suivi par Hamlyn et Shields, accepte la suggestion de Bywater et transpose la première phrase (b14-15) après la fin du passage, considérant que celui-ci est pour le reste entièrement consacré non pas à τὸ μὴ κατὰ ποσὸν ἀδιαίρετον mais bien à τὸ κατὰ (τὸ) ποσὸν ἀδιαίρετον36. Ce faisant, il marginalise le thème de l’intellection de l’indivisé selon l’εἶδος, dont on peut penser qu’il constitue au contraire le point d’orgue de tout le chapitre et que la discussion sur l’indivisé quantitatif avait pour principal but de l’introduire37. Sans accepter la transposition, Polansky aboutit au même résultat, puisqu’il considère que « ce qui est indivisé non selon la 35 Met. Θ 10, 1052a4-11 pourrait paraître pertinent dans ce contexte, mais ce passage me semble plutôt concerner les propositions à propos des êtres immobiles (géométriques, par exemple), qui sont toujours et nécessairement vraies ou fausses ; de plus, il nie seulement la possibilité d’un certain type d’erreur à leur sujet, à savoir une erreur « du point de vue du temps » (κατὰ τὸ ποτέ), que je ne peux pas explorer ici. 36 W.D. Ross (1963), p. 296-297. 37 Cf. [Simplicius], In DA 254, 11-15.

172

SYLVAIN DELCOMMINETTE

quantité mais selon l’εἶδος» renvoie tout simplement au continu mathématique, qui de fait n’est divisible que quantitativement, mais pas formellement38. Dans ce cas, Aristote ne mentionnerait tout simplement pas l’εἶδος lui-même comme objet indivisé de l’intellection dans ce chapitre, ce qui serait pour le moins paradoxal étant donné que dans d’autres contextes, il présente au contraire l’intellection indivisée comme celle qui porte sur l’εἶδος ou encore le τί ἦν εἶναι (Met. Δ 6, 1016b1-3 ; Ι 1, 1052a29-34), dont il y a tout lieu de penser qu’ils correspondent également aux ἀσύνθετα de Met. Θ 10, qu’Aristote explicite par les expressions τὸ τί ἐστιν (1051b26), τὰς μὴ συνθετὰς οὐσίας (1051b27) et τὸ ὂν αὐτό (1051b29). Bien plus, à la fin du présent chapitre, il nous dira qu’est toujours vraie l’intelligence qui porte sur « le ce que c’est selon le ce que c’est qu’être » (τοῦ τί ἐστι κατὰ τὸ τί ἦν εἶναι, 430b28), et il serait pour le moins étonnant qu’il n’y ait pas fait aucune allusion, ou même seulement une allusion furtive, dans ce qui précède. Formellement, et en m’inspirant essentiellement de l’excellent commentaire de ces lignes proposé par Rodier39, souverainement ignoré par Ross, je comprends la suite des idées de ce passage de la manière suivante. Après l’ἀδιαίρετον en acte, Aristote passe à l’ἀδιαίρετον en puissance, c’est-à-dire à ce qui est véritablement indivisible, et qui correspond à l’εἶδος, comme y insistaient déjà les commentateurs grecs. Celui-ci est pensé dans un temps et par « quelque chose de l’âme » qui sont indivisés non seulement en acte, comme les indivisés quantitatifs, mais aussi en puissance, ce qui les en distingue. On peut toutefois soutenir que ce par quoi on les pense40 et le temps en lequel on les pense est divisé par accident (κατὰ συμβεβηκός), ce qui est différent de la divisibilité en puissance (et a fortiori de la division en acte) : ce n’est donc pas à la manière de ceux-là (ᾗ ἐκεῖνα), scil. comme les continus ou les indivisés quantitatifs, que ces derniers sont divisés ; en réalité, ils sont divisés à la manière d’indivisés (ᾗ ἀδιαίρετα)41, c’est-à-dire en demeurant 38

R. Polansky (2007), p. 476. G. Rodier (1900), p. 476-478 et p. 481-482. 40 Il faut maintenir le ᾧ de 430b16, qui reprend le ἀδιαιρέτῳ de b15. Ross suit Vicomercatus et le corrige en ὅ, mais cette correction n’est exigée que parce qu’il déplace les lignes b14-15. 41 Les mots ἀλλ’ ᾗ ἀδιαίρετα (430b17) sont clairement les plus difficiles du passage. La solution de facilité consiste à suivre Torstrik et à les retrancher purement et simplement (d’après P. Siwek (1965), p. 337, ils sont omis dans certains manuscrits). J’avoue être tenté par la correction minimale proposée par Rodier, consistant à lire ἄλλῃ ἀδιαίρετα 39

DE ANIMA III 6

173

essentiellement indivisés. Car même dans ceux-ci (τούτοις, qui reprend le ἐκεῖνα qui précède), à savoir les indivisés quantitatifs, il y a quelque chose d’indivisé en ce nouveau sens, scil. selon l’εἶδος, qui n’est toutefois pas séparé de toute matière (nous avons vu que les longueurs mathématiques ont une matière intelligible), mais qui est la cause de l’unité aussi bien de la longueur que du temps nécessaire à son appréhension, selon les modalités expliquées précédemment ; et cela vaut pour tout continu en tant que tel. Tâchons à présent d’expliquer ce que tout ceci signifie. Tout d’abord, qu’est-ce que ce « quelque chose d’indivisé de l’âme » par lequel on pense l’εἶδος  ? Selon Philopon, suivi par la plupart des commentateurs et traducteurs anglo-saxons, il s’agirait d’une partie ou d’une faculté de l’âme (les deux notions étant équivalentes chez Aristote), à savoir le νοῦς42 ; selon Thémistius, suivi par les francophones, il s’agirait bien plutôt de la νόησις, c’est-à-dire de l’acte d’intellection lui-même43. La deuxième interprétation me paraît préférable, selon la logique du chapitre et du De Anima en général, consistant à partir de l’objet pour éclairer la nature de l’acte qui le saisit – et, seulement dans un troisième temps, la faculté correspondante. Elle est également en parfait accord avec ce que l’on peut lire à ce sujet dans la Métaphysique (Δ 6, 1016b13 ; Ι 1, 1052a29-31). Ce serait donc l’acte par lequel l’âme pense l’εἶδος qui serait non seulement indivisé en acte, mais absolument indivisible en puissance – même s’il est néanmoins divisible par accident, ce qu’il faudra encore expliquer. Qu’en est-il maintenant du « temps indivisé » en lequel on le pense ? Depuis les commentateurs grecs, on comprend généralement qu’Aristote se réfère ici à l’instant. Dans le cas présent, cette interprétation n’est pas exposée au même type de critiques qu’à propos des indivisés quantitatifs, car le temps indivisé dont il est ici question est censé être indivisé non seulement en acte, mais aussi en puissance, et l’instant semble effectivement répondre à cette caractérisation (à ceci près que, comme y insistent et à comprendre : ils sont indivisibles en tout autre sens (G. Rodier (1900), p. 481). Il me semble toutefois qu’on attendrait une particule pour introduire cette précision. Rodier propose également ἀλλ’ ᾗ αὐτὰ ἀδιαίρετα, Ross ἀλλ’ ᾗ ἐκεῖνα ἀδιαίρετα. 42 Philopon, In DA p. 550.20-21, 31-32 ; W.D. Ross (1963), p. 299 ; D. Hamlyn (1993), p. 62 ; R. Polansky (2007), p. 476 ; C. Shields (2016), p. 62. 43 Thémistius, In DA 110, 19 ; G. Rodier (1900), p. 481 ; J. Tricot (1934), p. 187 ; A. Jannone & E. Barbotin (1966), p. 83 ; R. Bodéüs (1993), p. 232 ; P. Thillet (2005), p. 169.

174

SYLVAIN DELCOMMINETTE

les commentateurs grecs44, selon Aristote lui-même, l’instant n’est pas à strictement parler un temps, en ce sens qu’il n’est pas une partie du temps, mais seulement une division du temps : cf. Phys. IV 11, 220a19, 21). E. Berti considère au contraire qu’Aristote ne peut vouloir dire cela, dans la mesure où l’intellection de l’εἶδος suppose toute une recherche par laquelle on passe des singuliers à l’universel qui les renferme, à la manière d’un tout composé de parties, ou encore par laquelle on rassemble les différents éléments composant sa définition (genre, différence spécifique), de sorte que le temps correspondant pourrait être considéré comme indivisé en acte, mais néanmoins divisible en puissance, exactement comme celui nécessaire à l’appréhension des longueurs mathématiques45. Tout en partageant entièrement le souci de Berti d’éviter d’attribuer à Aristote une conception « intuitive » de l’appréhension de l’essence, son objection ne me paraît pas ici pertinente, car elle semble confondre deux choses très différentes, correspondant aux deux parties du célèbre chapitre II 19 des Seconds analytiques (99b15-100b5 et 100b5-17) : la recherche de l’essence, qui effectivement prend du temps et correspond à l’ἐπαγωγή, et l’état cognitif qui en résulte, appelé νοῦς dans les Seconds analytiques, mais qui, en tant qu’il s’agit d’un acte, peut également être appelé νόησις. Dire que celle-ci s’effectue dans l’instant ne revient pas du tout à nier qu’elle ait été préparée par un long et patient travail, impliquant probablement des procédés aussi divers que ceux décrits dans les Topiques et la méthode de division platonicienne adaptée aux exigences proprement aristotéliciennes afin d’obtenir une définition en bonne et due forme. On peut d’ailleurs penser que l’un des efforts d’Aristote dans la Métaphysique, en particulier dans les chapitres Ζ 12, Ζ 17, Η 6 et Θ 10, consiste précisément à montrer que tout ce travail peut déboucher sur un acte simple, à savoir l’εἶδος en tant qu’il correspond à la dernière différence qui récapitule toutes les précédentes et implique par elle-même le genre à titre de « matière », rendant ainsi possible son appréhension par un acte simple de l’intelligence qui en définitive s’identifie purement et simplement avec le premier46. Et tout ceci ne revient pas non plus à faire de l’intellection au moment où elle se produit une « intuition » au sens purement passif et 44 Thémistius, In DA 110, 22-24 ; [Simplicius], In DA 254, 26-27 ; Philopon, In DA 549, 10-11, 23-27. 45 E. Berti (1978), p. 151 ; (1996), p. 397-398 et 402. 46 Sur cette question, je me permets de renvoyer à S. Delcomminette (2014), en particulier p. 116-118.

DE ANIMA III 6

175

réceptif du terme, puisqu’il s’agit au contraire d’un acte de l’intelligence par lequel elle s’identifie à son objet ; ni à nier qu’elle puisse s’exprimer dans une définition, ce qu’Aristote affirme explicitement en Met. Ι 1, 1052a29-31 où il écrit que les choses dont l’intellection est une et indivisée ont un λόγος un – mais ce λόγος, comme il le précise en Met. Θ 10, 1051b24-25, est de l’ordre de la φάσις et non de la κατάφασις. Quoi qu’il en soit, l’interprétation de Berti me paraît de toute façon exclue par le fait que ce texte me semble avoir précisément pour but d’opposer la divisibilité du temps en lequel s’opère l’intellection de l’εἶδος, qui est purement accidentelle, à celle de l’intellection de l’indivisé quantitatif, qui est quant à elle potentielle. Mais que peut bien signifier que l’intellection de l’εἶδος et le temps en lequel elle s’opère, quoique indivisés aussi bien en acte qu’en puissance, c’est-à-dire absolument indivisibles, sont néanmoins divisibles par accident ? Ici encore, l’explication de Rodier me semble parfaitement convaincante : certes, pour nous êtres temporels, l’intellection de l’εἶδος s’inscrit dans la durée, et par là même est divisible ; mais elle est néanmoins complète en chaque instant de cette durée, et ne subit aucune modification en elle47. Telle est la divisibilité par accident, qui n’altère en aucune manière l’indivisibilité essentielle de l’acte et du temps correspondant, et se distingue donc de la divisibilité en puissance de l’acte et du temps nécessaires pour appréhender une ligne mathématique, dont l’actualisation nous permet au contraire de constater qu’on ne pense pas la même chose (la même partie de la ligne) en chaque partie du temps correspondant, même si la question de savoir ce que l’on pense en chacune avant d’avoir effectué la division de manière actuelle ne peut recevoir de réponse. Ainsi, dire que l’on pense l’εἶδος dans l’instant ne signifie pas qu’il s’agirait d’une sorte d’étincelle subite et évanescente, qui s’éteindrait aussitôt qu’elle s’est produite : au contraire, une telle intellection peut durer un certain temps, et même doit nécessairement 47 G. Rodier (1900), p. 476. L’interprétation de Thémistius, In DA 110.27-31, mérite également d’être signalée : selon lui, la divisibilité accidentelle ici en question serait celle du temps nécessaire pour prononcer le mot exprimant l’εἶδος. Cependant, tout en partageant l’idée que l’intellection de l’εἶδος peut effectivement être exprimée par le langage – et je dirais même davantage dans un λόγος (une définition) que dans un mot –, je ne pense pas qu’Aristote irait jusqu’à identifier l’intellection elle-même avec son expression langagière, de sorte que cette interprétation ne me paraît pas permettre de rendre compte du fait que c’est bien l’intellection elle-même et le temps en lequel elle s’opère qui sont dits ici divisibles par accident.

176

SYLVAIN DELCOMMINETTE

durer quand elle se produit dans un être temporel, mais elle appréhende complètement et parfaitement son objet en chaque instant de cette durée. Le temps est pour elle un simple accident : il n’appartient essentiellement ni à l’acte lui-même ni à son objet. Dès lors, une telle pensée n’inclut pas la conscience du temps, qui suppose la perception du changement dans son objet, de sorte que dans ces moments-là, notre âme semble demeurer dans un maintenant unique et indivisible (cf. Phys. IV 11, 218b21-219a1) – bref, une telle expérience nous fait toucher à une certaine forme d’éternité, celle-ci étant comprise ici au sens d’une atemporalité. Dans ce cas, et dans ce cas seulement, on peut parler d’une coïncidence parfaite entre la pensée et son objet, tous deux également atemporels – du moins essentiellement, la première demeurant toutefois temporelle par accident. Aristote revient ensuite aux continus, mais en les mettant cette fois en rapport avec ce qu’il vient de dire au sujet de l’intellection de l’εἶδος – en ce sens, il n’y a pas simple retour en arrière, mais bien progression, puisque les éléments nouveaux intervenus entre-temps vont permettre d’éclairer ce qui a été dit précédemment. Il affirme en effet que même dans les continus, il y a quelque chose d’indivisé (τι ἀδιαίρετον) (430b17-18). Indivisé en quel sens ? Pas seulement au sens d’indivisé en acte, ce qui a déjà été dit dans la section précédente, mais au sens où « indivisé » a été pris dans les lignes immédiatement antérieures, c’està-dire au sens d’indivisé selon l’εἶδος. Cette interprétation est confirmée par la suite, puisque Aristote écrit que ce « quelque chose d’indivisé » n’est « sans doute pas séparé » (b18) : il ne peut donc désigner que l’εἶδος en tant qu’il est dans la matière. De fait, toute grandeur mathématique peut être considérée comme un composé de matière et d’εἶδος, bien que sa matière – à savoir le continu, comme il l’a rappelé plus haut (cf. DA III 4, 429b18-20) – soit une matière intelligible. Mais ce qui fait que cette grandeur – telle longueur, par exemple – est une, ce n’est pas sa matière, mais bien son εἶδος  : c’est ce dernier, en tant qu’indivisé (selon l’εἶδος, et donc absolument), qui produit l’unité (cf. ὃ ποιεῖ ἕνα, 430b18) de la longueur. En cela, elle est dans le même cas que n’importe quel composé de matière et d’εἶδος, au sein duquel, comme Aristote le montre en Met. Ζ 17 (1041b11-33), c’est toujours l’εἶδος qui est la cause non seulement de l’être, mais aussi de l’unité ; cependant, le présent passage nous permet de comprendre que cette unité du composé correspond à une simple indivision actuelle, qui s’accompagne d’une

DE ANIMA III 6

177

divisibilité potentielle (due à la matière) et se distingue donc de l’indivision absolue de l’εἶδος pris en lui-même. Reste que cette unité de la longueur mathématique est également ce qui rend possible son intellection dans un temps indivisé (en acte et non en puissance : il ne s’agit plus ici de l’instant), de sorte que l’indivisibilité de l’εἶδος est également responsable de l’unité du temps de l’intellection des grandeurs mathématiques. Donc même quand on ne pense pas l’εἶδος lui-même, mais l’εἶδος dans une matière, comme c’est le cas en mathématiques, son indivisibilité fait sentir ses effets. Et la même chose vaut pour tout continu, ajoute Aristote : en chacun, c’est l’εἶδος, en tant que producteur d’unité, qui rend possible leur appréhension en un temps un, bien que cette unité et celle du temps de son appréhension soient seulement indivision actuelle et non indivisibilité potentielle. Remarquons que cela permet de comprendre ce qu’Aristote annonçait dès le chapitre III 4, à savoir que « c’est par autre chose ou par se comportant autrement que l’on juge (ἑτέρῳ... ἢ ἑτέρως ἔχοντι κρίνει, 429b2021) » de la ligne droite et de son ce que c’est qu’être, c’est-à-dire de la ligne comme objet mathématique et de son εἶδος  : de fait, dans le premier cas, il s’agit d’une intellection indivisée en acte mais divisible en puissance et qui prend un certain temps, aussi petit soit-il, tandis que dans le deuxième cas, il s’agit d’une intellection indivisée à la fois en acte et en puissance et qui se déroule dans l’instant, selon les modalités expliquées précédemment. Il doit donc en aller de même pour la faculté à l’œuvre dans chacun des deux actes, à savoir l’intelligence : il ne s’agit sans doute pas de deux facultés différentes, mais en tout cas d’une faculté disposée de deux manières différentes, et c’est seulement lorsqu’elle pense l’εἶδος séparé de la matière qu’elle est elle-même séparée de la matière (cf. 429b21-22)48. 5. 430b20-26 Ἡ δὲ στιγμὴ καὶ πᾶσα διαίρεσις, καὶ τὸ οὕτως ἀδιαίρετον, δηλοῦται ὥσπερ ἡ στέρησις· καὶ ὅμοιος ὁ λόγος ἐπὶ τῶν ἄλλων, οἷον πῶς τὸ κακὸν γνωρίζει ἢ τὸ μέλαν· τῷ ἐναντίῳ γάρ πως γνωρίζει. δεῖ δὲ δυνάμει εἶναι τὸ γνωρίζον καὶ †ἐνεῖναι ἐν αὐτῷ†. εἰ δέ τινι μὴ ἔστιν ἐναντίον τῶν αἰτίων, αὐτὸ ἑαυτὸ γινώσκει καὶ ἐνεργείᾳ ἐστι καὶ χωριστόν. 48 Cela signifierait-il : qu’elle pense sans image ? Je n’oserais m’avancer jusque-là, mais voir L.P. Gerson (2005), p. 169-172.

178

SYLVAIN DELCOMMINETTE

Mais le point et toute division, ainsi que ce qui est indivisé de cette manière, est manifesté comme la privation ; et l’explication est semblable à propos des autres cas, par exemple : comment en vient-on à connaître le mal ou le noir ? en effet, on le connaît d’une certaine manière par le contraire. Mais il faut que ce qui connaît soit en puissance et possède en lui-même49. Mais s’il n’y a pas de contraire pour quelqu’une des causes, elle se connaît elle-même, est en acte50 et est séparée.

Aristote passe ici à une troisième forme d’indivisé, à savoir la limite qui divise une grandeur : le point comme division de la ligne, la ligne comme division du plan, le plan comme division du solide, et l’instant comme division du temps51. Ce type d’indivisé diffère des deux précédents : d’une part, il est indivisé non seulement en acte, mais également en puissance (il est véritablement indivisible) ; mais d’autre part, il ne s’identifie pas à l’εἶδος considéré sans matière – le point, par exemple, peut être « situé » sur la ligne (selon la définition aristotélicienne, il est une « unité avec position », Met. Δ 6, 1016b25-26, 30-31), et suppose donc l’étendue, même s’il n’est pas lui-même étendu. C’est sans doute en ce sens qu’il peut lui aussi être considéré comme une forme d’indivisé quantitatif (cf. Met. Δ 6, 1016b24-36) – mais il s’agit cette fois d’un indivisé quantitatif en puissance, puisqu’il est réellement indivisible. L’intellection qui le saisit doit donc elle-même être différente de celles qui saisissent les deux autres. Pour en rendre compte, Aristote fait usage du concept de privation (στέρησις). Contrairement à ce qu’affirme Pseudo-Simplicius52, il y a toutes les raisons de penser que le point etc. ne sont pas simplement comparés à des privations, mais sont bel et bien des privations pour Aristote. En tout cas, en Met. Ι 3, 1054a26-27, τὸ ἀδιαίρετον est explicitement considéré comme le contraire (ἐναντίον) de τὸ διαιρετόν ; or s’il est vrai que le contraire n’est pas strictement identique à la privation, cette dernière en est à tout le moins une espèce, et une espèce éminente, puisque Aristote considère qu’elle est la première contrariété (πρώτη... ἐναντίωσις, Ι 4, 1055a33), celle qui fonde toutes les autres, en ce sens que dans toute paire de contraires, l’un des deux termes peut être considéré comme la privation de son opposé 49 Je traduis la leçon retenue par Ross (ἐνεῖναι ἐν αὐτῷ), qui toutefois ajoute une crux. Voir ci-dessous. 50 Ou « est acte », selon certains manuscrits (ἐνεργεία plutôt qu’ἐνεργείᾳ), ce qui est peut-être mieux. 51 Cf. W.D. Ross (1963), p. 297. 52 [Simplicius], In DA 257, 1-6.

DE ANIMA III 6

179

(Γ 2, 1004b27 ; 6, 1011b18-20 ; Ι 4, 1055b17-29 ; Κ 3, 1061a19-20 ; 6, 1063b17-19 ; voir aussi Phys. I 7, 191a6-7). C’est ce qui explique que dans le présent texte, Aristote passe si facilement de la privation au contraire (430b23). Par ailleurs, aussi bien la privation que la contrariété diffèrent de la contradiction en ce qu’elles sont des oppositions dans un même genre53. De fait, le point n’est pas non-être absolu, mais privation de longueur ; la ligne, privation de surface ; la surface, privation de volume ; l’instant, privation de durée54. Or la privation est toujours connue à partir de ce dont elle est la privation : le mal à partir du bien, le noir à partir du blanc. Contrairement à Ross, je ne pense pas que cela signifie, par exemple, que nous ne devenions conscients du noir qu’en remarquant l’absence complète du blanc, ce dernier étant l’objet premier de la vue55 – ce qui me paraît tout simplement intenable d’un point de vue phénoménologique : il est clair que nous voyons le noir tout aussi directement que le blanc. C’est sans doute pour éviter ce genre de problème que Philopon soutient que le noir est un mauvais exemple, car il a sa forme propre et ne se réduit donc pas à la privation du blanc – à moins de comprendre le noir (μέλαν) au sens de l’obscur (σκότος), qui de fait peut être considéré comme la simple privation de lumière et est connu par cette dernière56. Le problème est qu’aussi étrange que cela puisse paraître, Aristote considère pour sa part sans la moindre ambiguïté que le noir est la privation du blanc (voir par exemple Met. Ι 2, 1053b30-31), ce qui découle d’ailleurs directement du fait qu’ils sont des contraires et que, comme nous venons de le voir, l’un des membres de tout couple de contraires est la privation de l’autre. Cependant, je ne pense pas du tout qu’Aristote prétende rendre compte ici de la manière dont nous devenons conscients du noir par la sensation : il cherche bien plutôt à expliquer la manière dont nous pouvons penser ce qu’est le noir, son essence, ou encore son εἶδος. De fait, comme il l’écrit dans la Métaphysique, « d’une certaine manière, des contraires, l’εἶδος est le même : en effet, l’οὐσία de la privation est οὐσία de l’opposé, par exemple la santé de la maladie, car la maladie est absence de la première » (τῶν ἐναντίων τρόπον τινὰ τὸ αὐτὸ εἶδος· 53 Comparer Cat. 6, 6a17-18 ; A.Po. I 4, 73b21-22 ; GC I 7, 324a2 ; Met. Γ 6, 1011b19-20 et Ι 4, 1055a27-28. 54 Comparer Philopon, In DA 552, 16-21. 55 W.D. Ross (1963), p. 298. 56 Philopon, In DA 547, 4-8, 552, 23-26.

180

SYLVAIN DELCOMMINETTE

τῆς γὰρ στερήσεως οὐσία ἡ οὐσία ἡ ἀντικειμένη, οἷον ὑγίεια νόσου, ἐκείνης γὰρ ἀπουσίᾳ ἡ νόσος, Ζ 7, 1032b2-5) – à tel point que « c’est la même formule qui montre la chose et sa privation (ὁ δὲ λόγος ὁ αὐτός δηλοῖ τὸ πρᾶγμα καὶ τὴν στέρησιν), même si ce n’est « pas de la même manière (πλὴν οὐχ ὡσαύτως) » (Θ 2, 1046b8-9). Or la saisie de l’εἶδος ou de l’οὐσία par son λόγος ne peut être le fait que de la pensée, et c’est bien d’elle qu’il s’agit dans tout notre chapitre. Remarquons que par cette explication, Aristote ramène l’intellection de ce nouveau type d’indivisé à celle du deuxième type : si nous pouvons penser un indivisé de ce type, c’est parce que nous pouvons penser l’εἶδος, dont il n’est que la privation. On voit une nouvelle fois combien l’intellection de l’indivisé selon l’εἶδος, loin d’être marginalisée, occupe une place centrale dans le chapitre, puisqu’elle fonde aussi bien celle de l’indivisé quantitatif en acte que celle de l’indivisé privatif. L’εἶδος est bien l’objet premier de la νόησις, l’intelligible par excellence et la cause de l’intelligibilité de tout le reste. Les dernières lignes du paragraphe sont difficiles et probablement corrompues. En 430b23-24, j’ai traduit la leçon retenue par Ross, qui semble également être celle lue par Pseudo-Simplicius et Philopon57 : ἐνεῖναι ἐν αὐτῷ, plutôt que ἓν εἶναι ἐν αὐτῷ, que l’on trouve dans la majorité des manuscrits et qui est retenue par Jannone. Ross entoure toutefois ces mots de cruces  ; et de fait, il pourrait sembler nécessaire d’ajouter par exemple ἐναντίον entre ἐνεῖναι et ἐν αὐτῷ pour obtenir le sens présupposé par ma traduction, comme propose de le faire Thillet en se basant notamment sur Bywater58. Mais quel est ce sens ? Tout d’abord, pourquoi faut-il que ce qui connaît soit en puissance ? Thémistius soutient que si la sensation n’avait pas une potentialité d’être active ou inactive mais était toujours active, elle ne percevrait jamais le noir, ni l’ouïe le silence ; de même, s’il n’y avait pas une intelligence appropriée à la fois à l’intellection et à la non-activité (à la non-intellection), elle ne pourrait pas penser des objets mauvais, ni aucun objet dépourvu de structure et de forme. Mais il y a une telle intelligence, à savoir l’intelligence en puissance, qui peut connaître les opposés par les opposés : les formes en étant active, les objets pensés par privation en

57 58

[Simplicius], In DA 257, 26 ; Philopon, In DA 552, 28. P. Thillet (2005), p. 389, n. 461.

DE ANIMA III 6

181

demeurant en puissance59. Cette interprétation nous met à mon avis sur la bonne piste, mais elle ne peut être reprise telle quelle, car elle revient à confondre la privation avec la puissance du contraire positif. Or la puissance correspond bien plutôt pour Aristote à la matière, qui peut être aussi bien l’εἶδος que sa privation (cf. Met. Λ 5, 1071a8-11) ; mais dès qu’elle devient l’une ou l’autre, elle s’actualise, que ce soit dans le sens de l’εἶδος ou de la privation. Il doit en aller de même pour l’intelligence : elle est tout autant en acte lorsqu’elle pense la privation que lorsqu’elle pense l’εἶδος. En revanche, la privation ne peut être en acte que comme actualisation d’une puissance : ainsi, comme l’explique Aristote en Met. Θ 9, 1051a17-19, le mal, en tant que privation du bien, est postérieur à la puissance, qui est indifféremment les deux, de sorte qu’il n’y a pas de mal en dehors des choses, scil. en dehors des composés de matière et d’εἶδος. Tel doit également être le cas de la pensée : seule une intelligence en puissance, qui peut passer à l’acte, peut penser aussi bien l’εἶδος que sa privation, pour autant que l’un des contraires, à savoir le contraire positif, soit présent en elle (ἐνεῖναι ἐν αὐτῷ) ; en revanche, une intelligence qui serait toujours en acte ne pourrait penser que le contraire positif, c’est-à-dire l’εἶδος. Cette interprétation permet également de comprendre les dernières lignes du présent paragraphe (430b24-26). On y voit généralement une allusion à l’intelligence divine qui, n’ayant pas de contraire, se connaît elle-même, est en acte (ou est acte) et est séparée. Berti considère au contraire qu’il s’agit des « formes des substances, c’est-à-dire un cas particulier des indivisibles selon la forme traités dans le passage qui précède » ; par conséquent, il adopte la lecture αὐτῷ du ms. M à la place de ἑαυτὸ retenu par les éditeurs à la ligne 25, et comprend qu’Aristote affirme que ces causes sont connues par elles-mêmes, à la différence des causes ayant un contraire – celles dont il a été question plus haut, scil. les indivisés selon la privation – dont l’intellection se produirait au moyen de ce contraire60. J’avoue ne pas comprendre comment la privation pourrait être considérée comme une cause ayant un contraire, à savoir l’εἶδος, tandis que l’εἶδος n’aurait pas quant à lui de contraire : si l’εἶδος est le 59 Thémistius, In DA 111, 26-34. Cette interprétation est reprise par G. Rodier (1900), p. 483 et 485-486. 60 E. Berti (1996), p. 398-401. Sa lecture de ce passage a évolué depuis son article précédent (1978, p. 146), bien que son interprétation demeure fondamentalement la même. Voir également l’interprétation similaire de R. Polansky (2007), p. 477.

182

SYLVAIN DELCOMMINETTE

contraire de la privation, la seconde n’est-elle pas le contraire du premier ? Certes, Aristote affirme dans certains contextes que l’οὐσία n’a pas de contraire61 ; mais je pense qu’il faut alors entendre l’οὐσία au sens de composé et non au sens d’εἶδος, qui a quant à lui un contraire, à savoir précisément la privation62. En revanche, Aristote affirme explicitement que dieu ou le bien en soi, qui est éminemment une cause, n’a pas de contraire (cf. Met. Λ 10, 1075b21-24), et cette manière de le désigner paraît dès lors parfaitement adéquate63. Quoi qu’il en soit, les principales raisons pour lesquelles Berti rejette l’interprétation traditionnelle sont d’une part qu’il ne voit pas pourquoi Aristote aurait dû faire allusion au premier moteur dans ce contexte, et d’autre part que ce premier moteur ne serait pas selon le Stagirite objet d’une intellection, mais seulement d’une démonstration quant à son existence64. Je ne m’avancerai pas sur le deuxième point, sinon pour faire remarquer que tout ce que l’interprétation traditionnelle prétend voir dans ce passage est que dieu lui-même se connaît, et non que nous pouvons le connaître ; or il est difficilement contestable que cette connaissance de dieu par lui-même est une intellection. Concernant le premier point, il me semble au contraire que cette précision s’insère parfaitement dans le contexte tel que je l’ai reconstruit. Aristote vient de dire que les indivisés selon la privation sont pensés à partir de leur contraire (l’εἶδος) et ne peuvent donc l’être que par une intelligence en puissance, toute puissance étant puissance des contraires ; il en conclut que ce qui n’a pas de contraire (dieu) ne peut être pensé par une telle intelligence, mais seulement par lui-même, et que ce faisant, il est purement en acte (ou même acte, si l’on préfère la leçon ἐνεργεία) et séparé de toute matière – puisqu’il s’agit alors d’une intellection en acte qui n’est pas l’actualisation d’une intelligence en puissance. La conséquence non formulée est que dieu ne peut penser la privation, et donc 61 Voir en particulier Cat. 5, 3b24-32. Pour d’autres références, voir H. Bonitz (1870), 544a60-b2. 62 Voir Met. Γ 6, 1011b18-19 : τῶν μὲν γὰρ ἐναντίων θάτερον στέρησίς ἐστιν οὐχ ἧττον, οὐσίας δὲ στέρησις, avec la note ad loc. de W.D. Ross (1924), I, p. 283 : « ‘(…) privation of the positive, substantial nature’ – more commonly in this connexion called εἶδος. » Sur cette question, je me permets de renvoyer à S. Delcomminette (2019). 63 Pour cette raison, j’avoue ne pas comprendre le problème que poserait le texte des manuscrits en 430b24-25 (εἰ δέ τινι μὴ ἔστιν ἐναντίον τῶν αἰτίων), où certains voudraient remplacer τῶν αἰτίων par τῶν ὄντων (Torstrik), d’autres par τῶν οὐσίων (Thillet), d’autres encore supprimer purement et simplement ces deux mots (Ross). 64 E. Berti (1996), p. 400.

DE ANIMA III 6

183

le mal65 ; or c’est là très certainement une conséquence qui importe au plus haut point à Aristote, lui qui écrit qu’il est absurde que l’intelligence divine pense certaines choses, et que ce qu’il y a de plus divin doit penser ce qu’il y a de plus divin et de plus honorable sans aucun changement (Met. Λ 9, 1074b23-27). 6. 430b26-31 Ἔστι δ’ ἡ μὲν φάσις τι κατά τινος, ὥσπερ ἡ κατάφασις, καὶ ἀληθὴς ἢ ψευδὴς πᾶσα. ὁ δὲ νοῦς οὐ πᾶς, ἀλλ’ ὁ τοῦ τί ἐστι κατὰ τὸ τί ἦν εἶναι ἀληθής, καὶ οὔ τι κατά τινος· ἀλλ’ ὥσπερ τὸ ὁρᾶν τοῦ ἰδίου ἀληθές, εἰ δ’ ἄνθρωπος τὸ λευκὸν ἢ μή, οὐκ ἀληθὲς ἀεί, οὕτως ἔχει ὅσα ἄνευ τῆς ὕλης. L’énonciation dit quelque chose de quelque chose, comme l’affirmation66, et est toujours vraie ou fausse. Mais ce n’est pas le cas de toute intelligence, mais celle qui porte sur le ce que c’est selon le ce que c’est qu’être est vraie, et ne dit pas quelque chose de quelque chose67 ; mais tout comme la vision du propre est vraie, mais que si le blanc est homme68 ou pas n’est pas toujours vrai, il en va ainsi de tout ce qui est sans matière.

Je n’ai pas grand-chose à ajouter sur cette conclusion. Aristote compare l’intellection des « ce que c’est selon le ce que c’est qu’être », c’est-à-dire des εἴδη « sans matière », à la sensation des sensibles propres, en ce qu’elle est toujours vraie, à la différence de l’intelligence dont la structure est prédicative et qui est susceptible d’être vraie ou fausse. Ce rapprochement a seulement trait à la question de la vérité et ne peut en aucun cas 65

Cf. Thémistius, In DA 111, 34-35. Le remplacement de κατάφασις par ἀπόφασις, proposé par Torstrik (suivi par Ross), qui implique qu’on traite φάσις comme synonyme de κατάφασις plutôt que comme notion générique, rend le texte plus clair, mais ne paraît pas strictement nécessaire. 67 Tricot, Barbotin et Thillet comprennent plutôt le segment καὶ οὔ τι κατά τινος au sens de « et non pas celle qui dit quelque chose de quelque chose » (scil. celle-là n’est pas toujours vraie). Cette traduction est possible, mais elle me paraît rendre le sens moins clair, bien qu’elle revienne en définitive au même que celle retenue ci-dessus : dans les deux cas, la structure de l’énoncé du ce que c’est (ou plus précisément du ce que c’est qu’être : la définition complète) est contrastée avec celle de la prédication. 68 Je ne comprends pas pourquoi Ross considère le texte des manuscrits comme « clairement impossible » (W.D. Ross (1963), p. 298) et entoure τοῦ ἰδίου ἀληθές, εἰ δ’ ἄνθρωπος τὸ λευκὸν de cruces. Bien qu’elliptique, ce texte me paraît tout à fait acceptable en l’état. 66

184

SYLVAIN DELCOMMINETTE

être utilisé pour soutenir que l’intellection des intelligibles est d’ordre « intuitif », ni même qu’elle est forcément non-discursive : la seule discursivité qui en est niée est celle qui suppose la structure prédicative. Par ailleurs, j’insiste une nouvelle fois sur le fait qu’Aristote ne dit pas que l’intellection de tout indivisé est toujours vraie, mais seulement celle qui concerne le deuxième type d’indivisé distingué dans les lignes précédentes. Qu’en est-il des indivisés mathématiques, que ce soient les indivisés en acte, mais pas en puissance, ou les indivisés selon la privation ? Aristote n’en dit rien. Dans les deux cas, on peut penser que l’intellection opère par une certaine composition, soit en unifiant une étendue intelligible soit en niant un certain genre. Mais cette composition n’est clairement pas d’ordre prédicatif. Introduit-elle un risque d’erreur ? Ou bien faut-il comprendre que ces objets sont plutôt en deçà du vrai et du faux, parce qu’ils ne sont ni des êtres sensibles ni des essences ? Le présent texte ne permet pas de trancher, et je laisse donc la question ouverte.

DE ANIMA III 7: THE ACTUALITY PRINCIPLE AND THE TRIGGERING OF MENTAL EPISODES Klaus CORCILIUS Für Michel

1. Torstrik’s Challenge It is hard to see what Aristotle’s De anima III 7 is about. The chapter starts out with a sentence that seems to be a verbatim quotation from DA III 5 while the rest seems to exhibit little thematic coherence, jumping back and forth between issues regarding sense-perception, pleasure and pain, desire, and, perhaps predominantly, thought. It also seems that parts of the chapter discuss issues that have been previously discussed in the De anima. There is certainly no obvious unifying theme in the text. This is why the chapter has puzzled interpreters from early on, some of whom openly admitted the lack of thematic unity. Most notably, Adolf Torstrik, in his famous 1862 edition of the De anima, claimed that the chapter divides into 8 sections that lack any thematic unity whatsoever both with each other and with the preceding and the following chapters. He further claimed that in view of the heterogeneity of these sections, not even a “very bad author (pessimum quidem scriptorem)” could have put them together into one single chapter (p. XXV) and that such confusion must have been the work of a later editor1. Torstrik’s radical verdict was received rather favorably 1 Heac laciniae neque inter se neque cum iis quae preacedunt vel sequuntur quidquam habent commune. A. Torstrik (1862), p. 199. A further reason given by Torstrik is that the first sentence of the chapter occurs in the exact same wording in III 5, 430a19 sq. Torstrik found strong words for the disunity of the chapter: Quantopere loci quos tanquam insiticios notavimus non solum ab universe hujus tractatûs tenore sed ab ipsâ re

186

KLAUS CORCILIUS

by later scholars. David Ross, for instance, in his widespread editions of the De anima, accepted the verdict (while opting for a slightly different grouping of the sections), as did editors, translators and commentators such as Förster, Siwek, Theiler, up until Christopher Shields in his recent translation and commentary. The resounding success of Torstrik’s claim is not difficult to explain. For he, apart from his brief arguments for the thematic disunity of the chapter2, provided also solid linguistic evidence to the effect that the parts of the chapter lack coherence even on a basic grammatical level. His most challenging points of criticism were that the μὲν-clause in 431a4 (φαίνεται δὲ τὸ μὲν αἰσθητὸν ἐκ δυνάμει ὄντος…) lacks a corresponding particle in the sequel of the text and that the protasis in 431a17-20 lacks an apodosis3. Perhaps due to these observations, later interpreters did not attempt unitary interpretations of the chapter, until, somewhat recently, Catherine Osborne (1998) and Ron Polansky (2007) have made explicit attempts at revising Torstrik’s disunity claim. Osborne argues that the unifying theme of the “central sections” of the chapter is the unity of the judgmental capacity of the intellect4, while Polansky (who does not seem to be aware of Osborne’s paper) argues that III 7 treats the necessary role of φαντασία as the causally initiating factor for all thinking5 and that the chapter thereby shows that the different subparts of the intellect pertain to one unitary faculty in the same way in which the quae tractatur abhorreant, facile quisque intelligit. Ea re tam manifesta est ut vere dici possit nullum ac ne pessimum quidem scriptorem res tam diversas potuisse miscere: ut haec incredibilis rerum rerum disparium confusio, rerum cognatarum distractio ei tribuendo sit qui haec post Aristotelem edidit. A. Torstrik (1862), p. 205, emphasis mine. Cp. the more cautious remarks of Torstrik’s teacher Trendelenburg: Quem singula communem finem expectant, non satis liquet, ut hic illic magis fortuito quam necessario filo contineri videantur. F.A. Trendelenburg (1862), p. 509. Similarly, G. Rodier (1900), p. 517-519. M. Burnyeat’s brief suggestion of how to fit DA III 7 into the overall argument of the DA, in spite of it being “a series of jottings”, is interesting, cf. M. Burnyeat (2001), p. 72. 2 Which, depending on one’s overall interpretation, one may agree with or not, see his commentarius p. 199-204. 3 A. Torstrik (1862), p. 199. 4 “[T]he way in which the intellect can be said to make additional judgments about the thoughts that it has without thereby presupposing a second faculty standing behind to make meta-judgments on the objects of thought.”, C. Osborne (1998), p. 434. 5 “[H]ow φαντασία has the further function of getting mind thinking in actuality…”, R. Polansky (2007), p. 481.

187

DE ANIMA III 7

sensory capacity, in spite of its various modalities, is but one capacity. Unfortunately, neither Polansky nor Osborne addresses, and let alone diffuse, Torstrik’s astute grammatical observations. Polansky and Osborne moreover exhibit a tendency to explain away the passages in III 7 that concern sense-perception and perceptual desire as mere “analogies” introduced by Aristotle in order to establish points about what he is actually interested in the chapter, which, according to them, is the operations of νοῦς. In this regard they are similar to other interpreters and commentators who regard the chapter as primarily concerned with νοῦς6. However, reducing Aristotle’s claims about perception and desire in III 7 to mere preparations for his claims about νοῦς fails to do justice to what the chapter offers in terms of Aristotle’s theory of desire and its relation to perception. It also fails to do justice to what seems to me the most striking feature of the chapter, which is the fact that it describes an ordered sequence of increasingly complex mental and cognitive achievements that ranges from the most basic case of simple perception to the highly complex theoretical contemplation of abstract objects. This admittedly very short discussion of the previous attempts at revising Torstrik’s disunity claim must suffice. We leave with an illustration of some examples of previous interpreters’ divisions of the chapter. Ways of dividing the chapter Torstrik

Ross

Shields

Osborne

Polansky

Below

431a1-4 431a4-7 431a8-16 431a16-17 431a17-20 431a20-b1 431b2-12

431a1-4 431a4-7 431a8-17

431a1-4 431a4-6 431a7-14 431a14-20

431a1-7 431a8-b1

431a17-20 431a20-b1 431b2-12

431a1-4 431a4-5 431a8-14 431a14-17 431a17-b1

431a20-b1 431b2-17

431b2-12

431b2-16

431b12-19

431b12-19

431b17-19

431b12-19

431b16-19

431a1-4 (T 1) 431a4-7 (T 2) 431a8-14 (T 3) 431a14-16 (T 4) 431a16-17 (T 5) 431a17-b1 (T 6) 431b2-10 (T 7) 431b10-12 (T 8) 431b12-19 (T 9)

6 Cf. Hicks ad loc. who writes that the chapter is concerned with the “operations of thought” and to that end “frequently draws on the analogy of sense-perception” R.D. Hicks (1907), p. 525, albeit without mentioning neither Torstrik’s disunity claim nor the missing particle corresponding to the μὲν in 431a4. Similarly, E. Barbotin (1966), p. 84: “intellect pratique”.

188

KLAUS CORCILIUS

My project here is to interpret DA III 7 as containing a continuous and coherent argument. The focus will lie on the question of the unity and coherence of the argument and less on the rich philosophical implications of this important and extremely dense chapter. I will argue that the unifying subject-matter of the chapter is neither the judgmental faculty of the intellect (Osborne) nor the necessary role of φαντασία for thought (Polansky), but rather the application of what I here call the Actuality Principle to the domain of what one may regard “embodied cognition” Under “embodied cognition” I here understand all sorts of cognitions specific of animals and humans, as Aristotle thinks of it, in the sense of all kinds of body-involving cognition, including concomitant and antecedent motor reactions. Hopefully, what I have in mind will become clear as we go along in the interpretation of the chapter. The Actuality Principle is the principle according to which everything that comes to be F comes to be so in virtue of the agency of something that already is F in actuality. I will have more to say about the precise formulation of the Actuality Principle below. More specifically, I will argue that DA III 7 purports to show that the triggering of every embodied form of cognition, i.e. every cognition of matter-involving objects, and every cognition-involving motor-process ultimately depends on the agency of some actually existing external object of perception. Aristotle starts with the case of basic perception and works his way up along an ascending sequence of increasing cognitive complexity. The series starts with basic perceptual desire, and goes on with the thinking of simple (uncombined) objects of basic perceptual desire; it continues with the perception of complex objects, to then discuss complex practical and theoretical thought (in the sense of combinations of a plurality of elements in judgments), to end with the thinking of abstract objects7. The chapter shows for all these cases, how they are set up either by an external actual object of perception or by a substitute for such an external 7 The very last cognitive state mentioned in the chapter is the thinking of separate substances, i.e. objects of thought that do not involve any matter (431b17-19, see below T 9). For this case of cognition Aristotle does not offer us an account of its triggering cause but simply asks the question whether it is possible to think such objects with a capacity that itself not separated (from matter). This, I take it, implies not only that the previous cognitive states are all in one way or the other matter-involving (hence ‘embodies cognition’) but also that he thinks of them as parts of a cognitive spectrum that spans from basic sense-perception to the thinking of separate substances. He thus fleshes out the claim made in III 4, 429b10-22 (ὅλως ἄρα ὡς χωριστὰ τὰ πράγματα τῆς ὕλης, οὕτω καὶ τὰ περὶ τὸν νοῦν) by offering a certain scala of cognitive states and their respective ways of actualization.

DE ANIMA III 7

189

object (a φάντασμα). The chapter also emphasizes the fact that the first initiating item of the series, the external object of perception, exerts its causal role as a complete actuality, which is to say that it acts as an unmoved mover. With this, the chapter shows how every form of human cognition and motor-reaction ultimately depends on an external moving (triggering) cause that acts as a complete actuality. The chapter thereby also shows how even very complex cognitive activities like associative desire or the thinking of abstract objects are triggered by the very things they are about (in a sense to be clarified below). The chapter thus vindicates Aristotle’s claim from previous chapters of the De anima that the object and the subject of cognition are somehow the same (“identity claim”), which is a claim that underlies his account of perception and thought as assimilation. So much for the argumentative goals of this chapter. However, I do not claim here to provide incontrovertible evidence for the argumentative unity of the chapter. I shall content myself with showing that the chapter can be given a plausible unitary interpretation, also on a grammatical level, if read as a chapter devoted to showing that the triggering causes of all sorts of animal and human cognitions and motor reactions are (in some sense to be specified) the same objects these states are about8. My hope is to thereby invert the burden of proof and to make it incumbent upon those who deny the thematic unity of the chapter to establish their theses against the unity claim. For the text of DA III 7, I will use the text and the critical apparatus of the still most reliable modern edition of the De anima, namely Aurel Förster’s Budapest edition from 19129. My translations will be largely based on Shields’ 2016 translation. 2. Getting started with perception Τ1

The Actuality Principle (431a1-4)

Τὸ δ᾽ αὐτό ἐστιν ἡ κατ᾽ ἐνέργειαν ἐπιστήμη τῷ πράγ[2]ματι. ἡ δὲ κατὰ δύναμιν χρόνῳ προτέρα ἐν τῷ ἑνί, ὅλως [3] δὲ οὐδὲ χρόνῳ· ἔστι γὰρ ἐξ ἐντελεχείᾳ ὄντος πάντα τὰ γι[4]γνόμενα.

8 In a way, then, one may think of fuction of the chapter as similar to that of the doctrine of recollection does in Plato’s account of the thinking of forms. The associative mechanism leading from the triggering of mental episodes described in DA III 7 to the actual thinking of objects of thought is described in Aristotle’s De memoria. 9 For a justification, see K. Corcilius (2017), p. LXIX-LXXIV.

190

KLAUS CORCILIUS

T2

Basic Perception (431a4-7)

φαίνεται δὲ τὸ μὲν αἰσθητὸν ἐκ δυνάμει ὄντος τοῦ [5] αἰσθητικοῦ ἐνεργείᾳ ποιοῦν· οὐ γὰρ πάσχει οὐδ᾽ ἀλλοιοῦται. [6] διὸ ἄλλο εἶδος τοῦτο κινήσεως· ἡ γὰρ κίνησις τοῦ ἀτελοῦς [7] ἐνέργεια, ἡ δ᾽ ἁπλῶς ἐνέργεια ἑτέρα ἡ τοῦ τετελεσμένου.10

The chapter’s first sentence makes a series of statements about potential and actual ἐπιστήμη and αἴσθησις: (i) (ii) (iii) (iv)

(v)

(vi)

(vii)

Actual ἐπιστήμη is identical with its object. In the individual thinker potential ἐπιστήμη is prior in time to actual ἐπιστήμη. Considered generally, i.e. in abstraction from individual thinkers, potential ἐπιστήμη is not even (οὐδέ) prior to actual ἐπιστήμη in time. is true because (γάρ) everything that comes into being comes into being through the agency of (ἐξ) something that is in actuality. (Actuality Principle in what follows). The object of perception (αἰσθητόν) is causally responsible for turning the capacity of perception (the αἰσθητικόν) from potentiality into actuality. Claim in (v) is true because (γάρ) x is neither changed, nor altered in the process, which is why the process is a different kind of change / different from change in kind. Claim (vi) is true because (γάρ) (ordinary) change is an actuality of what is incomplete, which is different from actuality without qualification, i.e. the actuality of what has already been brought to completion (and is therefore complete).

10 1 δ᾽ αὐτὸ LW: αὐτὸ δ᾽ CUVXy Sil Phl: δ᾽ αὐτὸ δ᾽ S 2 (ποτέρα U) (ἔνι ὧ ἔνι CV) ὅλως o Sil Phl c557,27 So 128,34: ἁπλῶς Phv557,27 3 (ἐντελεχείας WXy) (ὄντως X, [corr in -ος] y) 4 αἰσθητὸν LCUWy Phl Th 28,35: αἰσθητήριον SVX γρ. mgC¹: αἰσθητικὸν Sil A (ἐν X) 6 εἶδος τοῦτο κινήσεως L Sil cf p265,6: τοῦτο εἶδος κινήσεως CUVy Th 28,36: τοῦτο κινήσεως SX [ins εἶδος X³] 7 ἐνέργεια1 LSUX Sil c 265,12 Th 28,37: ἐνεργεία S: ἐνέργεια ἦν CVWy (ἑτέρου Sil A) ἡ2 b–X Sip265,15 Th 29,1: ἢ X: om L Sil A NB: o (= omnes codices) sine E! ο omnes codices significat h. e. libro primo ECSUVWXy, secundo praeterea Ρ, tertio L. Siglo a codd. EL, siglo b codd. CSUVWXy denotantur. o-y i. q. omnes codd. excepto uno y. Note that from the beginning of DA chapter 6 from 430a26 up until 431b16 towards the end of chapter 7 ο does not include codex Ε. Sil = lemma Simplicii, idem cum Philopono Sic = citatio Simplicii Sip = paraphrasis Simplicii Siv = varia lectio in Simplicio

DE ANIMA III 7

191

(i)-(iii) make a point about κατὰ δύναμιν ἐπιστήμη, namely that it transitions from potentiality into actuality ἐξ an actual object of ἐπιστήμη. Aristotle here may have in mind the teacher who teaches the person who is κατὰ δύναμιν ἐπιστήμων by bringing her into contact with the objects of ἐπιστήμη. This interpretative assumption would at least help us to see why, if considered in abstraction from individual thinkers, not even potential ἐπιστήμη is prior to actual ἐπιστήμη. (iv) purports to explain why this is so. There is a general principle according to which everything that comes into being (πάντα τὰ γιγνόμενα), including the process of becoming an actual ἐπιστήμων individual, comes to be ἐξ something in actuality. (i)-(iii) is a verbatim repetition of DA III 5, 430a20-21. This alone, as has been noted by scholars since antiquity, does not preclude the passage from making good sense in its own context here. Below, I will argue that this is indeed the case. However, first we need to make sense of the other claims contained in the above passages. (v) seems straightforward. But in (vi) there is a major unclarity. What is x, the grammatical subject of πάσχει and ἀλλοιοῦται? There are two relevant options: either the αἰσθητόν (a) or the αἰσθητικόν (b). On option (a) the αἰσθητόν causes the transition from potentiality to actuality in the αἰσθητικόν without itself undergoing any change, whereas on option (b) it is the αἰσθητικόν that transitions from potentiality to actuality without undergoing any change. Regarding (a), I should start by saying that it seems true for Aristotle. Surely, Aristotle thinks that (1) the objects of perception do not change by being perceived by an actual perceiver, (2) the causal efficacy of the αἰσθητόν in causing its own being perceived does not lead to a change in the αἰσθητόν, and that (3) the αἰσθητόν is the first unmoved mover of the series of changes that leads to sensory affection and perception, cf. DA III 12, 434b26-435a5: εἰ γὰρ μέλλει σῴζεσθαι [i.e. τὸ ζῷον], οὐ μόνον δεῖ [27] ἁπτόμενον αἰσθάνεσθαι ἀλλὰ καὶ ἄποθεν. τοῦτο δ᾽ ἂν εἴη, [28] εἰ διὰ τοῦ μεταξὺ αἰσθητικὸν εἴη τῷ ἐκεῖνο μὲν ὑπὸ τοῦ [29] αἰσθητοῦ πάσχειν καὶ κινεῖσθαι, αὐτὸ δ᾽ ὑπ᾽ ἐκείνου. ὥσπερ [30] γὰρ τὸ κινοῦν κατὰ τόπον μέχρι του μεταβάλλειν ποιεῖ, [31] καὶ τὸ ὦσαν ἕτερον ποιεῖ ὥστε ὠθεῖν, καὶ ἔστι διὰ μέσου ἡ [32] κίνησις, καὶ τὸ μὲν πρῶτον κινεῖ καὶ ὠθεῖ οὐκ ὠθούμενον, τὸ [33] δ᾽ ἔσχατον μόνον ὠθεῖται οὐκ ὦσαν, τὸ δὲ μέσον ἄμφω, [435a] πολλὰ δὲ τὰ μέσα, οὕτω καὶ ἐπ᾽ ἀλλοιώσεως, πλὴν ὅτι μένοντος [2] ἐν τῷ αὐτῷ τόπῳ ἀλλοιοῖ, οἷον εἰ εἰς κηρὸν βάψειέ τις, [3] μέχρι τούτου ἐκινήθη, ἕως ἔβαψεν· λίθος δὲ οὐδέν, ἀλλ᾽ [4] ὕδωρ

192

KLAUS CORCILIUS

μέχρι πόρρω· ὁ δ᾽ ἀὴρ ἐπὶ πλεῖστον κινεῖται [5] καὶ ποιεῖ καὶ πάσχει, ἐὰν μένῃ καὶ εἷς ᾖ.

We could therefore think of ἄλλο εἶδος τοῦτο κινήσεως in 431a6 in (viii) not as the kind of “motion” undergone by the αἰσθητικόν but as that εἶδος of change which is caused by the αἰσθητόν as an unmoved mover, as for instance, the motion set up by the visible object in the transparent environment11. Aristotle’s motivation to hold (a) can be explained by an epistemological consideration against the converse thesis: suppose the object of perception undergoes a change in virtue of setting up the motion in its environment that will lead to its being perceived by a perceiver. In that case, we wouldn’t perceive the objects as they are in themselves, but only as they change in relation to us12. On reading (a), by contrast, Aristotle can preserve the veridicality of perception. Option (a), therefore, clearly seems to him, to put it mildly, to be a viable possibility. Is (b) viable in the context of DA III 7? This is not so easy a question as one might think. Not infrequently, Aristotle attributes some kind of change to the senses themselves (as, for instance, in the continuation of the above passage from DA III 12: διὸ πάλιν οὗτος τὴν ὄψιν κινεῖ, ὥσπερ ἂν εἰ τὸ ἐν τῷ κηρῷ σημεῖον διεδίδοτο μέχρι τοῦ πέρατος, 435a8-10, cp. 422b3, and 426b31-427a1). However, his usage of these formulations seems to vacillate and one would like to avoid having to attribute to him the mutually inconsistent views that (i) the faculty of perception (the perceptual soul) is not a possible subject of change (which is a point Aristotle argues for at length in DA I), and that (ii) the soul is a subject of change in virtue of being affected by the sense object. In view of these difficulties, it is somewhat astounding that (b) is preferred by almost all commentators13. 11 There might be a contrast intended with II 5, 417b6-7 (ὅπερ ἢ οὐκ ἔστιν ἀλλοιοῦσθαι – εἰς αὑτὸ γὰρ ἡ ἐπίδοσις καὶ εἰς ἐντελέχειαν – ἢ ἕτερον γένος ἀλλοιώσεως) which is a claim not about the change involved in the object of perception, but about the change involved in the subject. 12 See also DA III 2, 426a5 (διὸ οὐκ ἀνάγκη τὸ κινοῦν κινεῖσθαι). Cf. Sens. 4, 441b20-21: flavor “is such as to alter the sense of taste into operation (ἀλλοιωτικὸν εἰς ἐνέργειαν),” and generally “for each of these [i.e. of the perceptible affections] is such as to bring about the perception, for they are all said [to be what they are] in virtue of their power to set up motion.” (ποιητικὸν γάρ ἐστιν ἕκαστον αὐτῶν τῆς αἰσθήσεως, τῷ δύνασθαι γὰρ κινεῖν αὐτὴν λέγεται πάντα, Sens. 445b7 sq.). The sense objects are what they are said to be in virtue of their causal effects on the sense-organs. The sense objects therefore do not change by thus affecting the sense organs. 13 For an exception, see S. Menn (1994), p. 110, n. 49.

DE ANIMA III 7

193

For textual support, they typically refer to a famous passage in DA II 5, 417b6-7, where Aristotle, in speaking about the transition from not perceiving to perceiving and its relation to ordinary chance, seems to make a similar point: ὅπερ ἢ οὐκ ἔστιν ἀλλοιοῦσθαι – εἰς αὑτὸ γὰρ ἡ ἐπίδοσις καὶ εἰς ἐντελέχειαν – ἢ ἕτερον γένος ἀλλοιώσεως

Here, the act of perceiving is denied as ordinary change and reassigned as either a different kind of change or different from change in kind. Thus, if we are committed to the impassivity of the sensory capacity (as opposed to the sense organs, see III 4, 429a29-b5), it seems that both (a) and (b) are true for Aristotle: neither the αἰσθητόν nor the αἰσθητικόν (strictly speaking) changes during the process of sense-perception; rather, for him, the process of sense-perception both begins and ends with an actuality of something complete, namely the actuality of the external αἰσθητόν on the one hand and the actuality of the αἰσθητικόν on the other. Which of the two readings should we prefer for our context, then? I shall base my decision upon my view of the overall argument of the chapter. In this regard option (a) clearly does better than (b). It does so for mainly two reasons. Firstly, (a), the reading according to which the object of perception (αἰσθητόν) does not change in virtue of setting up the change in its environment that will lead to its own being perceived, provides us with a context that explains why (vi) and (vii) occur in the first place in this context in DA III 7. This is because perception is a straightforward and unambiguous case of something coming ἐξ something, because perception “comes” in the relevant way from (ἐξ) the perceptual object. The ἐξ in this case points to the item that sets up a motion or a chain of motions as its first moving cause. This is the second meaning of ἔκ τινος εἶναι Aristotle distinguishes in Met. Δ 24. The one example he gives there is an offense leading to battle (1023a29 sq.: ἕνα [i.e. ἕνα τρόπον τὸ ἔκ τινος εἶναι λέγεται, KC] δ’ ὡς ἐκ τῆς πρώτης κινησάσης ἀρχῆς, οἷον ἐκ τίνος ἡ μάχη; ἐκ λοιδορίας, ὅτι αὕτη ἀρχὴ τῆς μάχης; “To come from something means (…) as from the first moving principle, e.g. what does the fight stem from? – From abusive language, because this is the source of the fight”14. On this interpretation, we can see why Aristotle speaks of 14

Trans. W.D. Ross (1924).

194

KLAUS CORCILIUS

perception and how perceptual motions are triggered by (“come from”) the actual external sense object as its first unmoved mover. He speaks of perception because the causation of perception by an external perceptual object is an instance of the Actuality Principle in (iii). Secondly, on reading (a), (v), (vi), and (vii) follow naturally from the Actuality Principle, as (v) validates the Principle for the case of basic perception and (vi) and (vii) explain why this is so. On option (b), by contrast, this is not so. It is not the case that perception comes “from” (ἐξ) the perceptual capacity as an unmoved mover. Hence option (b) disconnects the Actuality Principle from the claims about perception in the immediate sequel of (b). As far as the context is concerned, then, option (a) is preferable. But there is also an independent reason for rejecting (b) in favor of (a). This is that the chief textual parallel for (b), the famous passage in DA. II 5, quoted above, speaks of the transition into completion (εἰς αὑτὸ γὰρ ἡ ἐπίδοσις καὶ εἰς ἐντελέχειαν, 417b6-7), which is the process of coming into one’s own state of completion. But this, strictly speaking, is not what T 1 is about. T 1 makes a claim about the causal efficacy of the actuality of what is complete, i.e. of something that is not as of yet coming into its own state of completion (and therefore not fully completed) but of something that already is in that state and exercising it (cf. the perfect tense, ἡ δ᾽ ἁπλῶς ἐνέργεια ἑτέρα ἡ τοῦ τετελεσμένου in DA II 5, 417b7). This suggests the following picture. In sections (v)-(vii), Aristotle explains how the Actuality Principle (iv), which he seems to take for granted in the case of ἐπιστήμη (presumably on the basis of III 4, 430a3-5 and III 5, 430a19-21), holds also in the case of perception. He does so by showing how the actuality of perception causally “comes from” (ἐξ) an actually existent perceptual object, which moreover does not change in virtue of playing this causal role. Now, once the suggestion has been made that (v) – (vii) is an application of the Actuality Principle (iv) to the case of perception, there is a good motive to pursue this hypothesis further. If the Actuality Principle holds for πάντα τὰ γιγνόμενα, as Aristotle says it does (in iv), it would be natural for him to try to show how the principle holds also outside of perception in the cases of processes (γιγνόμενα) that lead to other mental events that either involve or somehow otherwise presuppose perception. This, I suggest, is indeed the motivation that pervades DA III 7. The chapter applies the Actuality Principle to the causation of embodied

DE ANIMA III 7

195

mental episodes of various sorts and in an ascending sequence of complexity, by showing how in all of these cases it is ultimately actual external objects of perception that act as triggers of the processes that lead to their actualization. The chapter applies the Actuality Principle along the following steps: T1 T2 T3 T4 T5 T6 T7 T8 T9

The granted case: ἐπιστήμη and the formulation of the general Actuality Principle (431a1-4) Basic perception (431a4-7) Basic (perceptual) desire (431a8-14) Simple practical thought of the objects of basic desire (431a14-16) A claim about thinking generally (431a16-17) Complex perception (431a17-b1) Complex practical thought (431b2-10) Complex non-practical thinking (431b10-12) Thinking of abstract objects and a final question (431b12-19)

The list makes it clear that from T 2 onwards Aristotle pursues a bottom-up strategy. He starts with basic perception and works his way up towards the more demanding kinds of embodied cognition, until he reaches the thinking of abstract objects, ending with a question concerning the thought of objects separated from magnitude. The suggestion, then, is that the question that pervades DA III 7 is how the actuality of external perceptual objects functions as triggers of episodes of perception, desire, and instances of practical and theoretical thinking. The case of complex perception (the perception of external objects and the discrimination of heterogeneous and co-specific perceptual input) in T 6, however, will prove a difficult case. That passage might or might not be part of the overall argument of the chapter. I will provide arguments for both its inclusion and exclusion but shall remain agnostic about which way to go. The Actuality Principle. It is important to note that the Actuality Principle is to be understood in a weak sense here that does not require the formal identity between the unmoved trigger and the content of the mental act it triggers in all respects. The forms of the triggering cause and the content of the mental act it triggers can come apart, and they will come apart as Aristotle progresses towards the more demanding kinds of mental events. Thus, while turning the αἰσθητικόν from being potentially F to being actually F happens through the triggering agency of an actual and (in respect of F) unchanging αἰσθητόν, F, there is no such

196

KLAUS CORCILIUS

complete formal identity between the agent and the patient of the triggering causal relation in the cases of processes that lead to thinking (from T 4 onwards including the first, but not the latter, part of T 6)15. DA III 7 seems first and foremost interested in the triggering cause of the processes that lead to the mental acts in question. The chapter gives no full account of the causation of the mental acts whose triggering causes it describes. Certainly, the triggering causes of the processes relevant in T 4 to T 9 are not sufficient to explain the actual presence of the contents of the mental events that they trigger, even if there must be some sort of formal connection between them. As I am going to argue towards the end of the chapter, the connection between the triggering causes of the kinds of thought relevant in DA III 7 and their content must be such as to bring about some sort of contact between the thinker and the object thought16. That contact, however, may be very frail, and certainly does not require complete formal identity17. 15

Compare the impressive list of parallel passages for Aristotle’s synonymy view of causation (i.e. the formal sameness of the effect and that ἐκ which the effect results) in τέχνη and the biological production of οὐσίαι adduced in Rodier’s commentary ad 431a3 (p. 490 sq.). However, if I am on the right track, what we are concerned with in DA III 7 is not synonymous causation of οὐσίαι either by τέχνη or by way of biological reproduction, but the triggering causes of processes that lead to something the form of which can be very different from their triggering causes, namely motor reactions and thoughts. In such cases, the relation between triggering causes and their effects certainly does not have to involve synonymy (the sight of a mouse can trigger an avoidance reaction in an elephant, and the φάντασμα of a piece of pie can trigger for instance a process that leads to the thought of Pythagoras’ theorem). Still, as I am going to argue below, there must be some connection (formal, causal, historical, representational etc.) between the triggering cause of a thought and what the thought is about for Aristotle; he just doesn’t require it to be complete formal sameness. A full account of the thought episode probably would have to involve the associative mechanism described in De Mem. 1 plus the account of thinking in DA III 4-5. 16 For a discussion of effecting change in something by the activity of something complete, see A. Anagnostopoulos (2017). 17 Thus, DA III 7 might also pursue the side interest of showing that mental episodes in animals and humans are not self-caused events in a radically spontaneous way, but have antecedent causes in nature (Phys. VIII 6, 259a4-12, cf. Phys. VII 2, 244b2-245a11, Ross). I very much thank Stephen Menn for sending me the notes of a paper on “Occasions, occasional causes, and occasionalism” he gave in honor of Dan Garber in Princeton 2014. In these notes, he suggests the following account of what seems a type of cause closely related to the kind of triggering cause relevant in DA III 7, namely procatarctic causes: “X is an occasion, or gives occasion, for Y to do Z {this neither entails nor contradicts X’s being an efficient cause to Y} the issue is not whether X resembles Z but whether it can explain Z. X is a merely occasional cause, occasioning Y to produce Z, when X insufficient to explain Z. X is a triggering cause, explaining why Y produces Z now rather than earlier or later.” It seems to me that DA III 7 is to some large extent, but

DE ANIMA III 7

197

Torstrik’s grammatical challenge To argue for a unitary reading of the chapter requires us to deal with Torstrik’s grammatical arguments. Starting with the first, Torstrik’s observation was that there is no particle that connects T 2 with T 3. He interpreted this as strong evidence for his disunity claim: since the μέν in T 2 (431a4 sq., φαίνεται δὲ τὸ μὲν αἰσθητὸν ἐκ δυνάμει ὄντος…) lacks a corresponding particle in the following clause in T 3 (431a8, τὸ μὲν οὖν αἰσθάνεσθαι ὅμοιον τῷ φάναι…), the two passages do not form a grammatical unity. But there might be a way to defuse the worry. For, supposing the above hypothesis about the unity of the chapter is correct, the clause in T 2, 431a4 sq., initiates a series of cases in which an external object of perception (either directly or indirectly) triggers a mental episode. Now, as it happens, the sentence that introduces the second item in this series of mental episodes in T 3 (basic perceptual desire) starts with a μέν as well, as it must, because it introduces a complex analogy that on its part consists in a series of steps (perception, pleasure and pain, and desire, see below ad T 3). So, in order to do justice to both of these demands according to school grammar, in T 3 Aristotle would have to write one δέ to connect with the μέν from the previous clause, and a μέν to get started with the first item of series of events leading to desire. But he cannot write δὲ μέν. It looks, therefore, as if he is forced to write μέν in both passages. But there is a further reason, compatible with the one just given, why he could have made use of what grammarians call “μέν solitarium” in T 2. See Denniston: The explanation of μέν solitarium in general, is either that the speaker originally intends to supply an answering clause, but subsequently forgets not entirely, concerned with procatartic causes of this kind. This is, first, because some of the triggering causes discussed in DA III 7 are sufficient for the effect. Thus, in perception the object of perception (X) is not a “merely occasional” cause because it is sufficient for Y (the perceiver), to produce Z (the perception of the object). But from there onwards (in desire, and practical and theoretical thinking), X and Z indeed do come apart (albeit, as I am going to argue below, in all the cases Aristotle considers in DA III 7 there must be some sort of contact-establishing connection between the triggering cause (X) of a thought and what the thought is about). Second, DA III 7 seems to entail the claim that the triggering causes of these mental events – since they all ultimately go back to external objects – act as actualities (unmoved movers, in relation to their effects), which seems irrelevant in the case of procatarctic causes. Still, the large overlap between the question of DA III 7, how embodied mental events come about ἐξ something in actuality, with the conception of procatartic causes is striking.

198

KLAUS CORCILIUS

his intention (…), or, far more frequently, that he uses μέν, like γε in contrast with something which he does not, even in the first instance, intend to express in words, or even (sometimes) define precisely in thought…18

If we go with the second and “far more” frequent option, Aristotle would use μέν in T 2 (431a4) in order to contrast the case of perception with something he does not express in words. The natural candidate for this is all the cases of mental episodes of which one might think that the Actuality Principle does not hold because we do not observe antecedent causes of them that act as complete and actualized items. These, of course, are the cases of perception-dependent cognition and motor reaction he will discuss in T 3 – T 9. The μέν in T 3 (431a8), by contrast, introduces actual perception as the first item in a series of events in order to show how even the more complex cases of mental events like desire etc. can be causally traced back to an actual external object of perception as their first triggering cause. This suggests, on a twofold basis, the following translation of the passage in 431a1-8: Actual knowledge is the same as its object. But potential knowledge is prior in time in the individual; however, considered generally it is not even prior in time; for all things that come to be do so from something that is in actuality. And at least in the case of the object of perception it is clear (φαίνεται δὲ τὸ μὲν αἰσθητὸν) that it makes that which can perceive actively so instead of potentially so; for it [acts in actuality and] is not affected or altered. Hence it is a different form of change; for change is an actuality of the incomplete, while the actuality without qualification is different, namely that of the complete. Now perceiving, to start with (τὸ μὲν οὖν αἰσθάνεσθαι), is like mere saying and thinking, but whenever it is pleasant or painful…19

18

J.D. Denniston (1934), p. 380. Torstrik’s second grammatical challenge is easier to deal with. His claim is that the protasis in T 6 (431a17-20) lacks an apodosis in what follows. However, this would be correct only if one accepts (as Torstrik seems to have done) Bonitz’s idea that there is no apodotic δὲ in Aristotle (H. Bonitz (1870) s.v. δέ). But there clearly is the apodotic δὲ in Aristotle. For apodotic δὲ generally see R. Kühner and B. Gerth (1904), II 2, § 532, p. 276; for examples especially in Aristotle see Denniston (1954), 177 sq. (to which one may add specifically in the DA III 3, 427a19 and III 10, 433b15). So we are justified in translating the sentence as do we below. 19

DE ANIMA III 7

199

3. The passages on basic desire and simple practical thought of its objects T3

Basic desire (431a8-14)

τὸ μὲν οὖν αἰσθάνεσθαι ὅμοιον τῷ φάναι μόνον καὶ νοεῖν· [9] ὅταν δὲ ἡδὺ ἢ λυπηρόν, οἷον καταφᾶσα ἢ ἀποφᾶσα διώ[10]κει ἢ φεύγει· καὶ ἔστι τὸ ἥδεσθαι καὶ λυπεῖσθαι τὸ ἐνερ[11]γεῖν τῇ αἰσθητικῇ μεσότητι πρὸς τὸ ἀγαθὸν ἢ κακόν, ᾗ τοι[12]αῦτα. καὶ ἡ φυγὴ δὲ καὶ ἡ ὄρεξις ταὐτό ἡ κατ᾽ ἐνέργειαν, [13] καὶ οὐχ ἕτερον τὸ ὀρεκτικὸν καὶ φευκτικόν, οὔτ᾽ ἀλλήλων οὔτε [14] τοῦ αἰσθητικοῦ· ἀλλὰ τὸ εἶναι ἄλλο.20 To be perceiving, then, is like bare saying or thinking; but whenever it is pleasant or painful, the soul, as if it were affirming or denying, pursues or avoids; and to feel pleasure and pain is to act with the perceptual mean in relation to what is good or bad insofar as they are such; and avoidance and pursuit when actual are the same. And what is capable of pursuing and what is capable of avoiding are not different, either from one another or from what is capable of sense-perception; but their being is different.

This is the next step in the unfolding of the series of triggering of increasingly complex forms of embodied cognition, namely the triggering of what I call basic desire. This is the desire for objects of perception in which the good or the bad the animal goes for just is the ἡδὺ ἢ λυπηρόν. In higher forms of desire, which is beyond this passage, what is good and what is pleasant may come apart. If the above interpretative hypothesis is correct, the question here is how such basic desire for perceptual objects is caused by something in actuality. The answer then would be that such episodes are triggered by external objects of perception. Aristotle illustrates this triggering process by way of at first sight seems a complex two-step analogy. (i) simple thinking or saying of something :: perceiving of objects (ii) affirmation / denial :: pursuit / avoidance when perceived objects are pleasant / painful

20 8 τῷ o–X Sil: τὸ X Phl D 10 (ἀποφεύγει C) καὶ2 LSX: ἢ CUVWy Sil 266,9 11 (post ἀγαθὸν eras ἧ ἀγαθὸν X) (ἢ κακὸν om CS, suppl mg C²) ἧ τοιαῦτα CSUWy Phc559,10: ἧ τοιαῦται V: ἧ τοιούτου L: om X: ἢ τὰ τοιαῦτα Sil c266,10.15 12 ἡ1 et ἡ2 b–W: om LW Sil cf Phl D δὲ LW Phl: δὴ SUX Sil: om CVy τὸ αὐτὸ LV, [in ras] X: ταὐτὸν C Sil: τοῦτο SUWy Phl (ἡ3 om V) 13 (τὸ] τι V) φευκτικόν b: τὸ φ. L Sil p 266,37 Th 113,27 NB: o sine E! c

200

KLAUS CORCILIUS

For Aristotle, technically speaking, the simple saying or thinking (φάναι μόνον καὶ νοεῖν) relates to simple terms21. Combinations of terms form sentences (λόγος), but sentences are not yet necessarily truth-apt (Int. 4). Affirmation and negation, by contrast, are forms of statements (ἀπόφανσις). Affirmative statements positively say something of something, negative statements deny that something holds of something (Int. 6), and both are necessarily true or false. Now if we distinguish between the combination of terms on the one hand and their affirmation / denial on the other, as Aristotle does in De Interpretatione, we get, in nice symmetry with the other side of the analogy, likewise a three-stage analogy: (i) Thinking or saying simple terms :: perceiving of objects (ii) Combining terms :: perceiving of pleasant / painful objects (iii) Affirming / denying :: pursuit/avoidance of pleasant / painful objects

But it is not important for my present concern whether we read the analogy as two- or as three-stage, as long as the reader sees that the passage makes a point about how pursuit/avoidance issue somehow as a consequence of the feeling of pleasure and pain, which in turn is somehow a consequence of the perception of objects that are either good or bad for the animal: (i) Perceiving an object (ii) Feeling pleasure or pain in perceiving that object (iii) Pursuing/avoiding that object

Animals feel pleasure when they perceive what is good for them and they feel pain when they perceive what is bad for them (or, as the text has it, when they “act with the perceptual mean in relation to what is good or bad”). The resulting sensory pleasure or pain is necessary and sufficient for sensory desire (indicated by the ὅταν)22. All of this, apart from giving us a natural teleological explanation of the fact that animals feel pleasure and pain and desire objects in their environment accordingly, show how an external object – like in the case of perception discussed in T 2 - can function as the unmoved mover of the animal’s motor reaction (see 21 See Int. 4; Met. 1051b24, although he himself often departs from that technical use, see H. Bonitz (1870), s.v. φάναι. 22 Qualifying the good or bad objects with “insofar as they are such”, scil. insofar as they are either good or bad (πρὸς τὸ ἀγαθὸν ἢ κακόν, ᾗ τοιαῦτα). The plural τοιαῦτα indicates that the ᾗ picks out the unity of the disjunction of perceived good or bad objects.

DE ANIMA III 7

201

DA III 10, 433b13-18). The case of basic desire differs from the case of perception in T 2, however, because the unmoved external object (the unmoved mover of the process) encounters an animal for which that object is either good or bad. This circumstance, that the animal’s body is such as to react physically in self-preserving ways to certain objects in its environment (“good and bad” have a biological meaning here), makes the animal discharge an amount of physical energy disproportionate to the input provided by the external object or its medium23. So much for the explanation of how actual objects of perception, by being perceived by perceivers, can trigger basic perceptual desires as unmoved movers. This validates the Actuality Principle beyond bare perception in the case of basic perceptual desire, which is a complex kind of embodied cognition. Note that threestage analogy Aristotle here uses will recur in later sections of the chapter whenever he talks about higher forms of motor-reactions24. This starts with the following passage. T 4 Simple practical thought of the objects of basic desire (431a14-16) τῇ δὲ διανοητικῇ ψυχῇ [15] τὰ φαντάσματα οἷον αἰσθήματα ὑπάρχει, ὅταν δὲ ἀγαθὸν [16] ἢ κακὸν, κατάφησιν ἢ ἀποφησιν καὶ φεύγει ἢ διώκει.25 The images belong to the thinking soul in the manner of perceptions, and whenever it is good or bad, one affirms or denies and avoids or pursues.

The passage builds on, and largely parallels, T 3. The difference is that Aristotle here speaks about how motor-reactions can be caused by the thinking of objects of desire instead of by their mere perception. I call this 23 The passage, even though it is mainly about the triggering of basic desire, entails a teleological explanation of the fact that animals feel sensory pleasure and pain and desire objects in their environment accordingly: pursuing pleasant and avoiding painful objects is the mechanism by means of which animals bring themselves in possession of what is good for them and avoid what is bad for them. Or, put otherwise: via pleasure / pain and corresponding pursuit / avoidance they survive, preserve, and maintain themselves on the level of perceivers. See DA III 12, 434b11-27, Sens. 436b15-437a1, PA II.17, 661a6-8, HA IX.1, 589a8-9. Teleologically speaking, pursuit and avoidance are motions of the living body in the service of the self-preservation of the animal (K. Corcilius (2008), I; K. Corcilius (2011); K. Corcilius and P. Gregoric (2013)). 24 He refers to the account given in T 2 also in EN 1139a21 sq. 25 14 δὲ o–S Sil Th 113,14: δὴ S Phl 16 (κακόν ἐστι SV) κατάφησιν ἢ ἀπόφησιν SUV Sil, corr bis in -φήσειν S²: φησὶν ἢ ἀπόφησι CX: φήσῃ ἢ ἁποφήσῃ L: καταφήσῃ ἢ ἀποφήσῃ y: κατάφασις ἢ ἀπόφασις W cf Php559,31 Th 113,18 καὶ b Sil: om L NB: o sine E!

202

KLAUS CORCILIUS

“basic practical thinking” because it describes thinking of objects, although short of deliberation about these objects (which happens in T7). The deviations from the basic model in T 3 are particularly interesting. There is a transition from ὅταν δὲ ἡδὺ ἢ λυπηρόν in T 3’s a9 to ὅταν δὲ ἀγαθὸν ἢ κακὸν in T 4’s a15-16, indicating that the thinking mind is oriented towards goodness as opposed to goodness as it appears in the merely biological guise of the pleasant or painful. Still, I think we would overburden Aristotle’s explanatory aims if we were to understand the passage as saying that a rational thinker always pursues what is actually good or actually bad. Rather, the point here seems to be simply that the thinking mind’s judgment of something’s being good or bad can issue in motor-reactions (desires) in the very same way in which the perception of something pleasant or painful in animals can issue in basic perceptual desire. This, I think, shows how Aristotle is pursuing a bottom-up strategy here in explaining how mental acts (in the present case desires) are triggered and that he is not concerned with ethical claims about what things we ought to pursue or avoid. Note also the interesting parallel (and difference) between κατάφησιν ἢ ἀποφησιν καὶ φεύγει ἢ διώκει here in T 4 and the οἷον καταφᾶσα ἢ ἀποφᾶσα διώκει ἢ φεύγει in T 3 (431a9-10). Concerning the objects of desire, the thinking mind issues judgments in the form of affirmation and denial which then lead to desire, whereas in non-rational animals and presumably in humans when they act non-rationally, perception leads to desire only as if they were affirming or denying (thus implying that they are not actually affirming or denying). The basic point of T 4, however, seems to be the exact same as in the previous section, namely – and ultimately – that it is an actual object of perception that triggers the thought of objects of desire, and therewith also the corresponding desire. Aristotle’s thesis is that images (or, perhaps more precisely, the products of the exercise of capacity of imagination, φαντάσματα) can operate as causal substitutes for external triggering causes26. Their images stand in for their actual presence. Aristotle puts his emphasis on this role of φαντάσματα as causal substitutes for external objects of perception for practical thought (and not so much on the proper nature of practical thought, as he will return to a more detailed analysis of complex, i.e. combined, practical and non-practical thinking later in T 7 and T 8).

26

DA III 3 429a4-6, MA 701b17-23, cp. 702a5 sq., Phys. 247a8 sq., Mem. 452a1 sq.

DE ANIMA III 7

203

The triggering role of φαντάσματα as causal substitutes of the acts of perception that brought them about – and ultimately as substitutes of the external objects that brought about these perceptions – does not have to fully embody the content of the mental episode it triggers. Goodness or badness, after all, are not perceptible qualities27. And as we will see in sections T 7 – T 9, in mental acts the forms of the triggering causes and what these acts are about (their content) can come widely apart. It is for this reason that Aristotle insists that φαντάσματα are necessary but not sufficient for thought28. This is confirmed in T 7 where Aristotle is careful not to say that φαντάσματα are thoughts, or the content of thoughts, but that the thinking mind thinks intelligible forms in them (τὰ μὲν οὖν εἴδη τὸ νοητικὸν ἐν τοῖς φαντάσμασι νοεῖ, see also DA I 1, 403a8-10; III 8, 432a12-13). Having shown in T 4 how thinking about the objects of basic perceptual desire is – ultimately – triggered by actually existing external objects of perception as well, Aristotle makes a general point about the triggering of all episodes of human thought in the following section. T5

A claim about thinking generally (431a16-17)

διὸ οὐδέποτε [17] νοεῖ ἄνευ φαντάσματος ἡ ψυχή.29 Which is why the soul never thinks without an image.

Why is it that thinking never occurs without an image? This becomes apparent if we read the statement as an application of the Actuality Principle, i.e. if we add to the claim in T 4 that thought thinks the objects of basic pursuit and avoidance using φαντάσματα instead of αἰσθήματα the further claim that these φαντάσματα trigger such episodes of thought en lieu of actual external objects of perception (thus allowing Aristotle to extending the Actuality Principle, according to which every embodied mental act is triggered by an actual object, beyond practical thought to human thought generally): P 1 φαντάσματα act as triggering causes of human thought in lieu of actual objects of perception P 2 Every mental act requires an actual object as a triggering cause (Actuality Principle) 27

See, for instance, Politics I 2, 1253a10-18. Mem. 1, 449b31-450a1; DA I 1, 403a9; III 3, 427b15; III 7, 431a16 sq., b2; III 8, 432a8-10. 29 17 φαντάσματος o cf Th 113,20: φαντασίας Sil c267,30 cf Ph 560 NB: o sine E! 28

204

KLAUS CORCILIUS

C

Every act of human thought requires a φαντάσμα as its triggering cause (“The soul never thinks without φαντάσματα”)

T 4 establishes that the thinking of the objects of desire can take the place of the actual perception of such objects, and that this is made possible by the stored images (φαντάσματα) of these objects, which ultimately go back to the external objects themselves. Hence, desires and intentional motor-reactions are triggered not only by the perception of actually existing external objects immediately at hand but also by their causal substitutes, the φαντάσματα of these objects. In the case of mental representation of the basic objects of desire as they are relevant in T 3 and T 4, the triggering causes will to some large extent coincide with their content (as they seem to do 100 % in the case of basic perceptual desire and to lesser degrees in the cases of the thinking of simple objects of desire). This will change in T 7 – T 9, albeit without affecting the validity of the Actuality Principle (which is not a principle that concerns the full extent of the content of the mental acts in question: it suffices for the validity of the principle if the thought of F is triggered by something that is something, or some part, of F in actuality30). 4. Complex Perception (431a17-b1): An ambiguous case It is not clear that the passage in T 6 is a part of the overall argument in DA III 7. To start with the arguments contra, the section does not seem to be indispensable for the overall argument of the chapter (even though it certainly fits into the overall scheme), while it does seem to be dispensable from a grammatical perspective. T 4 and T 5 connect seamlessly with what follows in T 7 where Aristotle’s treatment of complex practical thought takes up T 4’s treatment of the simple thinking of the objects of desire. Also, one must admit that the main point of the passage, i.e. to explain how one and the same act of perception can be both numerically one and at the same time have a multitude of intra- and multimodal features, does not seem to depend on the Actuality Principle in any obvious way. T 6 might therefore be regarded an insertion by a later editor. However, on the above hypothesis that the chapter is concerned with the pervasive validity of the Actuality Principle, we can at least give good 30

See below, p. 216.

DE ANIMA III 7

205

reasons for why either Aristotle or perhaps a later editor thought it appropriate to include the passage in its present context. First of all, the first sentence of the passage may very well be understood as bearing directly on this issue, taking up things exactly at that point where they were left in the case of basic perception in T 2. T 6 describes in broad strokes how the movements of the media set up by perceptual objects in two different sense modalities (sight and sound) lead to further movements in the peripheral sense organs until they reach the final point of their journey (the “extreme”, ἔσχατον), which is the perceptual mean (called a “boundary”, ὅρος, in 431a22. Note that the expression αἰσθητικὴ μεσότης from T 3, 431a11, is taken up by μεσότης in a19, which seems an important connection). The claim is that the perceptual mean responsible for the reception of the input from both sense modalities is numerically one and the same, while the being of each modality remain distinct (up until 431a20). The relation of this claim to the Actuality Principle consists in that it shows how complex perceptual input by one and the same complex perceptual object can cause a likewise complex and yet unified act of perception. This is explained by the numerical unity of the perceptual mean, which unity it exhibits in spite of being one common mean of the heterogeneous sense modalities. This combination of numerical and analogical unity enables the perceptual mean to receive simultaneous multi-modal input and to perceive this input as a unified complex sense impression. It also enables it to discriminate one kind of perceptual input from the other31: T 6 ὥσπερ δὲ ὁ ἀὴρ τὴν κό[18]ρην τοιανδὶ ἐποίησεν, αὕτη δ᾽ ἕτερον, καὶ ἡ ἀκοὴ ὡσαύτως, [19] τὸ δὲ ἔσχατον ἕν, καὶ μία μεσότης, τὸ δ᾽ εἶναι αὐτῇ [20] πλείω. Τίνι δ᾽ ἐπικρίνει τί διαφέρει γλυκὺ καὶ θερμόν, εἴ[21]ρηται μὲν καὶ πρότερον, λεκτέον δὲ καὶ ὧδε. ἔστι γὰρ ἕν [22] τι, οὕτω δὲ ὡς ὁ ὅρος. Καὶ ταῦτα ἓν τῷ ἀνάλογον [23] καὶ τῷ ἀριθμῷ ὄν, ἔχει πρὸς ἑκάτερον, ὡς ἐκεῖνα πρὸς ἄλ[24]ληλα· τί γὰρ διαφέρει τὸ ἀπορεῖν πῶς τὰ μὴ ὁμογενῆ κρίνει [25] ἢ τὰ ἐναντία, οἷον λευκὸν καὶ μέλαν· ἔστω δὴ ὡς τὸ Α τὸ [26] λευκὸν πρὸς τὸ Β τὸ μέλαν, τὸ Γ πρὸς τὸ Δ [ὡς ἐκεῖνα [27] πρὸς ἄλληλα]· ὥστε καὶ ἐναλλάξ. Εἰ δὴ τὰ ΓΔ ἑνὶ εἴη [28] ὑπάρχοντα, 31 The question was raised originally in III 2, 426b22 sq. There, however, Aristotle emphasized the numerical unity and simultaneity of such acts of discrimination, whereas here in III 7 he emphasizes the analogical unity of the different sense modalities in both the perceptual object and the perceptual mean (λεκτέον δὲ καὶ ὧδε, up until 431b2).

206

KLAUS CORCILIUS

οὕτως ἕξει, ὥσπερ καὶ τὰ ΑΒ, τὸ αὐτὸ μὲν [29] καὶ ἕν, τὸ δ᾽ εἶναι οὐ τὸ αὐτό, κἀκεῖνο ὁμοίως. ὁ δ᾽ αὐτὸς [431b] λόγος καὶ εἰ τὸ μὲν Α τὸ γλυκὺ εἴη, τὸ δὲ Β τὸ λευ[2]κόν.32 Now, in the same way in which the air brought the pupil into a certain state, and the pupil something else, in that same way hearing [brings something else in a certain state, and that again something else], still the ultimate point is one and one mean, though its being is more than one. And though we have already spoken of that by which one discriminates the sweet and the hot as different things, the difference must also be specified in this way. For there is something one, as a boundary is such. And these (i.e. sweet and hot) since they are one by analogy and in number stand in relation each to the other as those stand to one another. For what is the difference between raising the difficulty of how one discriminates what is not of the same genus and (of how one discriminates) contraries (within the same genus), e.g. white and black? Let A, the white, then, stand in the same relation to B, the black as C stands in relation to D, so that (the sameness of the relation holds) also alternando. If, therefore, CD were to belong to one thing, the same would hold as with AB as well, i.e. they would be one and the same, while their being would not be the same, and similarly with respect to those (CA). The same account would hold if A were the sweet and B the white.

Basically, Aristotle here grounds the possibility of the perception of unified and complex simultaneous perceptual input and the possibility of the discrimination of the content of one sense modality from another on a structural parallel between the object and the subject of perception. Both object and subject of perception are one in number and many in being, which is to say that the object of perception and the perceptual 32 ὥσπερ o Sil c269,19.28 Phl p560,9 So 138,5: ὡς scripsit Förster 18 αὕτη LSUVXy Sic 269,20; 270,17 So 138,6: αὐτὴ CW Sil A 19 (post ἓν add φεύγει ἢ διώκει y) αὐτῇ LCWy Sil: om SUVX [ins Xx] 20 (τὸ γλυκὺ Sil A) 21 (καὶ1 om W) (καὶ2, om y) ὧδε LSUVX Sil: νῦν CWy, cf Sip271,2 Php560,20 22 ὡς ὁ SUV: καὶ ὁ CXy: ἡ στιγμὴ καὶ ὁ W, ins ἡ στιγμὴ mgC²: καὶ L: «ὥσπερ ὁ» Php560,25: «ὥσπερ καὶ ὁ» Sip271,6 ἓν CUWy Sil p271,14; 272,7 Phl p560,31: ἐν LSVX καὶ τῷ C cf Php560,26: ἢ τῶ Ly: ἢ UVWX Sil c271,14 Phl: om S 23 ὄν Sil c271,14: ὂν o Phl ἑκάτερον o–X Phl: ἑκάτερα X sscr W¹ Sil c271,15.38 ὡς o Phl: ᾗ Sil c271,15 24 (διαπορεῖν W Phl) (πῶς om X, ins X³) μὴ LSUVX Sil Phlp561,5 So138,25: om CWy Sip272,3.7.21 25 (ὡς ἐν τῶ y) A .. B .. etc] πρῶτον .. δεύτερον .. etc Phc561,9.10 26 πρὸς, LUWXy Phc561,9: καὶ CSV, del et ins πρὸς C² (πρὸς]2 ἀέρος X1 sscr πρὸς X³) ὡς...27 ἄλληλα del Christ 27 (ὥστε] οὕτως W) τὰ LCWy: om SUVX, ins U² ἑνὶ o–Cy: ἐν Cy 28 (οὕτως] καὶ C) (ἕξει ex ἔχει y¹) τὰ LW corr U²: τὸ CSUVXy 29 καὶ ἕν LCWy Phc561,18: καὶ S: om UVX δ᾽1 LCVWy Phc561,18: post εἶναι SU: om X (κακεῖνος C, ς eras) 431b 1 καὶ LCWy: κἂν SUVX Sic272,28 NB: o sine E!

DE ANIMA III 7

207

capacity are both one in number (one substance, one capacity) and also one in analogy (many and heterogeneous perceptual features, many and heterogeneous perceptual modalities). For the perceptual capacity, heterogeneous numerical unity enables the subject to discriminate between a plurality of sense modalities and between different inputs within one sense modality. Read in the light of the foregoing, T 6 shows that complex perceptual episodes are triggered by actual external perceptual objects. And since T 7, unlike T 4 and T 5, will talk about complex thought which involves a plurality of φαντάσματα, T 6 may very well be regarded a necessary step in the sequence of applications of the Actuality Principle to different, and increasingly complex, embodied mental acts: only once it has been shown how complex perceptual content is triggered by likewise complex external actual objects is Aristotle in a position to discuss the cases of complex objects of human thinking. All of this, it seems to me, makes good sense and shows how T 6 is part of the increasingly complex ascending sequence of mental states. However, it still seems true that grammatically T 6 does not fit naturally in its context and that we probably wouldn’t miss it if it weren’t in the place where the manuscripts have it. This raises the question whether T 6 is in its right place, and whether its transposition can remedy the awkwardness that results from the fact that T 4 and T 7 seem to connect seamlessly. Once we contemplate this option the two most likely scenarios are to put T 6 in the immediate sequel of the treatment of basic perceptual desire in T 333, or right after the discussion of basic perception in T 2. 5. Complex thought, practical and theoretical (431b2-10) T7

Complex practical thought (431b2-10)

τὰ μὲν οὖν εἴδη τὸ νοητικὸν ἐν τοῖς φαντάσμασι νοεῖ, [3] καὶ ὡς ἐν ἐκείνοις ὥρισται αὐτῷ τὸ διωκτὸν καὶ φευκτόν, [4] καὶ ἐκτὸς τῆς αἰσθήσεως, ὅταν ἐπὶ τῶν φαντασμάτων ᾖ, [5] κινεῖται. οἷον αἰσθανόμενος τὸν φρυκτὸν ὅτι πῦρ, τῇ κοινῇ [6] γνωρίζει, ὁρῶν κινούμενον, ὅτι πολέμιος· ὅτε δὲ τοῖς ἐν τῇ [7] ψυχῇ φαντάσμασιν ἢ νοήμασιν, ὥσπερ ὁρῶν, λογίζεται [8] καὶ βουλεύεται τὰ μέλλοντα πρὸς τὰ παρόντα· καὶ ὅταν [9] 33

I thank Stephen Menn for suggesting this to me.

208

KLAUS CORCILIUS

εἴπῃ ὡς ἐκεῖ τὸ ἡδὺ ἢ λυπηρόν, ἐνταῦθα φεύγει ἢ διώκει, [10] καὶ ὅλως ἐν πράξει.34 That which is capable of thinking thinks the forms in images, and in the same way in which in these [i.e. in the αἰσθήματα] what is to be pursued and what avoided are distinguished for it, so too outside of perception, when it is with the images, it is moved. For example, when it has perceived the beacon, namely that it is fire, and sees it moving with the common sense, it recognizes that it is hostile. Sometimes, on the basis of images or thoughts in the soul, as if one where seeing, one calculates and plans what to do on the basis of things which are present. And whenever one says that the sweet or the painful is there, then one avoids or pursues here, and [this holds] generally in action.

Here, two cases of practical thought are distinguished. Case 1: forms as they are thought with the aid of φαντάσματα of things can trigger desires. This may involve complex chains of φαντάσματα and even conventional signs. Case 2: the triggering of desires by deliberation and practical thought (also with the aid of φαντάσματα). Case 1: intelligible forms are thought quite generally in the φαντάσματα of things (see T 4 and T 5, which is taken up by T 7 directly). And since some of these forms are forms of things to be pursued or avoided, the objects of desire will be given to the thinking part of the soul by way of φαντάσματα as well. This has the consequence that the thinking creature can be moved to act on inconclusive perceptual evidence or even in the absence of any external perceptual object of pursuit or avoidance. The example of the perception of the beacon illustrates this very clearly35. What is actually seen is just a moving fire; 34 3 (ἐν om y) ὥρισται LCWy Sil: ὥριστο UVX corr S Sic273,35 (αὑτῶ C: αὐτὸ y, sscr ω y¹) 4 αἰσθήσεως L Sil: αἰσθήσεως ὂν CSUVX: αἰσθήσεως ὢν Wy (ἢ X) 5 (κινεῖται corr in κινῆται X²) φρυκτὸν o Sil c274,15 Phc561,31 Th 114,1Q³: φευκτὸν Th 114,1 vulg (κοινῆ] κόρη ut v X, -ει X²) 6 ὅτε LC: ὁτὲ b–C Sil Phl τοῖς ἐν o Phl: ἐν τοῖς ἐν Sil 8 βουλεύεται LCWXy Sil: βούλεται SUV 9 ἡδὺ ἢ λυπηρὸν b–W Sil: ἡδὺ ἢ τὸ λυπηρόν L: λυπηρὸν ἢ ἡδὺ W (καὶ ἐνταῦθα y) NB: o sine E! 35 Cf. Thucydides, II, 94: “ἐς δὲ τὰς Ἀθήνας φρυκτοί τε ᾔροντο πολέμιοι καὶ ἔκπληξις ἐγένετο οὐδεμιᾶς τῶν κατὰ τὸν πόλεμον ἐλάσσων”. And III, 22 of the beacon as being an emergency signal calling allies to act, only to be obfuscated: “φρυκτοί τε ᾔροντο ἐς τὰς Θήβας πολέμιοι: παρανῖσχον δὲ καὶ οἱ ἐκ τῆς πόλεως Πλαταιῆς ἀπὸ τοῦτείχους φρυκτοὺς πολλοὺς πρότερον παρεσκευασμένους ἐς αὐτὸ τοῦτο, ὅπως ἀσαφῆ τὰ σημεῖα τῆς φρυκτωρίας τοῖςπολεμίοις ᾖ καὶ μὴ βοηθοῖεν, ἄλλο τι νομίσαντες τὸ γιγνόμενον εἶναι ἢ τὸ ὄν, πρὶν σφῶνοἱ ἄνδρες οἱ ἐξιόντες διαφύγοιεν καὶ τοῦ ἀσφαλοῦς ἀντιλάβοιντο.”

DE ANIMA III 7

209

however, in virtue of possessing stored φαντάσματα of what is to be pursued and avoided, the thinking animal can be moved to action even (καὶ) outside the direct causal influence of external objects of perception (ἐκτὸς τῆς αἰσθήσεως). This happens when the agent combines (or associates) in thought the forms of what is to be pursued or avoided with that inconclusive perceptual evidence (i.e. the moving fire). Since the moving fire is a conventional sign, the complexity of the mental act involved goes far beyond the mere representation of objects of perceptual desire, and also beyond the capabilities of non-intelligent animals. Still, the Actuality Principle holds. This is because of the causal substitute function of φαντάσματα (see T 4), i.e. because ultimately φαντάσματα go back to actually existing external perceptual objects. Thus, all mental acts of the given complex and seemingly complex kind are ultimately triggered by actually existing external objects. Case 2 goes beyond the mere combination of φαντάσματα of things to be pursued or avoided with actual perception in judgments. It describes how the intellect can move the agent not only on the basis of the representation of certain objects of pursuit and avoidance, which then are pursued or avoided (as in Case 1), but how the agent can, on the basis of such desires, rationally process a plurality of such objects by way of a preference calculus (minimal deliberation or “practical thinking” in what follows). Aristotle describes this kind of practical thinking by way of its minimal achievement: calculation viz. deliberation (λογίζεται καὶ βουλεύεται) happens minimally whenever the goodness or badness of some x is compared to the goodness of badness of some other object, y, where x(s) are objects of pursuit and avoidance immediately present to the agent by way of perception (παρόντα) and y(s) are future objects of pursuit and avoidance (τὰ μέλλοντα) present in her mind in the form of images or thoughts (ἐν τῇ ψυχῇ φαντάσμασιν ἢ νοήμασιν). Aristotle is here – as he often is in De Anima – interested in the simplest possible scenario36. He thus does not consider cases in which an agent deliberates about the relative goodness of x and y where both are objects of pursuit and avoidance present to perception here and now. And perhaps it is not necessary for him to consider such cases here, given that even then the calculation of which of the two is the greater good will involve some relation to the future; for otherwise such calculations will not need to 36

See n. 39 below.

210

KLAUS CORCILIUS

involve any deliberative effort, but only perception and desire. Typically, and minimally, the capacity of practical deliberation manifests itself in cases where some present good is weighed with some future good (abstention etc.). The agent who is capable of practical deliberation will then act on the basis of that calculus: “And whenever one says that the sweet or the painful is there (ἐκεῖ, i.e. in the future) then one avoids or pursues here (ἐνταῦθα, i.e. in the present, 431b8-9).” This is how on Aristotle’s theory we minimally come to be guided by our rational calculations37. Note that in b3 Aristotle is freely using ἡδὺ ἢ λυπηρόν instead of τὸ διωκτὸν καὶ φευκτόν. He does this, I suggest, because it is his general strategy in this chapter to explain the triggering of the different mental acts he investigates by way of considering minimal accounts of these acts. If this is correct, the type of goods practical deliberation is minimally concerned with will be objects of pursuit and avoidance such as they are relevant for basic perceptual desire in T 3, i.e. perceptually good and bad things. The main difference between Case 1 and Case 2 is firstly that while the former describes how the thought of an object of desire can lead to desire (cf. DA III 10, 433b10-13; DMA 6, 700b17-19; 7, 701b33-702a7), the latter shows how practical thought can refine such desires by subjecting them to a comparative (preference) calculus. This preference calculus of course presupposes the desire for the object to be pursued described in Case 1 (cf., for instance, DA III 10, 433a14-16; EN III 5, 1112b15-20). The second major difference is that the content of the mental acts described in Case 2 go significantly beyond perceptually good and bad things, including things in the future, as they are described in Case 1. Future things that are better than things present do not, and cannot, correspond to any given actual perceptual content (“better than” is a relation, 37 Another viable interpretation of the scenario in Case 2 has been suggested to me by Gabriel Richardson Lear (for which I would like to thank her warmly). She suggests that the scenario shows something slightly different, namely how an agent can be prompted to think about how to alter her current circumstances simply by imagining or thinking about other desirable circumstances without engaging in a comparative assessment. This is interesting and certainly not a possibility that one would like to rule out on Aristotle’s behalf. I still think, however, that this possibility can be taken care of by the cognitive machinery presented in Case 1, even if Aristotle doesn’t explicitly say so. A further motive is that the formulation λογίζεται καὶ βουλεύεται τὰ μέλλοντα πρὸς τὰ παρόντα in 431b7-8, even though compatible with her suggestion, seems to me to suggest some kind of calculus.

DE ANIMA III 7

211

see below). The agent in Case 2 acts mainly and decisively on the evidence provided by his rational preference calculus, perhaps supplemented by the inconclusive perceptual evidences from images of perceptually good and bad things as described in Case 1. This brings with it an additional emancipation from the direct causal influence of external objects, even though it remains true that it is them who ultimately trigger perception and the ensuing motor-reactions. Note that φαντάσματα now not only act as causal substitutes of external perceptual objects in thought, but also as representations of the proper content of thought (“as if one where seeing”: ὥσπερ ὁρῶν, b7, which, in contrast to Case 1, implies that one is not actually seeing or perceiving). This shows how rational human calculation can issue in motor reactions via the causal substitute function of φαντάσματα. From now on φαντάσματα are both: causal substitutes for the actual objects that trigger the mental episode in question (deliberation and the issuing motor reaction) and mental representations, not of the perceptual objects that originally brought them about (as in the previous cases), but of practical rational content. Presumably, this practical rational content will consist minimally in relative judgments of the form “y is a greater good than x” (which is a relation and thus does not correspond to anything perceptible). In a next step, Aristotle applies the idea of Case 2 to human action generally (καὶ ὅλως ἐν πράξει, 431b10)38. This makes excellent sense. It seems that we enter the sphere distinctive of human action, which is the one guided by contents of thought, only once we reach the capacity to rationally process the practical information we possess, and to let our actions be guided by it. And this begins – at least ontogenetically – at the point where our awareness of the “greater than” – relation between two perceptual goods determines our desires39. How does this relate to the Actuality Principle? In both, Case 1 and Case 2, what triggers the respective mental episode is a φάντασμα, of 38 David Ross changes ὅλως ἐν πράξει (“and generally in action”) to ὅλως ἓν πράξει (“and in general will do some one thing”, trans. Shields). However, that some one thing is done holds also for the previous cases of motor reactions. The point rather seems to be that with Case 2 we have arrived at what is distinctive of human action. Also, the term πράξεως in 431a10 seems to take up the noun πράξει in the very same line (and not, as Ross has it, the future verbal expression). See above. 39 Cf. the minimal account of deliberative φαντασία in III 11, 434a5-12, where Aristotle describes deliberation as applying a common standard to two different goods, arguing that this is a minimal achievement of calculation, λογισμοῦ ἤδη ἐστὶν ἔργον, 434a8.

212

KLAUS CORCILIUS

which Aristotle has shown in T 4 that it functions as a substitute for actual external objects of desire. However, I should repeat here that in both cases, gradually, the forms of the triggering causes of the mental episodes and what they are about come apart (see above, p. 203): while Case 1 is partially about future goods, Case 2 is specifically about the contents of human rational deliberation and therefore describes the minimal cognitive achievement required for the sphere of human action. The Actuality Principle holds in both cases. Now, in a further and rather bold step, Aristotle applies the foregoing to complex non-practical thought (the simple thinking of forms and thoughts without combination was already taken care of in T 4 and T 5, where it was said that the objects of the thinking soul are given to it by way of φαντάσματα): T 8 Complex non-practical thinking (431b10-12) καὶ τὸ ἄνευ δὲ πράξεως, τὸ ἀληθὲς [11] καὶ τὸ ψεῦδος, ἐν τῷ αὐτῷ γένει ἐστὶ τῷ ἀγαθῷ καὶ τῷ κα[12]κῷ· ἀλλὰ τῷ γε ἁπλῶς διαφέρει καὶ τινί.40 And also what is outside the sphere of action, the true and the false41, are in the same genus as the good and the bad, though to be sure they differ by being unqualifiedly [scil. true or false] in the one case and relative to someone [scil. good or bad] in the other.42

How are we to understand the difference and sameness between practical and non-practical thinking? Aristotle begins by stating that nonpractical thought falls in the same genus as practical calculation and deliberation. Presumably, what he has in mind is that non-practical thought consists in some kind of calculation like the one described in T 7 but is somehow ἄνευ πράξεως. (I refrain from using the expression “theoretical thinking”, since theoretical thought seems to involve further 10 τὸ2] καὶ τὸ y SilA 11 (τὸ om L) καὶ2 CSVy Sil c275,22: καὶ τῷ LUX: ἢ τῷ 12 τῷ SUVXy Sil: τὸ LCW γε L Sil: om b 41 Here, I depart from Shields’ translation: “And truth or falsity outside the sphere of action”, since it implies that there is truth and falsity in the kinds of practical thinking he just discussed in the previous sections. But there is no indication for this in the text. Also, what Aristotle says about practical truth in his ethical writings (EN VI / EE V, 1139a2227) does not match any of the aforementioned forms of practical thinking. The concept of practical truth describes a specifically ethical form of concordance of right desire with the true thought. 42 Again, I depart from Shields’ translation (“they differ in that in the one case they are to be pursued or avoided unqualifiedly and in the other relative to a person.”). See below. 40

W

DE ANIMA III 7

213

conditions for Aristotle which go much further than the mere assertion or denial in judgments about how things are.) However, he is quick to add to this somewhat astounding statement the important qualification that the true and the false are so without qualification, whereas the practically good and bad are so only relative to a given person. This last bit, to be sure, refers only to the kinds of practical thinking as they are discussed in the previous sections, namely the basic workings of practical thought and not to the optimal form that it takes in ethical thought. So much for the difference between these two kinds of thinking. But in what sense precisely are they supposed to fall under the same genus? As just mentioned, it would seem their generic sameness consists in the fact that both, practical and non-practical thought alike, work on the basis of a cognitive mechanism that is fundamentally the same. In T 7’s Case 2 complex practical thought was described as a rationally guided motor reaction in which objects of desire, as they are given in perception and φαντάσματα, are subjected to a preference calculus that evaluates present objects of desire in relation to future objects of desire (τὰ μέλλοντα πρὸς τὰ παρόντα, 431b8). So, both the objects of pursuit and avoidance and the result of the preference calculus are relative to the agent; the objects are things good and bad for the agent, and the calculus issues in a judgment about the relative preferability of present and future goods for the agent as well. If we now subtract the relativity from this procedure, we get a calculus in which non-practical objects as they are given to the thinking mind by way of φαντάσματα – these are objects which are not, or not primarily, objects of pursuit and avoidance for the thinking subject – are compared to other such non-practical objects; the outcome of this would be comparative judgments, in which identities and differences between the compared items are either affirmed or denied. The idea is that non-practical complex thought is a motor process very much like complex practical thought, only that in this case the comparative calculus is driven by the desire for the truth, instead of by the desire for one’s own relative benefit. As far as the comparative account of nonpractical thinking is concerned, Aristotle here might very well have picked up on Plato’s comparative account of human practical and theoretical thinking in the Theaetetus43. To Plato’s, Aristotle adds the idea 43 Καὶ τούτων μοι δοκεῖ ἐν τοῖς μάλιστα πρὸς ἄλληλα σκοπεῖσθαι τὴν οὐσίαν, ἀναλογιζομένη ἐν ἑαυτῇ τὰ γεγονότα καὶ τὰ παρόντα πρὸς τὰ μέλλοντα. ; “(…) in

214

KLAUS CORCILIUS

that such complex thinking is a motor reaction that involves desire and that is triggered by φαντάσματα (whereas Plato seems to regard thinking immediately as a certain kind of motion). In any case, with the account of complex practical and non-practical thinking given in T 7 and T 8, Aristotle has provided an account of the triggering (not the content) of such mental acts by φαντάσματα and thus – indirectly – by actually existing external objects. The Actuality Principle holds. 6. The thinking of abstract objects and a final question Now Aristotle moves on to the thought of abstract objects (431b12-19). T 9 τὰ δὲ ἐν ἀφαι[13]ρέσει λεγόμενα νοεῖ ὥσπερ ἂν εἰ τὸ σιμὸν, [14] ᾗ μὲν σιμὸν, οὐ κεχωρισμένως, ᾗ δὲ κοῖλον, εἴ τις ἐνόει ἐνεργείᾳ, ἄνευ [15] τῆς σαρκὸς ἂν ἐνόει, ἐν ᾗ τὸ κοῖλον· οὕτω τὰ μαθηματικά [16] οὐ κεχωρισμένα ὡς κεχωρισμένα νοεῖ, ὅταν νοῇ ἐκεῖνα. ὅλως [17] δὲ ὁ νοῦς ἐστίν ὁ κατ᾽ ἐνέργειαν τὰ πράγματα [νοῶν]. ἆρα [18] δ᾽ ἐνδέχεται τῶν κεχωρισμένων τι νοεῖν ὄντα αὐτὸν μὴ κε[19]χωρισμένον μεγέθους, ἢ οὔ, σκεπτέον ὕστερον.44

these [i.e. the concepts of beauty, ugliness, goodness, and badness as opposed to], above all, I think the soul examines the being they have as compared with one another by making a calculation within itself of past and present in relation to the future.” (Theaetetus 186a10-b1, translated by M.J. Levett, revised by M. Burnyeat, modified). Τὴν δέ γε οὐσίαν καὶ ὅτι ἐστὸν καὶ τὴν ἐναντιότητα πρὸς ἀλλήλω καὶ τὴν οὐσίαν αὖ τῆς ἐναντιότητος αὐτὴ ἡ ψυχὴ ἐπανιοῦσα καὶ συμβάλλουσα πρὸς ἄλληλα κρίνειν πειρᾶται ἡμῖν ; “But as regards their being [i.e. the being of hardness and softness as opposed to their perception] – the fact that they both are and also their opposition to one another, and the being, again, of this opposition, the soul itself attempts to judge them for us by going over them [viz. the hard and soft sensory objects], and comparing them in relation to one another.” (Theaetetus, 186b6-9, translated by M.J. Levett, revised by M. Burnyeat, modified). 44 12 ἐν LWy Sil c276,16 Phl bis, c566,10.12: om CSUVX Th 114,10 13 ἂν LCWy Sil Phl bis: om SUVX cf Php566,14 (εἰ om y) (σιμόν2 om L: σιμότης X) 14 (κεχωρισμένον W) ᾗ δὲ κοῖλον o–Xy SilA cf Th 114,17: ἧ δὲ κοῖδρη (sic) y: εἰ δὲ καμπύλον X: ᾗ δὲ καμπύλον Sip278,21 Phl p568,27 εἴ τις CSUVW Sip278,26: εἴ τι Ly Sic278,24Aa: om X: εἴ τι γε Sil A ἐνόει o–SV Sil: ἐννοεῖ SV Sic278,24Aa 15 (ἐννοῆ X) ἐν ᾗ LCWXy Sil c278,28: om SUV 16 οὐ κεχωρισμένα CUVy Sil c278,32 Phc566,17: οὐ κεχωρισμένα τῇ ὑποστάσει LW, mg C² cf Th 114,21: οὐ κεχωρισμένω S: τῶν οὐ κεχωρισμένων X: οὐ κεχωρισμένως Siv278,33 ὡς κεχωρισμένα SUVWXy [sscr -ω y¹] Sil c278,33 Phc566,17 cf Th 114,22: ὡσεὶ κεχωρισμένω C [corr in -α C²]: om L SilA: ὡς κεχωρισμένως Siv278,33 (νοεῖ] ἐνόει L) (ὅτε νοεῖ W) (ἐκεῖνος X) 17 νοῶν CSVWy Sil p279,9: om a UX Phl p566,24; 563,5; c37,27 18 ὄντα αὐτὸν a: αὐτὸν ὄντα b Sil Phl 19 (κεχωρισμεγεθους E, corr ins -μένον E²) (ὕστερον σκεπτέον y) NB: o sine E!

DE ANIMA III 7

215

One thinks the things called “in abstraction” just as if [one thinks] the snub, which while [one can] not [think it] separately [from its matter] insofar as it is snub, yet if one were to think it in actuality insofar as it is concave, one would think it without the flesh in which the concave is: in this way one thinks mathematical things, which one thinks as separate entities, though not existing as separate entities, whenever one thinks them. And in general, the capacity of thought, if actual, is the same as things thought45. We must consider later whether or not it is possible for one who is not oneself separated from magnitude to think something belonging to the class of things which are separated.

Like in the previous section, we must bring to bear in our interpretation the previous text, which, among other things, explained how φαντάσματα are necessary for complex (propositional) practical and theoretical thinking. T 9 explains how external objects viz. their φαντάσματα, if considered in certain ways, give rise to, i.e. trigger, episodes of simple thought of abstract mathematical objects. Thus, a φάντασμα of the snub can trigger our thought of the object “concavity” (τὸ κοῖλον). This happens if we think of the snub (which is concavity in nose matter), in abstraction / isolation from the matter it is in. The doctrine that such abstract objects are not, but nevertheless considered as, separate by thought is familiar from passages like DA I 1, 403b14-15, A.Po. I 13, 79a7-10, and Met. M 3. What is new here is the specifically psychological perspective under which Aristotle describes how we come to think them, namely by performing certain mental operations on our perceptions of external objects viz. on their φαντάσματα. The point is that for a thinker to think about the mathematical abstract object “concavity” is to perform a certain mental operation on her φάντασμα of a hylomorphic compound (the snub). Hence, φαντάσματα, and therewith indirectly actually existing external objects, set up the process that leads to mental episodes of abstract mathematical thought. Again, the Actuality Principle holds also in this case46. A final question. It remains to see how the last sentence of the chapter relates to the chapter’s overall goal. The first thing to note is that it 45 Again, I do not follow Shields’ translation. However, this time only for stylistic reasons: my translation aims at a closer rendering of the Greek construction. 46 Even though T 9 only seems to talk of mathematicals, the point seems generalizable. All abstractions, not only mathematical ones, can be subjected to this kind of analysis (cf. Aristotle’s degree scale account of abstraction in Phys. II 2).

216

KLAUS CORCILIUS

seems that we do not possess the text to which Aristotle’s forward reference refers (“we must consider later …”). The most promising candidate seems to be a previous passage in the DA, namely the very difficult section in III 4, 429b10-22, which ends with the statement that there is a direct parallel between the separability of things from matter with their corresponding modes of cognition: “And generally, therefore, as things are separable from matter so also the things in regard of thought.” (ὅλως ἄρα ὡς χωριστὰ τὰ πράγματα τῆς ὕλης, οὕτω καὶ τὰ περὶ τὸν νοῦν47). Although this passage does not explicitly take up our question in T 9, it strongly suggests a negative answer to it: it is not possible to think separate objects without something that is not itself separated48. But there is no room to fully engage with that question here. Our question is how the final section of T 9 relates to the previous sections. Let me start by recalling the fact that DA III 7 pursues a bottom-up strategy. There can be no doubt that the chapter describes an ascending sequence of increasingly complex cognitive and motor states (with or without T 6), and, if I am right, it also shows how each of these states is triggered by something that is – in some way – in actuality what these states are about. However, as far as intellectual achievements are concerned, the chapter has so far described only the triggering of cognitive and motor episodes in human minds that operate on the basis of perception and / or φαντάσματα, from which they extract their specific intellectual information (in different ways, either by practical deliberation or by making judgments about how things are and relate to each other). This culminates in the thinking of abstract objects, in which thinking is not exactly about extended magnitudes anymore, even though it is triggered by, and extracted from, certain aspects of them. In this way, all forms of cognition discussed so far are μὴ κεχωρισμένον μεγέθους: they are both caused by, and concerned with, external things, their relations, and their various aspects and properties. So, due to the fact that they all in one way or the other originate in perceptibles, it would be correct to say that for Aristotle all forms of cognition discussed so far in DA III 7 involve

47

This rendering differs from Shields’ translation. Note that the claim that something separate in νοῦς is required for the thinking of separate objects is compatible with the idea that it is impossible to think things outside of time without φαντασία (as some sort of enabling condition or triggering cause) in Mem. 450a7-11. 48

DE ANIMA III 7

217

magnitude (matter) both on the side of the cognizing mind and on the side of its objects. Now, if this is right and the triggering cause for all these magnitudeinvolving mental episodes is ultimately an external object, and the thoughts thereby triggered are in one way or another (however directly or remotely) concerned with features of external objects, then we have a story to tell about how the triggering cause and its cognitive effect relate: there is always something in the objects viz. their φαντάσματα that is – in whatever idiosyncratic ways – either identical to (actual knowledge in T 1, perception in T 2, and perhaps in complex perception in T 6 as well), or a part, or otherwise an aspect, of, the content of the cognition and the motor process it triggers. On that basis we can see how triggering cause and the mental episode it triggers positively connect: X triggers Y to produce thought Z because there is something in X (however remote that might be) that connects to what Z is about49. The Actuality Principle as validated so far showed how there is a way for X to provide some sort of contact with what Z is about. Insofar as this is the case, the application of the Actuality Principle to embodied cognition shows how it is somehow the πράγματα themselves that trigger their being thought about by the cognizing mind (the actual thought of them being δυνάμει in the mind50). This whole model, however, does not seem to apply, at least not in any clear and straightforward way, to the case of the thinking of objects entirely separate from magnitude. What perceptible features correspond to objects that are entirely separate from magnitude, in a way similar to the way in which external things and their φαντάσματα correspond to the mental states that are in one way or the other about them (as e.g. similarity, parthood, membership, etc.) or their features? This, I suggest, is the worry that underlies Aristotle’s question here in T 9, namely whether it is possible for a thinker whose mind is not separated from 49 The meaning of ἐν is of course unclear here; however, especially the contexts of desire and practical thought leave no doubt that ἐν cannot plausibly mean that the objects of thought are straightforwardly contained in perceptible forms. For if that were the case φαντάσματα would literally contain both their desirability and their intelligibles features, and this would make them sufficient to present the soul with the proper objects of thought. But this is denied in III 8, 432a12-13. Thus, φαντάσματα are not the proper objects, but the triggering causes, of thought episodes, and they are so because there is something “in” them that connects the thinking mind with its proper object. 50 See n. 15 above.

218

KLAUS CORCILIUS

magnitude to think something that it entirely separated from magnitude. The question then is how something not separate from magnitude can have contact with, and thereby be prompted to think about, something entirely separate from magnitude. And this, I submit, is a natural question for Aristotle to ask at this point51. 7. Conclusion I hope to have shown how DA III 7 can be given a unitary and coherent reading if it is understood as the application of the Actuality Principle to embodied mental episodes in an ascending series of complexity ranging from basic perception to the thought of abstract mathematical objects (but stopping short of the thought of separate substances). The question whether the discussion of complex perception in T 6 is an integral part of the otherwise remarkably coherent and unitary argument of the chapter remains an open question. But I hope to have shown that it is at least possible and not implausible to understand T 6 as a part of this argument. In any case, if the above hypothesis is broadly on the right track, then the validation of the Actuality Principle as carried out in DA III 7 shows how even such difficult cases as the seemingly radically spontaneous self-motions of animals and humans and even more so episodes of human thinking ultimately depend on actual external objects as their (relatively to them) unmoved triggering causes. The basic idea of the chapter then would be that even in these complex and indirect cases of causation of mental episodes the cognition of X is initiated by some kind of contact with an actual X, even if that actual X may be only very imperfectly related to what the proper thought of X is about. Certainly, X does not have to be formally identical in all respects to the thought it triggers. Still, the idea that the triggering cause establishes contact between an actual X and its thought would be preserved (and cast doubt upon as a possibility in the case of the thinking of objects entirely unrelated to magnitude). But there are several noteworthy philosophical upshots from the discussion of the text apart from that main line of argument. Perhaps most noteworthy is that Aristotle does seems to entertain the view that at least much of human discursive practical 51 Why doesn’t he answer the question here? One possible answer to this is that such cases of intellectual contact with objects separate from magnitude arguably would not fall under the scope of his theory of the soul.

DE ANIMA III 7

219

and theoretical thinking consists in some sort of motor-processes (T 7 and T 8). But there is no room here for a full engagement with the rich philosophical content of the chapter52.

52 I would like to express my thanks to Stephen Menn and Gabriel Richardson Lear for written comments on previous drafts of this chapter. I have benefitted from them. I also would like to thank Yue Lu and all participants of the discussion in Lille for their remarks and questions.

L’ÂME COMME LA MAIN : TRADUCTION ET COMMENTAIRE DU CHAPITRE III 8 Michel CRUBELLIER ἡ δὲ ψυχὴ ὑπάρχει τοιαύτη οὖσα οἵα δύνασθαι πάσχειν τοῦτο. Ce processus est rendu possible par la structure de l’âme.1

1. Traduction (à partir de III 7, 431b16) [431b16] Ὅλως δὲ ὁ νοῦς ἐστιν ὁ κατ’ ἐνέργειαν τὰ πράγματα. ἆρα δ’ ἐνδέχεται τῶν κεχωρισμένων τι νοεῖν ὄντα αὐτὸν μὴ κεχωρισμένον μεγέθους, ἢ οὔ, σκεπτέον ὕστερον. En somme, l’intelligence en acte est ses objets. Mais est-ce qu’il lui est possible ou impossible, tout en n’étant pas elle-même séparée de la grandeur, de penser quelque chose des réalités séparées ? Il faudra examiner cette question par la suite. [Chapitre 8] Νῦν δέ, περὶ ψυχῆς τὰ λεχθέντα συγκεφαλαιώσαντες, εἴπωμεν πάλιν ὅτι ἡ ψυχὴ τὰ ὄντα πώς ἐστι πάντα· ἢ γὰρ αἰσθητὰ τὰ ὄντα ἢ νοητά, ἔστι δ’ ἡ ἐπιστήμη μὲν τὰ ἐπιστητά πως, ἡ δ’ αἴσθησις τὰ αἰσθητά· πῶς δὲ τοῦτο, δεῖ ζητεῖν. [20] Pour l’instant, disons encore, en commençant par récapituler à propos de l’âme ce qui a été dit , que l’âme est d’une certaine façon toutes choses : en effet, les étants sont ou bien sensibles ou bien intelligibles, et la science est d’une certaine façon les objets de science et la perception est les objets sensibles. Mais de quelle façon, c’est ce qu’il faut rechercher. Tέμνεται οὖν ἡ ἐπιστήμη καὶ ἡ αἵσθησις εἰς τὰ πράγματα, ἡ μὲν δυνάμει εἰς τὰ δυνάμει, ἡ δ’ ἐντελεχείᾳ εἰς τὰ ἐντελεχίᾳ. τῆς δὲ ψυχῆς τὸ 1

A.Po. II 19, 100a13-14.

222

MICHEL CRUBELLIER

αἰσθητικὸν καὶ τὸ ἐπιστημονικὸν δυνάμει ταῦτά ἐστι, τὸ μὲν ἐπιστητὸν τὸ δὲ αἰσθητόν. Or donc, la science et la perception se divisent en leurs [25] objets, celle qui est en puissance en ses objets en puissance et celle qui est en acte en ses objets en acte ; et les capacités de percevoir et de connaître scientifiquement qui sont dans l’âme sont en puissance cela, , respectivement, ce qui est connaissable par la science et ce qui est sensible2. Ἀνάγκη δ’ ἢ αὐτὰ ἢ τὰ εἴδη εἶναι· αὐτὰ μὲν δὴ οὔ· οὐ γὰρ ὁ λίθος ἐν τῇ ψυχῇ, ἀλλὰ τὸ εἶδος· ὥστε ἡ ψυχὴ ὥσπερ ἡ χείρ ἐστιν· καὶ γὰρ ἡ χεὶρ ὄργανόν ἐστιν ὀργάνων, καὶ ὁ νοῦς εἶδος εἰδῶν καὶ ἡ αἴσθησις εἶδος αἰσθητῶν. Or il faut nécessairement que ce soit ou bien ces choses elles-mêmes ou bien leurs formes. Les choses elles-mêmes, certainement pas : car il n’y a pas de pierre dans l’âme, mais il y a sa forme. [432a] De sorte que l’âme est comparable à la main : en effet, la main est l’instrument d’instruments, cependant que l’intelligence est une forme de formes et la perception une forme de sensibles. Ἐπεὶ δὲ οὐδὲ πρᾶγμα οὐθὲν ἔστι παρὰ τὰ μεγέθη, ὡς δοκεῖ, τὰ αἰσθητὰ κεχωρισμένον, ἐν τοῖς εἴδεσι τοῖς αἰσθητοῖς τὰ νοητά ἐστι, τά τε ἐν ἀφαιρέσει λεγόμενα καὶ ὅσα τῶν αἰσθητῶν ἕξεις καὶ πάθη. Mais puisqu’en fait il n’existe, à côté des grandeurs (des grandeurs sensibles, selon l’opinion commune), aucun objet doté d’une existence indépendante, les [5] intelligibles sont dans les formes sensibles, aussi bien ceux qui sont définis par abstraction que tous ceux qui sont des manières d’être ou des propriétés des sensibles. Καὶ διὰ τοῦτο οὔτε μὴ αἰσθανόμενος μηθὲν οὐθὲν ἂν μάθοι οὐδὲ ξυνείη· ὅταν τε θεωρῇ, ἀνάγκη ἅμα φάντασμά τι θεωρεῖν· τὰ γὰρ φαντάσματα ὥσπερ αἰσθήματά ἐστι, πλὴν ἄνευ ὕλης. ἔστι δ’ ἡ φαντασία ἕτερον φάσεως καὶ ἀποφάσεως· συμπλοκὴ γὰρ νοημάτων ἐστὶ τὸ ἀληθὲς ἢ ψεῦδος. τὰ δὲ πρῶτα νοήματα τί διοίσει τοῦ μὴ φαντάσματα εἶναι; ἢ οὐδὲ τἆλλα φαντάσματα, ἀλλ’ οὐκ ἄνευ φαντασμάτων. Et c’est pour cela que quelqu’un qui ne percevrait rien ne pourrait rien apprendre ni rien comprendre ; et que, lorsqu’on réfléchit, il faut nécessairement que l’on fixe en même temps son attention sur une image ; en effet les images sont comme des perceptions, [10] avec cette différence qu’elles 2 Le texte des lignes b25-28 est mal assuré, mais les variantes ne me paraissent pas affecter le sens et la portée de l’argument (voir la note 22). J’ai suivi le texte édité par Förster : τέμνεται οὖν ἡ ἐπιστήμη καὶ ἡ αἵσθησις εἰς τὰ πράγματα, ἡ μὲν δυνάμει εἰς τὰ δυνάμει, ἡ δ’ ἐντελεχείᾳ εἰς τὰ ἐντελεχίᾳ. τῆς δὲ ψυχῆς τὸ αἰσθητικὸν καὶ τὸ ἐπιστημονικὸν δυνάμει ταῦτά ἐστι, τὸ μὲν ἐπιστητὸν τὸ δὲ αἰσθητόν.

L’ÂME COMME LA MAIN: DE ANIMA III 8

223

sont sans une matière. Pourtant, l’acte de former une image est autre chose que celui de déclarer ou de nier quelque chose, car le vrai et le faux consistent dans une combinaison de termes intelligibles. – Mais les premiers termes intelligibles, en quoi se distingueront-ils de ce qui n’est qu’une image ? – À moins de dire qu’aucun terme intelligible, en général, n’est une image, mais qu’ils ne vont pas sans images.

2. Commentaire 1. Le chapitre est assez bien délimité. La fin est nettement marquée, puisqu’on passe de l’étude des facultés de connaissance à celle des actions dirigées vers une fin, avec une transition explicite (les premières lignes du chapitre 9, 432a15-18). C’est moins évident en ce qui concerne le début, en particulier parce que le plan de la séquence qui précède n’est pas vraiment apparent ; la plupart des commentateurs considèrent le chapitre 7 comme une succession de remarques discontinues. Mais la première phrase de 8 annonce clairement une thèse : l’âme est d’une certaine façon toutes choses (431b21). Cette thèse est précisée, justifiée et développée à travers une argumentation cohérente et suivie jusqu’en 432a3. Il semble y avoir une discontinuité à cet endroit3 et la seconde partie du chapitre est consacrée à une autre thèse : les intelligibles sont dans les formes sensibles (432a4-5). Il est vrai que cette seconde thèse n’est pas entièrement étrangère à la première, puisqu’un trait remarquable de la première partie est que l’argumentation s’y déroule constamment (mais de façon parallèle) aux deux niveaux de la perception et de l’intelligence. De ce fait, la thèse de la seconde partie, en posant une relation substantielle entre l’intelligible et le sensible, apparaît comme un complément important à la première et donne sa pleine force à la thèse que « l’âme est d’une certaine façon toutes choses  » – c’est-à-dire à la fois les intelligibles et les sensibles. Mais il n’y a dans le texte aucun élément qui l’indique expressément : pas de transition explicite entre les deux parties (simplement reliées par un δέ), ni de conclusion qui ferait la synthèse, de sorte que l’interprétation que je défendrai ici garde un caractère conjectural. À vrai dire, les commentateurs ne se sont pas beaucoup préoccupés de chercher de la continuité et de la consécution dans ce chapitre, qu’ils 3 Par ailleurs, la jolie comparaison avec la main pourrait fournir une de ces conclusions littéraires qu’Aristote ne déteste pas. Dans ce cas, il faudrait se représenter les lignes 432a3-14 comme une sorte de note, un rappel ou une mise au point.

224

MICHEL CRUBELLIER

traitent en général comme une liste de propositions doctrinales acquises (ils y sont peut-être incités par l’annonce d’un « résumé » dans la première phrase). Il est vrai que, vu comme une liste de thèses, le passage pose moins de problèmes ; mais il est aussi moins instructif. Dans les pages qui suivent, je m’efforcerai au contraire de faire apparaître un ordre argumentatif. 2. La transition au début du chapitre III 94 présente le chapitre 8 comme la fin d’une section consacrée aux actes qui manifestent une capacité de « discernement » (τὸ κριτικόν), soit principalement la perception et la pensée réfléchie (διάνοια) ou l’intelligence (νοῦς). Cette analyse des actes de discernement a commencé de fait au chapitre II 5 ; mais il est à remarquer qu’à cet endroit elle n’a pas été présentée comme telle, puisqu’Aristote y annonçait seulement une étude de la perception. Ce n’est qu’au chapitre III 3 qu’il évoque la possibilité de regrouper sous une description commune la perception et les actes de connaissance intellectuelle ; encore ne s’agit-il que de légitimer la comparaison entre les deux : « on estime que comprendre (νοεῖν) et réfléchir (φρονεῖν) sont comme percevoir quelque chose, car dans l’un comme dans l’autre cas l’âme discerne tel des étants et en prend connaissance » (427a1921)5. Cette comparaison sert ensuite de fil conducteur au commencement de l’étude de l’intelligence (au chapitre 4), mais, même alors, elle n’a que le statut d’une hypothèse de travail6. En fin de compte, ce n’est qu’au terme de toute l’étude que la notion d’une capacité de discernement paraît reçue de plein droit. Il reste à se demander comment il faut se représenter cette capacité : est-elle vraiment unique, ou simplement unifiée ? ou n’est-elle qu’un trait commun 4 « On a défini par deux capacités l’âme qui caractérise les animaux, d’une part la capacité de discerner, ce qui est l’œuvre de la perception et de la pensée réfléchie, et d’autre part la production d’un mouvement selon le lieu. Au sujet de la perception et de l’intelligence, les analyses qui précèdent doivent suffire ; quant au principe qui produit le mouvement (…) », etc. (ἐπεὶ δὲ ἡ ψυχὴ κατὰ δύο ὥρισται δυνάμεις ἡ τῶν ζῴων, τῷ τε κριτικῷ, ὃ διανοίας ἔργον ἐστὶ καὶ αἰσθήσεως, καὶ ἔτι τῷ κινεῖν τὴν κατὰ τόπον κίνησιν, περὶ μὲν αἰσθήσεως καἰ νοῦ διωρίσθω τοσαῦτα, περὶ δὲ τοῦ κινοῦντος (…), 432a15-18). 5 Voir dans ce volume la contribution d’Annick Jaulin, p. 79. 6 Dans la formule de 429a13-14 : « si comprendre par l’intelligence est comparable à percevoir (…) » (εἰ δή ἐστι τὸ νοεῖν ὥσπερ τὸ αἰσθάνεσθαι), il me semble que le « si… » doit être pris au sérieux (en dépit de la présence du δή qui pourrait inciter à le rendre par « puisque… »).

L’ÂME COMME LA MAIN: DE ANIMA III 8

225

à deux facultés, la perception et l’intelligence, qui demeurent essentiellement distinctes ? La transition en amont (à la jonction des chapitres 7 et 8), si elle est assez nette, n’est pas aussi claire. Cela tient certainement, au moins en partie, à l’allure particulière du chapitre 7. Sans entrer trop avant dans les difficultés que présente ce chapitre7, j’avancerai que les petites sections qu’on y distingue généralement8 abordent (peut-être à partir de points de vue différents) les relations qui peuvent exister entre la perception, plus exactement entre diverses expériences perceptives, et des actes de l’intelligence humaine. Mais les dernières lignes du chapitre présentent un caractère différent : on y trouve une proposition générale introduite par ὅλως (431b16), puis une question dont Aristote déclare qu’elle doit être examinée « par la suite » (431b17-19). Je fais l’hypothèse que ces trois lignes, avec les deux premières lignes du chapitre 8, forment la transition qui introduit III 8, c’est pourquoi je les ai incluses dans la présente étude. 3. La proposition que l’intelligence en acte est identique à ses objets est une prémisse importante de la première partie du chapitre 8 (431b2223, voir plus loin § 6). Par ailleurs, elle n’est pas entièrement nouvelle au moment où elle est énoncée ici. Des formules assez semblables ont déjà été avancées dans les chapitres précédents : – en III 4 (429a27 sq.), pour proposer une interprétation aristotélicienne de la thèse (probablement platonicienne ou académique) que « l’âme est le lieu des formes » (mais, précise Aristote, cela ne vaut que pour l’âme intellective, et elle n’est qu’en puissance les formes) ; – vers la fin de III 4 (429b26-430a9), dans la discussion de la difficulté suivante : « l’intelligence ne doit-elle pas être elle-même intelligible ? et alors, qu’est-ce qui la distinguera des intelligibles ? » Aristote assume la thèse que l’intelligence est identique à ses objets, mais nie que cela impose de penser que ces objets (du moins ceux d’entre eux qui ont une matière) soient eux-mêmes des 7

Voir dans ce volume la contribution de Klaus Corcilius, p. 185-219. Les découpages donnés par les différents commentateurs varient assez peu ; pour ma part, je propose : 431a1-4, a47, a8-17, a17-20, a20-b1, b2-9, b10-12. 8

226

MICHEL CRUBELLIER

êtres intelligents ; ce qui revient à proposer une conception curieusement asymétrique de cette identité9 ; – on trouve enfin à deux reprises (en III 5, 430a19-21 et en III 7, 431a1-3) la phrase : « la science en acte est identique à son objet ; quant à la science en puissance, dans l’individu elle précède dans le temps, mais généralement parlant (ὅλως) elle ne précède même pas dans le temps »10. Il importe de souligner les différences qui existent malgré tout entre ces formules. On pourrait dire qu’il y a un schéma commun qui est l’affirmation d’une coïncidence entre un acte de connaissance et la réalité qu’il saisit. Mais les termes employés ne sont pas les mêmes à chaque fois : – du côté de l’objet, on a parfois, justement, la notion générale d’objet (πρᾶγμα ou τὰ πράγματα, le corrélat fonctionnel de la notion de connaissance11), mais le premier passage parle des formes  ; on trouvera aussi dans III 8 les choses ou les étants (τὰ ὄντα) ; – du côté du sujet, l’âme, l’intelligence ou la science  : on peut sans doute se permettre de négliger la distinction entre l’intelligence et la science. De fait, dès les premières lignes du chapitre 8, Aristote parle des « objets de science » et de « la science » là où on attendrait « les intelligibles » et donc « l’intelligence ». Cela pourrait se justifier en supposant que la connaissance scientifique est le fait de l’intelligence et qu’elle en représente l’activité la plus excellente12. Par contre il est certain que l’âme est autre chose que l’intelligence ; 9 « De sorte que l’intelligence n’appartient pas à ceux-ci [= les intelligibles qui ont une matière], car l’intellection de tels objets est une puissance sans matière ; mais l’intelligibilité appartient à celui-là [= l’intellect] » : ὥστ’ ἐκείνοις μὲν οὐχ ὑπάρξει νοῦς (ἅνευ γὰρ ὕλης δύναμις ὁ νοῦς τῶν τοιούτων), ἐκείνῳ δὲ τὸ νοητὸν ὑπάρξει (III 4, 430a7-9). 10 La répétition littérale est frappante, de sorte qu’on doit se demander si elle n’est pas due à un accident de la transmission du texte, par exemple une annotation marginale ultérieurement intégrée au texte. Il n’est cependant pas facile de décider quel est le contexte dans lequel elle s’insère mal, ou moins bien. En fait, les deux passages présentent une certaine continuité, et les deux sont difficiles. Dans III 5, le doublet est suivi de la formule ἀλλ’ οὐχ ὅτε μὲν νοεῖ ὅτε δὲ οὐ νοεῖ (« et il n’est pas vrai que tantôt pense et tantôt elle ne pense pas ») ; et dans III 7 il débouche sur l’affirmation que toute connaissance doit provenir d’une connaissance en acte (parce que « tout ce qui advient advient à partir d’une réalité en acte », ἐξ ἐντελεχείᾳ ὄντος πάντα τὰ γινόμενα, 431a3-4). 11 Comme dans Cat. 8, 11a24-34 (et aussi 12, 14b18-23). 12 Il faudrait quand même faire cette réserve qu’il existe d’autres activités caractéristiques du νοῦς ; sans parler de la distinction entre « science » et « acte d’intellection »

L’ÂME COMME LA MAIN: DE ANIMA III 8

227

– enfin, même la formule qui sert à exprimer cette coïncidence peut varier légèrement : à côté du verbe être employé au sens de « être identique à », on trouve aussi la formule « le lieu des formes » ; et surtout, dans la plupart de ces contextes, Aristote souligne que la façon dont nous devons concevoir cette identité demeure une question ouverte. 4. La thèse générale exprime une position qu’on peut appeler un optimisme épistémologique (en ce qu’elle équivaut à rejeter les arguments de type sceptique reposant sur la possibilité d’un hiatus ou d’une déformation entre la structure de la réalité et celle de l’expérience que nous en avons). Sa forme typique est celle qu’on trouve ici, à la fin de III 7, ou dans le doublet de III 5 / III 7 : l’intelligence (la science) en acte est (est identique à) son objet. On l’interprète classiquement – et, à mon avis, de façon correcte – en disant que la connaissance scientifique se réalise par la saisie complète de l’essence de l’objet, laquelle consiste dans sa forme (qui représente tout ce qu’il y a de véritablement réel dans l’objet). Comme d’autre part le début du chapitre III 4 a établi (429a15-27) que, pour concevoir adéquatement son objet, l’intelligence ne doit comporter aucune détermination préexistante (qui parasiterait la connaissance), il s’ensuit qu’au moment où elle connaît et comprend un objet donné, l’intelligence est pleinement identique à cet objet. La formulation de III 5 / III 7 : τὸ αὐτό ἐστιν ἡ κατ’ ἐνέργειαν ἐπιστήμη τῷ πράγματι, est donc la plus précise et la moins contestable. Elle devient problématique déjà lorsqu’on veut l’étendre à une pluralité d’objets, voire à tous les objets d’une intellection possible, comme c’est le cas ici, en 431b17 : même si je peux m’identifier ainsi, tour à tour, à n’importe quel objet intelligible (moyennant, probablement, l’acquisition préalable de certaines connaissances indispensables), cela ne signifie pas que mon intelligence « soit » tous ces objets, ou alors dans un sens affaibli qui peut s’exprimer en langage aristotélicien par la restriction « en puissance »13. Et, comme nous le verrons, l’extension de cette dans différentes listes de modalités de la connaissance (en particulier EN VI, 1139b15-17, 1141a3-8 et suivants, ou la fin de Seconds Analytiques II 19). 13 C’est peut-être le sens de la correction imposée en 427a27-29 à la thèse platonicienne que « l’âme est le lieu des formes » – mais « en puissance seulement et non pas en entéléchie ».

228

MICHEL CRUBELLIER

thèse à l’âme tout entière (dans le chapitre 8) est encore plus problématique et en change la signification. 5. La question posée en 431b18-19 est : Est-ce qu’il est possible ou impossible que , tout en n’étant pas elle-même séparée de la grandeur, pense quelque chose des réalités séparées ?

Comme je serai amené à m’y référer à plusieurs reprises dans la suite, je la nommerai la question de la connaissance du séparé. Il faut éclaircir d’abord la formule « non séparé de la grandeur », qui n’est pas fréquente dans le corpus (on attendrait plutôt « séparée / non séparée de la matière » ou « du corps »). Le terme μέγεθος figure dans la liste des « sensibles communs » (voir II 6, 418a17-18 et III 1, 425a15-19) et semble désigner quelque chose comme l’étendue ou l’extension (il faut normalement préciser : « la grandeur de quelque chose »14). Bien que cette notion suppose ou semble supposer, dans son usage courant, l’existence d’une réalité corporelle qui serait son substrat, elle représente une certaine forme d’abstraction puisqu’elle n’implique ni une composition physique ni des propriétés spécifiquement sensibles (poids, température, couleur, etc.)15. Je proposerai plus loin (§ 13) une raison qui peut expliquer l’emploi de μέγεθος dans ce contexte. Une question importante et difficile est de savoir si, lorsqu’on envisage que l’intelligence ne soit « pas séparée de la grandeur », on veut 14 Mais à vrai dire la même précision devrait être donnée pour « la matière » chez Aristote. 15 On peut être troublé par le fait que dans la section noétique de De Anima III, le terme μέγεθος apparaît à plusieurs reprises, et de façon assez marquée, en relation avec la question de savoir si, comment et jusqu’à quel point l’intelligence est « distincte » (χωριστός) – il faut probablement sous-entendre « des autres actions de l’âme » et éventuellement des facultés qui les produisent). Aristote demande si elle est distincte « selon l’étendue » (κατὰ μέγεθος) ou « selon la notion » (κατὰ λόγον, 429a11-12) ; une question voisine de celle-ci est posée à propos du principe de l’action orientée au chapitre III 9, 432a19-20). À la réflexion, il ne me semble pas qu’il faille établir un lien entre ces deux emplois du mot. Dans le cas de la distinction des actions ou des facultés, « selon la grandeur » se réfère probablement à un modèle qu’on pourrait dire additif : une fonction A est distincte de la fonction B « selon la grandeur » si, en supposant que A vienne à disparaître, on peut ou doit concevoir que B demeurerait (alors que si elles diffèrent seulement par la notion, la suppression de A impliquerait du même coup la disparition de tout ce qui sert de support à la fonction B).

L’ÂME COMME LA MAIN: DE ANIMA III 8

229

parler [a] de l’extension du corps et des phénomènes corporels qu’on pourrait se représenter comme le support objectif (et plus ou moins causal) d’un acte d’intellection, ou bien [b] de l’extension des images ou d’autres éventuels contenus de pensée qu’on peut se représenter comme les supports intentionnels ou subjectifs de l’acte d’intellection. La version [a] apparaît contredite par un passage célèbre du chapitre II 1, qui dit que certaines « parties de l’âme » « ne sont l’entéléchie d’aucune du corps »16 – une affirmation qui ne peut concerner que l’intelligence et qui ne s’accorde pas bien avec le fait qu’ici la condition ὄντα αὐτὸν μὴ κεχωρισμένον μεγέθους semble introduite de façon non problématique17. Au contraire la version [b], non seulement évite cette difficulté, mais permet de comprendre pourquoi Aristote introduit la question à cet endroit : il résulte des analyses qui précèdent que l’activité de l’intelligence présuppose des représentations perceptives (perceptions effectives ou images résultant de perceptions18 passées) sur lesquelles l’intelligence s’exerce – et cela, sinon nécessairement et toujours, du moins dans un grand nombre de situations, qui ont été évoquées et expliquées dans le chapitre 718, et peut-être déjà dans d’autres passages à partir du début de III 4. Dans cette hypothèse, la thèse que l’intelligence en acte est ses objets implique notamment qu’elle partagera avec les objets qui ont été envisagés (perceptions, images et reconstructions abstraites), la propriété d’être étendue. – Pour bien entendre cette proposition, il ne faut pas perdre de vue qu’on ne parle pas ici de l’intelligence comme capacité ou faculté, mais d’actes particuliers d’intellection, qui en tant que tels ont un contenu déterminé. Il est donc naturel qu’à cet endroit Aristote en vienne à se demander quel peut être le statut d’un acte d’intellection qui porte ou porterait sur une réalité dépourvue de « grandeur » et indépendante de toute grandeur. On peut se demander pourquoi il ne l’aborde pas immédiatement et de front. Une explication serait que, comme on l’a souvent relevé, il semble bien que cette question l’embarrasse19 ; une autre (non exclusive de la précédente), qu’il pense qu’il doit d’abord revenir, pour en préciser la signification, sur la prémisse l’intelligence n’est pas elle-même indépendante de la grandeur. 16

II 1, 413a7. Il est vrai que puisque notre passage formule une question, on ne doit pas exclure absolument qu’il introduise une condition contrefactuelle. 18 431a14-15, 16-17, 431b2. 19 Cf. II 1, 413a6-9 ; Met. Λ3, 1070a25-26. 17

230

MICHEL CRUBELLIER

De fait, le νῦν δὲ qui ouvre le chapitre 8 forme une opposition assez nette avec le σκεπτέον ὕστερον qui clôt le chapitre 7 ; et par ailleurs la question du lien entre la connaissance intellectuelle et la grandeur revient dans le chapitre 8 à partir de la ligne 432a3. J’y reviendrai, donc, moi aussi. 6. πάλιν (« encore », 431b21) a deux emplois possibles dans le grec d’Aristote : il peut être compris soit (a) comme signifiant « encore une fois », « à nouveau », soit (b) comme signifiant « ce qu’il reste à dire », ou « en passant à l’étape suivante ». Le sens (b) est le plus fréquent chez Aristote et il me semble qu’il convient mieux ici, car en toute rigueur la thèse que « l’âme est d’une certaine façon toutes choses » est nouvelle, même si elle a été d’une certaine façon annoncée par des déclarations antérieures (cf. § 3). On pourrait objecter que πάλιν, entendu en ce sens, serait redondant avec νῦν δὲ ; la nuance, à mon avis, tient au fait que νῦν indique seulement le temps présent alors que πάλιν indique l’ordre rationnel de la recherche : « à ce point, pour pouvoir avancer, il nous faut établir que –  ». Συγκεφαλαιώσαντες est généralement, et assez naturellement, traduit par « en récapitulant ». Il reste à savoir ce qui est récapitulé, de quelle façon et dans quel but. Une indication est donnée par la phrase πῶς δὲ τοῦτο, δεῖ ζητεῖν : « de quelle façon, c’est ce qu’il faut rechercher » (b23-24) : cette formule indique que ce qui va suivre20 est une nouvelle recherche. De sorte que la récapitulation se réduit aux trois lignes 431b21-23. Autant dire qu’elle n’a pas le caractère d’un résumé scolaire ou didactique. Elle rappelle simplement – sans revenir sur les étapes de leur établissement – deux résultats importants : la science est d’une certaine façon les objets de science et la perception est les objets sensibles. À partir de ces deux propositions, et de la prémisse supplémentaire : tout étant est soit sensible, soit intelligible (admise ici sans justification particulière), on peut conclure – de façon non-rigoureuse – que « l’âme », dans la mesure où elle englobe les deux fonctions de la perception et de l’intelligence, « est en quelque sorte toutes choses ». Cette conclusion doit être affinée et mieux établie. C’est ce que fera le chapitre 8, en 20 On pourrait objecter que la formule « il faut le rechercher » indique que cette recherche devra être faite à l’avenir, sans préciser l’échéance ; mais il me paraît assez net que ce qui commence à la ligne 24 concerne précisément la façon dont on peut dire que « l’âme est toutes choses ». J’espère pouvoir le montrer dans les pages qui suivent, en précisant en quel sens je l’entends.

L’ÂME COMME LA MAIN: DE ANIMA III 8

231

explicitant notamment ce que signifie la modalisation πως (« d’une certaine façon »), employée aussi bien à propos des intelligibles qu’à propos des sensibles, et en précisant la façon dont les capacités de percevoir et de comprendre se coordonnent et se complètent au sein de « l’âme ». – Dans cette première phrase, on peut aussi remarquer la formule περὶ ψυχῆς τὰ λεχθέντα συγκεφαλαιώσαντες, et la place saillante qu’y occupe le syntagme περὶ ψυχῆς, « à propos de l’âme ». On pourrait la comprendre de façon minimaliste ou neutre, en considérant que ces résultats qui vont être rappelés ont en effet été établis « à propos de l’âme ». Mais on peut aussi défendre l’idée que précisément ils n’ont pas été établis au niveau de l’âme tout entière, mais successivement et séparément pour la perception, puis pour l’intelligence (il faudrait même préciser, comme nous allons le voir, au niveau de l’intelligence et de la perception en acte et non en puissance). La récapitulation en question revient donc à rappeler deux résultats importants (κεφάλαια) et à les combiner en les rapportant cette fois à l’âme prise comme un tout. Cette façon de reprendre l’analyse au niveau de l’âme produit un déplacement par rapport à la conception de l’intellect, au moins telle qu’elle a été indiquée dans la question de la connaissance du séparé : dans les lignes qui vont suivre on ne considérera pas toutes les activités de l’intelligence, mais seulement celles qui ont un lien avéré avec l’âme définie comme entéléchie première du corps organisé ; et d’autre part on travaillera sur la connexion entre l’intelligence et la perception au sein d’une vie animale unique. 7. Les lignes 431b24-26 introduisent une nouvelle prémisse : la science et la perception se divisent en leurs objets – prémisse qui va de soi si on considère les actes de connaissance scientifique ou de perception21 ; mais Aristote l’étend aux puissances correspondantes, ce qui lui permettra de conclure que « les capacités de percevoir et de connaître scientifiquement qui sont dans l’âme » (désignées par les adjectifs en -ικος) doivent être, en un certain sens, la même chose que ce que sont, au niveau de leurs objets, la possibilité d’être perçu ou celle d’être connu scientifiquement (indiquées par des adjectifs en -τος)22 ; ce qui ne va 21 Elle signifie que « la science physique », par exemple, se subdivise dans toutes les propositions spéciales concernant des objets naturels (chacune de ces propositions peut être appelée « une science » dans le vocabulaire d’Aristote). 22 Le texte des lignes 431b24-28 n’est pas parfaitement fixé ; mais les variantes affectent peu le sens de l’argument. Elles portent sur trois points : (1) les termes indiquant

232

MICHEL CRUBELLIER

pas de soi. De fait, on trouve au début du livre III (III 2, 425b26-426a26) une discussion fine et précise dans laquelle Aristote, tout en partant du principe que l’acte du sensible et celui de la sensation sont en réalité un seul et même acte, souligne que leur « être » (τὸ εἶναι) n’est pas le même et que cette identité ne s’étend pas aux puissances : Ἐπεὶ δὲ μία μέν ἐστιν ἐνέργεια ἡ τοῦ αἰσθητοῦ καὶ τοῦ αἰσθητικοῦ, τὸ δ’ εἶναι ἕτερον, ἀνάγκη ἅμα φθείρεσθαι καὶ σώζεσθαι τὴν οὕτω λεγομένην ἀκοὴν καὶ ψόφον, καὶ χυμὸν δὴ καὶ γεῦσιν, καὶ τὰ ἄλλα ὁμοίως· τὰ δὲ κατὰ δύναμιν λεγόμενα οὐκ ἀνάγκη· ἀλλ’ οἱ πρότερον φυσιολόγοι τοῦτο οὐ καλῶς ἔλεγον, οὐθὲν οἰόμενοι οὔτε λευκὸν οὔτε μέλαν εἶναι ἄνευ ὄψεως, οὐδὲ χυμὸν ἄνευ γεύσεως. τῇ μὲν γὰρ ἔλεγον ὀρθῶς, τῇ δ’ οὐκ ὀρθῶς· διχῶς γὰρ λεγομένης τῆς αἰσθήσεως καὶ τοῦ αἰσθητοῦ, τῶν μὲν κατὰ δύναμιν τῶν δὲ κατ’ ἐνέργειαν, ἐπὶ τούτων μὲν συμβαίνει τὸ λεχθέν, ἐπὶ δὲ τῶν ἑτέρων οὐ συμβαίνει. Puisque l’acte du sensible et celui de la capacité de percevoir sont un seul et même acte, cependant que leur être est différent, il est nécessaire que soient détruits ou conservés ensemble ce qu’on appelle l’audition et la résonance, ainsi que la saveur et le goût , et de même pour les autres sens. Mais lorsque ces termes sont entendus selon la puissance, ce n’est pas nécessaire. Pourtant les anciens philosophes de la nature se sont exprimés de façon incorrecte, lorsqu’ils ont exprimé l’opinion que rien n’est blanc ni noir s’il n’y a pas de vision , et qu’il n’existe pas de saveur si elle n’est pas goûtée. En un sens c’était juste, et en un autre sens ce ne l’était pas : car, comme la perception et le sensible s’entendent en deux sens, d’une part selon la puissance et d’autre part selon l’acte, ce les produits de la division (on trouve δυνάμεις en face de τὰ δυνάμει et ἐντελεχείας ou τὰς ἐντελεχείας en face de τὰ ἐντελεχείᾳ) ; (2) à la ligne 27 on peut lire, selon l’accentuation, ταῦτα ou ταὐτὰ, et on trouve ταὐτόν dans la branche E de Ross (a de Förster) ; (3) ἐπιστητὸν est remplacé par ἐπιστημονικὸν dans U et αἰσθητὸν par αἰσθητικὸν dans U et X. Pour commencer par ce dernier point (3), on peut écarter les versions qui comportent ἐπιστημονικὸν ou αἰσθητικὸν, puisqu’elles reviennent à poser l’identité entre des termes identiques. La variation (1) a moins d’importance (il ne serait pas très gênant d’appeler « puissances » les X-en-puissance et « actes » les X-en acte). En ce qui concerne la variation (2), il n’y a pratiquement pas de différence entre ταὐτόν et ταὐτὰ, dans la mesure où le pluriel est de toute façon à interpréter de façon distributive, comme l’indique la division τὸ μὲν… τὸ δὲ… qui suit. L’accentuation ταῦτα, bien qu’on ne la trouve dans aucun manuscrit, s’accorderait particulièrement bien avec l’interprétation proposée ci-dessus ; mais ταὐτὰ et même ταὐτόν, dès lors qu’on exclut l’idée que la capacité de percevoir puisse être dite identique à celle de connaître scientifiquement (ce que tout le contexte contredit), devront être compris comme renvoyant aux πράγματα de la ligne 25, ce qui nous conduit au même sens (dans une forme grammaticalement moins correcte). Enfin la correction de Hayduck adoptée par Ross (l’ajout de devant ἐπιστητὸν et devant αἰσθητὸν) facilite peut-être le texte, mais n’est nullement indispensable. – Dans ma traduction, j’ai suivi, comme Förster, le texte de la branche E (a).

L’ÂME COMME LA MAIN: DE ANIMA III 8

233

qu’ils disent est bien ce qui se produit pour ceux-ci, mais pas pour les premiers. (III 2, 426a15-25)

Par ailleurs, on trouve dans les Catégories la présentation d’une relation semblable entre les propositions vraies et les états de fait qui leur correspondent : bien que les unes et les autres « s’impliquent réciproquement » (ἀντιστρέφεται), les états de fait occupent une position d’antériorité, « car c’est parce que le fait existe ou n’existe pas que la proposition est dite vraie ou fausse » (Cat. 12, 14b21-22). Est-il impossible de concilier notre passage de III 8 avec ces textes ? Non, si on remarque d’une part qu’Aristote ne parle que d’une assimilation « en puissance » (δυνάμει ταῦτά ἐστι, 431b27), et surtout, d’autre part, que la formule qu’il emploie conserve l’asymétrie et l’espèce d’antériorité des objets sur les capacités qui peuvent les discerner : lorsqu’il dit que « les capacités de percevoir et de connaître scientifiquement (…) sont en puissance cela », cela n’implique pas que réciproquement les objets de connaissance doivent « être » ces capacités, en quelque sens que ce soit. Mais alors, qu’est-ce que cela veut dire concrètement ? Dans le cas de « l’être-perceptible » (c’est-à-dire, pour les sensibles propres, le fait d’être visible, audible, tangible, odorant ou sapide), être perceptible, c’est avoir la capacité d’affecter l’organe sensoriel en l’affectant d’une valeur précise comprise à l’intérieur d’une échelle définie par un couple de contraires : tel degré de température entre le chaud et le froid, le rouge (ou telle nuance de rouge) entre la blancheur et le noir23. C’est cette notion assez subtile qu’Aristote indique dans la suite du chapitre 8 (431b28-29) en précisant que lorsqu’il dit que la capacité de percevoir est identique au perceptible, cela doit s’entendre non de l’objet perceptible concret, mais de « sa forme ». Je reviendrai sur le sens exact à donner à cet emploi du mot εἶδος. 23 Il doit exister quelque chose de semblable pour les sensibles communs, mais ce n’est pas une structure aussi nettement définie. Aristote propose un schéma assez complexe, dans lequel la perception et la reconnaissance du mouvement (ou du changement, κίνησις) joue un rôle spécial : « toutes ces choses [= les sensibles communs], nous les percevons au moyen du mouvement (ainsi nous percevons la grandeur par le mouvement – et donc la figure aussi, car la figure est une sorte de grandeur; nous percevons que quelque chose est en repos par le fait que cela ne se meut pas ; et le nombre par la négation du continu) – et au moyen des sensibles propres » (III 1, 425a16-19). Par ailleurs, il est probable que les sensibles communs jouent un rôle essentiel dans la structuration de l’expérience sensible dont je parlerai plus loin (§ 9 et 10), structuration grâce à laquelle le sensible lui-même, pourrait-on dire, devient intelligible.

234

MICHEL CRUBELLIER

Le correspondant de cette propriété au niveau de la capacité de percevoir, c’est, d’une part, la possibilité d’être ainsi affecté, donc la réceptivité à un ensemble déterminé de propriétés : les couleurs pour la vue, les degrés de température pour le toucher ; et d’autre part la possibilité de repérer chacune de ces valeurs à partir d’une norme (« moyenne »24) unique, capacité sans laquelle être affecté par l’objet sensible, même selon une qualité perceptible, se réduirait à une forme de passivité physique ordinaire25. On remarquera que le couple αἰσθητικόν / αἰσθητόν se distingue du couple « acte de perception » / « effectivement perçu » par le fait que la perception en acte est toujours la perception d’un sensible particulier, alors que la capacité de percevoir ou d’être perçu est générale (la couleur dans l’objet visible, et la sensibilité à la couleur, du côté du sujet qui voit)26. Cette différence s’explique par le fait que dans le cas des capacités, on se situe au niveau de l’âme entendue comme entéléchie première du corps dans sa totalité, entéléchie première qui peut se réaliser au travers d’un très grand nombre d’entéléchies secondes (cette structure est proprement ce que nous appelons la vie de l’animal). On remarquera aussi que de ce fait il n’y a pas entre les capacités et les objets-en-puissance qui leur correspondent, une identité aussi nette ni, en fait, aussi stricte, que dans le cas des actes (ce qui explique le πως). 8. Dans le cas de l’être-intelligible, Aristote n’a pas donné autant d’indications, c’est le moins qu’on puisse dire. Il aborde cependant la question dans un passage du chapitre III 4, sur la base d’une analogie entre l’intelligence et la perception – la même analogie que nous trouvons à l’œuvre dans la première partie du chapitre 8. Elle jouait déjà un rôle important au chapitre II 5, où Aristote introduit son hypothèse des trois degrés de réalisation d’une capacité psychique (Dreistufenlehre) pour servir de cadre général à l’étude des perceptions. Au chapitre III 4, conformément à cette hypothèse, il envisage un développement de l’intelligence en deux stades, dans lequel le stade préliminaire est l’analogue de l’aptitude générale à percevoir. On peut se représenter ce stade préliminaire comme l’état ou la disposition d’une 24

Voir II 11, 423b7, 424a1-10 ; II 12, 424a25-b2. C’est ce qui se produit pour les plantes, voir II 12 424a32-b3. 26 On trouve des considérations semblables (appliquées à la science) dans Métaphysique M 10 ; voir § 10. 25

L’ÂME COMME LA MAIN: DE ANIMA III 8

235

intelligence mature ou éduquée27, par opposition d’une part à l’aptitude générale à la connaissance intellectuelle telle qu’elle existe chez un enfant ou un être humain non éduqué, et d’autre part à sa pleine réalisation dans un acte d’intellection. Mais l’analogie conduit aussi à formuler une difficulté embarrassante : dans le cas de l’intelligence, il n’existe pas de structures corporelles qui correspondraient à ce que sont, pour la perception, la réceptivité à une gamme particulière de qualités et l’existence d’une norme « moyenne ». Non seulement nous ne connaissons rien de tel, mais, on l’a vu, Aristote avance un argument très fort pour dire que dans le cas de l’intelligence il ne peut pas exister de telles conditions déterminées28. Sur la façon dont l’intelligence se réalise comme entéléchie première, puis comme acte d’intellection, le chapitre III 4 donne trois indications sommaires, mais importantes : Τὸ μὲν γὰρ αἰσθητικὸν οὐκ ἄνευ σώματος, ὁ δὲ χωριστός. ὅταν δ’ οὕτως ἕκαστα γένηται ὡς ὁ ἐπιστήμων λέγεται ὁ κατ’ ἐνέργειαν (τοῦτο δὲ συμβαίνει, ὅταν δύνηται ἐνεργεῖν δι’ αὑτοῦ), ἔστι μὲν οὖν καὶ τότε δυνάμει πως, οὐ μὴν ὁμοίως καὶ πρὶν μαθεῖν ἢ εὑρεῖν· καὶ αὐτὸς δὲ αὑτὸν29 τότε δύναται νοεῖν. La capacité de percevoir n’existe pas sans corps, alors que l’autre [= l’intelligence] est séparée30. Et chaque fois qu’elle acquiert certaines déterminations31, de la façon dont nous disons qu’il est « savant » en acte – ce qui a lieu lorsqu’il peut actualiser son savoir spontanément –, 27

Ce qui complique un peu le schéma, puisqu’à côté du modèle d’une intelligence adulte universelle, Aristote semble bien envisager des compétences intellectuelles spécialisées, qu’il décrit comme « la possession de la [= d’une certaine] science » : c’est le cas de la compétence du lecteur (un exemple récurrent), mais on peut aussi envisager, par exemple, la capacité de « voir » une figure géométrique et de reconnaître les possibilités de construction qu’elle offre, capacité qui résulte à la fois d’une pratique constante de la géométrie et de la connaissance de certains théorèmes fondamentaux. Cette complication ne remet pas en cause le modèle des trois stades, qu’on doit considérer comme stylisé. 28 Car si l’intelligence avait quelque détermination propre antérieure à son activité de connaissance, cette activité s’en trouverait perturbée ou biaisée. 29 En conservant le texte des manuscrits, sans la correction de Ross (δι’ αὑτοῦ au lieu de ...δὲ αὑτὸν). 30 Si on fait l’hypothèse qu’Aristote ne s’exprime pas de façon simplement négligente lorsqu’il écrit à la fin du chapitre 7 que l’intellect « n’est pas séparé de la grandeur » et, ici, qu’il « est séparé » (sous-entendu : « du corps »), il faut poser qu’il y a une différence cruciale entre séparé du corps et séparé de la grandeur (voir ci-dessus § 5). 31 ἔκαστα est à prendre dans un sens strictement distributif : quand l’intelligence « devient x » ou « devient y ».

236

MICHEL CRUBELLIER

même alors, certes, elle est en puissance en un certain sens, mais pas de la même façon qu’elle l’était avant de comprendre et de découvrir, c’est-à-dire qu’à ce moment elle est capable de se penser elle-même. (III 4, 429b5-9)

Pour l’intelligence, s’actualiser consiste à « se penser elle-même », une formule qui s’éclaire dès qu’on la rapproche de la thèse qui nous occupe ici32 (« l’intelligence est ses objets »). Pour en être capable, elle doit se trouver dans un certain état déterminé qui n’est désigné que par la formule très abstraite ὅταν … ἕκαστα γένηται, éclairée cependant par la comparaison « au sens où on dit que quelqu’un est savant en acte »33. Aristote indique aussi en passant que cet état résulte d’un apprentissage, lequel est spécifié ici par les deux verbes μαθεῖν (acquérir la compréhension de quelque chose) et εὑρεῖν (découvrir du nouveau). 9. Aux lignes 431b28-432a3, la célèbre comparaison de l’âme avec la main a pour but de préciser les deux notions de l’αἰσθητόν (le sensible) et de l’ἐπιστητόν (le « connaissable par la science »). La prémisse il faut nécessairement que ce soit ou bien ces objets euxmêmes ou bien leurs formes n’est pas autrement justifiée ; on peut penser qu’elle s’explique par le fait que la seule chose, en-dehors de l’objet lui-même, qui puisse donner lieu à une identification, c’est la forme. La prémisse il n’y a pas de pierre dans l’âme paraîtra sans doute, quant à elle, immédiatement évidente ; mais il ne faudrait pas que cette évidence risque de masquer la forme précise de l’inférence à laquelle elle donne lieu, ni par conséquent la portée précise de la conclusion. L’inférence, ici comme aux lignes 431b24-28, va du niveau de l’acte à celui de la puissance. « Il n’y a pas de pierre dans l’âme » lorsque nous effectuons un acte de perception ou de connaissance scientifique concernant cette pierre, mais seulement la forme de la pierre. Par conséquent, l’αἰσθητόν ou l’ἐπισθητόν – qui sont les termes qui nous intéressent ici, et qui sont des puissances – sont la puissance de produire dans l’âme, respectivement, la forme sensible ou la forme intelligible de cette pierre, c’est-à-dire le type de qualités perceptibles ou de connaissances vraies 32 Et qui, comme on l’a vu, est également présente dans le contexte immédiat de ce chapitre III 4 via l’allusion à la thèse académique que l’âme « est le lieu des formes ». 33 De sorte que « devenir x », dans ce contexte, signifie en fait acquérir sur x (sur le cerveau de la sangsue, par exemple) un savoir virtuel que l’on sera désormais capable d’actualiser spontanément (on est devenu un « spécialiste de x »).

L’ÂME COMME LA MAIN: DE ANIMA III 8

237

qui caractérisent cette pierre, de telle sorte qu’elle pourra être perçue ou scientifiquement connue. C’est en ce sens qu’il faut comprendre la comparaison avec la main, « instrument d’instruments ». La même idée est introduite sous une forme plus développée au livre IV des Parties des animaux  : Celui qui est le plus intelligent (φρονιμώτατος), en effet, se servira plus correctement du plus grand nombre d’instruments ; or la main semble bien être, non pas un seul instrument mais plusieurs, car c’est comme si elle était un instrument pour instruments (ὄργανον πρὸ ὀργάνων). (687a19-21) La main en effet devient griffe, pince, corne, ainsi que lance, épée et toute autre sorte d’arme ou d’instrument : elle sera tout cela du fait qu’elle est capable de tout saisir et de tout tenir. (687b3-5)

L’idée est à la fois que la main vaut plusieurs instruments et qu’elle peut servir à plusieurs instruments – parce qu’elle permet de les manier ; la comparaison implique que chacune des deux formes de connaissance, la perception comme l’intellection, occupe, à son niveau propre, une position de ce genre. La comparaison suggère donc que les « formes » dont on parle ici sont des moyens qui permettent (facilitent, rendent plus précise) la connaissance des choses ; et et qui le permettent (a) parce que, comme la main, elles ne sont pas spécialisées et (b), à chaque fois, pour un nombre quasi-illimité d’objets (donnés ou simplement possibles). Dans cet usage, il serait légitime de rendre εἶδος (dans sa première occurrence) par « structure ». Dans le cas de la perception, on peut raisonnablement supposer que ces structures sont du type de celles que j’ai mentionnées au § 7 et qui comportent à la fois un espace de variété qualitative et une norme par rapport à laquelle chaque impression reçue particulière pourra être évaluée, identifiée et reconnue. On dirait assez naturellement en français que ces structures permettent de mettre en forme les données perceptives. Ces « formes de sensibles » sont donc autre chose que les formes de chaque substance concrète particulière dont parle Met. Z, par exemple. Surtout, si la forme dont il est question ici est comparable à la main, c’est une forme pour les sensibles (voir § 10). – Je suppose ici que le génitif adnominal ὄργανον ὀργάνων équivaut, en moins précis, à l’expression ὄργανον πρὸ ὀργάνων utilisée dans les PA., avec les deux valeurs mentionnées ci-dessus (un instrument qui équivaut à plusieurs parce qu’il est approprié au maniement de tous). Si au contraire on veut

238

MICHEL CRUBELLIER

soutenir que l’adnominal est employé ici de façon intentionnelle et appropriée, alors il faut dire que c’est dans l’expression εἶδος εἴδων qu’il est employé à dessein, pour poser, entre la structure de l’intelligence et les formes dont elle est la forme, une relation encore plus étroite qu’entre la main et les instruments qu’elle manie (ὄργανον ὀργάνων résultant alors, par une sorte de mouvement rhétorique, du parallélisme avec εἶδος εἴδων) ; les deux interprétations ne s’excluent pas nécessairement. Ou, pour dire la même chose autrement : Aristote écrit ici que la perception (et de même l’intelligence) « est une forme » à la façon dont la main « est un outil », mais précisément la main n’est pas un outil. De ce point de vue, la thèse que la perception est « une forme d’objets sensibles » (εἶδος αἰσθητῶν), que donnent les manuscrits en 432a334, est tout à fait recevable et va dans le même sens que la remarque de II 12, 424a18-19, que la perception est ce qui est « capable de recevoir les sensibles sans leur matière » (τὸ δεκτικὸν τῶν αἰσθήτων ἄνευ τῆς ὕλης)35. 10. L’analogie suggère qu’il doit exister quelque chose du même genre pour les connaissances intellectuelles : des structures qui en facilitent la saisie et la compréhension. C’est une supposition tout à fait raisonnable, et les exemples que j’ai mentionnés plus haut (les conventions de l’écriture alphabétique, les constructions de base de la géométrie accompagnées de l’intuition du géomètre exercé) peuvent donner une idée de ce que peuvent être ces structures. Peut-on aller plus loin et en proposer une description plus générale ? Aristote ne dit rien sur ce sujet dans l’Âme, mais on trouve des indications précieuses dans un passage de la Métaphysique36 : La science, tout comme le fait de savoir (ἡ γὰρ ἐπιστήμη, ὥσπερ καὶ τὸ ἐπίστασθαι), est double ; et il y a savoir en puissance et savoir en acte. La puissance, qui est comme la matière de l’universel et qui est indéfinie, porte sur l’universel et l’indéfini ; l’acte est déterminé et porte sur un objet 34 Que Förster supprime, sans doute pour faire dire au texte que la perception est elle aussi « une forme de formes », mais de « formes sensibles » cette fois (comme dans la leçon de CE au début de II 12). Sur ce point, voir plus loin § 13 et § 15. 35 De sorte que le texte majoritaire de SUX peut être conservé en face de la variante de CE (τῶν αἰσθητῶν εἴδων), sans doute induite par l’idée que si c’est sans matière, alors c’est une forme. 36 Contexte : Aristote cherche à préciser (en fait, pour en restreindre la portée) la thèse qu’il n’y a de science que de l’universel (1087a10-15).

L’ÂME COMME LA MAIN: DE ANIMA III 8

239

déterminé. C’est un certain acte qui porte sur un certain objet ; mais par accident la vue voit la couleur universelle, parce que cette couleur qu’elle voit est de la couleur ; et ce que regarde le lecteur, cet A-ci, est un A. (Met. M 10, 1087a15-21)

Le savoir en puissance consiste en un ensemble de dispositions qui rendent capable de reconnaître, dans les objets ou les situations déterminés qu’on pourra être amené à considérer, des essences ou des relations universelles et des déterminations nécessaires (comme le lecteur qui reconnaît dans cette configuration particulière un A, alors que pour un Barbare ou un illettré ce seront seulement « des traits » ou tout au plus « un son, mais je ne sais pas lequel »). C’est une connaissance virtuelle, décrite ici comme « la matière de l’universel » parce qu’on peut la formuler dans des énoncés universels (du type « la somme des angles de tout triangle est égale à deux droits »), ou encore parce qu’elle pourra se réaliser dans tous les énoncés formulables à propos de triangles particuliers et dans toutes les utilisations qu’on en fera pour résoudre des problèmes particuliers37. À cet endroit, un lecteur vigilant me demandera peut-être pourquoi je me réfère à ce passage de la Métaphysique, où le savoir en puissance est décrit comme une matière, pour éclairer le texte de l’Âme où l’intelligence est appelée une forme. C’est qu’il s’agit, justement, d’une « forme de formes », c’est-à-dire d’une forme à partir de laquelle l’âme pourra élaborer des formes ; à ce titre elle peut être appelée une puissance et même une matière. Ou, pour retirer à ces manières de dire ce qu’elles peuvent avoir de paradoxal, disons qu’il s’agit très exactement d’une entéléchie première, qui est forme et acte par rapport aux pures possibilités qui la précèdent, mais puissance par rapport aux réalisations qu’elle rend possibles. Il reste cependant des questions à ce sujet : – étant donné que l’activité de l’intelligence n’est pas l’actualisation d’une « partie » du corps de l’animal (c’est-à-dire d’un organe ou d’un appareil au sens anatomique), il faut qu’elle s’appuie sur d’autres activités ou d’autres fonctions – à savoir les activités perceptives38 ; 37

Met. M 10, 1086b33-37. Principalement, mais peut-être pas exclusivement : en effet, quoique le chapitre 8 ne parle que de la perception, un passage de III 7 (431b8-12) suggère que les conduites de poursuite et de fuite pourraient jouer un rôle spécifique de point d’appui pour l’affir38

240

MICHEL CRUBELLIER

– étant donné que l’âme est essentiellement une, il faut comprendre comment la capacité de connaissance intellectuelle s’intègre avec les autres capacités dans la vie d’une âme unique. Dans l’interprétation que j’en propose, la deuxième partie du chapitre apporte à ces deux questions des réponses qui aboutissent à une modification et une densification de la relation entre l’intelligence et la perception : Aristote montre que l’intelligence n’est pas seulement analogue à la perception, mais qu’elle en dépend ; et il donne des indications quant à la façon dont elle s’appuie sur la perception. 11. La thèse centrale de la deuxième partie est l’affirmation que « les intelligibles sont dans les formes sensibles » (432a4-5). Mais afin de déterminer plus précisément la signification et la portée de cette phrase, je dois commencer par une mise au point à propos de la prémisse qui l’introduit, et qui me semble avoir été souvent mal comprise : ἐπεὶ δὲ οὐδὲ πρᾶγμα οὐθὲν ἔστι παρὰ τὰ μεγέθη, ὡς δοκεῖ, τὰ αἰσθητὰ κεχωρισμένον (432a3-4). En effet, les traducteurs comprennent le plus souvent39 la formule οὐθὲν ἔστι παρὰ τὰ μεγέθη … τὰ αἰσθητὰ comme signifiant « aucun objet n’existe en dehors des grandeurs sensibles », ce qui revient à dire que les grandeurs sensibles sont absolument les seuls objets qui existent – une proposition qui reviendrait à nier l’existence d’objets tels que le premier moteur de l’univers et les autres moteurs célestes immobiles, et qui en outre s’accorderait mal avec la présence, dans son voisinage immédiat, de la question de la connaissance du séparé, laquelle présuppose qu’on reconnaisse que l’hypothèse de réalités « séparées de la grandeur » mérite au moins d’être prise en considération. Curieusement, plusieurs traducteurs et commentateurs40 ne semblent même pas gênés par cette incohérence. Ceux qui la perçoivent cherchent à s’en tirer en s’appuyant sur le ὡς δοκεῖ de la ligne 432a4 et en supposant qu’Aristote invoque (contre les Platoniciens) une opinion commune mation et la négation. Par ailleurs, les chapitres III 9-11 montrent que l’intelligence est un élément constitutif de la formation des conduites finalisées, au moins chez l’être humain. 39 Seules les traductions allemandes de Seidl et de Corcilius rendent παρὰ τὰ μεγέθη τὰ αἰσθητὰ – correctement, selon moi – par « neben den wahrnehmbaren Größen ». 40 Ainsi Hett, Tricot, Hicks, Polansky.

L’ÂME COMME LA MAIN: DE ANIMA III 8

241

à laquelle lui-même n’adhère pas41. Mais je ne connais pas d’exemple d’un tel double jeu chez Aristote (qui d’ailleurs le considérerait sans doute comme une faiblesse facilement détectable)42. En fait, la difficulté disparaît dès lors qu’on comprend que παρὰ + Acc. signifie proprement « à côté de –  » ; c’est-à-dire, non pas simplement en dehors de – , mais en dehors de – et en même temps dans une relation de proximité spéciale avec – . La phrase signifie alors qu’il n’y a pas lieu de redoubler chacune des grandeurs sensibles par une entité prétendument « séparée », c’est-à-dire dotée d’une existence distincte43. Pour le dire autrement : considérons un univers d’objets S bien délimité, constitué – disons-le en première approximation – des objets d’une connaissance intellectuelle possible. La prémisse signifie que dans cet univers, il n’existe pas d’autres objets que des grandeurs sensibles (dans une acception sans doute assez large de cette expression, j’y reviens au § 13). Ainsi entendue, elle n’entre pas en conflit avec la doctrine du premier moteur de l’univers et n’est pas non plus dissonante avec la question de la connaissance du séparé. Aristote ne fait rien d’autre que rappeler son rejet des formes séparées des Platoniciens44. Il reste une difficulté pour cette interprétation, précisément la présence de l’incise ὡς δοκεῖ. Si la prémisse exprime une thèse aristotélicienne cardinale, pourquoi l’affecter de cette modalisation qui semble la signaler comme n’étant (au minimum) pas entièrement certaine ? Mais il n’est pas nécessaire 41 Ainsi Thillet, Bodéüs, Hicks et, semble-t-il, Rodier (p. 524). Thillet énonce très directement le présupposé sur lequel repose cette interprétation : « Pour l’opinion naïve, il n’y a que du sensible. Naturellement, Aristote ne saurait accorder cela purement et simplement. Il ne le peut que dans la mesure où cela le conforte dans son point de vue immanentiste : les formes sont immanentes au réel, et non pas transcendantes comme le veut le platonisme, il en tire ici argument » (Thillet, commentaire ad loc., p. 392-393 ; c’est moi qui souligne). 42 Par ailleurs, ἐπεὶ a une valeur nettement assertorique. Même si on tente de faire valoir que l’expression ὡς δοκεῖ, à l’inverse, traduit une certaine prise de distance (ce que je n’accorde pas, voir sur ce point le § 13), il reste que la proposition principale – exprimant la conclusion inférée de cette prémisse – n’est pas du tout modalisée (ἐν τοῖς εἴδεσι τοῖς αἰσθητοῖς τὰ νοητά ἐστι, et non pas ἔσται ou εἴη ἀν, par exemple), de sorte que l’on doit estimer que c’est bien l’intention assertorique qui l’emporte. 43 C’est ainsi, me semble-t-il, qu’il faut rendre l’emploi absolu de κεχωρισμένον en 432a4. 44 On pourrait exprimer la même idée autrement, en disant que la limitation à cet univers d’objets particulier est la suite normale du fait que le chapitre 8 se situe au niveau de l’âme, cf ci-dessus § 6.

242

MICHEL CRUBELLIER

de supposer que ὡς δοκεῖ modalise la prémisse tout entière45. Du fait de sa position en incise, entre τὰ μεγέθη et l’épithète τὰ αἰσθητὰ, on peut aussi envisager que cette référence à l’opinion concerne l’idée que les grandeurs dont on parle ici sont des grandeurs sensibles. En paraphrasant légèrement, cela donne : « il n’existe pas d’entités dotées d’une réalité indépendante à côté des grandeurs (, conformément à l’opinion commune : sensibles) ». C’est l’équivalent d’une sorte de principe de parcimonie : pour tout le monde, pour un jugement non prévenu, des « grandeurs » sont des entités sensibles ; et il n’est ni raisonnable ni nécessaire de postuler qu’il existe d’autres grandeurs, immatérielles ou non-sensibles. 12. La thèse de la seconde partie paraît s’inférer clairement et nécessairement de la prémisse moyennant quelques compléments et simplifications. On aurait ainsi la déduction suivante : (a) il n’y a aucune réalité dans l’univers S qui soit en dehors des grandeurs sensibles (432a3-4) (b) (c*) – les « intelligibles » ne sont pas en dehors des grandeurs sensibles (c) les « intelligibles » sont dans les formes sensibles (432a4-5)

La prémisse (a) est tirée du texte. La prémisse (b) est indispensable, mais devra être clarifiée en précisant ce que sont ces « intelligibles », et de quelle façon ils appartiennent à S. La conclusion (c) est elle aussi dans le texte ; mais ce qu’on peut déduire de (a) et (b) est en réalité (c*). On passerait de (c*) à (c), d’une part en admettant (ce qui à première vue ne fait pas difficulté) que si les intelligibles ne sont pas à l’extérieur ils sont à l’intérieur, et d’autre part en admettant qu’on peut passer des « grandeurs sensibles » de la ligne 4 aux « formes sensibles » de la ligne 5. 45 Il ressort des commentaires de Simplicius et de Philopon qu’il a dû exister dans l’Antiquité tardive des éditions dans lesquelles cette phrase était ponctuée ainsi : ἐπεὶ δὲ οὐδὲ πράγμα οὐθὲν ἔστι παρὰ τὰ μεγέθη / ὡς δοκεῖ τὰ αἰσθητά / κεχωρισμένον, de sorte que la modalisation ne portait pas sur la phrase entière. L’interprétation de Philopon (568.17-18) est aberrante ; celle de Simplicius (284.13-29) est plus intéressante : il pense que les sensibles (ou plus précisément l’opinion courante sur le statut des sensibles) sont introduits comme un exemple d’objets non séparés de la grandeur. On se retrouve ainsi assez près de l’interprétation que je défends ci-dessus ; mais en plaçant ainsi τὰ αἰσθητά dans l’incise, et donc en dehors de la protase, on ne peut plus comprendre la place centrale donnée au sensible dans l’apodose « les intelligibles sont dans les formes sensibles ».

L’ÂME COMME LA MAIN: DE ANIMA III 8

243

13. Je commence par ce dernier point. Les grandeurs sensibles sont les faits et les objets qui définissent l’univers S. Il est probable qu’il faut entendre ici la notion de « grandeur » en un sens assez large46 : on peut supposer que ce sont les diverses réalités sensibles qui peuvent faire l’objet d’une connaissance scientifique, physique ou mathématique. La possibilité de prendre comme prémisse une proposition portant sur cet ensemble de réalités signifie probablement que pour Aristote l’existence de ces connaissances scientifiques est un fait indiscutable, de même de l’affirmation que ces grandeurs peuvent être perçues – mais dans une moindre mesure, puisqu’il doit, sur ce point, faire appel à l’opinion générale. L’emploi du terme « grandeurs » à cet endroit s’explique d’ailleurs sans doute par le fait qu’il est neutre au regard de leur statut d’objets sensibles ou d’objets de connaissance intellectuelle. De ce fait, on pourrait choisir de les extraire du monde physique et de les traiter comme des réalités stables et absolues, indépendantes de la perception : c’est ce que fait Platon (ce qui le conduit à traiter leurs contreparties sensibles comme de simples images passagères). Aristote fait le choix inverse, de les maintenir dans le sensible (ce qui l’engage à montrer comment elles peuvent faire l’objet d’une connaissance intellectuelle)47. Les objets concernés sont des réalités singulières, qui ne sont pas toutes dotées d’une existence indépendante (c’est-à-dire « séparées » au sens de l’usage absolu du mot) : c’est le cas tout particulièrement des objets mathématiques48, mais sans doute aussi d’un grand nombre d’objets de 46 Dans la mesure où « la grandeur » semble ici inclure l’ensemble des objets d’une connaissance physique ou mathématique, on pourrait la considérer comme une sorte de représentant de toute la liste des sensibles communs. – Il est vrai qu’il n’y a pas de texte d’Aristote permettant d’appuyer cette supposition, mais on pourrait la justifier en remarquant que la grandeur est, parmi les sensibles communs, celui qui est le plus apte à être représenté comme un objet stable offert à la perception (c’est un caractère qu’elle partage avec la figure, mais celle-ci peut être analysée en éléments – côtés, angles, etc. – caractérisés chacun par sa grandeur, cf. 425a18). 47 La même explication peut valoir pour l’emploi de « grandeur » dans la question de la connaissance du séparé au chapitre 7. 48 Le terme μέγεθος sert couramment à désigner les grandeurs mathématiques continues (Mugler p. 280-281), mais Aristote l’emploie aussi, assez souvent, pour parler de la grandeur de phénomènes physiques (un exemple caractéristique dans Phys. IV 11, 219a10-18. – Par ailleurs je soutiendrai plus loin (§ 14) que les lignes 432a5-6 corroborent cette façon de définir l’extension de S. – La question du statut ontologique des êtres mathématiques chez Aristote demanderait un exposé assez long, mais on peut se contenter ici d’une observation simple : les objets naturels sont souvent caractérisés par

244

MICHEL CRUBELLIER

la science physique, qui sont des aspects ou des traits, ou encore des parties de certains objets naturels séparés, ou des événements affectant de tels objets. Il est cependant important que chacun de ces objets qui n’ont pas d’existence indépendante puisse être rattaché clairement à un ou des objets dotés d’une existence indépendante. L’autre formule (« les formes sensibles ») est plus floue. Ce qu’elle a en commun avec l’expression « grandeurs sensibles », c’est qu’elle vise les objets sensibles non pas directement en eux-mêmes (comme des entités concrètes ayant forme et matière et prises comme des touts, selon le sens ontologique de σύνολον49), mais en tant qu’ils peuvent être connus, reconnus ou pensés. Il y a cependant, entre ces deux formulations, deux différences significatives. D’une part, la notion de « grandeur » est a priori plus restrictive que celle de « forme » : toute grandeur sensible a nécessairement une forme ; mais le terme de « grandeur » opère une sélection parmi les différents objets sensibles et leurs différents aspects. Il n’en retient que ce qui est scientifiquement connaissable, alors qu’a priori la notion de « forme sensible » convient à tout ce qui peut être perçu. On rencontre ici la seconde différence : dans « grandeurs sensibles », « sensible » est une détermination qui s’ajoute à la notion de grandeur (« ὡς δοκεῖ » met ce point en évidence, en soulignant que l’attribution du prédicat « sensibles » demande un acte spécial de jugement), alors que dans l’expression « formes sensibles », les deux termes sont dans une relation essentielle : ces formes sont les formes de l’êtresensible des sensibles, elles sont ce par quoi ils sont sensibles. Il y a deux façons de concevoir de telles formes sensibles : ou bien c’est la « forme sensible » propre de chaque réalité perceptible50, ou bien ce sont les formes qui structurent la possibilité de percevoir. La première interprétation s’accorderait avec la formule du chapitre II 12 des déterminations mathématiques ou mathématisables qui leur sont essentielles (la Lune est sphérique, l’être humain a deux mains et trente-deux dents, etc.) ; mais, considérés en eux-mêmes, la sphère ou les entiers naturels 2 et 32 ne sont pas essentiellement liés à la Lune ou à la structure de l’être humain, et à ce titre ils n’en sont pas « des manières d’être ou des propriétés ». 49 Le terme σύνολον ne figure pas dans le seul passage ontologique de l’Âme, le chapitre II 1 sur la définition de l’âme, dans lequel cette notion est désignée comme « la substance au sens du composé » (οὐσία οὕτως ὡς συνθέτη, 412a16) et comme « ce qui est de la forme et de la matière » (τὸ ἐκ τούτων, 412a9) 50 Il conviendrait de dire « les formes sensibles », car beaucoup d’objets auront une forme visuelle, une forme auditive, etc.

L’ÂME COMME LA MAIN: DE ANIMA III 8

245

(« la perception est ce qui a la capacité de recevoir les sensibles sans leur matière »51), la seconde plutôt avec la description de la perception comme « forme de sensibles » lors de la comparaison avec la main. J’ai une préférence pour la seconde, notamment parce qu’elle assure une meilleure continuité entre les deux parties du chapitre 8 ; et aussi parce qu’on peut en tirer une conception intéressante (plus riche) de la relation entre l’intelligence et la perception (§ 15 et 16). Les deux interprétations ne sont d’ailleurs pas incompatibles. 14. Les « intelligibles » (τὰ νοητά) de 432a6 sont assez bien définis, même si c’est seulement en extension, par la mention de deux sousclasses : d’un côté les termes qui sont « dits par abstraction » (τὰ ἀφαιρέσει λεγόμενα), c’est-à-dire constitués par un acte de dénomination spécifique52, qui isole certaines propriétés sans que celles-ci appartiennent essentiellement à une classe d’êtres bien déterminée ; de l’autre, « ceux qui sont des manières d’être ou des propriétés des sensibles » (ὅσα τῶν αἰσθητῶν ἕξεις καὶ πάθη). On retrouve ici l’extension de la notion de grandeur entendue dans le sens large envisagé ci-dessus (§ 12)53. Il est assez naturel, au fond, que ces deux notions coïncident (à savoir, celle des grandeurs sensibles et celle des « intelligibles » de 432a6), puisque la discussion porte sur l’existence et le statut (séparé ou non-séparé) d’entités qui incarnent ce qui est scientifiquement connaissable dans les grandeurs sensibles (mais naturellement, la signification des deux expressions, ou le point de vue à partir duquel elles visent leurs référents, ne sont pas les mêmes). On pourrait à bon droit compléter : Mais puisqu’en fait il n’existe aucun objet doté d’une existence indépendante à côté des grandeurs (…) sensibles, les intelligibles sont dans les formes sensibles. (432a3-5) 51 Surtout si on accepte le texte de e et des commentateurs anciens : τὸ δεκτικὸν τῶν αἰσθητῶν εἴδων. 52 C’est ainsi qu’il faut entendre « définis par abstraction » dans ma traduction. 53 On remarquera au passage que le texte ne mentionne pas une possible troisième classe d’intelligibles (celle à laquelle, dira-t-on peut-être, on devait le plus s’attendre), à savoir les essences des espèces et des genres (humain ou animal, par ex.), qui correspondraient si bien aux Formes platoniciennes les plus éminentes. On peut admettre qu’Aristote, s’exprimant d’une façon relâchée, les inclut dans la catégorie des « manières d’être et propriétés ». Si au contraire on veut soutenir que c’est à dessein qu’il n’en parle pas ici, j’avoue que je n’ai pas d’explication claire de ce choix. Il faudrait peut-être le rapprocher de la façon dont, à la fin du chapitre, il place à part les « intelligibles premiers », qui sont des termes simples pensés en eux-mêmes, par opposition aux jugements (cf. § 18).

246

MICHEL CRUBELLIER

Les « intelligibles » sont donc les contenus intelligibles qui correspondent aux « grandeurs sensibles », et à ce titre ils font partie de l’univers S, de sorte que l’ajout de notre prémisse complémentaire (b) est légitime. Il reste à préciser en quel sens on peut dire qu’ils sont « dans les formes sensibles ». Si on s’en tient à la signification la plus faible de ἐν + Dat., c’est-à-dire le fait d’être situé quelque part, c’est assez facile : dire que les intelligibles sont « dans » les formes sensibles, c’est dire qu’on ne peut pas y avoir accès autrement qu’à travers ces formes – et donc à travers la perception de certains objets, comme on peut d’ailleurs le voir avec les faits qu’Aristote introduit à titre de confirmation dans les lignes suivantes (432a7-10). C’est plus difficile si on veut donner à ἐν + Dat.54 la valeur plus forte de contenu dans quelque chose. Il me paraît impossible de soutenir que les intelligibles sont une partie ou un trait des formes sensibles, si du moins on entend par là une partie ou un trait existant en acte ; car alors il suffirait de saisir la forme sensible pour avoir en même temps la connaissance intellectuelle correspondante : par exemple, le contenu intelligible que représente la connaissance d’une éclipse serait donné dans la « forme sensible » de l’éclipse. Or il est certain que pour Aristote la connaissance scientifique contient plus et autre chose que la simple saisie de cette forme sensible55 (la reconnaissance de la cause du phénomène et de sa nécessité, voire de sa conformité à une régularité générale et universelle56). Il faut donc penser que les intelligibles sont contenus « dans » les formes sensibles, c’est en puissance ; et que l’activité propre de l’intelligence consiste à les en dégager – ou, plus précisément, à élaborer des 54 On ne trouve pas dans le corpus d’analyse méthodique et complète des différents sens et usages de ἐν ; ce qui s’en approche le plus, ce sont deux passages du livre IV de la Physique  : IV 3, 210a14-24 (dans l’analyse du lieu) et IV 12, 221a12 sq. (dans une discussion sur le temps). Il semble qu’Aristote y privilégie la valeur contenu dans. – A moins de considérer que le sens de matière ou condition de possibilité (210a20-21) se rapproche de l’interprétation que j’avance ici pour la valeur de situation : il ne serait pas absurde de dire que les formes sensibles sont la matière de l’intellection. 55 Même si on entend par là l’expérience d’une perception précise et instruite – par exemple la perception d’un humain qui reconnaît que ce qui se passe est une éclipse (par opposition à celle d’une bête qui assiste simplement à l’obscurcissement temporaire de la pleine Lune). 56 Voir A.Po. I 31, 87b38-88a5, et II 2, 86a26-28 (même dans ce second texte, qu’on oppose parfois à celui de I 31, il y a un passage de la perception du fait à la reconnaissance de sa cause).

L’ÂME COMME LA MAIN: DE ANIMA III 8

247

connaissances rationnelles à partir d’éléments contenus dans les formes sensibles. Cette interprétation permet de proposer l’esquisse d’un modèle des rapports entre la perception et l’intelligence et du développement de celle-ci à partir de celle-là. 15. Les bases de ce modèle sont les suivantes : – Il y a une analogie d’ensemble entre la capacité de percevoir et la capacité de connaissance intellectuelle ; chacune des deux a son entéléchie première dans des structures (appelées « formes » en 431b28-432a3) qui existent à la fois dans l’âme et dans les objets (respectivement « sensibles » et « intelligibles ») ; – il y a une asymétrie, ou une coordination, à l’intérieur de cette analogie : la perception met en forme – et rend ainsi connaissables – les événements physiques qui résultent du contact entre des faits physiques externes et les appareils perceptifs d’un animal ; l’intelligence met en forme des connaissances qui résultent de l’activité perceptive (mais pas uniquement de l’activité perceptive présente). La formule qui clôt la première partie du texte : ...καὶ ὁ νοῦς εἶδος εἴδων καὶ ἡ αἴσθησις εἶδος αἰσθητῶν (432a2-3) exprime à la fois l’analogie et l’asymétrie57. Il convient de souligner que dans cette interprétation les connaissances issues de l’activité perceptive ne sont pas simplement des contenus (tels que les différentes qualités sensibles propres de cette cerise individuelle, par exemple), mais aussi des relations d’ordre et des principes de mise en ordre : les sensibles communs sont perçus, et il ne fait pas de doute que pour Aristote beaucoup de bêtes58 perçoivent l’avant et l’après, le dedans et le dehors, le beaucoup et le peu, le semblable et le différent, etc., et reconnaissent sans doute aussi des types d’objets qui sont comme 57 C’est une raison de plus de résister à la tentation de faire de la perception aussi une forme de formes, et d’accepter la relative étrangeté de εἶδος αἰσθητῶν (même chose pour le passage déjà cité du début de II 12). 58 Même s’il existe aussi des animaux très imparfaits qui ne sont sensibles qu’à une qualité sensible propre, ainsi les insectes qui sont attirés par la lumière ou qui au contraire la fuient, etc. Aristote fait référence à ces animaux au début du chapitre III 11 et remarque que, de même qu’ils se meuvent « de façon indéfinie » (ἀορίστως), ils doivent connaître des formes « indéfinies » de désir et de représentation (433b31-434a5).

248

MICHEL CRUBELLIER

des proto-universaux. Cette conception est indispensable59 à la construction grâce à laquelle Aristote écarte la thèse de la réminiscence dans le dernier chapitre des Seconds Analytiques. 16. Mais avec cette conception, ne court-on pas le risque d’effacer tout simplement la distinction entre la connaissance perceptive et la connaissance intellectuelle ? alors que dans tant d’autres passages60 Aristote insiste au contraire sur la nécessité de cette distinction et sur l’autonomie de l’intelligence. Où est donc le critère ? À partir de quel niveau les principes d’ordre cessent-ils d’être immanents à la perception pour devenir des normes propres à l’intelligence ? C’est ainsi que Seconds Analytiques II 19 a troublé beaucoup de commentateurs, parce qu’Aristote commence par décrire la genèse de la connaissance des principes à partir de la perception, de la mémoire et des images et qu’il finit par écrire que cette connaissance ne peut être conçue que comme un acte de l’intellect. – Mais on trouve, dans ce chapitre même, l’indication très précise d’un point d’articulation entre la connaissance sensible (au sens large) et la connaissance intellectuelle : De l’expérience, ou de l’universel fixé entièrement dans l’âme (ἐκ πάντος ἠρεμήσαντος τοῦ καθόλου ἐν τῇ ψυχῇ), de cet un à côté des multiples, qui se retrouve identique dans toutes ces choses-là, vient le principe (ἀρχή) de l’art et de la science. (100a6-8)

Le point culminant de l’ἐμπειρία, c’est « l’universel fixé entièrement dans l’âme » – par opposition à ce qui serait une reconnaissance approximative et passagère comme on peut souvent l’observer dans le comportement des bêtes. Ce point extrême de la connaissance de type sensible est aussi le point de départ (ἀρχή) de la science, à partir du moment où il est reconnu et pensé à l’aide de normes propres à l’intelligence, à savoir, ici, l’identité et l’universalité : « l’universel fixé (ou : « en repos ») dans l’âme » est significativement repris par la formule : « ce qui se retrouve identique dans toutes ces choses ». 59 On retrouve ici l’alternative posée à la fin du § 13 ; si on choisit de concevoir les « formes sensibles » comme la forme propre de chaque expérience perceptive, l’activité de l’intelligence apparaîtra miraculeuse et en tout cas mystérieuse, car nous n’aurons aucune indication sur la nature et l’origine des pouvoirs qui lui permettent, à partir de nos expériences perceptives particulières, d’accéder à une connaissance vraie, universelle et nécessaire. 60 DA I 4, 408b18-19 ; II 1, 413a7-9 ; III 4, 429a29-b5 ; cf. aussi Met. Λ 3, 1070a25-26.

L’ÂME COMME LA MAIN: DE ANIMA III 8

249

On pourrait formuler un peu autrement l’objection précédente et sa solution, en revenant à l’analogie avec la perception. Ce qui fait que l’animal perçoit les qualités des objets physiques au lieu de simplement subir leur action (comme font les plantes), c’est que ses appareils sensoriels possèdent une médiété innée. – Qu’est-ce donc qui jouera le même rôle dans le cas de l’intelligence ? – On peut conjecturer que ce sont des formes de pensée dotées d’une force normative, c’est-à-dire que l’être humain possède une sorte de sens inné de l’identité, de l’universalité et de la nécessité (à quoi il faut sans doute ajouter notre intérêt propre pour la vérité, j’y reviens au § 18). Pour prolonger cette conjecture, on peut encore concevoir que la mise en œuvre de telles formes soit le fait de « l’intellect agent », comme le suggère la fameuse comparaison avec la lumière : … ὁ δὲ τῷ πάντα ποιεῖν, ὡς ἕξις τις, οἷον τὸ φῶς· τρόπον γάρ τινα καὶ τὸ φῶς ποιεῖ τὰ δυνάμει ὄντα χρώματα ἐνεργείᾳ χρώματα. … une intelligence qui est parce qu’elle produit toutes choses, à la façon d’une disposition, comme la lumière – et en effet la lumière, d’une certaine façon, transforme en couleurs en acte des couleurs qui n’existent qu’en puissance. (III 5, 430a15-17)

La perception des bêtes les plus parfaites peut comporter un réseau de relations d’ordre entre des objets perçus et mémorisés, la constitution de quasi-universaux et un certain nombre d’anticipations. Ces bêtes sont naturellement capables de les utiliser ; mais on peut dire qu’au milieu de ces connaissances virtuelles, elles circulent comme à tâtons dans l’obscurité ; alors que l’intelligence éclaire d’un seul coup l’ensemble du décor. 17. Il n’est sans doute pas possible – en tout cas, pas ici – d’aller plus avant sur une question à propos de laquelle Aristote lui-même n’a donné que quelques indications discontinues ; je reprends donc la lecture du chapitre 8. Les dernières lignes (432a7-14) avancent deux confirmations à l’appui de la thèse de la seconde partie. Elles ne sont pas de même nature. Alors que la seconde énonce un fait d’expérience (de notre expérience subjective de sujets pensants), la première décrit une expérience de pensée : « quelqu’un qui ne percevrait rien ne pourrait rien apprendre ni rien comprendre ». Il paraît clair que la supposition d’un être humain

250

MICHEL CRUBELLIER

qui ne percevrait absolument rien est contrefactuelle ; pour le reste, comme les commentateurs le relèvent généralement, cette thèse reprend des affirmations faites ailleurs, sous une forme plus limitée61. La seconde confirmation est plus intéressante, car elle soulève une question et donne ainsi l’occasion d’une dernière mise au point sur la nature propre de la connaissance intellectuelle : « lorsqu’on réfléchit, il faut nécessairement que l’on fixe en même temps son attention sur une image » (ὅταν τε θεωρῇ, ἀνάγκη ἅμα φάντασμά τι θεωρεῖν, 432a8-9). – On ne doit d’ailleurs pas exclure que la classe de ces « images » sur lesquelles peut s’appuyer la pensée théorique ne se limite pas à des représentations d’objets sensibles absents, mais qu’elle inclue aussi des schémas qu’on pourrait dire abstraits, comme lorsqu’on utilise une spatialité conventionnelle pour se représenter les relations entre des termes de logique formelle ou d’ontologie. Plus que tout le reste de notre expérience perceptive, les catégories spatiales offrent des modèles à la fois suffisamment universels et suffisamment variés, qui se prêtent à une représentation objective précise, y compris pour des relations extrêmement abstraites ; c’est peut-être pour cette raison qu’Aristote juge que notre intelligence est « inséparable de la grandeur »62. Le fait que « l’âme ne pense (νοεῖ) jamais sans image » a déjà été mentionné au chapitre 7 (431a16-17), avec cette justification que « les images sont, pour l’âme qui réfléchit, comme des perceptions ». On peut compléter l’idée exprimée dans ce passage, en s’appuyant sur les lignes qui précèdent (431a8-14) : les images jouent pour l’âme qui réfléchit le rôle tenu, pour l’âme qui agit, par certaines perceptions, c’est-à-dire que lorsque l’animal éprouve du plaisir ou de la douleur, il apprend à poursuivre ou à fuir les perceptions correspondantes ; et lorsqu’il agit d’une âme (humaine) qui recherche la connaissance, il en va de même en ce qui concerne le vrai et le faux – avec cette différence que la recherche 61 Avant tout A.Po. I 18, 81a38 : si une perception manquait, nécessairement un savoir scientifique manquerait aussi, qu’il nous serait impossible d’acquérir – il faut entendre par là, comme le note Pellegrin (p. 376, n. 1), la perception d’un objet ou d’un fait particulier ; voir aussi Met. A 9, 993a7 : « les choses dont il y a sensation, comment quelqu’un qui ne possède pas la sensation pourrait-il les connaître ? » 62 Ce n’est pas, chez Aristote, une position simplement théorique. L’intelligibilité spécifique des rapports spatiaux et son corrélatif, l’intelligence propre de la vision, jouent un rôle central dans les traités biologiques, comme A.L. Carbone (2011) l’a montré récemment dans un livre admirable. Il y a là une perspective de recherche qui pourrait être étendue avec profit à d’autres domaines.

L’ÂME COMME LA MAIN: DE ANIMA III 8

251

de la vérité ne se borne pas à des situations actuelles données, mais s’étend à tous les objets de connaissance possibles. On trouve dans III 8 un rappel de cette justification, mais sans référence (du moins pour commencer) à la recherche de la connaissance : « les images sont comme des perceptions, avec cette différence qu’elles sont sans matière » (ὥσπερ αἰσθήματά ἐστι, πλὴν ἄνευ ὕλης, 432a910). Cette remarque peut surprendre parce qu’elle sous-entend que les perceptions comportent une matière, alors que le passage classique, déjà cité, de II 12, 424a18-19, pose que la perception « reçoit les sensibles sans leur matière ». Si on ne veut pas admettre qu’il y a là un simple relâchement de l’expression, induit par des contextes différents, il faudra dire que les perceptions sont toujours données avec une matière, ainsi que je l’ai fait dans ma traduction. 18. Cette justification conduit Aristote à s’interroger sur la différence entre l’imagination (φαντασία, l’activité63 qui produit des images, φαντάσματα) et l’intelligence – ce qui est une façon d’aborder indirectement l’objection soulevée au § 16 : qu’est-ce qui distinguera la connaissance proprement intellectuelle de la connaissance de type perceptif, représentée ici par les φαντάσματα ? Il répond à la question en se référant à une activité caractéristique de l’intelligence : énoncer une proposition64 et ainsi s’engager en direction du vrai. – Mais Aristote attribue aussi aux images la capacité d’être dites vraies ou fausses65 (même si elles sont plus souvent fausses que vraies). – Oui, mais elles sont vraies ou fausses, pourrait-on dire, simplement de fait. Aristote leur attribue les étiquettes vrai ou faux, mais refuse de leur appliquer les verbes « dire vrai » ou « dire faux » (ἀληθεύεσθαι / ψεύδεσθαι) : il envisage cette possibilité en 428a1-4, mais conclut que l’imagination ne peut ni être l’une des puissances qui disent le vrai ni dériver de puissances 63 Le parallèle avec φάσις / ἀπόφασις confirme que c’est bien l’activité d’imaginer qui est visée ici. 64 Selon EN VI, 1139b14-17, c’est un acte que le νοῦς partage avec d’autres puissances de l’âme : la compétence technique (τέχνη), la science (ἐπιστήμη), la prudence (φρόνησις) et la sagesse (σοφία) ; mais nous avons vu que le chapitre III 8 considère une conception large de l’intelligence, qui inclut au moins la science et peut-être d’autres (par ex. la compétence technique) ; surtout, une telle activité n’est pas attribuée à l’imagination. 65 L’affirmation est répétée à plusieurs reprises au cours du chapitre III 3 : 428a1-4, 428a11-12, 428a18, 428b2-9.

252

MICHEL CRUBELLIER

qui disent le vrai, et cela probablement parce qu’elle ne dit pas. Quand nous disons qu’une image est « fausse » ou « vraie », cette qualification porte en fait sur un jugement que nous portons à partir de cette image. De sorte que la capacité de dire vrai est un critère pertinent pour distinguer la connaissance intellectuelle de l’imagination. Cette capacité comprend deux éléments66 : d’un côté une visée ou une intention spécifique, que j’ai désignée ci-dessus comme notre « engagement » en direction de la vérité, de l’autre la forme linguistique ou logique du jugement67 comme une synthèse d’un type particulier, qu’Aristote nomme ici συμπλοκή, « combinaison » et dont il a déjà indiqué au chapitre III 6 (430a27-b6) qu’elle est l’œuvre de l’intelligence. L’argument du chapitre 8 part de ce second élément : « l’acte de former une image est autre chose que celui de déclarer ou de nier quelque chose, car le vrai et le faux consistent dans une combinaison de termes intelligibles ». Mais on peut alors soulever une objection (a12-14) : si la différence entre l’intelligence et l’imagination tient au fait que l’intelligence énonce des jugements, ne pourra-t-on pas dire malgré tout que les termes qu’elle combine (les « intelligibles premiers », πρῶτα νοήματα) ne sont pas réellement différents des images ? Même le critère de la combinaison de termes pourrait ne pas être discriminant, puisque dans certains cas (ainsi l’exemple classique du diamètre apparent du Soleil, III 3, 428b2-4) la fausseté de l’image est en fait celle d’un jugement prédicatif68 dont les deux termes (le sujet, le Soleil, et le prédicat, large d’un pied) sont donnés dans l’image. La réponse est que la forme d’unité étroite que représente la « combinaison » ne peut pas exister entre des images : on peut bien assembler des images entre elles, comme lorsqu’on imagine ou représente un bouc-cerf ou un centaure, mais le résultat de cet assemblage est une autre image, qui 66 Ces deux constituants pourraient être désignés comme étant, l’un, l’élément éthique, et l’autre l’élément logique de l’activité de l’intelligence. – L’existence de l’élément éthique permet de comprendre l’importance des passages où Aristote établit un lien entre les deux couples {affirmer, nier} et {poursuivre, fuir} (III 7, 431b10-12 ; cf. EN VI, 1139a27-29). 67 J’admets, sans entreprendre de le justifier ici, que l’emploi de vrai pour caractériser des termes simples est un emploi marginal ou dérivé (cf. Met. Θ 10, 1051b17-32 ; E 4, 1027b18-29). Ce point de vue peut d’ailleurs trouver une certaine confirmation dans la marche de l’argumentation à la fin du chapitre 8, puisque celle-ci va du jugement vrai au « terme intelligible », et non l’inverse. 68 Et non pas simplement des jugements d’existence, comme lorsqu’on dit qu’il existe ou n’existe pas de centaures ou de bouc-cerfs.

L’ÂME COMME LA MAIN: DE ANIMA III 8

253

n’est pas plus susceptible de dire vrai ou faux que celles dont elle est composée. Il faut donc admettre qu’aucun des termes simples sur lesquels portent les actes de l’intelligence – pas plus ceux que nous considérons isolément que ceux qui sont engagés comme sujet ou prédicat dans une proposition que nous affirmons – ne peut être une simple image : Τὰ δὲ πρῶτα νοήματα τί διοίσει τοῦ μὴ φαντάσματα εἶναι; ἢ οὐδὲ τἆλλα φαντάσματα, ἀλλ’ οὐκ ἄνευ φαντασμάτων. Mais les premiers termes intelligibles, en quoi se distingueront-ils de ce qui n’est qu’une image ? – À moins de dire qu’aucun terme intelligible, en général, n’est une image, mais qu’ils ne vont pas sans images. (432a12-14)

À la ligne 432a13, je conserve le texte donné par les manuscrits : τἆλλα, au lieu de corriger en ταῦτα comme le font bon nombre d’éditeurs et de traducteurs modernes (et déjà médiévaux). Ce qui les a gênés est apparemment qu’ils ne voyaient pas à quels « autres νοήματα » pourrait renvoyer ἄλλα : ce ne sont pas les πρῶτα νοήματα qu’on vient de mentionner, mais cela pourrait difficilement être les intelligibles complexes que représentent les jugements (pour lesquels la confusion avec des images serait difficilement concevable). Cependant, le remplacement de τἆλλα par ταῦτα aboutit, comme le remarque Hicks (p. 547), à attribuer à Aristote un argument indigent puisque, pour éviter que les intelligibles simples ne soient confondus avec des images, il se contenterait d’affirmer que « ce ne sont pas des images ». Il est plus simple de supposer qu’on a ici la valeur classique de ἄλλα qui indique que le cas qu’on vient de considérer s’inscrit dans une classe plus large, incluant les jugements : c’est l’interprétation de Hicks. Le point faible de celle-ci est celui que je viens d’indiquer : a priori on ne serait pas tenté de réduire les jugements à des images. Mais on pourrait aussi penser que, par τὰ ἄλλα, Aristote se réfère à d’autres termes simples, ceux que l’intelligence considère isolément (par opposition à ceux qui constituent les éléments d’une proposition) et qu’il entend ainsi souligner que les « images » sur lesquelles l’esprit qui réfléchit paraît « fixer son attention » sont en fait les supports d’une activité propre et autonome de l’intelligence.

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III 9 LES PARADOXES DE L’ÉLIMINATION Pierre-Marie MOREL

Il arrive que certains textes d’importance apparemment secondaire, des textes de passage comme on en trouve très souvent chez Aristote, aient des conséquences théoriques inattendues, y compris lorsque leurs conclusions explicites sont négatives. C’est le cas de la section III 9, 432b7-433a8 du Traité de l’âme. Au début de ce chapitre, dans les lignes qui précèdent le passage en question, Aristote commence par rappeler les deux éléments de base de la définition de l’âme animale. En tout animal, il y a (1) une faculté discriminante ou « critique », qui peut être la pensée (διάνοια) ou la sensation ; et (2) une faculté « par laquelle s’accomplit le mouvement local » (432a15-17). Estimant en avoir assez dit sur le principe critique, parce qu’il a traité longuement de la sensation et de l’intellect, il envisage maintenant le principe moteur. Il se demande donc « quelle de l’âme il peut bien être » (τί ποτέ ἐστι τῆς ψυχῆς)1. Deux possibilités se présentent : ou bien (2.1) il s’agit d’une « unique partie de l’âme » (ἕν τι μόριον αὐτῆς)2, ou bien (2.2) il s’agit de l’âme entière. La première hypothèse se dédouble : cette partie peut être séparable ou isolable, ou bien en grandeur (2.1.1), c’est-à-dire physiquement, ou bien conceptuellement, en raison (2.1.2). Aristote ajoute en 432a17-22 une alternative subsidiaire : ou bien (2.1a) le principe moteur est l’une des parties qu’on évoque habituellement et dont on a déjà parlé, ou bien (2.1b) il s’agit d’une partie spéciale qui n’est pas comprise dans cet ensemble. 1 L’ajout du terme « partie » est ici une pure commodité de traduction. Comme on va le voir, il ne doit pas donner à penser qu’Aristote impute la responsabilité du mouvement à une « partie » psychique distincte. 2 Voir DA II 2, 413b14-15.

256

PIERRE-MARIE MOREL

Ainsi, pour identifier le principe moteur, on suppose deux choses : tout d’abord, qu’il y a une faculté distincte de la faculté critique ; ensuite, qu’il n’y a pas d’autre alternative, pour identifier la partie motrice, que d’imputer le mouvement, soit à l’âme tout entière (2.2), soit à une partie de celle-ci (2.1). On peut déjà écarter l’hypothèse (2.2), puisque certains animaux, on le verra, n’ont pas le mouvement local. Reste donc l’hypothèse d’une partie spéciale (2.1), à laquelle nous pourrions imputer la responsabilité de ce type de mouvement. Or cela pose incontestablement problème : « mais cela pose directement la difficulté de savoir en quel sens il faut parler de parties de l’âme et combien il y en a » (ἔχει δὲ ἀπορίαν εὐθὺς πῶς τε δεῖ μόρια λέγειν τῆς ψυχῆς καὶ πόσα) (432a22-23). Suit un véritable réquisitoire contre l’idée même de « parties » de l’âme (432a23-b7) : elles sont impossibles à dénombrer, difficiles à classer et difficiles également à diviser. L’âme ne doit pas être mise en pièces. L’allusion polémique à Platon est manifeste. Celui-ci nous a égarés en expliquant le fonctionnement de l’âme par ses divisions, et en rapportant chaque faculté à des parties distinctes opérant par elles-mêmes3. La critique sera reprise en III 10, 433b1 sq., mais Aristote a déjà eu l’occasion de rejeter l’hypothèse de parties distinctes en II 24. Il maintenait une séparation « en raison » des parties, ou plus exactement des facultés ou puissances de l’âme, justifiant ainsi qu’on les isole ou les sépare, non pas réellement, mais conceptuellement. La critique de la fragmentation de l’âme ne porte donc pas simplement contre une éventuelle séparation corporelle des parties  : elle concerne la définition et l’organisation des facultés. Pour le dire autrement, Aristote ne se contente pas de remettre en cause une moriologie corporelle ou physique de l’âme (la thèse de la séparation « en grandeur ») au profit d’une théorie des facultés, dans laquelle ces dernières seraient encore fonctionnellement distinctes les unes des autres. Il montre qu’une telle théorie serait également inadéquate. Toutefois, DA III 9 va plus loin, en rejetant également l’idée selon laquelle chaque opération correspondrait nécessairement à une faculté et une seule. Comme je voudrais le montrer, il y a ici 3 Voir par exemple D.W. Hamlyn (1993), p. 150 ; T.K. Johansen (2012), p. 246-247 ; C. Shields (2016), p. 349, qui note qu’Aristote engage ici le débat, non seulement avec Platon, mais également avec sa propre conception de l’âme bipartite, telle que l’on peut la déceler dans les traités de philosophie pratique. 4 DA II 2, 413b29.

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

257

un véritable changement de paradigme, de sorte que nous serons nécessairement dans l’impasse pour répondre à la question initialement posée, à savoir : à quelle partie ou faculté de l’âme devons-nous imputer en propre le mouvement local ? Il faut pour cette raison tout reprendre depuis le départ pour aborder enfin le cœur de la question5. C’est ce qui se passe en 432b7, ainsi que le montre la proposition de transition καὶ δὴ καὶ περὶ οὗ νῦν ὁ λόγος ἐνέστηκε, que l’on peut rendre, avec R. Bodéüs, par « ainsi donc, pour revenir au propos où l’on était resté à l’instant (…) »6, phrase qui indique à la fois une progression de l’argument et la reprise d’un propos antérieur. La section qui commence alors (III 9, 432b7-433a8), bien qu’elle paraisse au premier abord peu intéressante, essentiellement négative et purement transitoire – surtout si on la compare aux pages, décisives, qui suivent aussitôt et qui constituent le chapitre 10 –, a précisément pour fonction de rectifier les termes du débat et de poser d’une manière enfin adéquate la question de l’origine psychique du mouvement local. On pourrait soutenir qu’en un sens l’ὄρεξις est la première faculté motrice et qu’elle peut donc être identifiée au principe moteur7. L’insistance du chapitre 10 sur le rôle du désir invite d’ailleurs à aller dans cette direction. On verra cependant que, selon les dernières lignes de notre texte (et le chapitre 10 le confirmera), ce ne peut être qu’une solution partielle : le désir a bien un rôle de premier plan dans la motricité animale, mais il ne suffit pas à la produire par lui-même ou en propre8. Plus fondamentalement, la question, telle qu’elle est implicitement posée en III 9, oblige à reformuler le problème des parties ou facultés. Elle demande en effet que l’on substitue, à une conception distributive des facultés (à telle faculté, correspond telle opération ou action9), une conception coopérative ou combinatoire des facultés (à telle faculté, correspond telle fonction dans une opération ou action complexe mobilisant plusieurs facultés). Selon ce nouveau paradigme, les facultés œuvrent, non pas séparément, mais en 5

G. Rodier (1900), p. 533. R. Bodéüs (1993). Dans le même esprit, voir la traduction de E. Barbotin (1966) : « mais revenons donc à l’objet présent de notre propos ». 7 Par exemple D.W. Hamlyn (1993), p. 152 : « desire is the cause of movement in animals. » 8 Voir R. Polansky (2007), p. 526. 9 Au sens où l’exercice d’une fonction naturelle peut être désignée, chez Aristote, par le terme πρᾶξις. 6

258

PIERRE-MARIE MOREL

vue d’une opération ou action globale10 : elles sont fonctionnellement corrélées. Ainsi, pour reprendre le dispositif de départ, on devra dire que les deux facultés initiales, la faculté critique (1) et la faculté motrice (2), bien qu’on puisse les distinguer, ne sont pas dissociées en fait et que la seconde implique aussi la première, de sorte qu’on ne pourra plus se contenter de l’alternative initiale, à savoir : (2.1) ou (2.2). Il faudra envisager une troisième solution (2.3) : la motricité s’explique par un concours de parties et de facultés. Pour l’établir, Aristote recourt ici à une méthode d’élimination qu’il pratique volontiers, et dont il attend généralement des résultats positifs. C’est le cas par exemple en Éthique à Nicomaque I 6, 1097b34 sq., quand il s’agit d’identifier la vie la meilleure pour l’homme, et où il apparaît qu’elle consiste en une activité de l’âme en accord avec la vertu la plus élevée. Dans le Traité de l’âme lui-même, il l’emploie en III 3, 428a1-429a2 à propos de la définition de la φαντασία : elle n’est ni sensation, ni opinion accompagnée de sensation, ni science ou intellect, ni une combinaison (συμπλοκή : 428a25-26) d’opinion et de sensation. Il en résulte, par élimination, qu’elle est « le mouvement qui se produit sous l’effet de la sensation en acte »11. La comparaison avec ce dernier passage est doublement instructive. Elle montre en effet, en premier lieu, qu’une faculté ou un type de mouvement cognitif peut être rapporté à une unique faculté, ou bien rapporté à une « combinaison » de facultés. L’idée que la capacité motrice de l’animal dépende d’une telle coopération, au lieu d’être attribuée à une unique faculté, représente donc un cas de figure parfaitement admissible. Le texte de DA III 3, en second lieu, donne l’exemple d’une élimination positive, en produisant une définition correcte par épuisement d’hypothèses successives. Sur ce point, la comparaison conduit à opposer les deux textes. En III 9, on éliminera toutes les facultés ou parties qui ne sont pas par elles-mêmes motrices, mais la procédure, dans ce cas précis, ne permettra pas d’identifier un cas heureux, qui résisterait à l’élimination en excluant tous les autres. On devra conclure, au contraire, à l’absence de tout cas qui résisterait totalement à 10 Ainsi qu’Aristote l’explique à propos de la vie ou action complète de l’organisme, qui est la résultante et la fin de toutes les « actions » particulières des parties qui le composent. Voir PA I 5, 645b14-20 et P.-M. Morel (2007), p. 158-163. 11 DA III 3, 429a1-2. Je remercie Klaus Corcilius pour avoir attiré mon attention sur les similarités méthodologiques entre ce texte et celui de III 9. Elles ont été également remarquées par C. Shields (2016), p. 349.

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

259

l’élimination : aucune faculté ou partie de l’âme n’échappe à cette dernière, parce qu’aucune n’est cause par elle-même du mouvement local. Notre passage va donc excéder son modeste propos initial : il n’a pas seulement pour objet de mettre à l’écart les facultés que l’on pourrait à tort supposer motrices par elles-mêmes, mais encore de contribuer à une certaine idée de l’unité de l’âme, au-delà de ses différentes facultés ou fonctions12. La théorie du mouvement animal, telle qu’elle sera présentée dans la suite du traité, présuppose la méthode d’élimination paradoxale qui est ici à l’œuvre, ainsi que la théorie des facultés qu’elle implique13. 1. L’objet de la recherche : le moteur du mouvement local proprement dit (432b7-14) Aristote commence par délimiter l’objet qui est en discussion. En premier lieu, on recherche, non pas seulement un organe auquel attribuer le mouvement, mais plus vaguement un principe moteur, un κινοῦν. Rien ne dit en effet que le moteur soit une seule faculté, une seule partie ou un seul organe. Il se peut, comme on l’a dit et ainsi que la suite le confirmera, que le moteur de l’animal soit un principe complexe impliquant plusieurs facultés et plusieurs organes. En second lieu, ce n’est pas tout mouvement, mais le mouvement local, par opposition aux mouvements internes de croissance et de dépérissement. D’une part, ce mouvement est commun à tous les animaux, tandis que le mouvement local, comme on le verra, n’appartient qu’aux animaux capables de locomotion. D’autre part, on a déjà donné (en II 4) la cause des mouvements internes de ce type : il s’agit de la partie végétative-nutritive qui, de fait, est commune à tous les vivants14. La mention de la respiration et du sommeil et de l’éveil a pour but essentiel de les écarter de la recherche, car ils représentent un type de mouvement distinct du mouvement local, qui seul nous intéresse désormais. On peut toutefois se demander pourquoi ces deux cas conduisent à 12 Ces enjeux sont bien notés par R. Polansky (2007), p. 502 et C. Shields (2016), p. 348. 13 K. Corcilius (2008), p. 250, constate lui aussi que l’argumentation par élimination de DA III 9, bien qu’elle aboutisse à un résultat essentiellement négatif, participe à l’élaboration de la théorie de la locomotion animale. Je voudrais faire un pas de plus en montrant que cette aporie implique nécessairement l’explication du mouvement par la combinaison du désir et d’une faculté cognitive. 14 DA II 4, 415a22-416b31.

260

PIERRE-MARIE MOREL

une « grande difficulté » (ἔχει γὰρ καὶ ταῦτα πολλὴν ἀπορίαν). Il est clair que ces états ou fonctions dépendent principalement de la faculté végétative-nutritive15. Ce n’est donc pas sur ce point qu’il y a difficulté. En revanche, je ferais volontiers l’hypothèse qu’Aristote estime ici qu’il n’est pas toujours aisé de les rapporter à tel ou tel organe, d’autant que certaines de ces opérations relèvent de plusieurs parties ou facultés16. C’est en tout cas vrai du sommeil et de la veille. Le De somno pose la question de la partie dont dépend le sommeil dès la première page17 et il montre que, s’il est vrai que le sommeil et la veille dépendent de la faculté végétative-nutritive, ils impliquent également la faculté sensitive : être éveillé, c’est sentir18, tandis que le sommeil est une sorte d’impuissance, de lien ou d’immobilisation de la sensation. Le sommeil se définit donc par référence à la faculté sensible et à la faculté nutritive conjointement : ce n’est pas n’importe quelle impuissance de la sensation, mais une impuissance de la sensation due à l’exhalaison produite par l’absorption de nourriture19. En d’autres termes, le sommeil pose des difficultés du même ordre que le mouvement local. Quoi qu’il en soit, ce n’est pas ici le propos d’Aristote, puisqu’il s’agit proprement d’expliquer le mouvement local de l’animal et d’en identifier le principe moteur. Il ajoute une précision : on cherche « le moteur, pour l’animal, du mouvement de locomotion » (τὸ κινοῦν τὸ ζῷον τὴν πορευτικὴν κίνησιν) (432b13-14). Il lève ainsi toute ambiguïté : les mouvements internes des fluides et du souffle interne ou connaturel pourraient à la rigueur s’expliquer par une sorte de mouvement local interne20, comme le mouvement qui draine les impressions sensibles depuis les organes sensoriels périphériques vers la région cardiaque, mais ce que nous recherchons, c’est le moteur du mouvement local qu’accomplit l’animal en son entier, c’est-à-dire le moteur de sa πορεία (entendue au sens large de locomotion ou progression, et non pas au sens de « marche », par opposition aux autres modes de locomotion, comme la nage ou le vol). 15 Le DMA donne d’ailleurs, comme exemples de mouvements non volontaires, « le sommeil, l’éveil, la respiration et les autres mouvements semblables » en DMA 11, 703b9. 16 Voir en ce sens R. Polansky (2007), p. 524. 17 PN, De somno 1, 453b12-13. 18 PN, De somno 1, 454a5. 19 PN, De somno 3, 456b17-19. 20 Comme le montre S. Berryman (2002).

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

261

2. Élimination de la faculté nutritive (432b14-19) Aristote pose ici son premier résultat (432b14-15) : le moteur que l’on cherche n’est pas la faculté végétative-nutritive. Il invoque deux arguments. Un argument fonctionnel et un argument de fait relevant de la zoologie au sens d’étude du vivant en général. L’argument zoologique vient en second, mais je commencerai par lui, car c’est de loin le plus simple des deux : si le moteur était la faculté végétative-nutritive, les plantes seraient capables de locomotion ; or, de facto, elles ne le sont pas ; donc le moteur recherché n’est pas la faculté végétative-nutritive. L’argument fonctionnel, fondé sur l’analyse des fonctions ou opérations requises pour qu’il y ait mouvement local, consiste à établir que le mouvement local vise toujours une certaine fin. Il suppose, corrélativement, des facultés qui ne sont pas directement liées à la faculté végétative-nutritive : le désir et l’imagination (φαντασία). On comprend donc que la fin en question – par opposition aux mouvements relevant de la faculté végétative-nutritive qui, eux aussi, ont une fin naturelle – est un but à atteindre, une fin intentionnelle. Il s’agit de mouvements qu’Aristote qualifierait de « volontaires », conformément à la distinction déjà évoquée de DMA 11. Indirectement, c’est là un pas de plus en direction de cette idée que la locomotion implique à la fois et indissociablement une faculté motrice au sens strict – c’est-à-dire une faculté qui n’a pas d’autre fonction que de mouvoir l’animal – et une faculté de représentation ou faculté cognitive. Ce passage présente également une difficulté d’ordre textuel, puisqu’on peut lire en 432b16, καὶ ἢ μετὰ φαντασίας ἢ ὀρέξεώς ἐστιν21, ou bien : καὶ μετὰ φαντασίας καὶ ὀρέξεώς ἐστιν. La seconde lecture, défendue notamment par Ross22, se justifie sans doute par ce qui va suivre dans le texte du DA. Le chapitre 10 commence en effet par l’affirmation très nette, en 433a9, selon laquelle le moteur est double, intellect (ou φαντασία) et désir. Notons cependant que, même là, Aristote emploie la tournure disjonctive (ἢ ὄρεξις ἢ νοῦς), que la plupart des commentateurs comprennent comme ayant un sens conjonctif23. 21 Par exemple : W.S. Hett (1936), A. Jannone (1966), R. Polansky (2007) et visiblement R. Bodéüs (1993). 22 W.D. Ross (1961). 23 Voir en effet DMA 6, 700b15-18.

262

PIERRE-MARIE MOREL

Il n’est donc pas nécessaire de corriger le texte en 432b16 pour se mettre en conformité avec la doctrine, puisque celle-ci s’accommode très bien de la formulation disjonctive, là même où elle opte pour une relation en réalité conjonctive. Il est toutefois possible qu’en 432b16, et si l’on veut prendre à la lettre le sens disjonctif de καὶ ἢ μετὰ φαντασίας ἢ ὀρέξεώς ἐστιν, Aristote soit encore volontairement imprécis ou évasif, car son propos n’est pas alors de décider quelle est précisément la faculté motrice, mais de montrer que la faculté végétative-nutritive ne peut pas l’être, puisque le changement de lieu implique toujours un but, une intention de mouvement. Or il suffit pour cela d’invoquer soit la φαντασία, soit l’ὄρεξις, soit les deux. C’est ce qu’indique clairement la phrase suivante : « aucun ne se meut, sinon par force, s’il ne désire pas ou ne fuit pas » (οὐθὲν γὰρ μὴ ὀρεγόμενον ἢ φεῦγον κινεῖται ἀλλ᾿ ἢ βίᾳ, 432b17-18). Disons donc qu’à ce point du texte il n’importe pas encore de distinguer entre φαντασία et ὄρεξις, mais seulement d’éliminer de la compétition la faculté végétative-nutritive. 3. Élimination de la faculté sensitive (432b19-26) L’élimination de la partie sensitive est moins aisée que celle de la faculté végétative. On sait en effet que la sensation juge ou discrimine, qu’elle fait donc partie des facultés « critiques » évoquées dans le schéma initial de III 9. Or, grâce à cette aptitude, mais aussi parce qu’elle est liée au plaisir et à la peine par appréciation de l’excès et du défaut24, elle a une dimension effectivement pratique, au sens où elle est orientée vers le mouvement intentionnel et, de ce fait, vers un certain type de mouvement : « quand c’est agréable ou pénible, comme si on affirmait ou niait, on poursuit l’objet ou on s’en détourne » (ὅταν δὲ ἡδὺ ἢ λυπηρόν, οἷον καταφᾶσα ἢ ἀποφᾶσα, διώκει ἢ φεύγει)25. La sensation semble donc prendre en charge l’inclination et l’aversion dont l’évocation, dans la partie précédente, permettait d’éliminer la faculté nutritive, qui en est précisément dépourvue. On pourrait donc penser que la sensation est la faculté motrice recherchée. Pourtant, elle ne résiste pas à la méthode d’élimination, et cela en vertu de deux arguments. Le premier est à nouveau un argument zoologique 24 25

DA III 2 ; III 7, 431a11 ; PN, De Sensu 2, 436b10 sq. DA III 7, 431a9-10.

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

263

de facto, tiré de la simple prise en compte des caractéristiques de certains genres ; le second est un argument plus spéculatif, qui relève de la représentation théorique de la finalité naturelle. L’argument de facto est le suivant : de nombreux animaux, comme les éponges et les coquillages, sont fixes et privés de la locomotion26, non pas par accident mais par nature – d’où l’expression διὰ τέλους, « tout au long de leur vie » –, et cela bien qu’ils sentent. On en infère donc que la sensation n’implique pas la locomotion. L’argument téléologique consiste à rappeler, en renfort de l’argument zoologique, le principe bien connu « la nature ne fait rien en vain ». La formulation de cet argument est ici assez elliptique mais elle peut être reconstruite ainsi : (a) les animaux en question, qui sont des animaux complets et non pas atrophiés, ont la sensation mais n’ont pas d’organes servant à la locomotion ; (b) or, « la nature ne fait rien en vain » ; (c) donc la locomotion ne leur est pas essentielle ; (c’) donc la possession de la sensation (inhérente à la nature de l’animal) n’implique pas la locomotion. Au-delà de la présente recherche du principe moteur, il s’agit d’une application instructive du principe téléologique selon lequel « la nature ne fait rien en vain », et cela pour au moins deux raisons. D’une part sa formulation, au lieu d’établir l’omniprésence de la détermination téléologique, réserve une place à la nécessité (432b21-22 : ἡ φύσις μήτε ποιεῖ μάτην μηθὲν μήτε ἀπολείπει τι τῶν ἀναγκαίων), rappelant ainsi que la réalisation de la finalité naturelle suppose aussi les moyens nécessaires, ce qu’Aristote appelle ailleurs nécessité « hypothétique » ou « conditionnelle »27. D’autre part, il va clairement dans le sens d’une interprétation locale et déflationniste du principe « la nature ne fait rien en vain ». Celui-ci, en effet, ne renvoie pas ici à une téléologie globale ou cosmique, mais à l’organisation des parties et des facultés à l’intérieur des 26 Voir DA I 5, 410b19-20, ou encore, par exemple : HA I 1, 487b6-10 ; VIII 1, 588b10-17 (texte qui se fonde sur le cas des animaux fixes pour établir que le passage des végétaux aux animaux est « continu ») ; PA IV 7, 683b4-9. 27 Comme le signale déjà G. Rodier (1900), p. 534. La bibliographie sur ce point est abondante, même si notre texte est rarement convoqué dans ce dossier. On se reportera, entre autres, à P. Pellegrin (1990).

264

PIERRE-MARIE MOREL

limites définies par le genre ou l’espèce. Si la nature « ne fait rien en vain » et si elle réalise toujours « le meilleur », ainsi qu’Aristote l’affirme souvent de manière conjointe, ce n’est pas en vertu d’un lien cosmique qui reconduirait tous les êtres à une fin une et commune, mais parce qu’elle œuvre dans le cadre des propriétés qui sont celles de l’espèce ou du genre. Elle réalise donc le meilleur, non pas absolument, mais relativement aux possibilités de l’organisme et au genre de vie, au βίος, de l’animal concerné28. Le De incessu animalium (De la locomotion des animaux) mentionne à plusieurs reprises le principe en question, et il le fait toujours dans la perspective d’une téléologie locale, opératoire au niveau de l’espèce ou du genre. Je peux me contenter ici de citer l’un des passages les plus explicites de ce traité : Pour commencer l’examen, procédons comme nous avons souvent l’habitude de le faire dans notre travail de naturaliste, en considérant la manière dont les choses se passent dans toutes les opérations de la nature. L’une de ces caractéristiques est que la nature ne fait rien en vain mais, en chaque espèce animale, en réalisant toujours le meilleur selon ce que permet son essence. C’est pourquoi si telle réalisation est préférable, elle est aussi et par là même conforme à la nature (ἡ φύσις οὐθὲν ποιεῖ μάτην, ἀλλ’ ἀεὶ ἐκ τῶν ἐνδεχομένων τῇ οὐσίᾳ περὶ ἕκαστον γένος ζῴου τὸ ἄριστον· διόπερ εἰ βέλτιον ὡδί, οὕτως καὶ ἔχει κατὰ φύσιν). (IA 2, 704b12-18)

Ce texte me paraît en outre légitimer, dans ma reconstruction de l’argument de DA III 9, l’insertion de la première conclusion (c) : la nature ne fait rien en vain signifie en effet, en l’occurrence, que la conformation de l’espèce (l’organisation de ses parties et de ses facultés) est à la fois limitée, car conditionnée par les possibilités (ἐκ τῶν ἐνδεχομένων) fixées par l’essence, et optimale, car elle réalise le meilleur (τὸ ἄριστον) par rapport au τέλος défini par cette essence. C’est donc l’essence de l’animal concerné qui sert de critère pour décider si telle partie ou faculté est nécessaire. On peut dès lors conclure, en vertu d’une application indirecte du principe « rien en vain », que si certaines espèces sont privées d’organes locomoteurs, c’est tout simplement qu’elles n’en ont pas besoin. Il me paraît important de noter que DA III 9, dont l’objet premier est la locomotion animale, est en parfaite concordance avec le traité qu’Aristote a lui-même consacré à cette question. L’argumentation de III 9 ne se 28 Voir en ce sens, et pour un état récent des débats actuels sur la finalité naturelle et sur la signification de cette formule, D. Henry (2013) ; P.-M. Morel (2016b).

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

265

contente pas d’une méthodologie abstraite et de principes généraux. Ainsi que l’allusion aux explications des Parva naturalia sur le sommeil et la respiration l’avait déjà montré, et ainsi que les références à venir au Mouvement des animaux le confirmeront, notre passage a manifestement pour toile de fond les acquis concrets des enquêtes qui sont à l’articulation de la psychologie et de la biologie. Corrélativement, il apparaît que la considération des seuls facteurs formels – relatifs à la forme qu’est l’âme –, ne suffit pas à rendre compte de la locomotion et de la manière dont le désir meut le corps. Il faut encore rendre compte des conditions matérielles, en l’occurrence corporelles, sans lesquelles le désir serait incapable de mouvoir l’animal. Les opuscules de psycho-biologie, Parva naturalia et De motu animalium, sont donc le prolongement logique et nécessaire – sinon dans un ordre chronologique, tout au moins d’un point de vue systématique – de l’argumentation de DA III 9-1029. Pour conclure sur cette section du chapitre, il est clair que la nature du genre animal, parce que celui-ci suppose la sensation sans impliquer la locomotion, donne un double argument, de facto et de jure, pour éliminer la faculté sensible de l’enquête sur le principe moteur. 4. Élimination de la faculté intellective (432b26-433a6) On pourrait ensuite soutenir la candidature de la faculté intellective, au nom du même type d’arguments que ceux qui avaient servi à défendre celle de la faculté sensible. DA III 7, une fois encore, doit être convoqué. Il y est dit, en effet, que « l’âme réfléchissante » dispose d’images, de 29 T.K. Johansen (2012), p. 252-257, souligne également la corrélation avec l’étude des aspects matériels de la locomotion, mais il semble attribuer une sorte d’autonomie théorique à l’étude proprement psychologique (par opposition aux phénomènes communs à l’âme et au corps) de la manière dont le désir, en tant qu’affection de l’âme, met en mouvement. Voir notamment p. 255 : « the cause of locomotion in this sense can be described in terms of what is true of the animal by virtue of its soul. » Je considère pour ma part qu’il n’y a pas d’explication proprement « psychologique » (au sens littéral et exclusif, s’il en est un) du mouvement, y compris dans le DA, indépendamment des états ou phénomènes « communs à l’âme et au corps », et qu’aucune affection de l’âme, à l’exception de l’intellect considéré en tant que séparé, ne peut de toute façon constituer une cause exclusivement formelle. Je dois, faute de pouvoir développer ici ma position, renvoyer à P.-M. Morel (2007), p. 20-31 (sur la corrélation DA-MA-PN) ; 71-89 (sur l’unité des facultés et le statut des opérations et affections communes à l’âme et au corps). L’analyse fameuse de la colère et des affections de l’âme entendues comme « formes engagées dans la matière » (voir DA I 1, 403a24 sq.) suffit cependant à rappeler, et à mon sens à justifier, ce principe général d’interprétation.

266

PIERRE-MARIE MOREL

φαντάσματα, qui sont semblables aux impressions sensibles, les αἰσθήματα. Aristote ajoute : « quand un bien ou un mal est affirmé ou nié, on fuit ou poursuit »30. Plus loin, il précise même que : Τὰ μὲν οὖν εἴδη τὸ νοητικὸν ἐν τοῖς φαντάσμασι νοεῖ, καὶ ὡς ἐν ἐκείνοις ὥρισται αὐτῷ τὸ διωκτὸν καὶ φευκτόν, καὶ ἐκτὸς τῆς αἰσθήσεως, ὅταν ἐπὶ τῶν φαντασμάτων ᾖ, κινεῖται. L’intellection pense les formes contenues dans les images, et étant donné que, dans ces formes, par son intermédiaire, est déterminé ce qui est à poursuivre et à fuir, et cela même en dehors de la sensation, quand s’applique aux images, on se meut. (DA III 7, 431b2-5)

Aristote ne dit pas ici que l’intellection se contente de représenter et de définir ce qui est à poursuivre ou à fuir : il affirme que l’on se meut, qu’il y a véritablement mouvement, du moment que l’on pense, grâce aux images ou dans les images, ce qui est à poursuivre ou à fuir. Pourquoi ne pas dire, dans ces conditions, que l’intellect, éventuellement l’intellect en tant qu’il est lié à la φαντασία, est le principe moteur que nous cherchons, tout au moins chez les animaux rationnels ? Le chapitre 9 montre une fois encore sa pertinence et sa nécessité dans l’économie générale de l’argumentation sur le mouvement animal : par rapport au chapitre 7, il se situe incontestablement à un niveau d’exigence plus élevé et apporte des précisions cruciales par rapport au schéma initial. Il va en effet montrer que l’intellect (νοῦς), ou la partie calculatrice (τὸ λογιστικόν), tombe également sous le coup de l’élimination. Tout d’abord, l’intellect spéculatif, théorétique, ne porte pas sur les πρακτά, les faisables ou événements pratiques, parce qu’il ne se prononce pas sur les objets de désir ou d’aversion, alors que le mouvement de locomotion les implique toujours (432b27-29). De fait, ajoute Aristote, la représentation par l’intellect de ce qui est à rechercher ou à fuir ne signifie pas qu’il « commande » (κελεύει) de rechercher l’objet ou de le fuir, puisque l’on peut par exemple considérer par la pensée quelque chose d’effrayant ou d’agréable, sans pour autant que la pensée « commande d’avoir peur »31 (ou, doit-on également penser, de jouir de tel plaisir). Ce cas ne fait donc même pas question : une telle activité n’est pas par elle-même motrice. ὅταν δὲ ἀγαθὸν ἢ κακὸν φήσῃ ἢ ἀποφήσῃ, φεύγει ἢ διώκει, (DA III 7, 431a15-16). G. Rodier (1900), p. 535, suggère, s’inspirant de Thémistius, de traduire φοϐεῖσθαι par « fuir ». 30

31

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

267

Cela donne en tout cas à Aristote l’occasion d’invoquer les aspects physiologiques de la motricité animale en précisant : « c’est le cœur qui se meut et, s’il s’agit de ce qui est plaisant, quelque autre partie » (432b31-433a1). Par rapprochement avec DMA 11, on peut supposer que l’expression « quelque autre partie » fait allusion au mouvement des organes sexuels. Cette phrase, si elle n’est pas totalement innocente, est en tout cas apparemment anodine. Pourtant elle pose une difficulté : faut-il comprendre que, lorsque la pensée figure un objet de peur ou de plaisir sans commander aucun mouvement, il n’y a pas véritablement mouvement de poursuite ou de fuite, mais qu’il peut y avoir un mouvement organique localisé dans une partie déterminée (la tachycardie et l’érection évoquées en DMA 11) ? Ou bien faut-il entendre que, tandis que l’intellect est incapable d’initier aucun mouvement, certaines parties du corps en sont capables ? Cette seconde hypothèse annoncerait DA III 10, 433b19-22, qui fait manifestement allusion au cœur comme principe organique par l’intermédiaire duquel le désir meut le corps, doctrine qui est, comme on sait, au centre du MA. Toutefois l’hypothèse ne fonctionne pas pour les organes sexuels qui, en réalité, ne meuvent qu’eux-mêmes. Aristote rappelle à ce propos, en DMA 11, 703b21, l’image du Timée, selon laquelle le sexe est, chez le mâle, doté d’une relative autonomie de mouvement, comme s’il était un animal distinct32. Il semble donc plus probable que la première hypothèse soit la bonne, et que le cœur soit ici évoqué, non pas comme organe principal du désir et de la locomotion, mais pour les mouvements qui l’affectent en propre. La concordance avec le texte de DMA 11 sur les mouvements involontaires est assez claire et confirme cette lecture : Λέγω δ’ ἀκουσίους μὲν οἷον τὴν τῆς καρδίας τε καὶ τὴν τοῦ αἰδοίου πολλάκις γὰρ φανέντος τινός, οὐ μέντοι κελεύσαντος τοῦ νοῦ κινοῦνται. J’entends par « involontaires », par exemple, le mouvement du cœur et celui du sexe (car souvent, quand quelque chose nous apparaît, ces mouvements se produisent, sans toutefois que l’intellect les ait ordonnés). (DMA 11, 703b5-8)

La situation est sans doute un peu plus compliquée dans le détail, car il n’est pas exact de dire que la pensée et la sensation ne produisent aucun mouvement. Elles produisent des altérations qui peuvent être à 32

Platon, Timée 91 b-c.

268

PIERRE-MARIE MOREL

l’origine des émotions, comme le DMA le montre à plusieurs reprises33. Il leur suffit même de produire une « forme » pour produire un effet physiologique34. Toutefois, le point sur lequel Aristote veut insister, en DA III 9 comme en DMA 11, est sans doute le suivant : l’intellect peut représenter un objet et même un mouvement possible relativement à cet objet, sans pour autant ordonner ou commander un tel mouvement. Certaines parties du corps peuvent à cette occasion se mouvoir spontanément, mais, si tel est le cas, la cause de leur mouvement n’est pas un commandement de l’intellect. À ce point de l’analyse, on pourrait encore objecter que l’on vient de parler de l’intellect pris dans sa fonction purement théorique, et non pas de l’intellect envisagé dans sa fonction pratique, c’est-à-dire l’acte par lequel l’intellect « prescrit » et par lequel « la réflexion dit de fuir ou poursuivre » (433a1-2)35. Ne pourrait-on admettre que celui-ci est effectivement moteur et qu’il est la faculté que nous cherchons ? Aristote écarte cette option au nom de l’argument de l’ἀκρασία : Καὶ ἐπιτάττοντος τοῦ νοῦ καὶ λεγούσης τῆς διανοίας φεύγειν τι ἢ διώκειν οὐ κινεῖται, ἀλλὰ κατὰ τὴν ἐπιθυμίαν πράττει, οἷον ὁ ἀκρατής. En outre, même si l’intellect prescrit et que la réflexion dit de fuir ou de rechercher telle chose, il ne meut pas pour autant, mais on agit alors en vertu de l’appétit, comme l’intempérant. (DA III 9, 433a1-3)

Je ne peux examiner trop longuement la question, comme on sait très disputée, de l’ἀκρασία. Quoi qu’il en soit, ce n’est pas la définition même de l’intempérance qui est en jeu ici : Aristote se contente de prendre l’ἀκρατής comme exemple. Cet exemple prouve que l’intellect 33

Voir P.-M. Morel (2016a). Voir par exemple : DMA 11, 703b20 ; DMA 7, 701b13-24 : « chez l’animal, la même partie peut devenir plus grande ou plus petite et changer de configuration, car les parties s’accroissent du fait de la chaleur, et rétrécissent du fait du froid et se modifient. Or les images, les sensations et les pensées produisent des altérations. Les sensations, en effet, constituent aussitôt des espèces d’altérations, et l’imagination et la pensée ont la puissance de leurs objets. D’une certaine manière, en effet, la forme, représentée par la pensée, du chaud ou du froid, ou de l’agréable ou de l’effrayant possède effectivement la qualité de chacun des objets considérés, c’est pourquoi l’on tremble et l’on a peur à leur seule pensée. Or ces états sont tous des affections et des altérations, et quand il y a des altérations à l’intérieur du corps, certaines parties deviennent plus grandes et d’autres plus petites. » 35 Qu’il s’agisse ici de l’intellect pratique ne fait pas de doute, comme le montre G. Rodier (1900), p. 535. 34

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

269

peut ordonner d’agir (et non pas seulement, comme précédemment, considérer froidement une action possible sans rien prescrire), sans que s’ensuive un mouvement conforme à ce qui a été ordonné. C’est donc que le mouvement s’accomplit, non pas « en vertu de l’intellect », mais « en vertu de l’appétit », κατὰ τὴν ἐπιθυμίαν. Il en va, dit la suite du texte, comme de celui qui possède la science médicale, le médecin en titre, qui ne soigne pas lui-même, mais délègue cette tâche à ses subalternes. Il faut donc « distinguer, de la science, un autre dont dépend l’action conforme à la science » (433a4-6 : ὡς ἑτέρου τινὸς κυρίου ὄντος τοῦ ποιεῖν κατὰ τὴν ἐπιστήμην, ἀλλ᾿ οὐ τῆς ἐπιστήμης). La métaphore n’est pas d’une clarté parfaite36. En un sens, c’est le médecin en titre qui décide et qui est donc celui dont « dépend » l’action au premier chef, même s’il n’agit pas lui-même. Il me semble toutefois qu’Aristote veut dire ici que le médecin décide, à la fois intellectuellement et pratiquement, de ce qu’il faut faire, et tient donc la place de la science, mais que l’action conforme à la science – en insistant sur « action », l’action proprement dite – dépend d’une autre intervention. Le jeu sur κύριος, dans l’expression κυρίου ὄντος τοῦ ποιεῖν, et si l’on met l’accent sur τοῦ ποιεῖν, revient dans ce cas à dire la chose suivante : l’action de soigner est conforme à la science médicale, mais le ποιεῖν, le fait même d’agir, de produire la santé, suppose un autre facteur. La science médicale ne suffit pas à produire la santé. De même, l’intellect, qu’il soit théorique ou pratique, n’est pas le principe moteur que nous cherchons, parce qu’il n’implique jamais de manière nécessaire le mouvement de locomotion. 5. Élimination de la faculté désirante ? (433a6-8) Enfin, d’une manière inattendue, après avoir éliminé la faculté végétative-nutritive, la sensation et l’intellection, et après n’avoir accordé à ces deux dernières qu’une partie seulement de la responsabilité du mouvement, et alors que l’on s’attend à l’absolu triomphe du désir, qui reste 36 R. Polansky (2007) semble supposer que ce qui détermine effectivement le passage à l’action de soigner, c’est le désir du médecin, par opposition à sa connaissance purement abstraite de ce qu’il faudrait faire. Il est certain qu’Aristote veut montrer ici que le désir est la faculté qui, s’ajoutant à la pensée, fait agir. Je ne sais pas, pour ma part, s’il faut aller jusqu’à distinguer ici, dans l’esprit du médecin, entre sa connaissance de la médecine et son désir de l’exercer effectivement.

270

PIERRE-MARIE MOREL

seul en course, Aristote formule une restriction à propos de ce dernier, en 433a6-8 : Ἀλλὰ μὴν οὐδ᾿ ἡ ὄρεξις ταύτης κυρία τῆς κινήσεως· οἱ γὰρ ἐγκρατεῖς ὀρεγόμενοι καὶ ἐπιθυμοῦντες οὐ πράττουσιν ὧν ἔχουσι τὴν ὄρεξιν, ἀλλ᾿ ἀκολουθοῦσι τῷ νῷ. Mais le désir non plus n’a pas le mouvement sous sa dépendance : les gens tempérants, quand ils désirent, c’est-à-dire éprouvent des appétits, ne font pas ce dont ils ont le désir, mais suivent l’intellect.

On doit sans doute comprendre que le désir n’a pas la maîtrise absolue du mouvement, car il est toujours possible, au moins pour l’individu tempérant, d’agir différemment. Le désir n’est donc pas seul à déterminer le mouvement : l’intellect a sa part et, selon les cas, la sensation ou la φαντασία peuvent également avoir la leur, comme le montrent le DMA et le chapitre 10 du livre III du DA. Le désir ne meut pas s’il n’est pas instruit de ses buts par une faculté cognitive. En d’autres termes, l’ὄρεξις elle non plus ne peut prétendre être la cause propre et suffisante du mouvement local. On pourrait objecter que le cas de la tempérance est un cas limite, qui ne concerne pas, par exemple, le mouvement local des animaux autres que l’homme. De même le désir en question n’est-il sans doute que l’ἐπιθυμία, la forme basique, pulsionnelle de l’ὄρεξις. Dans les lignes qui suivent (III 10, 433a23-25), Aristote montre que l’ἐπιθυμία est combattue, non pas par l’intellect seul, mais par l’intellect aidé de la βούλησις, c’est-à-dire par une autre forme de désir. On pourrait donc comprendre 433a6-8 de la manière suivante : le désir pulsionnel n’est pas nécessairement déterminant, car les gens tempérants obéissent à l’intellect, mais cela ne signifie pas que le désir en général (une certaine forme d’ὄρεξις tout au moins) ne soit pas le principe moteur. Malgré cela, j’observe qu’Aristote se contente ici d’opposer l’intellect au désir et que, de toute façon, la doctrine exposée dans la phrase suivante (433a9) est claire : il faut, pour qu’il y ait mouvement local, deux moteurs : non seulement une composante « orectique », mais aussi une composante « critique » ou cognitive. L’argument est en outre suffisant : il suffit d’un cas (ici celui des pulsions dominées par la force intérieure du tempérant) pour montrer que le désir ne peut pas être tenu, inconditionnellement, pour l’unique principe moteur de l’animal. Si le rôle du désir est prééminent dans la locomotion, ainsi que DA III 10 va le confirmer, ce n’est donc pas parce qu’il en serait par lui-même

MOUVEMENT ANIMAL ET THÉORIE DES FACULTÉS EN DA III

271

la cause suffisante, mais parce qu’il est la seule faculté dont la fonction soit directement et nécessairement liée au mouvement local – y compris quand celui-ci est empêché ou défendu par l’intellect –, tandis que les autres facultés envisagées ont d’autres fonctions et ne sont pas directement ni nécessairement liées à la locomotion. La méthode d’élimination a donc porté ses fruits de manière paradoxale : aucune des facultés ou parties de l’âme ne peut être par ellemême la cause suffisante du mouvement local, même si l’ὄρεξις en est la cause principale, mais c’est précisément ainsi qu’il faut expliquer un tel mouvement. Le chapitre 9 a donc permis de reformuler le problème du moteur en ajoutant une troisième alternative : si le mouvement local ne dépend ni d’une unique partie ou faculté, ni de l’âme tout entière, c’est qu’il dépend d’un dispositif complexe dans lequel différentes facultés coopèrent. On passe d’une hypothèse fondée sur l’idée de distribution (à telle faculté, telle opération ou action), à une solution qui repose sur l’idée d’une coopération des facultés (à telle faculté, telle fonction, dans une opération ou action complexe mobilisant plusieurs facultés)37. Les enjeux ne sont pas négligeables, non seulement pour l’analyse psychologique de la motricité, mais encore pour la métaphysique de la forme qui fonde la conception aristotélicienne de l’âme : que l’âme soit une forme signifie qu’elle n’est ni une somme de parties, ni même l’ensemble de ses propres facultés, mais l’unité indivisible de leur mutuelle organisation. La responsabilité particulière du désir fera l’objet du chapitre 10, mais le chapitre 9, grâce notamment à la question de l’intempérance, en donne déjà une idée. Le désir ne pourrait pas mouvoir s’il n’était pas instruit par une faculté cognitive, informé de ce qu’il faut faire ou de ce 37 Voir R. Polansky (2007), p. 512 : « The purpose for making some objection to the desiderative power is to complete the challenge to other faculties and to suggest that no single faculty of soul solely on its own causes voluntary animal motion but a combination of the cognitive, that is, φαντασία or thought, and the desiderative leads to animal motion. » Par extension et par transposition dans le domaine de la philosophie pratique, ainsi que me l’a fait remarquer Charlotte Murgier, on voit qu’il y a concordance avec les observations d’Aristote sur le principe de l’action, la προαίρεσις, puisque celle-ci n’est pas simplement désir, mais « désir délibérant » ou « intellect désirant » (EN III 5, 1113a11 ; VI 2, 1139b4-5). La combinaison du désir et de la pensée prend dans ce cas la forme d’une double spécification (je remercie David Charles de me l’avoir signalé), celle d’un intellect qui se définit comme désirant, mais aussi celle d’un désir qui se spécifie comme désir raisonnant.

272

PIERRE-MARIE MOREL

qui est désirable. Il n’est donc que l’un des éléments du dispositif global qui explique le mouvement animal. Il a cependant une responsabilité spécifique, bien signalée par la construction κύριος τοῦ ποιεῖν : c’est du désir, et de lui seul, que dépend le « faire » en tant que tel, le passage à l’acte, distingué de toutes les informations, y compris prescriptives, qui peuvent le précéder et l’orienter. C’est même là sa tâche propre, puisque les autres facultés impliquées dans l’explication du mouvement ont d’autres fonctions – des fonctions cognitives essentiellement – que de concourir à la locomotion. Le DMA donnera une claire illustration de la fonction spécifique du désir, en se plaçant sur le plan physiologique : le désir a comme principe organique le cœur, point de départ du souffle interne et, par son intermédiaire, du mouvement des membres. Or le cœur ne pourrait mouvoir les autres parties, et notamment les membres locomoteurs, sans bénéficier des informations données par les facultés cognitives, même s’il est le seul organe à pouvoir accomplir un tel travail.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE AT THE END OF ARISTOTLE, DE ANIMA III 10 Christof RAPP

1. Introducing the cliffhanger passage In De Anima III 9-10 Aristotle discusses the psychic movers, νοῦς (thought, understanding) and ὄρεξις (desire). This discussion is triggered by the conviction that the soul is characterized not only by perceiving and thinking, but also by its role in the locomotion of animals. Starting in chapter 5 of De Anima II, the discussion has dealt with various kinds of perception, with φαντασία and with thinking. After the completion of these discussions in chapter 6 or 7 of Book III, it seems to be the right occasion to finally turn to the capacity for locomotion and to the moving factors within the animal. At the end of this discussion, towards the end of De Anima III 10, Aristotle makes an intriguing reference to a different project, namely the project of accounting for the functions that are common to body and soul. This remark seems to suggest that the envisaged endeavour will be something like a “psycho-physical” inquiry, i.e. an inquiry that deals with the above-mentioned attributes or functions insofar as they belong to a living body, possibly brought about in specific regions of such a body (and not only as, say, purely psychic capacities). In the particular case of animal locomotion, the treatment of such common functions or of such functions qua being common, De Anima III 10 says, is intended to explain by what instrument (ὄργανον) the animal is moved. Since our particular chapter, close to the end of De Anima III, cannot properly enter into this kind of discussion, Aristotle gives a short preview of what we can expect with respect to this account of the instrument of movement. In this sketchy summary at the end of De Anima III 10, the “cliffhanger passage” as we

274

CHRISTOF RAPP

shall call it, Aristotle tells us that bodily movement originates where an end point and a starting point coincide, as in a joint or hinge. This seems puzzling at first glance, since, although joints are obviously needed to facilitate the bodily movement of animals, we usually do not expect an explanation to appeal to joints as the origin or to make them essential to such an explanation. Also, this reference to joints follows an ambitious account of movement in general, according to which movement implies an unmoved mover, a moved mover, the instrument of movement and the thing moved. How exactly do joints fit into this model? Does the unmoved mover use a joint as an instrument? How is such an instrument related to the intermediate moved mover? One might well think that this reference to joints or hinges cannot seriously be meant to fully capture the upshot of Aristotle’s announcement that he will talk about functions that are common to body and soul. A not unusual approach is to view the reference to joints as merely metaphorical or as part of a comparison that aims to illustrate a more abstract idea, e.g. that a single thing can be two in being, just like the coinciding starting point and end point within a joint. It seems to be worthwhile, hence, to dedicate a full paper to this passage and to try to get clear about the role of joints and movers in Aristotle’s cliffhanger passage. 2. The immediate context (De Anima III 9-10) The two chapters of the third book of De Anima that focus on the identification of the movers or the moving factors in animal locomotion, chapters 9-10, give rise to many questions and are indeed the subject of many scholarly debates. This is not the place to enter into any of these debates; we will rather confine ourselves to setting the stage for the cliffhanger passage at the end of III 10, which has not so far received the same amount of scholarly attention as the rest of these chapters. It is widely agreed that of the two chapters, chapter 9 is dedicated to a rather aporetic (some may prefer the term “dialectical”) discussion of questions concerned with the movers of animals: if we look for these movers among the parts of the soul, one might immediately raise the question of how we can distinguish parts of the soul at all. There is the risk, Aristotle points out, that we tear apart the soul in too many parts and end up with a fragmented soul. Also, some capacities that at first glance may seem to be plausible candidates for the role of the mover turn out not to

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

275

be decisive (κύριον) for whether the animal moves or not: some sort of thinking, to begin with, is thought to move human beings, but akratic, i.e. weak-willed, agents are actually moved in accordance with what they desire and against what they think. Similarly, non-rational desires such as appetites are thought to be movers, but enkratic, i.e. self-controlled, agents are not immediately set in motion by their appetites, but act in accordance with what reason says. By the beginning of chapter III 10, at any rate, it seems clear that there are two remaining candidates for what moves the animal, νοῦς and ὄρεξις, if we conceive of them in a broad enough sense. Ὄρεξις should be taken in the generic sense in which it comprises several types of desire, both rational desire (βούλησις) and non-rational desire, namely appetite (ἐπιθυμία) and passion or spirited desire (θυμός). Νοῦς (thinking, understanding) is similarly used here in a relatively generous sense which allows of the possibility that φαντασία can take the place of νοῦς; this is required if one wishes to cover all possible cases of animal movement, since sub-human animals do not possess νοῦς in the strict sense of the word. And even human beings occasionally act and move contrary to correct reasoning or knowledge. In these latter cases φαντασία becomes an ersatz-νοῦς: like νοῦς, φαντασία functions in these cases as a discriminatory (κριτικόν) faculty; it can grasp, identify or distinguish possible objects of desire. Strictly speaking, νοῦς and ersatz-νοῦς play a role as movers only if they present practicable goods, i.e. goods that can be reached or realized by moving and acting, while the objects of theoretical νοῦς are of no interest to those inquiring into the causes of movement and action. A significant part of chapter III 10 deals with what might seem to be a sort of competition between ὄρεξις and νοῦς. Which of them deserves to be called the mover of the animal? In the end, it seems that there is no clear winner, since, in a way, both are required for making the animal move; still there is a certain asymmetry between the contributions of ὄρεξις and νοῦς, in that νοῦς never imparts movement without ὄρεξις, while one kind of ὄρεξις, ἐπιθυμία, can impart movement without or even against thinking / reasoning (λογισμός). In another sense, the competition between ὄρεξις and νοῦς is solved by the reference to a third player: the practicable or realizable good, i.e. the object of desire. It can come in two forms: as either a real good or as a merely apparent good. Either way, it is clear that this practicable

276

CHRISTOF RAPP

good (ἀγαθὸν πρακτόν) is a mover. To the extent that the practicable good (either real or apparent) is the proper object of desire (ὀρεκτόν), it is also clear that desire (ὀρεκτικόν) is a mover, since the object of desire moves the animal only by being desired (without an animal’s corresponding faculty of desire, one could not even speak of an object of desire). And since the object of desire must be thought of or imagined by an animal in order to impart movement, it also becomes clear how and why νοῦς or ersatz-νοῦς is involved: an animal could not desire an object without grasping, understanding, perceiving or imagining such an object. Thus, once we acknowledge that it is the object of desire, i.e. the practicable good or the apparent practicable good, that moves the animal by involving the discriminatory and desiderative faculties of the animal’s soul (in quite different ways), it seems that we are confronted with a mover that is itself unmoved. For the object of desire does not move by being itself moved, it rather moves as something unmoved. And it is at this point in the discussion that our perplexing cliffhanger passage appears. 3. Imparting movement: the general scheme (DA III 10, 433b13-18) We subdivide the crucial passage into three sections: the really perplexing and troublesome section is the third one; but this third section must be read in the context of the two previous sections, which, unsurprisingly, provide problems and questions of their own. First, Aristotle gives a general scheme for the initiation of movement and tries to map some of the crucial players from the previous discussion in chapter III 10 onto this general scheme. Ἐπεὶ δ᾽ ἔστι τρία, ἓν μὲν τὸ κινοῦν, δεύτερον δ᾽ ᾧ κινεῖ, ἔτι τρίτον τὸ κινούμενον, τὸ δὲ κινοῦν διττόν, τὸ μὲν ἀκίνητον, τὸ δὲ κινοῦν καὶ κινούμενον, ἔστι δὴ τὸ μὲν ἀκίνητον τὸ πρακτὸν ἀγαθόν, τὸ δὲ κινοῦν καὶ κινούμενον τὸ ὀρεκτικόν (κινεῖται γὰρ τὸ κινούμενον ᾗ ὀρέγεται, καὶ ἡ ὄρεξις κίνησίς τίς ἐστιν, ἡ ἐνεργείᾳ1), τὸ δὲ κινούμενον τὸ ζῷον· Since there are three things, first, what initiates motion, second, that by which it initiates motion, and further, third, what is moved, and that which 1 This is how W.D. Ross (1961) reads the Greek text; A. Torstrik (1862) reads ἐνεργείᾳ and P. Siwek (1965) aspirates and accentuates as ἢ ἐνεργείᾳ.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

277

initiates motion is twofold, in the one instance being unmoved and in the other initiating motion while being moved, there is: something unmoved, the good concerned with what can be done; something initiating motion while being moved, the faculty of desire (for what is moved is moved insofar as it is desiring, and desire, when in actuality, is a kind of motion); and what is moved, the animal.2

A natural and widespread reading of this section is as follows. There are three factors in the initiation of movement: (1) that which imparts the movement, (2) that by means of which it (= the mover from (1)) imparts movement and (3) that which is moved. Immediately afterwards item (1) from this scheme, i.e. the mover, is subdivided into two different factors. The subdivision is introduced by τὸ δὲ κινοῦν διττόν. Sometimes this is translated as a comment about homonymy: “the ‘that which imparts movement’ is ambiguous”; but semantic ambiguity is hardly Aristotle’s main concern here. Rather, he wants to point to the fact that there are two kinds of movers or factors that move, in that they impart motion in quite different ways. One mode of being a mover is to impart movement while being unmoved, another mode of being a mover is to impart movement through being itself moved. On the reading we are considering now, this amounts to saying that (1) subdivides into two items (1a) an unmoved mover and (1b) a moved mover. This reading can be presented in the following, now fourfold, scheme: The fourfold scheme: What initiates motion (M) What initiates motion as something unmoved (UM) What initiates motion as something moved (MM) That by means of which it initiates motion (BW) What is moved, i.e. the moved body (MB)

What is the purpose of introducing such a scheme? The obvious suggestion is that Aristotle wishes to subsume the results reached about animal locomotion under a more general scheme that is meant to hold also of different kinds of movement. Compliance with this more general scheme would lend additional support to the findings about animal movement. 2 Aristotle, De Anima III 10, 433b13-18, the English translation is taken from C. Shields (2016).

278

CHRISTOF RAPP

At the same time, such a general scheme, which has been attested by other types of movement, could confirm the completeness of the account of animal movement given so far or could instead point to gaps in it. At any rate, in order to complete this project, Aristotle first has to fill in the general scheme with the moving factors that he had earlier identified for the peculiar case of animal movement. According to the text quoted, it is easy to fill in three of four items from the list, as Aristotle does: (UM) (MM) (BW) (MB)

The realizable good The faculty of desire ? The animal

This reading follows the order of the text down to line b18. It is surprising that the text seems to skip item (BW) from the general scheme. However, if we glance at the very next line we read that “the instrument by means of which desire imparts movement is already bodily” (ᾧ δὲ κινεῖ ὀργάνῳ ἡ ὄρεξις, ἤδη τοῦτο σωματικόν ἐστιν)3. This seems to pick up on the wording by which factor (BW) was originally introduced: “that by means of which it imparts movement (δεύτερον δ᾽ ᾧ κινεῖ)”. Aristotle clearly adds a further specification here of item (BW), namely that it is a bodily instrument. What remains to be explained, then, is the fact that Aristotle mentions the bodily instrument as a candidate for position (BW) only after he has introduced the animal or the animal’s body as item (MB). Why is it, in other words, that he mentions the animal and the bodily instrument in the reverse order and leaves it to the reader to see that the bodily instrument is meant to fill position (BW) for the case of animal movement? One might indeed see a problem in this reverse order; still, it is in a way obvious and trivial that in the context of animal motion the animal is the moved item within the fourfold scheme, while the notion of a bodily instrument calls out for further comment, which is actually provided in the following lines. This might be the reason why Aristotle mentions them in the reverse order: he mentions the obvious factor first, in order to focus on the less obvious factor, the introduction of which leads to a discussion of several lines. It seems, then, that we are justified in assuming a fourfold scheme and that Aristotle 3

Aristotle, DA III 10, 433b19.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

279

is actually able to fill each position of this scheme for the special case of animal movement. The fourfold scheme filled in: (UM)

The realizable good

(MM)

The faculty of desire

(BW)

The bodily instrument by means of which the animal is moved (see lines 433b19-21) The animal

(MB)

It also seems possible now to determine the role of this scheme at the end of the discussion of moving factors in chapter III 10. It provides a model which accommodates the roles of both the object of desire (ὀρεκτόν) and the faculty of desire (ὀρεκτικόν). In this respect, it can be read as a summary and generalization of the preceding discussion. At the same time, factor (BW) of the general scheme draws the reader’s attention to the fact that there is a question that is absent from the preceding discussion: by means of which instrument is the animal moved? Answering this kind of question would go beyond the limits of the project of De Anima, since De Anima does not go into the bodily details relevant to psychic capacities. This is why factor (BW) provides the occasion for referring the reader to a different kind of project, the main ideas of which are summarized in the subsequent section. Seen from this perspective, the application of the general fourfold scheme to animal motion indicates what is missing in the account of animal movement, namely the treatment of factor (BW), insofar as it requires a bodily instrument. So far, so good. As already indicated, it is most probably the purpose of the general scheme to show that, broadly conceived, the explanatory ingredients of animal movement are in line with the explanation of other kinds of movement. In principle, it would serve this purpose if we could show that one and the same scheme explains animal motion and other kinds of motion. Aristotle does not explicitly refer his reader to other texts, but there are pertinent parallels. One famous passage in which we find the scheme just sketched is in the eighth book of the Physics. It reads: Τρία γὰρ ἀνάγκη εἶναι, τό τε κινούμενον καὶ τὸ κινοῦν καὶ τὸ ᾧ κινεῖ. τὸ μὲν οὖν κινούμενον ἀνάγκη κινεῖσθαι, κινεῖν δ’ οὐκ ἀνάγκη· τὸ δ’ ᾧ κινεῖ, καὶ κινεῖν καὶ κινεῖσθαι (συμμεταβάλλει γὰρ τοῦτο ἅμα καὶ κατὰ τὸ αὐτὸ τῷ κινουμένῳ ὄν·) […] τὸ δὲ κινοῦν οὕτως ὥστ’ εἶναι μὴ ᾧ κινεῖ, ἀκίνητον.

280

CHRISTOF RAPP

For there must be three things – the moved, the mover, and that by which it moves4. Now the moved must be in motion, but it need not move anything else; that by which (the mover) moves must both move something else and be itself in motion (for it changes together with the moved, with which it is in contact and continuous) […] and the mover – that is to say, that which causes motion in such a manner that it is not merely the instrument of motion – must be unmoved.5

The similarities with the scheme from De Anima III 10 we considered so far are striking. However, here we find a threefold, not a fourfold scheme. The difference is that here, the factor by which the mover imparts movement (i.e. (BW) according to our previous rendering) is straightforwardly identified with what imparts movement while being itself moved. Factors (MM) and (BW) from our previous scheme seem to coincide, so that we get the following threefold scheme: The threefold scheme: (i) What imparts motion as something unmoved (UM) (ii) That by which the mover imparts movement (BW), i.e. what imparts motion as something moved (MM) (iii) What is moved (MB)

It is tempting, as we saw, to think that in De Anima III 10, 433b13-18, Aristotle employs a fourfold scheme, which he arrives at by subdividing the position of the mover in an initially threefold scheme: the moving factor (M) is subdivided into an unmoved and moved mover. Recently, Christopher Shields has argued that we need not assume that such a subdivision has taken place: It is sometimes thought that Aristotle [in De Anima III 10 scil.] departs from his earlier treatments by further subdividing (i)6 […]. The division in the current passage [De Anima III 10, 433b13-18], however, is not subordinate to (i). Instead, Aristotle is allowing that while both the good sought and the faculty of desire initiate motion, only the faculty does so while being itself in motion. Thus, the object of desire initiates motion without itself being in motion.7 4 Note that the Revised Oxford Translation translates the phrase τὸ ᾧ κινεῖ straightforwardly as “the instrument of motion”. 5 Aristotle, Physics VIII 5, 256b14-20. Greek text from W.D. Ross (1950), the translation is based on the Revised Oxford Translation, J. Barnes (1984). 6 In Shields numbering, to be sure, (i) corresponds to (M) in our scheme. 7 C. Shields (2016), p. 362.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

281

I take Shields’ remarks to mean, firstly, that De Anima III 10, 433b1318 does not employ the fourfold scheme as outlined in the previous section, but rather the threefold scheme from the Physics and, secondly, that Aristotle’s comment that we interpreted as a subdivision of (M) (or of (i) in Shields’ numbering) should rather be taken to imply that (ii), i.e. desire (and, at the same time, that by which the mover moves) is the kind of mover that initiates motion while being itself in motion. Shields’ remarks originate from an important observation concerning the faculty of desire (ὀρεκτικόν). For it may seem awkward that Aristotle nominates the ὀρεκτικόν as a mover, since the ὀρεκτικόν in Aristotle consistently refers to the capacity or faculty to desire, while a capacity (δύναμις) as such, namely in the state of a mere capacity or in potentiality (δυνάμει), does not move, strictly speaking, anything. It seems to be exactly for this reason that Aristotle in lines 433b17-18 is keen to emphasize that it is the activated capacity that is involved in the initiation of movement, not the mere disposition One advantage of Shields’ reading is immediately clear: it would allow us to construe the scheme of De Anima III 10 as fully congruent with that of Physics VIII. Accordingly, we would have to reassign the moving factors of animal movement. Indeed, Shields suggests: There are, contends Aristotle, three [such factors, scil.]: (i) what initiates motion, identified as the good to be accomplished; (ii) that by which one8 initiates motion, the activity of the faculty of desire; and (iii) what is moved, the animal.9

This could be arranged in the following alternative scheme: The threefold scheme applied to De Anima III 10: (i) What moves (the animal) as something unmoved, i.e. the object of desire. (ii) That by which it moves (the animal), what imparts motion as something moved, i.e. the activity of the faculty of desire. (iii) What is moved, i.e. the animal. 8 It is not clear to me why Shields says “one” here, since the implicit subject of this formulation in Aristotle seems to be the mover, while “one” might be taken to introduce a different subject, perhaps the agent or the animal. 9 Ibid. By contrast, R. Polansky (2007) treats factor (ii) as “that by which desire moves the animal [and] some bodily part as organ or instrument (b19-22)” (p. 522), taking the διττόν remark to subdivide position (i) of the threefold scheme, effectively yielding the fourfold scheme as outlined above.

282

CHRISTOF RAPP

It seems then that there are at least two different ways of construing our passage in De Anima III 10, 433b13-18 – either, roughly, along the lines of the fourfold or along the lines of the threefold scheme. The really controversial question, however, is not whether (UM) and (MM) from the De Anima-passage are derived from a subdivision of the mover (M) in the De Anima-passage or of (i) in the Physics passage, since there, in the Physics, the mover (τὸ κινοῦν) is straightforwardly identified with the unmoved mover alone, while the commonality between the unmoved mover and the intermediate factor that imparts movement by being moved – namely that both are movers in some sense – is not really emphasised, but is only indirectly acknowledged, namely when Aristotle adds that that by which (the mover) moves must both move something else and be itself in motion10. The more substantial issue is whether the intermediate factor that imparts movement by being moved, i.e. (MM), in the De Anima-passage is to be identified with that by which movement is imparted (BW). This identification follows by necessity, if we try to equate the De Anima-scheme to the Physics-scheme (and Christopher Shields, for one, explicitly suggests this identification in the above quoted text). If one actually goes for such an identification, the next thing to consider would be whether desire, which fills the (MM) slot, is the one and only factor in play that instantiates the “by which”-relation or whether there could be more of them. Clearly it cannot be the only such factor. For, if it were, there would be no place anymore for the bodily instrument “by means of which desire imparts movement” (433b19). However, there are good reasons for thinking, as we will see, that it is the very purpose of our cliffhanger passage to motivate an additional discussion of this bodily instrument. A slightly different way to put the same issue would be this: is (BW) meant to refer to the only instrument of motion? After all, the notion of “instrument” seems to be a natural way to interpret the instrumental dative in the formulation τὸ ᾧ κινεῖ, by which the “by which”-relation is introduced. But again, if we suppose that there is just one such instrument and if we assume that, in accordance with the Physics-passage, the moved mover plays the role of the instrument of motion, the application of the 10 Also, the Physics-passage acknowledges that there is a genuine mover apart from the unmoved one, in that it characterizes the unmoved mover in a negative way by saying that it does not impart movement in such a manner as would depend on its being moved.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

283

Physics-scheme to De Anima would pose the problem that there would be no more space for another instrument in addition to the faculty to desire. These worries could be addressed either by allowing more than one item to play the role of (BW) or by loosening the tie between (BW) and the targeted “instrument”. However, this would lead to the duplication of the “by which”-relation, which, at first glance, seems to be an unwelcome consequence: first ὄρεξις is that by which the (unmoved) mover moves and then there is a bodily instrument by which ὄρεξις moves. Still, there are indications that this duplication need not be taken as a vicious one. Most notably, the mover of the Physics-passage that uses the moved mover as something by which motion is imparted is the unmoved mover, while in De Anima 433b19 it is desire that uses something else, the bodily ὄργανον, as means by which it moves the body. The different subjects in the two cases clearly indicate that when the moved mover is mentioned in the Physics as that by which the unmoved mover moves, this is a different instance of the “by which”-relation than in the De Anima-passage, where a bodily organon is described as that by which ὄρεξις imparts movement to the body. Also, when Aristotle refers in De Anima 433b19 to the instrument by which desire moves, he says: ᾧ δὲ κινεῖ ὀργάνῳ ἡ ὄρεξις, which amounts to saying: “the instrument or organ by which ὄρεξις imparts movement”. Calling this item the instrument or organ by which desire imparts movement certainly marks a difference from the unqualified “by which”-relation. Clearly, the Greek organon can refer to all kinds of instruments and can, hence, accompany any occurrence of the instrumental “by which”; but if it refers, as in our context, to parts of the body (which seems to be confirmed in the same sentence, which actually says that such an ὄργανον would be ἤδη σωματικόν – already bodily), it could highlight a quite specific relation, namely the relation between a capacity of the soul and a specific part of a living body and not the general instrumental relation that was prevalent in the Physics passage. In this sense the peculiar hylomorphic-psychophysical context of our passage seems to require going beyond the much more general scheme from the Physics, insofar as the passage asks for a more specific kind of instrument of the transmission of movement, i.e. an instrument by which movement is transmitted from the desiring capacity of the soul to certain bodily limbs. What is the philosophical upshot of all this? In the threefold scheme of the Physics the “by which”-relation characterizes the intermediate

284

CHRISTOF RAPP

role of the moved mover (MM), while in DA III 10 Aristotle clearly uses the “by which”-relation (in part at least) to address a more specific question, namely the identification of a bodily part or organ that desire (the intermediate role of which is not called into question) uses as an instrument. Accordingly, it seems that the fourfold scheme functions better within DA III 10, in that it is apt to set the stage for the question of how the soul moves the body, as it is put in De Motu Animalium 611, while, with regard to the question of animal motion, the threefold scheme from the Physics more generally accounts for the different factors that are involved in the initiations of animal motions. 4. Traces of the general scheme in De Motu Animalium Before we move on to the next section of our cliffhanger passage, let us pause here to register several traces and variants of the general scheme in the De Motu Animalium. This treatise is dedicated to the investigation of the common cause of animal movement (698a4) and thus suggests itself as pursuing a project that is at least akin to that of De Anima III 9-10. Right from the beginning the treatise tries to embed the specific question of animal locomotion into the broader context of moving factors and preconditions of movement in general. Although neither the fourfold nor the threefold general scheme of movement explicitly appears, there are clear indications that a general scheme with the elements (UM), (MM), (BW) and (BM) is presupposed. Indeed, the role of (UM) is invoked already in the course of the treatise’s first eleven lines (DMA 1, 698a8-11): here Aristotle points out that the principle of self-movers is the unmoved and that the first mover is necessarily something unmoved. In the ensuing chapters Aristotle establishes that animal movement requires two different sorts of unmoved points, firstly a resting point within the animal, as is given in each and every joint, and secondly an external unmoved platform against which the animal supports itself in order to move forward. The introduction of these two unmoved prerequisites of animal movement leads to the digressive discussion of whether the same requirements also hold of the 11 DMA 6, 700b9-10: λοιπὸν δ’ ἐστὶν θεωρῆσαι πῶς ἡ ψυχὴ κινεῖ τὸ σῶμα. In general, all quotes from the Greek text of DMA are taken from Primavesi’s edition in C. Rapp/O. Primavesi (forthcoming).

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

285

movement of the cosmos as a whole. The precise function of this latter discussion within the De Motu Animalium is a notorious subject of scholarly dispute12. At any rate, it is remarkable that, when Aristotle finally returns to the more specific question of animal movement in chapter 6 of the treatise, he explicitly recalls the role of the realizable good as the first (unmoved) mover13: it is “the object of desire and of thought that imparts movement first, though not any object of thought, but the goal of what is achievable in action”14. A few lines later Aristotle offers a slightly simplified version of the general scheme from De Anima III 10: Τὸ μὲν οὖν πρῶτον οὐ κινούμενον κινεῖ, ἡ δὲ ὄρεξις καὶ τὸ ὀρεκτικὸν κινούμενον κινεῖ. τὸ δὲ τελευταῖον τῶν κινουμένων οὐκ ἀνάγκη κινεῖν οὐθέν. The primary mover then causes movement without itself being moved, but desire and the faculty of desire cause movement while being moved. But it is not necessary that the last of the moved things move anything else.15

In these lines Aristotle clearly refers to (UM), to (MM) and to (MB); in addition, he identifies desire as the factor that in the case of animal movement plays the role of (MM): (UM) (MM) (MB)

Desire and / i.e. the faculty to desire

All this obviously brings us close to the scheme from DA III 10; however, while De Anima III 10 also refers to the “by which”-relation (BW) as well as to a bodily instrument or ὄργανον, DMA 6 does not. The absence of these elements from DMA 6 need not indicate a substantial doctrinal difference; perhaps it is due to the peculiar context and agenda of this chapter that there is no occasion to mention (BW) as it was used 12 See e.g. M.C. Nussbaum (1978), p. 6 and p. 109; Nussbaum goes as far as to argue that because of this digression on cosmological issues DMA includes an “interdisciplinary” project. 13 See above in section III: “the fourfold scheme filled in”. 14 DMA 6, 700b23-25: ὥστε κινεῖ πρῶτον τὸ ὀρεκτὸν καὶ τὸ διανοητόν. οὐ πᾶν δὲ διανοητόν, ἀλλὰ τὸ τῶν πρακτῶν τέλος. 15 DMA 6, 700b35-701a2 (Unless otherwise indicated translations from the DMA are mine). The final remark about the last factor that needs not move anything else seems to be reminiscent of Physics VIII 5, 256b14-20.

286

CHRISTOF RAPP

in DA III 10. And indeed we will find the two so far missing elements in the course of chapter 10 of DMA, which is dedicated to the peculiar role that the connate πνεῦμα is supposed to play within the initiation of animal movement. The discussion of the connate πνεῦμα is, as it were, sandwiched by two clear references to the general scheme; at the beginning of this chapter, Aristotle states: Κατὰ μὲν οὖν τὸν λόγον τὸν λέγοντα τὴν αἰτίαν τῆς κινήσεως ἐστὶν ἡ ὄρεξις τὸ μέσον, ὃ κινεῖ κινούμενον. ἐν δὲ τοῖς ἐμψύχοις σώμασι δεῖ τι εἶναι σῶμα τοιοῦτον. On the basis of the account stating the cause of motion then, desire is the intermediate thing, i.e. that which causes movement while being moved. In animate bodies, however, there must be some body of that sort.16

In this passage Aristotle refers his readers to “the account stating the cause of motion (τὸν λόγον τὸν λέγοντα τὴν αἰτίαν τῆς κινήσεως)”; it is in virtue of this very account that desire can be identified as the intermediate item, i.e. as the factor that moves while being moved. And the kind of account that makes desire an intermediate item is the general scheme that refers to (UM), (MM) and (MB). We can leave open whether Aristotle wants to refer us to the abbreviated version of this scheme that was provided in DMA 6 or to other versions provided in other treatises; at any rate, it is clear that he thinks of this scheme as a standing account of the causal factors of motion and that he wishes to apply this account within the project of DMA too. The second sentence in this quotation is meant to build the bridge to the topic of the ensuing discussion, namely the discussion of the connate πνεῦμα; it presents itself as the application of a general account to the more specific case of animal motion. The formulation that in animate bodies there must be “some body of that sort” is not immediately clear. Of what sort is the required body: is it meant to be something that moves by being moved or is it meant to be the body by which desire imparts movement? And does this amount to saying that such a body is required in addition to desire or that desire must, in some way, be “embodied”? Obviously, the ambiguity of this formulation is related to philosophically challenging questions concerning the relation between the psychic capacities and the body; and, apparently, Aristotle does not attempt to clarify these questions here. Still, a few 16

DMA 10, 703a4-6.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

287

paragraphs later, at the end of the discussion of the connate πνεῦμα, Aristotle gives another reference to the general scheme that will help us to clarify the envisaged role of the announced “body”: ᾯ μὲν οὖν κινεῖ κινουμένῳ μορίῳ ἡ ψυχή, εἴρηται, καὶ δι’ ἣν αἰτίαν. It has now been explained by which moved part the soul imparts movement, and through what cause.17

This concluding remark is important for a number of reasons. To begin with, it seems to correspond to the announcement made at the beginning of DMA 6, that it remains to inquire how the soul moves the body18. Obviously, Aristotle assumes that this task has now been completed. More specifically, our concluding remark mentions a “moved part”, which seems to pick up on the announcement made in the previous quotation; it seems to be safe then to interpret “part” as the “bodily part” or the “body” that has been announced at the beginning of the chapter. Since this (bodily) part is said to be moved, we are now in a better position to understand the ambivalent formulation from the beginning of the chapter, namely that there must be “some body of that sort”: at least, it seems clear by now that Aristotle at the beginning of the chapter wished to refer to a bodily part that moves by being moved. Furthermore, it is clear from the context – since the last two quotes sandwich, as we said, the discussion of the connate πνεῦμα – that the connate πνεῦμα is the moved (bodily) part by which the soul imparts movement. Last but not least, the quotation uses the instrumental dative that we encountered in our previous discussion of the general scheme in DA III 10: the moved part by which the soul imparts movement. Given that our concluding passage responds to the opening of the chapter where “the account which states the cause of motion” was explicitly appealed to, it is safe to interpret the occurrence of this instrumental dative as a reference to position (BW) within the general scheme from DA III 10. This again allows us to conclude not only that all particular elements of the general scheme ((UM), (MM), (BW), (MB)) in DA III 10 are found in the DMA, but also that the DMA construes (BW), just as we have argued DA III 10 does, as a reference to the bodily part or instrument by which the soul or a specific faculty of the soul moves the body 17 18

DMA 10, 703a28-29. DMA 6, 700b9-10.

288

CHRISTOF RAPP

(and not, as in Physics VIII, as a reference to the moved mover by which (UM) moves (MB)). To conclude, the treatise De Motu Animalium, for one, presupposes and makes use of “an account stating the cause of motion” that equals or is quite similar to the general scheme, as unfolded in DA III 10. It is interested not only in the common cause of animal locomotion, but also in the more specific question of how the soul (or a specific faculty of the soul) moves the body. In the course of investigating this latter question, it inquires into the bodily part by which the soul imparts movement to the animal as a whole. All this prepares the stage for explaining the most-significant cross-reference of our cliffhanger passage. 5. The bodily instrument and the reference to the treatment of erga that are common to body and soul (DA III 10, 433b19-21) We turn back, finally to our cliffhanger passage in DA III 10. Recall that, after having introduced the general scheme and after having filled in all positions with the exception of position (BW), Aristotle hastens to add a few remarks on just this point, i.e. the “by which”-relation, and uses this occasion to refer his readers to a different project or treatise: ᾯ δὲ κινεῖ ὀργάνῳ ἡ ὄρεξις, ἤδη τοῦτο σωματικόν ἐστιν – διὸ ἐν τοῖς κοινοῖς σώματος καὶ ψυχῆς ἔργοις θεωρητέον περὶ αὐτοῦ. The instrument by which desire initiates motion is already something bodily; accordingly, it is necessary to examine it among the functions common to body and soul.19

The opening phrase, “the instrument by which,” we have argued, picks up on the previously introduced fourfold scheme and, in particular, the second factor Aristotle introduces: “second, that by which it [the mover] initiates motion …”. We have already commented on the ambiguity of the rendering “the instrument / ὄργανον by which”, i.e. that it can be taken as a neutral explication of the instrumental dative or as introducing an actual part or organ of the body. Here, we get the explicit assertion that the instrument by which motion is imparted is something bodily, i.e. a part or organ of the animal’s body. And, apparently, the 19

DA III 10, 433b19-21, transl. Shields.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

289

point of this assertion is to explain why the discussion of this instrument cannot reasonably belong to the project of De Anima, but must belong to a different kind of enterprise. This different enterprise is characterized as one that, unlike the De Anima itself, deals with functions or attributes that are “common to body and soul”; again, this recalls the opening of the Parva Naturalia, where, in the first chapter of De Sensu, Aristotle announces that he will deal with those attributes of animals that are “common to body and soul, such as perception, memory, passion (θυμός), appetite and in general desire as well as pleasure and pain”20. One worry concerning this remark is this: as a matter of fact, certain attributes or functions that are common to body and soul are discussed in the course of De Anima; indeed, Aristotle states in DA I 1 that most of the functions he is going to discuss “are not without the body”21. Sense-perception, in particular, which is mentioned as the first example for attributes that are common to body and soul, is one of the main topics of De Anima. Presumably, then, we should not take this formulation in a merely extensional sense; rather Aristotle seems to be thinking of a certain respect in which, or perspective from which, these attributes or functions are dealt with, namely: insofar as they are common to body and soul. And indeed, when we glance through the treatises of the Parva Naturalia we find that many of them are interested in particular parts or regions of the body that are associated either with perception or nourishment or the place of vital heat. Some of them, for example, converge in highlighting the role of the heart for the functions in question. This interest in particular parts of the body and in particular physiological processes might be taken to exemplify what it means to inquire into such functions insofar as they are common to body and soul, especially since the De Anima, which is not the right place for such an inquiry, actually refrains from going into too much anatomic detail. Next, we should face the fact that the Parva Naturalia does not actually include a treatise that deals with the anatomic preconditions of animal motion, whereas the treatises De Motu Animalium and De Incessu Animalium do provide anatomic and physiological details of animal motion, but have not been transmitted within the Parva Naturalia. Now, since the heading Parva Naturalia is, of course, not Aristotelian, and 20 21

De Sensu 1, 436a6-10. DA I 1, 403a7.

290

CHRISTOF RAPP

since the composition of this collection might be due to some amount of historical contingency, one might ponder the possibility that one of the two treatises that actually engage with the bodily conditions of animal motion might have been treated by Aristotle as part of a collection including the other treatises of the Parva Naturalia22. This possibility might be supported by appeal to the quoted lines from De Sensu, where Aristotle mentions, among the functions common to body and soul to be dealt with, appetites and desire. For since desire, ὄρεξις, is not explicitly dealt with in any of the treatises that have come down to us as part of the Parva Naturalia, it is tempting to think that Aristotle himself associated an additional treatise with the general project of the Parva Naturalia – namely a treatment of the function of desire insofar as it is common to body and soul. Now, De Incessu Animalium does in fact deal with the anatomic side of animal locomotion and animal locomotion is the immediate effect of episodes of desire; however it does not really deal with desire itself and thus does not deal with desire “insofar as it is common to body and soul”. Hence, the fact that Aristotle speaks of a treatment of desire as part of the general project of treating functions common to body and soul, may apply to De Motu Animalium, but not to De Incessu Animalium. The idea would be then that the cross-reference of DA III 10 to a different kind of project is indeed a reference to the project of De Motu Animalium and that, by describing this project as an inquiry into certain functions insofar as they are common to body and soul Aristotle clearly specifies the project of De Motu Animalium as part of or at least akin to the more general project of the treatises collected in the Parva Naturalia. In addition to the opening of the De Sensu, the last line of De Motu Animalium suggests that the treatise should be seen as embedded into the general project of the Parva Naturalia: So then, concerning the parts of each animal, and concerning soul, and again concerning perception and sleep and memory and shared animal motion, we have explained the causes…23

In addition to these considerations, which were meant to narrow down the scope of the cross-reference of DA III 10 and to characterize the 22 For a full discussion of the relation between De Motu Animalium and the Parva Naturalia, see M. Rashed (2004) and A. Laks (forthcoming). 23 DMA 11, 704a3-b1.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

291

nature of the specific endeavour Aristotle has in mind, we have already encountered a much more substantial link between DA III 10 and De Motu Animalium – namely, first, the prominence of the general fourfold scheme in the latter treatise and, second, the fact that as one of the main results of the entire treatise, DMA 10 claims to have identified the instrument by which the soul moves the body, which is remarkable given that the identification and treatment of such a bodily instrument is the very purpose of the cross-reference we are dealing with. 6. The preview of the account of the bodily instrument (DA III 10, 433b21-27) As for the question of the bodily instrument by or through which motion is initiated, our cliffhanger passage first refers the reader to another project or treatise and then gives a brief preview of some of the main theses of this other project or treatise: Νῦν δὲ ὡς ἐν κεφαλαίῳ εἰπεῖν, τὸ κινοῦν ὀργανικῶς ὅπου ἀρχὴ καὶ τελευτὴ τὸ αὐτό – οἷον ὁ γιγγλυμός· ἐνταῦθα γὰρ τὸ κυρτὸν καὶ τὸ κοῖλον τὸ μὲν τελευτὴ τὸ δ᾽ ἀρχή (διὸ τὸ μὲν ἠρεμεῖ τὸ δὲ κινεῖται), λόγῳ μὲν ἕτερα ὄντα, μεγέθει δ᾽ ἀχώριστα. πάντα γὰρ ὤσει καὶ ἕλξει κινεῖται· διὸ δεῖ, ὥσπερ ἐν κύκλῳ, μένειν τι, καὶ ἐντεῦθεν ἄρχεσθαι τὴν κίνησιν. For now, though, to summarize: That which initiates motion instrumentally is where the starting point and the end point are the same, for instance, in a hinge – since here the convex is the end point and the concave the starting point (for which reason the one is at rest and the other is moved), and though differing in account, they are inseparable in magnitude. For all things are moved by pushing and pulling; consequently, it is necessary, just as in the case of a circle, for something to remain fixed and for the motion to begin from there.24

Aristotle leaves it at that. What we get is a very condensed list of some main theses and motifs of the envisaged25 inquiry into functions 24 DA III 10, 433b21-27, based on Shields’ translation. I changed Shields’ translation “something initiates motion instrumentally when the starting point and the end point are the same” into “That which initiates motion instrumentally is where (ὅπου) … are the same.” 25 By “envisaged” I do not want to say that this other project is necessarily one to be carried out in the future; in principle, the reference to another treatise might also be a back-reference to an existing work.

292

CHRISTOF RAPP

that are common to body and soul. The first sentence of the preview introduces “that which moves instrumentally” (τὸ κινοῦν ὀργανικῶς) as the subject of the following remarks; the reference is to a mover that moves something in a way instruments do. Although this is an unusual and ambiguous rendering, the immediate context makes it clear that the reference is to the moved mover and, more specifically, to the instrument, namely the bodily instrument identified in 433b19, by which the non-bodily moved mover initiates movement. With regard to this instrumental mover the preview consists of the following claims: (i) it is located where the starting point and the endpoint are the same and this sameness of the starting point and the endpoint is just like in a hinge or joint (οἷον ὁ γιγγλυμός), (ii) for here, in the hinge or joint the convex and the concave are the starting point and the end point respectively, (iii) therefore (διὸ) the one of them is at rest, while the other is moved or moves itself. (iv) The relation between the two parts of the hinge, the convex and the concave part or the starting and the end point respectively, is further characterized as being different in account (λόγῳ μὲν ἕτερα), but inseparable in size or extension (μεγέθει δ᾽ ἀχώριστα). (v) Next we are told somewhat unrelatedly that everything is moved by pushing and pulling – which is obviously meant to say that pushing and pulling is facilitated by the aforementioned internal differentiation of the hinge or the hinge-like parts into a moving and a resting part. This kind of connection is confirmed by the next and final thought, that (vi) motion arises where something (for example the resting point within a hinge or joint) is at rest, just like in a circle. In several respects the content of this preview is somewhat surprising, if not confusing. It is first of all surprising because of what it does not mention: the soul, a particular organ (or so it seems), the psycho-physical nature of desire, etc. It is also surprising that most of this preview consists in remarks on the hinge or joint. Is the main purpose of this preview to refer us to particular joints within the animal’s body? Also, if there are many such joints, is the soul meant to use all of them to move the body? Furthermore, Aristotle introduces the key notion of γιγγλυμός as an example or comparison, for he says “just like e.g. the hinge” (οἷον ὁ γιγγλυμός). And even if we can make good sense of the dominant role of the hinge-theme within this passage, we are faced with a second, quite different comparison: the comparison with a circle.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

293

Confusing as it may be, the preview section introduces motifs and claims that we also find in the De Motu Animalium. If the De Motu Animalium actually takes up the main motifs and claims from our current passage, this provides a nice confirmation of the suspicion formulated in the previous sections, that it is the project carried out in De Motu Animalium to which Aristotle wants to refer us in the cliffhanger passage. In fact, a century ago, Werner Jaeger26 was the first to highlight the parallels between the cliffhanger passage in DA III 10 and DMA; this was part of his attempt to refute those scholars, starting with Valentin Rose in the 19th century27, who questioned the authenticity of De Motu Animalium, partly because they had difficulties determining the systematic place of this treatise within the Aristotelian oeuvre (since, after all, it seems to overlap with the biological treatises, but also to include cosmological and even metaphysical parts). Now, the authenticity of DMA is no longer an issue; nor is the agreement between DA III 10 and DMA really controversial anymore28. However, what is still worth exploring is how these two texts mutually shed light on one another29. For on the one side we need the full account provided in De Motu Animalium in order to make sense of the scattered remarks on hinges and end- and starting points in the preview passage, while on the other side the preview passage is of greatest importance in that it offers a summary and thus a candid view of what Aristotle himself takes to be the most important results of the project carried out in De Motu Animalium; this latter direction of mutual interpretation seems to be particularly striking, since modern readers of DMA would certainly highlight quite different aspects of the treatise or would highlight them under different descriptions, as for example the practical syllogism in DMA 7, the argument for cardiocentrism in DMA 9 or the introduction of the connate πνεῦμα in DMA 10. It might be worthwhile, then, to go through the main topics of the preview one by one. (i) Where the starting point and the endpoint are the same. That in the anatomy of an animal the joint is the starting point of one limb, and the endpoint of another is a claim that actually pervades the treatise 26

See W. Jaeger (1913). See V. Rose (1854). 28 For a recent survey see C. Rapp (forthcoming), section 3; see also M.C. Nussbaum (1978), p. 3-12. 29 For a recent contribution to this kind of comparison see A. Laks (forthcoming). 27

294

CHRISTOF RAPP

DMA30. Also, Aristotle says in different contexts within DMA that these points are one in a certain sense31, while different in another. The Greek word γιγγλυμός either means the joint or the hinge of the door. However, the context concerns the animal bodies, not the interior of a building. Hence, it is most probably meant to refer to the joint, which is as a matter of fact, a favourite topic of De Motu Animalium. One might object that Aristotle only says that the case of the bodily instrumental mover is like the case of a γιγγλυμός, so that γιγγλυμός could still designate the door hinge and not the anatomic part; we will however suggest a different explanation for the “like” at the end of this discussion. (ii) The convex and the concave. It seems clear that an internal differentiation within the hinge or joint facilitates the phenomenon that one part moves or is moved, while another part is at rest. Again this is a claim that recurs several times in DMA32. It is only here in DA III 10, though, that Aristotle describes these two different points within the joint as the convex and the concave. Obviously, this is meant to illustrate that the starting point and the endpoint actually intertwine as though they were just one point. Anatomically speaking, Aristotle’s description seems to apply to at least two types of joints, the ball-and-socket joint (as in the hip) and the hinge-like joint (as in the elbow). There are other types of joints, however, in which the starting point and the end point are not really related as the convex and the concave (for example the saddle joint or the pivot joint); perhaps Aristotle was not aware of this anatomic detail, or perhaps he found the description in terms of convex and concave more suitable for illustrating the spatial unity of these two points within a joint. (iii) One part is at rest, while the other is moved. That movement requires something that is unmoved or at rest, is a central thesis of DMA. In DMA 1 this idea is first illustrated with regard to the internal resting point within the animal and within the animal’s joints. It is because something must be at rest, when something else is to be moved, that the animal has joints33. In DMA 2 Aristotle argues that the movement of 30 See DMA 8, 702a22-23: ἡ δὲ καμπὴ ὅτι ἐστὶν τοῦ μὲν ἀρχὴ τοῦ δὲ τελευτή, εἴρηται – “It has been said that the joint is the starting point of one thing and the end point of one other thing.” 31 See DMA 1, 698b1, DMA 8, 702a30, DMA 9, 702b30. 32 See most notably DMA 1, 698a16-17, 21-22, DMA 10, 703a12-14. 33 See DMA 1, 698a17.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

295

animals also requires an external resting point, i.e. some sort of platform that offers resistance and against which the animal can support itself. In DMA 4 he summarizes this theorem of the internal and external resting point by saying: Τὰ γὰρ τοιαῦτα πάντα ἀνάγκη καὶ ἐν αὑτοῖς ἔχειν τὸ ἠρεμοῦν καὶ ἔξω πρὸς ὃ ἀπερείσεται. For all such things [animals, scil.] must have something stationary both in themselves and outside themselves against which they push off.34

In DMA 8-9 Aristotle tries to locate the ultimate origin of movement within the body. For this purpose he repeats and stresses the insights from DMA 1 on the need of an internal resting point: Ὅταν γὰρ κινῆται ἐντεῦθεν, ἀνάγκη τὸ μὲν ἠρεμεῖν τῶν σημείων τῶν ἐσχάτων, τὸ δὲ κινεῖσθαι· ὅτι γὰρ πρὸς ἠρεμοῦν δεῖ ἀπερείδεσθαι τὸ κινοῦν, εἴρηται πρότερον. For whenever the animal is moved from here [from the joint, scil.], one of these endpoints must be at rest, while the other must be moved; for it has been stated before that what imparts movement must press against something at rest.35

At a later stage of the discussion, after it has already been established that the origin of movement must be in the middle of the animal’s body, Aristotle uses the same pattern of argument again in order to show that if the middle region of the body serves as the origin of movement, there must be something that is at rest in that region as well: there must be something at rest, Aristotle says, that imparts movement, yet is not moved36. Here, in DMA 9, this argument is used to make the point that the soul, which is unextended, must be located in the middle region of the body and must be unmoved, in order to impart movement to both sides of the body. Finally, after having introduced the connate πνεῦμα Aristotle relates the πνεῦμα to the soul by the use of the same pattern: Τοῦτο δὲ πρὸς τὴν ἀρχὴν τὴν ψυχικὴν ἔοικεν ὁμοίως ἔχειν ὥσπερ τὸ ἐν ταῖς καμπαῖς σημεῖον, τὸ κινοῦν καὶ κινούμενον, πρὸς τὸ ἀκίνητον. 34 DMA 4, 700a18-20, the translation is taken from B. Morison’s English translation of the DMA in C. Rapp and O. Primavesi (forthcoming). 35 DMA 8, 702a24-27. 36 DMA 9, 702b34-35.

296

CHRISTOF RAPP

But it seems that this [the connate πνεῦμα, scil.] stands to the psychic origin just like the point in the joint, the one that imparts movement and is moved, stands to the unmoved one.37

It follows from this analogy that the connate πνεῦμα must be adjacent to the heart, where the “psychic origin” of movement is located. Since only a few lines later Aristotle claims that it is now clear what the instrument is by which the soul moves the body, it is abundantly clear to say that this idea, that a moved or moving part entails a corresponding resting point, truly pervades the De Motu Animalium from the beginning to the end. (iv) Different in account, but inseparable in size. There is no verbally identical formulation in De Motu Animalium, but it seems that in this treatise Aristotle nevertheless describes a similar idea with different terminology. That the two points within a hinge or joints are different in account most probably refers to the fact that one is the starting point, while the other is the endpoint – an idea that is often repeated in the course of DMA. In one section of DMA Aristotle speaks of two endpoints38, which however is no objection to the claim that they are different in account, for one of them is the endpoint of x, while the other is the endpoint of y. The formulation that they are inseparable in size or magnitude has no counterpart in DMA; clearly it is meant to emphasise the unity of these points in spite of their numerical and definitional distinctness. In DMA Aristotle has adopted a different model for accounting for the unity and distinctness of these two points, for he keeps saying that they are one and two in potentiality and in actuality39, i.e. when for example the arm is straight, the endpoint of the upper arm and the starting point of the lower arm are actually one, while potentially two, whereas, when the arm is bent, they are two in actuality and one in potentiality. Of course, we can only speculate as to why he changed the terminology; one obvious advantage of putting this relation in terms of actuality and potentiality consists in the wider applicability of this latter model (for example, it is applicable to bodily regions and parts that are not related as the convex to the concave). 37

DMA 10, 703a12-14. DMA 9, 702b16. 39 See e.g. DMA 1, 698a19-21, 698a27-b1, DMA 8, 702a30-31, DMA 9, 702b25, 702b30. 38

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

297

(v) Everything is moved by pushing and pulling. The idea that motion needs some kind of pushing was implicitly presupposed or explicitly mentioned in all the passages that argue for the necessity of resting points, for it is by pushing against these resting points that the moved parts move themselves. In addition, there is a crucial passage in DMA 10 that points out that the thermically incited alterations of the connate πνεῦμα, i.e. its contractions and expansions, must result in pushing and pulling40. (vi) Movement derives from where something is at rest, like in a circle. In DMA 1 Aristotle introduces the idea that there must be an unmoved starting point of motion (DMA 1, 698b1-2, 4-7) by referring to a geometrical circle; joints are used, he says, like the κέντρον41, the centre of the circle, is used when someone draws the circumference with the help of a compass-like instrument, fixing it at the centre of the circle. In DMA 1 Aristotle explicitly refers to the movement of a diameter along the circumference, which seems to suggest that its origin in the centre remains at rest, while the endpoint on the circumference is moved42. In Aristotle’s time the centre of a circle was also used for illustrating the fulcrum of a lever, which comes close to the role that the resting point is supposed to play in Aristotle’s account43. The above comparison makes clear that DMA is indeed the project that DA III 10 envisages. There are many profound parallels, whereas the differences concern only minor or terminological points44. So much is clear, but how does this help us to understand what Aristotle took to be the nature and the main result of the envisaged project of DMA? 7. The nucleus of the envisaged account As we said, the preview puts a perplexing emphasis on the joint / hinge motif, apparently because the joint or hinge represents an object that is 40 See DMA 10, 703a19-21: τὰ δ’ ἔργα τῆς κινήσεως ὦσις καὶ ἕλξις, ὥστε δεῖ τὸ ὄργανον αὐξάνεσθαί τε δύνασθαι καὶ συστέλλεσθαι – “The products of this motion are pushing and pulling, with the result that the organ can both be increased and contract.” 41 698a18: ὥσπερ γὰρ κέντρῳ χρῶνται ταῖς καμπαῖς …. 42 See DMA 1, 698a22-23: ὥσπερ ἂν εἰ τῆς διαμέτρου …. 43 See C. Rapp (forthcoming), section 9. 44 Clearly there is another treatise in the Aristotelian oeuvre that deals with joints and resting points, De Incessu Animalium, however this latter treatise does not address the question of how the soul moves the body, nor does it concern the cause of animal locomotion in general.

298

CHRISTOF RAPP

internally differentiated or subdivided; and this latter thought is of interest, again, because the initiation of movement (in one way or the other involving pushing and pulling) is facilitated by the presence of a moved part supporting itself against an unmoved part. If so much is conceded, one might wonder whether the joint or hinge mentioned in the preview is just meant as a simile or comparison (perhaps for all kinds of configuration where we get a moved and an unmoved part) or whether the preview is meant to prepare the reader for the thought that the bodily instrument we are looking for is a joint or a joint-like bodily part within the animal. Indeed, the rendering “like e.g. a joint (οἷον ὁ γιγγλυμός)” seems to give preliminary support for the simile-reading. One recent version of the simile-reading has been suggested by Christopher Shields (which we treat here as representing the most promising version of the simile-reading): When contending that the starting point (archê) in intentional action is the same as its end point (teleutê), Aristotle means that for a particular action, the goal first moves the faculty of desire as its final cause and then also serves as the terminus of the action, that in which it is completed. His illustration serves not to explicate how the first moving cause is also the terminus, however, but rather how one and the same state of affairs can yet differ in being, insofar as it is both the starting point and the end point.45

This suggestion seems to make two interrelated claims: first that in an intentional action the goal that triggers the desire and, thus, is the starting point of the action, is at the same time the terminus by which the action is completed, and second that the illustration given (i.e. the joint / hinge) just serves to show how the same state of affairs can be one and same, while still being distinguishable in being, as when one and the same point serves as both starting point and endpoint. As for the latter claim, it seems clear that the identity of the starting point and the endpoint is an important implication of the joint / hinge; it is less clear, though, whether this is the main reason why the joint / hinge is invoked. The first claim seems to be that the preview ultimately aims at a thesis that is not about joints or bodily parts at all, but about the double role of an intended goal in the initiation and execution of an action. Although Aristotle would probably subscribe to the thesis that the starting point and the endpoint of an intentional action are in this sense the same, it seems unlikely that this is 45

C. Shields (2016), p. 362-363.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

299

Aristotle’s main concern in the given context. To begin with, if DMA is, as we argued in the previous sections, the reference of this preview passage, it would be strange to assume that, although DMA takes pains to discuss the structure of real flesh-and-bones joints, the preview of this treatise uses them only as an analogous illustration. Second, it would remain unclear how the identity of the starting point and the endpoint of an intentional action could shed any light on the corporeal instrument by which the soul moves the body. Finally, Shields’ reading seems to presuppose a peculiar understanding of the conjunction ὅπου in section 433b21-27; indeed he translates the first sentence of this passage as “something initiates motion instrumentally when the starting point and the end point are the same”, rendering ὅπου as “when”. If we translate ὅπου, however, as “where” (which, anyway, seems to be the more natural translation), it is clear that the reference is to a certain place – namely a place in the animal’s body. This spatial meaning is important, not only in the parallel passages of De Motu Animalium, but also in the cliffhanger passage of DA III 10; for example, in the end of that passage Aristotle clearly refers to the point at rest, from which motion originates, with the locative term ἐντεῦθεν. Let us conclude, then, that the attempt to read the occurrence of the joint / hinge in the preview passage as a mere illustration, simile or analogy is not quite satisfactory (given that even the most promising version of this kind of reading, the one suggested by Christopher Shields, turned out to be unconvincing). This is due primarily to the fact that the reference to real joints, i.e. joints as parts of animal bodies, pervades the treatise DMA, which again has been shown to contain the project announced in our cliffhanger passage. It remains to elaborate on the alternative, i.e. the view that the mention of joints46 in the preview is not just an illustration of a different, non-anatomic state of affairs. As already mentioned, the preview seems to provide the unique opportunity to hear Aristotle himself commenting on what he takes to be the main results of DMA and, more specifically, what he takes to be the nature and the role of the instrument by which the soul or a particular capacity of the soul is supposed to move the body. Following the path of 46 The observation that the word γιγγλυμός can also mean hinge and not only joint seems to be the proverbial straw at which non-literalist readings try to grasp. Having ruled out such readings now, there is no need any more to consider the alternative “hinge”.

300

CHRISTOF RAPP

literal, non-analogous, readings of this passage we are unambiguously told that joints are the place where such a bodily initiation of movement takes place. Again, this does not seem to be fully satisfactory, for example because there are many such joints in animal bodies (i.e. the elbow, the hip, the knee, the shoulder, the ankle, the wrist, and the like) and it seems unlikely that all of them stand in the same relation to the (one) moving principle of the soul (if, say, the movement in the ankle is dependent on the movement in the knee and the movement in the knee is dependent on the movement in the hip, not all of these joints can have the same relation to the first origin of movement within the animal’s body). Also, the question about the bodily instrument of motion was put in the singular, so that we have no reason for expecting a multitude of such instruments. This seems to be one of the worries that has driven commentators to resort to simile-readings. However, it is a plain fact that Aristotle considers bodily joints to be places from which movement originates owing to their internal subdivision into a starting point and an endpoint. In this sense, joints are starting points of movement in a thoroughly concrete and literal sense and are not merely meant as an illustration. It is important to observe, though, that on Aristotle’s account, joints are not the only bodily regions that are internally subdivided into a starting point and an end point. As we said, Aristotle refers to joints and their peculiar structure (internal subdivision; starting point and endpoint are the same in a sense, while different in another sense; movement is originated when one of them is at rest, while the other is moved and supports itself against the resting one) throughout the treatise DMA. However, he applies this structure not only to ordinary joints, but also to other bodily regions. Above all, he argues (i) that if the middle region of the animal’s body is to serve as an origin of movement, this region too must contain something at rest that imparts movement47, (ii) that in order to move both sides of the body simultaneously the soul must reside as something unmoved within the middle region of the body, just as the unmoved part within a joint is adjacent to the moved part48, and (iii) that the connate πνεῦμα stands in a similar relation to the psychic origin as the moved point within a joint to the unmoved one49. This amounts to saying that 47 48 49

DMA 9, 702b12 sq. DMA 9, 702b34-703a3. DMA 10, 703a11-13.

JOINTS AND MOVERS IN THE CLIFFHANGER PASSAGE

301

joints are not only themselves origins of movement, but also serve as a model for accounting for the structure of other origins of movement within animal bodies50 – and this holds most notably of the most central origin of movement within the body, which is not at the same time the endpoint of something else51. It is, as it were, in a joint-like region of the body where the moving principle of the soul imparts motion with the help of a particular kind of instrument, and saying that this region is “joint-like” means precisely that there is an internal differentiation in the sense outlined above, i.e. that this part or region is one or unified in one sense, while two in another, and that one part is at rest, while the other moves by supporting itself against the unmoved one. On this reading, the formulation “like e.g. a joint (οἷον ὁ γιγγλυμός)” does not mean that the joint is only an illustration; it just means that the joint is only one example among others that instantiate the structure of being internally subdivided into a starting point and an endpoint and into a resting and a moving part. If this is the main point of the preview it refers us to the context of an argument that is unfolded between the second half of DMA 8 (from 702a21) and the end of DMA 1052. The preview in DA III 10 might be important for the interpretation of DMA, then, because it tells us that it is here, in the context of chapters 8-10 (and not, say, in the more often quoted chapters 6 and 7 of DMA), that we find the main result of the DMA’s project. Still, there seems to be a more important conclusion to draw from the preview passage. The preview introduces the nucleus of the envisaged account by referring to joints and thus makes clear that we have to read the main results of these later chapters of DMA in the light of the discussion of the structure of joints and unmoved platforms from the earlier chapters of the treatise. The argument of chapters 8 to 10 culminates first in the claim that the soul is a resting point within the animal’s 50 Is this a concession to simile-readings? No, because joints are not only illustrations, but are themselves origins of movement and literally incorporate the decisive structure of having a point that is a starting point and an endpoint. Yes, insofar as joints are at the same time used as a model for the more generally applicable structure of comprising a starting point and an endpoint in such a way, that one of them is moved, while the other is at rest. Still (even if one grants that a “model” is not too far from a simile or analogue), this particular model is not used for any configuration that displays a sort of identity of the starting point and the endpoint, but just for bodily regions or parts. 51 DMA 8, 702b6-7. 52 For a reconstruction of this argument see P. Gregoric (forthcoming).

302

CHRISTOF RAPP

body (end of chapter 9) and next in the introduction of the connate πνεῦμα as the particular stuff by which motion is imparted and transmitted to various parts of the body (chapter 10). Chapter 10 of DMA concludes the main argument of these chapters by saying that it has now been said by which part the soul imparts movement53. One might thus wonder why the preview does not straightforwardly mention the connate πνεῦμα as the main invention of the envisaged project. It is precisely with respect to this latter point that the cliffhanger passage of DA III 10 might help to change the usual perspective. For if we take the hint in the preview seriously, it seems that Aristotle is not so much interested in promoting the particular kind of stuff that will do the job in the end; rather he seems eager to emphasize his analysis of the general structure that facilitates movement (i.e. the subdivision of a starting point and endpoint, that they are one in a sense, but two in another sense, and that one part must be at rest, while the other moves). It is this general structure that defines certain constraints for the anatomic structure that is apt to initiate and to transmit the resulting motion. In particular there are constraints for the kind of stuff that is responsible for amplifying and transmitting the initially feeble alteration that is initiated by perception and desire. The claim that the connate πνεῦμα fits the bill is mentioned almost in passing. The insight that of all kinds of stuff it is the connate πνεῦμα that happens to have all the required properties, is only borrowed from other, more physiological treatises, while the argumentative effort of these chapters aims to establish the thesis that the initiation of movement must derive from a central point in the body that displays the aforementioned general structure. For this reason, it is neither by mistake nor out of negligence that the preview in DA III 10 does not mention the connate πνεῦμα; it is not mentioned because the emphasis both of the preview and of the treatise DMA itself is on the thesis that the origin of movement must be located within an anatomic environment that displays the aforementioned structure; and it is this general structure that connects the specific case of animal locomotion with other kinds of movement. This synoptic interest, – to connect the account of animal locomotion with other kinds of movement once again – is quite typical of the project of the DMA, which, from the beginning, introduces general kinetic principles that are likely to apply to various kinds of motion. 53

DMA 10, 703a28-29.

THE MOVER(S) OF RATIONAL ANIMALS DE ANIMA III 11 IN CONTEXT Jennifer WHITING

Any attempt to understand De Anima III 11 as a self-contained unit – even one drawing on points previously established – is misguided. For the opening lines (=[1*]) are clearly parenthetical. And what follows them (=[2]) picks up where the final lines of III 10 (=[0*]) left off 1. [0*] Ὅλως μὲν οὖν, ὥσπερ εἴρηται, ᾗ ὀρεκτικὸν τὸ ζῷον, ταύτῃ αὑτοῦ κινητικόν· ὀρεκτικὸν δὲ οὐκ ἄνευ φαντασίας· φαντασία δὲ πᾶσα ἢ λογιστικὴ ἢ αἰσθητική. ταύτης μὲν οὖν καὶ τὰ ἄλλα ζῷα μετέχει. In general then, as previously said, insofar as an animal is capable of desire, it is on account of this capable of moving itself. But it is not capable of desire without φαντασία. And all φαντασία is either calculative or perceptual. Of this, then, even the other animals have a share. (433b27-30) [1*] (Σκεπτέον δὲ καὶ περὶ τῶν ἀτελῶν τί τὸ κινοῦν ἐστιν, οἷς ἁφῇ μόνον ὑπάρχει αἴσθησις, πότερον ἐνδέχεται φαντασίαν ὑπάρχειν τούτοις, ἢ οὔ, καὶ ἐπιθυμίαν. φαίνεται γὰρ λύπη καὶ ἡδονὴ ἐνοῦσα, εἰ δὲ ταῦτα, καὶ ἐπιθυμίαν ἀνάγκη. φαντασία δὲ πῶς ἂν ἐνείη; ἢ ὥσπερ καὶ κινεῖται ἀορίστως, καὶ ταῦτ’ ἔνεστι μέν, ἀορίστως δ’ ἔνεστιν.) (We must inquire also about the imperfect , what is the mover of these, to whom only αἴσθησις by means of touch belongs, and whether it is possible for φαντασία to belong to them or not, and also appetite . For pain and pleasure are present in them. And if these , it is necessary that appetite is 1 The translations are my own. Passages [1*] to [5*] provide a complete translation of De Anima III 11. The asterisks indicate the final version of my translation; lack of an asterisk, as in [2] here, indicates an as yet incomplete translation with further decisions to be made and defended in what follows.

304

JENNIFER WHITING

as well. But how could φαντασία be present in them? Or is it the case that just as they move indeterminately, so too these are present in them, but indeterminately?) (433b31-434a6) [2] Ἡ μὲν οὖν αἰσθητικὴ φαντασία, ὥσπερ εἴρηται, καὶ ἐν τοῖς ἄλλοις ζῴοις ὑπάρχει, ἡ δὲ βουλευτικὴ ἐν τοῖς λογιστικοῖς. (πότερον γὰρ πράξει τόδε ἢ τόδε, λογισμοῦ ἤδη ἐστὶν ἔργον· καὶ ἀνάγκη ἑνὶ μετρεῖν· τὸ μεῖζον γὰρ διώκει· ὥστε δύναται ἓν ἐκ πλειόνων φαντασμάτων ποιεῖν). Perceptual φαντασία, as previously said, belongs on the one hand even among the other animals, whereas deliberative belongs on the other hand among rational animals2. (For whether one will do this or that is already the work of calculation. And it is necessary to measure by one , for pursues the greater ; so is able to make one out of several φαντάσματα.) (434a6-10)

As [1*] indicates, Aristotle has been discussing what accounts for the self-generated movements of animals. He is specifically concerned to explain locomotion, as distinct from the sorts of movements in place to which the creatures mentioned in [1*] are limited. These creatures do not locomote. So whatever sort of φαντασία might be required to explain their movements in place is not relevant here, where the question is what accounts for locomotion. Aristotle is thus content to toss off a few speculative and parenthetical remarks before returning to his official business3. 2

Two points here. First, I use the awkward “on the one hand … on the other hand …” here so as to flag the μὲν / δὲ structure for future reference. For I see in that structure the key to making sense of what follows, especially [3] and [4]. Second, I suspect that Aristotle uses ἐν τοῖς ἄλλοις ζῴοις and ἐν τοῖς λογιστικοῖς (rather than simple datives without ἐν) so as to signal that deliberative φαντασία may not belong to all so-called rational animals and that even perceptual φαντασία may not belong to all animals (which is precisely the possibility envisaged in [1*]). See Politics I 13, with its rare use of τὸ βουλευτικόν (on which more below): “the slave is generally without the deliberative capacity, while the female has but it is not in control , and the child has it but it is incomplete.” [1260a12-14] 3 These speculative remarks are not especially germane to our topic, but I take the idea to be that stationary animals, in virtue of their limited sense of touch, experience pleasure and pain, and so a rudimentary form of desire, but not one that involves determinate representations of objects towards or away from which they might locomote. These creatures are affected by the tangibles associated with nutrition – hot, cold, wet, and dry, and perhaps also flavor (which Aristotle treats as a kind of tangible that supervenes on the former). So these creatures experience hunger (which is an ἐπιθυμία for what is dry and hot) and thirst (which is an ἐπιθυμία for what is wet and cold). But what they experience are more like mere sensations than representations of objects. So their survival depends not on the representation

THE MOVER(S) OF RATIONAL ANIMALS

305

Seeing the parenthetical nature of [1*] might tempt us to read III 11 as a relatively self-contained unit that starts with [0*] rather than [1*]. But that too would be misguided. For that would suggest that the chapter is mainly about the two forms of φαντασία and their proper extensions. But – as I shall argue here – III 11 aims primarily to resolve an aporia that emerges towards the end of III 9 and is articulated at the start of III 10. This is an aporia about what “part” of soul is responsible for the second of the two capacities that are supposed to distinguish the souls of animals from those of plants – namely, the locomotive capacity. The first of these capacities, the κριτική or “cognitive” capacity, has been examined at length in II 5-III 8, where Aristotle discusses αἴσθησις, φαντασία, and νοῦς qua cognitive capacities4. Νοῦς and φαντασία figure again in the discussion of locomotion to which he turns at the start of III 9, but they figure thenceforth only insofar as they play a role in accounting for locomotion. If there is any relatively self-contained unit here, it is what we now know – thanks to the editors who divided the De Anima into chapters – as III 9-115. I stress this point because I worry that taking III 11 to be concerned primarily with the two forms of φαντασία and their proper extensions has led many readers, including prominent translators, to misread a key sentence. This is [3][a], which launches a highly compressed and anaphoric argument that I translate (with some gaps to be filled in as we proceed and some crucial bits in bold) as follows. [3] [a] [b] [c] [d]

καὶ αἴτιον τοῦτο τοῦ δόξαν μὴ δοκεῖν ἔχειν, ὅτι τὴν ἐκ συλλογισμοῦ οὐκ ἔχει, αὕτη δὲ ἐκείνην, διὸ τὸ βουλευτικὸν οὐκ ἔχει ἡ ὄρεξις6.

of perceptual objects as pleasant (and so to be pursued) or painful (and so to be avoided), but rather on pleasant and painful sensations together with rudimentary ἐπιθυμίαι for the persistence or cessation of these very sensations. Their movements are generally reactions to sensations, like scratching where there is an itch – for example, closing in response to noxious stimuli, opening to let warmth or moisture in, and so on. No determinate representations are required to explain such “indeterminate” movements. 4 Aristotle uses νοῦς, which is typically rendered “intellect” or “thought”, sometimes to refer to the relevant capacity of thought and sometimes to refer to the activity of thought. 5 The chapter divisions are not due to Aristotle but were established, in some cases following previous editions, in the Basle edition of 1531. 6 Most editors place a semi-colon here. But this simply encourages the standard (but in my view mistaken) reading of [4][a]. So I prefer a period (and, eventually, parentheses).

306

JENNIFER WHITING

[a] And this is the reason why _______ seems not to have / entail δόξα – [b] because _______ does not have / entail the < > that comes from reasoning, [c] but this has / entails that. [d] Whence ὄρεξις does not entail the deliberative capacity7. (434a10-12) [4] [a] [b] [c] [d] [e] [f] [a] [b] [c] [d] [e] [f]

νικᾷ δ’ ἐνίοτε καὶ κινεῖ [τὴν βούλησιν], ὁτὲ δ’ ἐκείνη ταύτην, ὥσπερ σφαῖρα, ἡ ὄρεξις τὴν ὄρεξιν, ὅταν ἀκρασία γένηται· φύσει δὲ ἀεὶ ἡ ἄνω ἀρχικωτέρα καὶ κινεῖ· ὥστε τρεῖς φορὰς ἤδη κινεῖσθαι. But sometimes _______ conquers and moves ________ , and sometimes that this, just as a sphere , desire desire whenever akrasia occurs. But by nature, the higher is always more of an ἀρχή and moves 8. So it [scil., the rational animal] is moved with three different φοραί. (434a13-16)

[5*] τὸ δ’ ἐπιστημονικὸν οὐ κινεῖται, ἀλλὰ μένει. ἐπεὶ δ’ ἡ μὲν καθόλου ὑπόληψις καὶ λόγος, ἡ δὲ τοῦ καθ’ ἕκαστον (ἡ μὲν γὰρ λέγει ὅτι δεῖ τὸν τοιοῦτον τὸ τοιόνδε πράττειν, ἡ δὲ ὅτι τόδε τοιόνδε, κἀγὼ δὲ τοιόσδε), ἢ δὴ αὕτη κινεῖ ἡ δόξα, οὐχ ἡ καθόλου, ἢ ἄμφω, ἀλλ’ ἡ μὲν ἠρεμοῦσα μᾶλλον, ἡ δ᾽οὔ. 7 Aristotle tends to use ὄρεξις as a generic term for desire, covering both ἐπιθυμία (or “appetite”, a form of non-rational desire) and βούλησις (or “wish”, a form of rational desire associated with calculation). As he explains in DA III 10, in lines that are highly relevant to our topic (lines that follow immediately on [E] below): “And in fact, νοῦς evidently does not move without ὄρεξις, for βούλησις is ὄρεξις. And whenever an animal is moved in accordance with calculation [λογισμός], it is moved in accordance with βούλησις. But ὄρεξις also moves against calculation, for ἐπιθυμία is a kind of ὄρεξις.” [433a23-26] This explains the point in [3][d]: because ἐπιθυμία is a form of ὄρεξις that does not entail the presence of a deliberative capacity, ὄρεξις, simply as such, does not entail the presence of a deliberative capacity (or, as I suggest below, of deliberative φαντασία). 8 Aristotle uses ἀρχή both in its specifically political sense (in which it refers to governing agents, such as monarchs and oligarchs) and in a more general sense (in which it refers to the principles or sources of phenomena such as motion or growth). In other words, not all principles in the general sense are governing principles. But (as I argue below) [4][e] is plausibly read as saying that higher principles are principles in a stricter sense than lower ones, by which I think he means that they are governing principles in a way that lower ones are not.

THE MOVER(S) OF RATIONAL ANIMALS

307

Τὸ ἐπιστημονικόν, however, is not moved, but remains . But since the one < δόξα> is a universal supposition or statement, while the other < δόξα> is of the particular – for the one says that it is necessary for such to do such , while the other says that this is such or that I am such – either this δόξα [scil., the particular one] moves , not the universal one, or both but the one is rather at rest, the other not9. (434a16-21)

There is much here that remains to be resolved. But one point should be clear: if we do not understand [3], which introduces δόξα, then we cannot easily understand how the argument is supposed to reach its conclusion in [5], which is about the role of δόξα in moving rational animals. Our working hypothesis should thus be that δόξα – or some item closely associated with it – functions here as a kind of “middle term” that connects [3] with [5]. But before we attempt to fill in the missing details, let us take quick look at the general context in which the argument of III 11 appears. 1. The general context: De Anima III 9-10 At the start of III 9, Aristotle explicitly sets aside the cognitive capacities by which the souls of animals are taken to differ from those of plants and turns to the locomotive capacities that are supposed to be distinctive of complete and non-mutilated animals. The official question is: what in an animal soul is responsible for the animal’s locomotion? Is it some one “part” that is separable either in magnitude or in account from other “parts” of soul? Or is it the whole soul? And if it is some one part, is it one proper to locomotion and additional to those previously 9 In the official account of φαντασία back in III 3, Aristotle uses ὑπόληψις (here rendered “supposition”) as a generic term that applies to ἐπιστήμη, δόξα and φρόνησις, as well as their opposites. Ἐπιστήμη (generally “scientific knowledge”) and φρόνησις (generally “practical wisdom”), being forms of knowledge, cannot be false, but δόξα (“belief”) can be either true or false. Tὸ ἐπιστημονικόν refers to whatever part of soul has the relevant sort of ἐπιστήμη. Insofar as the ἐπιστήμη in question here seems to be practical rather than theoretical, Aristotle may be departing from the apparently canonical use at EN 1139a6-15, where τὸ ἐπιστημονικόν is said to grasp things whose principles cannot be other than they are and is contrasted with τὸ λογιστικόν (which is explicitly associated with deliberation and said to be of things that are – or whose principles are – capable of being other than they are). But I think it worth exploring on some other occasion whether Aristotle speaks this way here because he thinks that at least some principles of the form “it is necessary for such to do such ” are on a par with the necessary truths grasped in scientific inquiry.

308

JENNIFER WHITING

mentioned? Or is it one of these – e.g., the αἰσθητικόν, the ὀρεκτικόν, or νοῦς? In sum, what is the mover [τὸ κινοῦν] of an animal? It is clear by the end of III 9 that Aristotle’s investigation is not simply about the movements of animals from one place to another simply as such. His references to ἀκρασία and ἐγκράτεια, and to practical νοῦς, reveal that his question extends to distinctively human forms of activity, including those in which virtues of character are exercised10. The conclusion of III 11, about the role of δόξα in moving rational animals, is thus focused on a special case of the more general question raised at the start of III 9. This conclusion is more specifically part of Aristotle’s response to an aporia articulated in III 10, about whether there is in the case of rational animals a single mover on a par with the one he sees in the souls of nonrational animals, namely τὸ ὀρεκτικόν. Let’s take a closer look at how the aporia arises in III 9 and at how Aristotle approaches it in III 10. Aristotle begins his search for the mover by eliminating possible candidates. [A] Ὅτι μὲν οὖν οὐχ ἡ θρεπτικὴ δύναμις, δῆλον· ἀεί τε γὰρ ἕνεκά του ἡ κίνησις αὕτη, καὶ μετὰ φαντασίας καὶ ὀρέξεώς ἐστιν· That it is not the nutritive capacity is plain, because this motion [scil., locomotion] is always both for the sake of something and together with φαντασία and ὄρεξις. (432b14-16)

Aristotle appeals here to the case of plants, which have the nutritive capacity but do not locomote, and not simply because they lack the requisite organs. For as he goes on to explain, if the nutritive capacity were a mover, then – given his views about how nature works – complete and non-mutilated plants would have locomotive organs. But the crucial point is that locomotion does not occur without φαντασία and ὄρεξις. And for these αἴσθησις is required. One might thus suppose that τὸ αἰσθητικόν is what moves the animal. But Aristotle makes a parallel move to eliminate τὸ αἰσθητικόν. There exist complete and non-mutilated animals that have αἴσθησις but are stationary: and given Aristotle’s views about how nature works, these animals would have locomotive organs if τὸ αἰσθητικόν were for locomotion. 10 The idea, explained in passages [B] and [C] below, is of subjects who are either uncontrolled by reason or controlled by it. But there is, as we shall see, an important difference between the sort of control exercized by reason in an enkratic agent and the sort of control exercized by reason in a genuinely virtuous agent.

THE MOVER(S) OF RATIONAL ANIMALS

309

Aristotle then turns to the calculating capacity or part [τὸ λογιστικόν] and, more generally, νοῦς. He quickly dismisses theoretical νοῦς and – after a famous remark aimed to answer an objection and / or raise a laugh – turns to cases where νοῦς issues commands. [B] Ὁ μὲν γὰρ θεωρητικὸς οὐθὲν θεωρεῖ πρακτόν, οὐδὲ λέγει περὶ φευκτοῦ καὶ διωκτοῦ οὐθέν, ἀεὶ δὲ ἡ κίνησις ἢ φεύγοντός τι ἢ διώκοντός τί ἐστιν. ἀλλ’ οὐδ’ ὅταν θεωρῇ τι τοιοῦτον, ἤδη κελεύει φεύγειν ἢ διώκειν, οἷον πολλάκις διανοεῖται φοβερόν τι ἢ ἡδύ, οὐ κελεύει δὲ φοβεῖσθαι, ἡ δὲ καρδία κινεῖται, ἂν δ’ ἡδύ, ἕτερόν τι μόριον. For theoretical νοῦς contemplates nothing doable in action; nor does it say anything about what is to be avoided or pursued, but the movement is always either fleeing something or pursuing something. Nor, whenever it [scil., νοῦς] contemplates some such thing, does it immediately command to pursue or avoid , but frequently it thinks something fearful or pleasant but does not command fear (although the heart moves) or if pleasant, (although some other part ). (432b27-433a1) Ἔτι καὶ ἐπιτάττοντος τοῦ νοῦ καὶ λεγούσης τῆς διανοίας φεύγειν τι ἢ διώκειν οὐ κινεῖται, ἀλλὰ κατὰ τὴν ἐπιθυμίαν πράττει, οἷον ὁ ἀκρατής. καὶ ὅλως δὲ ὁρῶμεν ὅτι ὁ ἔχων τὴν ἰατρικὴν οὐκ ἰᾶται, ὡς ἑτέρου τινὸς κυρίου ὄντος τοῦ ποιεῖν κατὰ τὴν ἐπιστήμην, ἀλλ’ οὐ τῆς ἐπιστήμης. Further, even when νοῦς commands and thought says one should avoid something or pursue , does not move but acts according to appetite, as for example the ἀκρατής ; and generally we see that the one who has medical does not cure, since something else – rather than the knowledge – is in control of his acting according to this knowledge11. (433a1-6)

We might be tempted to conclude that, at least in this sort of case, it is ὄρεξις – of which ἐπιθυμία is a species – that moves the animal. But Aristotle problematizes this conclusion by turning to the converse case, that of the ἐγκρατής: [C] Ἀλλὰ μὴν οὐδ’ ἡ ὄρεξις ταύτης κυρία τῆς κινήσεως· οἱ γὰρ ἐγκρατεῖς ὀρεγόμενοι καὶ ἐπιθυμοῦντες οὐ πράττουσιν ὧν ἔχουσι τὴν ὄρεξιν, ἀλλ’ ἀκολουθοῦσι τῷ νῷ. 11 In Metaphysics IX 5, Aristotle claims that with rational capacities, which can produce contrary effects, it is ὄρεξις or προαίρεσις that determines what the agent does. In other words, it is ὄρεξις or προαίρεσις (which is a special kind of ὄρεξις) that will determine whether or not an agent uses her ἐπιστήμη. Φρόνησις, however is different: because it is inseparable from ethical virtue, its subject cannot fail to act on it. See Nicomachean Ethics VI/Eudemian Ethics V, chapters 12-13.

310

JENNIFER WHITING

But surely ὄρεξις is not in control of the movement. For enkratic subjects, who desire and have appetites, do not do what they desire but follow νοῦς . (433a6-8)

Here ends III 9, in the midst of an emerging aporia, one due to the fact that rational animals are subject to the three kinds of movement [φορά] to which [4][f] refers: the akratic, enkratic, and what I shall call the “normative” φορά (on which more below)12. The aporia arises from a pair of appearances that seem to point to different answers to the question what moves a rational animal: whereas it is ὄρεξις (in the form of ἐπιθυμία) that appears to move the ἀκρατής (who acts against νοῦς), it is νοῦς that appears to move the ἐγκρατής (who acts against ὄρεξις). And it may seem that Aristotle settles for a disjunctive conclusion according to which what moves the animal is in some cases ὄρεξις and in other cases νοῦς. For as he says in the opening lines of III 10: [D] Φαίνεται δέ γε δύο ταῦτα κινοῦντα, ἢ ὄρεξις ἢ νοῦς, εἴ τις τὴν φαντασίαν τιθείη ὡς νόησίν τινα· πολλοὶ γὰρ παρὰ τὴν ἐπιστήμην ἀκολουθοῦσι ταῖς φαντασίαις, καὶ ἐν τοῖς ἄλλοις ζῴοις οὐ νόησις οὐδὲ λογισμὸς ἔστιν, ἀλλὰ φαντασία. It appears at any rate that these two move : either ὄρεξις or νοῦς, if one counts φαντασία as a sort of νοῦς. For many follow their φαντασίαι against ἐπιστήμη, and in other animals there is neither thinking nor calculation, but φαντασία13. (433a9-12) 12 Nothing in my argument depends on this, but I suspect that Aristotle uses the term φορά here (rather than κίνησις or even πρᾶξις) because there are (as De Caelo I 3 argues) three forms of φορά and because these three forms are appropriate metaphors for the forms of behavior with which III 11 is concerned. Circular motion, being perfect, corresponds to what I call the “normative” φορά; and the two forms of linear motion, one towards and one away from the center, correspond to ἐγκράτεια and ἀκρασία, which are naturally represented as moving in opposite directions, one towards what νοῦς commands and the other away from it. 13 The second conjunct (“and in other animals…”) gives the reason for counting φαντασία as a sort of νοῦς [ὡς νόησίν τινα]. But we need not read this as claiming that φαντασία is strictly speaking kind of νοῦς. This may be a tis alienans. On the minimalist reading that I prefer, the point is simply that φαντασία plays in the behavior of non-rational animals the role that νοῦς is supposed (teleologically speaking) to play in the behavior of rational animals. But the first conjunct (“many people follow their φαντασίαι against ἐπιστήμη”) makes it clear that there are cases in which φαντασία ends up playing in rational animals the role that νοῦς should, teleologically speaking, play – namely, cases of ἀκρασία. Aristotle says, of course, that the operations of νοῦς will themselves involve φαντάσματα [431a16-17]. So what is envisioned here is presumably a case where non-noetic or perceptual φαντάσματα dominate, perhaps occluding or overpowering the noetic φαντάσματα that should, teleologically speaking, determine the agent’s behavior. This is precisely the sort of conflict with which III 11 is concerned – namely, conflict between perceptual and deliberative φαντασία.

THE MOVER(S) OF RATIONAL ANIMALS

311

But it would be premature to take Aristotle’s conclusion to be disjunctive. For that would not sit easily with the remainder of III 10, where Aristotle makes it clear that νοῦς (construed broadly so as to include φαντασία) can no more move the ἐγκρατής in the absence of ὄρεξις than ὄρεξις can move the ἀκρατής in the absence of νοῦς (again construed broadly so as to include φαντασία). Aristotle clearly thinks that ὄρεξις and νοῦς must at least co-operate. But there are signs that he adopts a stronger view, at least in cases where the behavior of rational animals goes as it is supposed (teleologically speaking) to go. In these cases he seems to think that νοῦς and ὄρεξις somehow function as one. I have argued elsewhere that Aristotle treats τὸ ὀρεκτικόν and τὸ φανταστικόν (as well as τὸ αἰσθητικόν) as labels for one and the same “part” of soul, and that he uses different labels in different explanatory contexts14. When this part is functioning in a cognitive capacity, he tends to use τὸ αἰσθητικόν or τὸ φανταστικόν; when it is functioning in its locomotive capacity, he tends to use τὸ ὀρεκτικόν15. I read Aristotle’s argument for the unity of these so-called “parts” as parallel to his argument for the unity of the “common sense”. If sight and taste, for example, did not function as one in perceiving honey, it would be as if one person saw the gold color and another tasted the sweet flavor, and there would be no subject that perceived a single object as both gold and sweet. Similarly, if my ὄρεξις and my αἴσθησις (or my ὄρεξις and my φαντασία) did not function as one, it would be as if one person desired honey and another saw (or remembered) some honey: no action to secure any honey would be taken. Aristotle goes on in III 10 to make a parallel point about the need for ὄρεξις and practical νοῦς to function as one. [E] Ἄμφω ἄρα ταῦτα κινητικὰ κατὰ τόπον, νοῦς καὶ ὄρεξις, νοῦς δὲ ὁ ἕνεκά του λογιζόμενος καὶ ὁ πρακτικός· διαφέρει δὲ τοῦ θεωρητικοῦ τῷ τέλει. καὶ ἡ ὄρεξις ἕνεκά του πᾶσα· οὗ γὰρ ἡ ὄρεξις, αὕτη ἀρχὴ τοῦ πρακτικοῦ νοῦ, τὸ δ’ ἔσχατον ἀρχὴ τῆς πράξεως. Both of these then – νοῦς and ὄρεξις – are capable of moving with respect to place. I mean the νοῦς which is for the sake of calculating, i.e., practical νοῦς. For it differs from theoretical νοῦς by its end. 14 See J. Whiting (2002) for more detailed arguments for the views summarized in the next few paragraphs. 15 I suspect that he avoids τὸ κινητικόν because he wants to avoid giving the impression that the “part” in question is itself a subject of locomotion.

312

JENNIFER WHITING

And all ὄρεξις is for the sake of something. For that of which there is ὄρεξις, this is the starting point [ἀρχή] of practical νοῦς [viz., of practical reasoning]. And the last thing [scil., the conclusion of practical reasoning] is the starting point [ἀρχή] of action. Ὥστε εὐλόγως δύο ταῦτα φαίνεται τὰ κινοῦντα, ὄρεξις καὶ διάνοια πρακτική· τὸ ὀρεκτὸν γὰρ κινεῖ, καὶ διὰ τοῦτο ἡ διάνοια κινεῖ, ὅτι ἀρχὴ αὐτῆς ἐστι τὸ ὀρεκτόν. καὶ ἡ φαντασία δὲ ὅταν κινῇ, οὐ κινεῖ ἄνευ ὀρέξεως. ἓν δή τι τὸ κινοῦν, τὸ ὀρεκτικόν. εἰ γὰρ δύο, νοῦς καὶ ὄρεξις, ἐκίνουν, κατὰ κοινὸν ἄν τι ἐκίνουν εἶδος· So it is reasonable that these two, ὄρεξις and practical thought, appear to be the movers . For the object of desire moves and on account of this thought moves , because the starting point of this [scil., thought] is the object of desire. And φαντασία, whenever it moves , does not move it without ὄρεξις. The mover is indeed some one thing, τὸ ὀρεκτικόν, since if two things – νοῦς and ὄρεξις – moved , they would move according to some common εἶδος16. (433a13-22)

Aristotle’s point here is that, just as in non-rational animals αἴσθησις and ὄρεξις work together as one, so too in rational animals νοῦς and ὄρεξις work together as one – at least when things go as they are supposed, teleologically speaking, to go. Akratic and enkratic behavior are of course all too common. But the unified operation of νοῦς and ὄρεξις is nevertheless the norm for rational animals17. We can see what might be involved in νοῦς and ὄρεξις working together as one if we consider how, in non-rational animals, αἴσθησις and ὄρεξις work together as one. The idea is explained back in III 7, where Aristotle argues that the αἰσθητικόν and the ὀρεκτικόν are in some sense the same part or capacity of soul, even if they “differ in being” in the sense that the account of what it is to perceive is different from the account of what it is to desire. Aristotle’s argument for the identity of these parts or capacities appeals to the fact, as he sees it, that whenever an object is perceived as pleasant or painful, the soul of the perceiver, “as if asserting or denying pursues or flees” that object. The idea, I think, is that perceiving an object as pleasant involves desiring to pursue that object. In a non-rational animal, then, the perceptual 16 As I explain in “Locomotive Soul”, the argument here is similar in form to the argument at the end of De Anima I 5 for the unity of the entire soul. 17 It is what produces the third of the three kinds of φορά mentioned in [4], the normative kind (as distinct from the akratic and enkratic kinds).

THE MOVER(S) OF RATIONAL ANIMALS

313

appearance [τὸ αἴσθημα] of an object of as pleasant is inseparable from a desire to pursue that object, and whatever object appears perceptually most pleasant to the animal will be the object of the animal’s strongest – and perhaps even its only – desire. Matters are more complicated in the case of rational animals for whom at least some of the φαντάσματα employed in thought are no less separable from desires than the αἰστθήματα just mentioned are. For immediately upon concluding that the αἰσθητικόν and the ὀρεκτικόν are the same, Aristotle adds “to the thinking soul, φαντάσματα function like αἰσθήματα, and whenever says or denies good or bad, it pursues or avoids ”. The idea here seems to be that whenever the soul thinks of some x as good or bad, which it cannot do without having a φάντασμα of x as good or bad, the soul is disposed either to pursue x or to flee it. This suggests that having a φάντασμα of x as good can no more be separated from having a desire for x than having an αἴσθημα of y as pleasant can be separated from having a desire for y. And it is this inseparability that gives rise to the sort of conflict among ὀρέξεις that opens the door to the possibility of ἀκρασία and ἐγκράτεια. These phenomena arise because the φαντάσματα that come about in perception may differ in content from the φαντάσματα that come about in thought and each sort of φάντασμα is inseparable from its own ὀρέξεις. Perceptual φαντασία may present x as pleasant and thus to-be-pursued, while deliberative φαντασία may present x as unhealthy and thus to-beavoided. So a creature with both forms of φαντασία may end up being subject to conflicting ὀρέξεις18. For example, perceptual φαντασία may dispose a subject to go for some object that deliberative φαντασία disposes her to reject; or perceptual φαντασία may dispose her to flee some object that deliberative φαντασία disposes her to pursue. Aristotle recognizes the seeds of such conflict in his official account of φαντασία back in III 3, where the relation between φαντασία and δόξα is in question. As he says there, an object of which one has true ὑπόληψις can appear falsely to some subject in the sense that it can appear to that subject to be in some way other than it in fact is. For example, the sun may appear to a subject to be a foot across even though 18 Deliberative φαντασία may on its own generate conflicting ὀρέξεις, but this is not the sort of case with which III 11 is concerned. So I leave it aside.

314

JENNIFER WHITING

that subject is persuaded – i.e., believes – that it is larger than the inhabited world [428b2-4]. Similarly, one good (e.g., a pleasure) may appear to a subject to be greater than another even though that subject believes – presumably as result of reasoning – that it is in fact smaller. In sum, the valence of an object’s presentation in perceptual φαντασία may differ from the valence of its presentation in deliberative φαντασία; and if each presentation involves its own ὄρεξις the agent will be subject to conflicting ὀρέξεις. III 11 is concerned with precisely this phenomenon. So let us return to it. 2. Getting [2] and [3] right We left off approaching [3], with some pieces of [2] yet to be determined. One problem with [3] is that it is not clear what the subject of τοῦ δόξαν μὴ δοκεῖν ἔχειν is supposed to be. What is it that seems not to have or entail δόξα? There is also a question about what noun to understand with τὴν ἐκ συλλογισμοῦ in [b]. Is the point, as seems grammatically smoothest, about the kind of δόξα that comes from reasoning? Or is the point, as many translations have it, about the kind of φαντασία that comes from reasoning? Recent translations in English, French, and German take the subject of τοῦ δόξαν μὴ δοκεῖν ἔχειν to be τὰ ἄλλα ζῷα (from the end of [0], picked up again in [2][a]). This makes it difficult to understand δόξαν, as seems most natural, with τὴν ἐκ συλλογισμοῦ in [b]. For the claim would then be that the other animals seem not to have δόξα because they do not have the kind of δόξα that comes from συλλογισμός. This is a somewhat odd claim. Is lack of this kind of δόξα really what lies behind the common belief that the other animals lack any sort of δόξα? Mightn’t they have, as Aristotle thinks rational animals have, a form of δόξα that comes simply from αἴσθησις? It is also unclear what this claim would contribute to the present argument. So translators who take τὰ ἄλλα ζῷα to be the initial subject tend to read Aristotle as saying that the other animals do not have the kind of φαντασία that comes from reasoning19. These translators face 19 K. Corcilius (2017); C. Shields (2016); E. Barbotin (1966); H. Seidl (nach W. Theiler) (1995).

THE MOVER(S) OF RATIONAL ANIMALS

315

two problems: first, to determine the referent of ἐκείνην [“that”] in [3] [c]; and second, to show what role the resulting claim plays in the overall argument. Shields, in his new Clarendon translation, follows Bywater (1888) and solves both problems in one fell swoop: he excises the universally attested αὕτη δὲ ἐκείνην. Corcilius, in his new German translation, is more circumspect: he takes the point to be that the other animals seem to lack δόξα because they lack deliberative φαντασία whereas deliberative φαντασία entails perceptual φαντασία20. But it is not clear what role the claim that deliberative φαντασία entails perceptual φαντασία would play in this context. For it is hard to see how this claim is supposed to be connected either with what precedes it (the claim that the other animals seem to lack δόξα because they lack the kind of φαντασία that comes from reasoning) or with the διὸ-claim that follows immediately upon it (“whence ὄρεξις does not entail the deliberative capacity”). But if, as I suggest, Aristotle is contrasting deliberative with perceptual φαντασία and saying that perceptual φαντασία does not entail the sort of δόξα that comes from reasoning, then we can read the claim that perceptual φαντασία does not entail the sort of δόξα that comes from reasoning as explaining the διὸ-claim. We need only assume what Aristotle takes to be true of perceptual φαντασία in non-rational animals – namely, that it can sometimes, by itself and without any help from δόξα, produce ὄρεξις. From this it follows that ὄρεξις does not entail the presence of the deliberative capacity. For this reason, and others to be spelled out below, I take the initial subject of [3] to be ἡ αἰσθητικὴ φαντασία and I understand δόξα, as seems most natural, with τὴν ἐκ συλλογισμοῦ in [3][b]. We can easily read [2] as setting up this reading of [3] if we read the remarks in [2][c] and [d] as parenthetical remarks about deliberative φαντασία21. We can thus read Aristotle as introducing the μὲν / δὲ pair 20 “Dies ist auch die Ursache dafür, dass (die andere Lebewesen) keine Meinung zu haben scheinen, weil sie (nämlich) nicht die aus einer vergleichenden Überlegung hervorgegangene (Vorstellung) haben, diese aber jene (wahrnehmungsmässige Vorstellung hat).” 21 Hicks parenthesizes in this way, but his motivation for doing so is misguided. He says that he puts the parentheses here in order to make it clear what he takes the antecedent of τοῦτο at the start of [3] to be – namely, the fact that the other animals have only perceptual φαντασία. But the τοῦτο clearly looks forward to the ὅτι clause in [3][b]. So there is no need for parentheses to do the job that Hicks assigns to his. Still, someone who reads [2][c]-[d] as I propose – i.e., as making parenthetical remarks about deliberative φαντασία – may for that reason adopt the parentheses proposed by Hicks.

316

JENNIFER WHITING

in [2][a] and [b], and then making some parenthetical remarks about the δὲ item before turning in [3] to compare the μὲν item to the δὲ item (and comparing them in a way that will help him to resolve the aporia with which he is concerned). And if this is what he is doing, then we can reasonably take the αὕτη [‘this’] in [3][c] as referring to the proximate, δὲ-item, whose operations were described in the parenthetical remarks at the end of [2]. In other words, we can read the sequence as follows: [2*] [a] Perceptual φαντασία [ἡ μὲν οὖν αἰσθητικὴ φαντασία], as previously said, belongs even among the other animals, [b] whereas deliberative [ἡ δὲ βουλευτικὴ] belongs among rational . [c] (For whether one will do this or that is already the work of λογισμός. [d] And it is necessary to measure by one , for it [scil., deliberative φαντασία] pursues the greater ; so it is able to make one 22 out of several φαντάσματα.) [3] [a] And this is the reason why it [scil., perceptual φαντασία] seems not to involve δόξα: [b] because it does not involve the kind of δόξα that comes from reasoning, [c] but this involves that . [d] (Whence ὄρεξις does not involve the deliberative capacity.)

It might seem that I should, given my emphasis on the μὲν / δὲ structure, take ἡ αἰσθητικὴ φαντασία as the referent of ἐκείνην [“that”] in [3][c]. But I do not think that required. It would be natural for Aristotle to express the thought that the μὲν-item seems not to entail some x that the δὲ-item entails by saying that the μὲν-item seems not to entail x whereas the δὲ-item entails that. And reading [3][c] this way – i.e., as saying that deliberative φαντασία (unlike perceptual φαντασία) entails the sort of δόξα that comes from reasoning – serves not only to license the διὸ-claim in [3][d]; it also allows us to make better sense of [4] than commentators have hitherto made of it. Or so I shall argue. But first let me propose a minor emendation of [3][d] that is not strictly necessary to my overall reading but attracts me for two reasons: it resolves 22 I find it difficult to suppress the thought that taking δόξα to be suggested by the context would render its introduction in the next line less abrupt. But the default here is φάντασμα.

THE MOVER(S) OF RATIONAL ANIMALS

317

a problem that many commentators have had and it makes the argument (at least as I read it) much smoother. Commentators have been puzzled by the reference in [3][d] to the deliberative capacity [τὸ βουλευτικόν]. For not only is the substantive τὸ βουλευτικόν comparatively rare in Aristotle’s corpus: it is also unclear why Aristotle should speak of the capacity here. But there would be no puzzle if Aristotle had written τὴν βουλευτικήν instead of τὸ βουλευτικόν. Indeed τὴν βουλευτικήν, with φαντασίαν understood, is just what we might expect here, given that Aristotle is talking about ὄρεξις and δόξα, not τὸ ὀρεκτικόν and τὸ δοξαστικόν23. I take the point here to be, as [5] suggests, about the roles played by token desires and beliefs in producing the behavior of rational animals. In other words, I read [3][d] as claiming that there can be individual ὀρέξεις even when deliberative φαντασία is not in play. I also read [3][d] as parenthetical, since it contributes nothing to the argument, whose focus is φαντασία, not ὄρεξις24. If this is right, then we should not take ἡ ὄρεξις, as most translators do, as the subject of νικᾷ and κινεῖ in [4][a]. For there are plausible alternatives that allow us to make better sense of the overall argument. One is ἡ βουλευτικὴ φαντασία; the other, closely associated with the first, is ἡ ἐκ συλλογισμοῦ . The main advantage of the latter is that it keeps δόξα explicitly in play, thus explicitly setting up the conclusion in [5]. But the idea is the same either way, since deliberative φαντασία is now firmly associated with the kind of δόξα that comes from reasoning. And my appeal to the μὲν / δὲ structure leads me to favor ἡ βουλευτικὴ φαντασία. Hence: [3*] [a] And this is the reason why perceptual φαντασία seems not to entail δόξα – [b] namely, because it does not entail the δόξα that comes from reasoning. [c] But this [scil. deliberative φαντασία] entails that [scil. the δόξα that comes from reasoning], [d] (whence ὄρεξις does not entail deliberative )25. 23 Aristotle could easily have chosen to talk about these capacities had he wanted to. For the relevant substantives are (unlike τὸ βουλευτικόν) common in his corpus. 24 Aristotle often does this: he will use a διὸ-claim to confirm a point he has just made, without taking the διὸ-claim itself as part of his argument. 25 Please note that my argument for taking ἡ βουλευτικὴ φαντασία as subject in [4][a] does not depend on my proposal to read τὴν βουλευτικήν in place of τὸ βουλευτικόν in

318

JENNIFER WHITING

[4] [a] And sometimes it [scil. deliberative φαντασία26] conquers and moves 27.

There is no conclusive argument for this on purely grammatical grounds, so any decision in its favor must depend in part on the overall sense afforded by this reading as compared with other possible ones. Let me pause then to explain why I find the standard way of reading [4][a] unsatisfactory. Readers who are happy with the proposed reading of [4][a] and want simply to follow the argument as I interpret it to its conclusion may skip section 3. Anyone still tempted after reading section 4 to take ἡ ὄρεξις as subject of νικᾷ and κινεῖ can read section 3 later. 3. Optional digression: problems with the standard reading of [4][a] The main problem with taking ἡ ὄρεξις as subject of νικᾷ and κινεῖ is that doing so makes it difficult to align the elements of [4][a]-[e] with the three φοραί mentioned in [4][f]. To see this, we should note that there are, schematically speaking, three ways to read [4]. The differences between them turn on where in [a]-[e] each of the three φοραί mentioned in [f] makes its appearance. The φοραί are, as I have said, the enkratic, the akratic, and the “normative”. The normative φορά involves some analogue of the unity of ὄρεξις and φαντασία that characterizes the behavior of non-rational animals. I call it “normative” because it is the kind of φορά we find when things go as they are supposed, teleologically speaking, to go. But things do not always go as they are supposed, teleologically speaking, to go: rational animals sometimes engage in both enkratic and akratic behavior. Still, we must keep in mind that these forms of behavior are anomalies in Aristotle’s cosmos: neither beasts nor heavenly bodies are subject to such vagaries. The norm, as we have seen, is the unity of νοῦς (broadly conceived so as to include φαντασία) and ὄρεξις. Let us return then to [4], keeping this norm in mind. [3][d]. Because I am treating [3][d] as parenthetical, I am taking the subject of [3][c] as the default subject in [4][a]. 26 Alternatively: . 27 Someone might object to the fact that my reading takes the first occurrence of νικᾷ as intransitive while taking the remaining occurrences as transitive. But I do not think it out of the question that Aristotle would do this. One might, as an alternative, take the initial occurrence as transitive with something like πάντα as its implicit object.

THE MOVER(S) OF RATIONAL ANIMALS

319

It is clear that two of the three forms of φορά – the enkratic and akratic – are introduced in [a]-[e]. But it is an open question where (if at all) the normative form is introduced: it is possible that it is not explicitly mentioned here but simply in the background provided by III 10, where the relevant form of unity is discussed at length. This possibility yields the first of the three schematic ways to read [4], which involves seeing in [4][a]-[e] only the akratic and enkratic φοραί. Let’s call this the “anomalous cases” schema. The other ways of reading [4] are, by contrast, “exhaustive”: they include the normative form along with the two anomalous forms. One locates each of the three φοραί in [a]-[d] and then takes [e] to make either a meta-point or parenthetical remarks about what happens φύσει: let’s call this the “independent [e]” schema. The other locates the normative case in [e] and the anomalous cases in [a]-[d]: let’s call this the “normative [e]” schema. The anomalous / exhaustive choice-point has not been explicitly articulated in any place I know. I assume that most commentators unreflectively adopt some version of the exhaustive reading. This, I suspect, is why so many end up proposing to alter the received text: they have difficulty lining [4][a]-[e] up with each of the three φοραί mentioned in [f], so they feel the need to add something. Trendelenburg conjectures a δέ at the start of [d] – yielding ἡ δ’ ὄρεξις τὴν ὄρεξιν – because this would allow us to locate the third form of φορά in [d], which is explicitly concerned with the akratic φορά. How exactly the rest should be sorted will depend on whether or not we follow the manuscript tradition in which τὴν βούλησιν follows νικᾷ δ’ ἐνίοτε καὶ κινεῖ in [4][a]. If we read τὴν βούλησιν, [a] would introduce the akratic φορά and so overlap with [d], in which case we would need either to locate the normative case in [e] or to read [4] as covering only the anomalous cases. But either way, the case for adding δέ is undermined. If we locate the normative case in [e], the relevant δέ awaits us there; if we read [4] as covering only the anomalous cases, then there is no need for another δέ. Either way, we should leave the text alone. It should be noted that τὴν βούλησιν is independently awkward – and not just because of the word order. There is no parallel for κινεῖν τὴν βούλησιν elsewhere in Aristotle’s corpus. And perhaps more importantly, the object of κινεῖ in III 9-11 is generally (though sometimes only implicitly) τὸ ζῷον, as in [5] (towards which [4] is working). So it

320

JENNIFER WHITING

is tempting to follow the manuscripts in which τὴν βούλησιν does not appear and to take τὸ ζῷον to be the implicit object of κινεῖ. This is the reading I prefer. But some commentators think Aristotle must have written τὴν βούλησιν in [a] because they think it required in order to make sense of ὁτὲ δ’ ἐκείνη ταύτην in [b]. Their reading runs as follows: [4] [aO/B] sometimes ὄρεξις conquers βούλησις [bO/B] but sometimes this conquers that .

On this reading, [a] introduces the akratic φορά while [b] introduces the enkratic one. But since [d] is explicitly concerned with the akratic φορά, [a] and [d] once again overlap, and Aristotle seems to be jumping around. Hicks (ad loc.) cites Simplicius 310, 28 in support of reading [d] as using ἀκρασία in a broad sense that includes ἐγκράτεια as well as ἀκρασία, and so as providing a résumé of the enkratic and akratic cases. This would remove the appearance of jumping about. But I think we can do better. I do not agree with Hicks that τὴν βούλησιν is required in order to make sense of ὁτὲ δ’ ἐκείνη ταύτην in [b]. But I can easily imagine an early editor or scribe adding it because he (like Hicks) thought it (or something like it) required in order to make sense of [b]. For other changes have been proposed by commentators struggling to make sense of [b], some indeed quite radical. Cornford proposed to read ὁτὲ δὲ κινεῖ γ᾽ αὑτήν in place of [b]’s ὁτὲ δ’ ἐκείνη ταύτην. Αnd Bywater, citing αὕτη δὲ ἐκείνην back in [3], proposed to add ὁτὲ μὲν αὕτη ἐκείνην before ὁτὲ δ’ ἐκείνη ταύτην. I suspect that there lies behind Bywater’s radical proposal a knee-jerk tendency to assume that any ἐκείνη ταύτην requires a reciprocal αὕτη ἐκείνην28. But I think we can make good sense of [4] as it stands if we give up the idea that ἡ ὄρεξις from [3][d] is the subject of νικᾷ δ’ ἐνίοτε καὶ κινεῖ in [4][a]. I cannot discuss in further detail what I find unsatisfactory with the various ways I have tried to make sense of III 11 while understanding 28 I suspect that assumption also drives Seidl’s queer translation of 433a12-13. Though he renders ἐκείνη appropriately as “jener”, he interprets it as referring to the immediately preceding τὴν βούλησιν, for which Aristotle would surely have used αὕτη (as in Bywater’s proposal to add ὁτὲ μὲν αὕτη ἐκείνην before ὁτὲ δ’ ἐκείνη ταύτην).

THE MOVER(S) OF RATIONAL ANIMALS

321

ἡ ὄρεξις as the subject of [4][a]. So let me simply spell out what seems to me the best way to go if we take ἡ ὄρεξις as subject, and then explain how I think we can do better. If we go with the manuscripts in which τὴν βούλησιν does not appear, then we can achieve a moderately satisfactory result provided we are willing read ἐκείνη in [b] as referring back (as ἐκείνη often does) to something in the preceding bit of text. My preferred candidate is ἡ φαντασία βουλευτική because I think that most natural given the μὲν / δὲ structure of the passage. But I am willing to allow that ἐκείνη might refer back to ἡ ἐκ συλλογισμοῦ , which is associated with deliberative φαντασία. Either way, the idea is the same: sometimes deliberative φαντασία – or the δόξα that results from it – conquers an opposed ὄρεξις. Here, then, is the best sense I can make of the passage with ἡ ὄρεξις as subject: [4OR] [a] Sometimes conquers and moves [b] but sometimes that conquers this

[c] just as a sphere [d] and ὄρεξις ὄρεξις whenever ἀκρασία occurs. [e] But by nature, the higher is always more of an ἀρχή and moves . [f] So it [scil., the rational animal] is moved with three different φοραί.

This allows us to read [a] as introducing the normative case, [b] as introducing the enkratic case, and [d] (with Trendelenburg’s conjectured δέ) as introducing the akratic case. We can then read [e] as making a meta-point or simply a parenthetical remark about what happens φύσει. In this case, we should probably read the enigmatic reference to the operations of some sort of sphere as applying only to the enkratic φορά, and we should probably read Aristotle as referring to a heavenly sphere rather than a ball in some game. If this is right, then Aristotle is comparing the operations of deliberative φαντασία on ὄρεξις to the ways in which he takes higher heavenly spheres to control the movements of lower ones30. But that raises the question why, given [4][e], the control Alternatively: . On the ancient controversy about how to interpret ὥσπερ σφαῖρα, see Hicks ad 434a13. Commentators tend to take σφαῖρα to refer either (as Themistius took it) to a 29 30

322

JENNIFER WHITING

of the higher spheres over the lower ones is not invoked in the normative case. And while we might answer that there are in the normative case no opposed movements to be controlled from above, I think we can do better – and without having to insert so much as a single δέ. 4. Getting [4] right The key is to read [4] as turning – after [3*]’s comments about perceptual φαντασία – to deliberative φαντασία, which is introduced in [3][c]. Hence: [4*] [a] And sometimes it [scil., deliberative φαντασία] conquers and moves 31 [b] but sometimes that [scil., perceptual φαντασία] conquers this [scil., deliberative φαντασία] [c] just as a sphere [d] desire desire whenever ἀκρασία occurs. [e] But by nature, the higher is always more of an ἀρχή and moves . [f] So it [scil., the rational animal] is moved with three different forms of movement [φοραί].

One immediate advantage of [4*] is that it treats [4][b]-[d] as a coherent unit focused on ἀκρασία, which allows us to make decent sense of the remark about the sphere along the lines suggested by Simplicius: the idea is that in all cases of ἀκρασία the strongest desire defeats the weaker in something like the way in which the larger of two balls in some game – or the ball moving with the greatest force – knocks the other out of the way or at least off its course. The idea, I think, is roughly as follows. The subject reasons her way to some belief – for example, that she should not consume alcohol when she is caring for children – and thus comes to have a rational desire (i.e., a βούλησις or προαίρεσις) never to drink when children are in her care. heavenly sphere or (as Simplicius took it) to a ball in some game. Hutchinson argues, quite plausibly, for a third view – namely, that Aristotle is adopting something like the account of orderly and disorderly motions found in Plato’s Timaeus 43-44. It is difficult to be sure what Aristotle has in mind, but I am presently inclined to side with Simplicius both because Aristotle seems to restrict the comparison to the case of ἀκρασία and because he does not mention the sphere in connection with the normative case. But this comparison deserves more attention than I give it here. 31 See note 27.

THE MOVER(S) OF RATIONAL ANIMALS

323

We must suppose here that the subject does not simply come to have an idle belief, without any connection to desire. The akratic case is, by hypothesis, one in which the subject’s coming to have a belief involves or is at least associated with her coming to have a corresponding βούλησις. She may frequently – perhaps even for the most part – abstain from alcohol when she is caring for children. Yet on some occasions when children are in her care, the opportunity to have a drink presents itself and appears so attractive that she simply helps herself without attending to (let alone changing) her belief that she should not be doing so. In this case, we should imagine that having the drink appears to the subject so pleasant that it dominates her practical outlook in something like the way in which certain objects can dominate her visual field, not necessarily by occluding other objects but simply by standing out for some reason – by being, for example, so bright in color that she does not notice other, less brightly colored objects in her visual field. In such cases, she may see the surrounding objects without really attending to them or recognizing the sort of import they normally have for her. So she simply acts on what is, at present, her strongest desire, which operates here like a larger ball (or one moving with greater force), mechanically knocking a smaller ball (or one moving with lesser force) out of its way. But why should Aristotle restrict this simile to the case of ἀκρασία? Why would he not treat ἐγκράτεια in the same way? The answer, I think, is that ἐγκράτεια involves operations of reason in a way that ἀκρασία does not. And the relevant operations of reason involve activities of deliberative φαντασία. For, as Aristotle says at EN 1150b21 sq., there are some people who, because they foresee [προαισθόμενοι καὶ προϊδόντες] what is to come and prepare themselves in accordance with their reasoning [προεγείραντες ἑαυτοὺς καὶ τὸν λογισμὸν], are not overcome by their passion (whether it be pleasant or painful). The idea, I think, is that ἐγκράτεια involves some form of deliberate and conscious resistance in the face of whatever desires are opposed to the agent’s βούλησις. We might thus contrast enkratic behavior with akratic behavior by saying that whereas reason goes on holiday in ἀκρασία, reason works overtime in ἐγκράτεια. If this is right, then enkratic behavior does not necessarily involve the stronger desire overpowering the weaker: βούλησις may be weaker than ἐπιθυμία in a phenomenological sense but able to resist the force of ἐπιθυμία because βούλησις is reinforced by the operations of deliberative

324

JENNIFER WHITING

φαντασία. There is nothing outré here: visualizing the future benefits of resisting present temptation is a common strategy for resisting strong desires that tempt us to act against our βουλήσεις. But for this, the image of one moving ball knocking another out of the way is not quite right. We need something other than a moving ball, something capable of keeping a forcefully moving ball from reaching its destination: something like a goalie’s mitt moving in anticipation of what is to come. What is crucial here is the element of anticipation. Ἐγκράτεια is not simply a matter of one desire, a βούλησις, overpowering another in a mechanistic way, as one moving ball might overpower another. If βούλησις is to win out over opposed desires, βούλησις must rely on cognitive powers such as those involved in visualization: the subject must both imagine what is likely to come about and picture appropriate ways of responding to it, perhaps in conjunction with the probable rewards of responding in the appropriate ways. This, I submit, is why it is only ἀκρασία that [4*] assimilates to one ball conquering another. So I propose to read [4*], schematically speaking, as follows. [a] is about ἐγκράτεια, where deliberative φαντασία is victorious over some wayward ὄρεξις (which cannot, of course, be βούλησις). [b] then introduces ἀκρασία by pointing out that the previously discussed form of φαντασία [ἐκείνη = perceptual φαντασία] sometimes conquers this [ταύτην = deliberative φαντασία]; in this case, [b] explains, we find one ὄρεξις conquering another. This leaves [e] to cover the normative case. In other words, I propose to go with a “normative [e]” schema. I read φύσει δὲ ἀεὶ in [e] as introducing a different kind of principle, one that does not involve knocking things about in something like the way one ball knocks another about. When this sort of principle is effective there is no opposition of the sort involved in ἐγκράτεια, so there is no need to rely on the sort of reinforcement deliberative φαντασία is capable of providing. This sort of principle governs the movements of the rational animal in a way comparable to that in which the unified principle responsible for the behavior of non-rational animals governs their movements. There is no conflict of principles requiring one to overpower another. That is why Aristotle calls this sort of principle ἀρχικωτέρα. It does not simply overpower other movers like itself, as one ball overpowers another in a game or one desire overpowers another in cases of ἀκρασία. Rather, this sort of principle governs the behavior of a rational animal in

THE MOVER(S) OF RATIONAL ANIMALS

325

something like the way in which a wise ruler governs the behavior of willing subjects (i.e., subjects whose will is shaped by their appreciation of the contents of the ruler’s wisdom as distinct from merely trusting in the wisdom of whatever the ruler prescribes). In this case, the animal is moved by practical νοῦς, though not, of course, in the absence of ὄρεξις. What we have here is what Aristotle elsewhere calls προαίρεσις and characterizes as either ὀρεκτικὸς νοῦς or ὄρεξις διανοητική32. These are not genuine alternatives, but rather two different ways of describing the same state of soul, a state that can be characterized equally well as a kind of thought and as a kind of desire: neither characterization is privileged. The idea here is that in the normative case – the case of the φρόνιμος or practically wise agent – the subject’s ὀρεκτικόν is shaped by the principles that inhabit her practical νοῦς. One might even say that the subject’s ὀρεκτικόν is en-formed by the principles that constitute its practical νοῦς. This is the hylomorphic conception of φρόνησις that I, following John McDowell, elsewhere attribute to Aristotle. Φρόνησις stands to the properly formed ὀρεκτικόν as Aristotle takes the soul to stand to the organic body; and just as the capacities that constitute the soul (e.g., capacities to reproduce, perceive, and locomote) cannot exist apart from a body with properly formed organs, so too φρόνησις cannot exist apart from properly formed desires. It is, as John McDowell puts it, “the properly moulded state of the motivational propensities in reflectively adjusted form”33. 5. Conclusion This brings us to [5], which I read as completing the discussion of the normative case that is introduced in [4][e]. On this account, the aporia articulated at the start of III 10 is resolved in [4][a]-[d], where Aristotle indicates (very roughly) how the conjunction of perceptual and deliberative φαντασία gives rise to the possibilities of enkratic and akratic behavior, and then turns in [4][e] and [5] to what he sees as the norm in 32 EN 1139b4-5: διὸ ἢ ὀρεκτικὸς νοῦς ἡ προαίρεσις ἢ ὄρεξις διανοητική, καὶ ἡ τοιαύτη ἀρχὴ ἄνθρωπος. 33 I take the idea of hylomorphic φρόνησις, or hylomorphic virtue, from John McDowell (1998). See J. Whiting (2019).

326

JENNIFER WHITING

the animal world, i.e., the unified operation of νοῦς (broadly construed so as to include φαντασία) and ὄρεξις. For rational animals, the norm involves δόξα, both universal and particular, moving the animal. Universal δόξαι arise from reasoning; particular δόξαι arise sometimes from reasoning, sometimes directly from perception. When there is no conflict between deliberative and perceptual φαντασία, the universal and particular δόξαι work together automatically, in ways that Aristotle describes in chapters 7 and 8 of de Motu Animalium. For example, when the subject thinks the universal (that every man ought to walk) along with the particular (that he himself is a man), straightaway he walks – provided nothing interferes34. The proviso is Aristotle’s signal that he means to be describing the normative case: it tacitly acknowledges the existence of the deviant forms of φορά to which [4][f] refers. The form of the proviso suggests that there is an explanatory asymmetry between the normative φορά and its deviant counterparts. When things go as they are supposed (teleologically speaking) to go, the subject makes the relevant inferences more or less automatically and acts straightaway (again more or less automatically). But things do not always go as they are supposed (teleologically speaking) to go. In some cases, the presence of a strong appetite, for example, an appetite to φ, can result in a φάντασμα of φ-ing as so pleasant that the subject acts straightaway, before her standing belief that she ought not φ in situations of the present sort is activated – that is, before any φάντασμα of φ-ing in that situation as harmful (or in some other way inappropriate) is formed. This is the form of ἀκρασία that Aristotle calls “impetuous” [προπέτης] and opposes to the “weak” [ἀσθενής] form. In other cases, the φάντασμα of φ-ing in that situation as in some way inappropriate may be formed, with the result that there is conflict between the perceptual φάντασμα of φ-ing as pleasant (and so to-be34 Aristotle’s use of δόξα is like our use of “belief”: sometimes it refers to a propositional attitude (e.g., my belief that φ), sometimes it refers to the propositional content of such an attitude (i.e., what it is that I believe – namely φ). The present question is whether (as in the normative case) a propositional attitude with the appropriate content both actually comes about and is effective, or whether (as in the other φοραί) the propositional attitude either fails to come about or comes about but fails to bring about the appropriate action because something else prevents it from operating as it is supposed (teleologically speaking) to do. For more on this issue, and the issues that arise in the next few paragraphs, see sections 6-8 of M. Pickavé and J. Whiting (2008).

THE MOVER(S) OF RATIONAL ANIMALS

327

pursued) and the doxastic φάντασμα of φ-ing as in some way appropriate (and so to-be-avoided). How the conflict is resolved will presumably depend on the strength of the sort of reinforcements afforded by the subject’s powers of deliberative φαντασία. If they are sufficiently strong, her action will count as enkratic; otherwise, her action will presumably exemplify “weak” ἀκρασία35. In a subject with practical ἐπιστήμη – or what Aristotle usually calls φρόνησις – the universal δόξαι are stable and not subject to being dislodged either by perceptual φαντασίαι or by particular δόξαι arrived at by means of corrupt reasoning36. This explains why [5] describes τὸ ἐπιστημονικόν as remaining at rest37. In the φρόνιμος, the universal δόξαι remain constant while the particular δόξαι vary from one situation to another. And when, via either reasoning or perception, the φρόνιμος comes to have some particular δόξα that instantiates (so to speak) one of her universal δόξαι, then straightaway she acts – though not of course without ὄρεξις. For the δόξαι in question involve φαντασίαι that are themselves inseparable from ὀρέξεις38. In this (the normative) case, 35 The passage in which Aristotle distinguishes “impetuous” from “weak” ἀκρασία is the one (mentioned above) where he speaks of subjects who, because they foresee what is coming and prepare for it in accordance with reasoning, are not overcome by their passion: EN 1150b19-29. See also, De Anima III 7, where Aristotle allows that: “Sometimes, by means of the φαντάσματα and νοήματα in the soul, , as if seeing, calculates and deliberates [λογίζεται καὶ βουλεύεται] in relation to the things that are present what things are going to come about. And whenever it says that there will be the pleasant or painful, here it flees or pursues – and generally in action.” [431b6-10] But this is only sometimes. No such calculation or deliberation occurs in cases of “weak” ἀκρασία. 36 Aristotle’s use of τὸ δ’ ἐπιστημονικόν in this context is admittedly odd. See EN 1142a23 sq.: ὅτι δ᾽ἡ φρόνησις οὐκ ἐπιστήμη, φανερόν . . . But it is not without parallel: see the translation of EN 1145b22-31 in section 1 of M. Pickavé and J. Whiting (2008), along with the discussion in section 8 of that paper. We might also deal with the oddity by taking Aristotle to be thinking, along the lines suggested in note, that the ἕξις (or condition) of someone who grasps the reasons for universals of the form “it is necessary for such to do such ” is comparable to that of someone who grasps the principles involved in the sort of demonstrative reasoning characteristic of theoretical sciences, both natural and mathematical. In other words, he may be thinking of ἐπιστήμη in a generic way. 37 I suspect that he is also alluding to the idea, prominent throughout de Motu (and elsewhere in his work), that movement generally involves something that remains at rest and something that itself moves. But I do have not the space to pursue this point here. 38 We might wonder here whether Aristotle thinks of vicious subjects as analogous to virtuous ones in having universal δόξαι that remain constant while their particular do vary from one situation to another. Aristotle may allow that at least some vicious agents

328

JENNIFER WHITING

νοῦς and ὄρεξις work together as they do in other animals: they work together as one39. The other φοραί – the enkratic and akratic – are the peculiar province of so-called rational animals.

achieve such stability, but his general view seems to be that vicious agents lack the sort of stability that he takes to characterize virtuous agents. Even so, it is hard to be sure what he would say here, since his discussions of this issue – see especially Nicomachean Ethics IX 4 – do not clearly distinguish universal from particular δόξαι. We might also wonder why Aristotle does not speak in [4][f] of four φοραί. For I think it quite probable that he would, if he were explicitly considering the question, treat vicious behavior as involving a fourth kind of φορά. But I cannot pursue this question here. 39 I was honored to be included in the Lille conference celebrating the career of Michel Crubellier, with whom I once had the pleasure of co-teaching in Berlin and from whom I have continued to learn since that time. I am thus pleased to contribute this essay in honor of him. I would like to thank Christian Wildberg and members of the Lille audience, especially David Charles and Robert Howton, for helpful discussion of these issues. Charles Brittian, as usual, deserves mega-thanks.

WHY DE ANIMA NEEDS III 12-13 Robert HOWTON

1. On the whole, Aristotle’s De Anima (DA) is a cohesive work. Its aim, announced in its initial chapter, is “to consider and ascertain [the soul’s] nature and essence, then the attributes belonging to it” (402a7-8). Throughout most of DA, Aristotle keeps this explanatory objective in view, justifying each step in his exposition by reference to its contribution to an account of soul’s nature and essence1. To this general rule, however, DA’s final chapters are a striking exception. The opening lines of DA III 12 mark an abrupt shift from the topic of the immediately preceding chapters, yet Aristotle makes no attempt to connect the issues he explores in that chapter and its sequel to DA’s broader account of soul’s nature and essence. It is therefore a pressing question for commentators how, or even whether, DA III 12-13 fit with the investigation of soul contained in the foregoing chapters of DA. Are these chapters the “impossible non-sequitur” they appear to be2, or are they instead an integral step in Aristotle’s attempt to ascertain soul’s nature and essence? What, if anything, do they contribute to the account of soul Aristotle develops in DA? I intend to show that the final chapters of DA are not only integral to that work, but more central to Aristotle’s account of soul than even the most optimistic commentators have acknowledged. On the interpretation 1 Aristotle’s attention to this expository detail is clearest in DA II 3-4, where he shifts focus to psychic capacities on the grounds that an account of them will also be the “most appropriate” (οἰκειότατος) account of soul (415a12-13); but see also the references to the explanatory objective e.g. in I 2, 403b20-24, II 1, 412a3-6, and II 2, 413a20-22. 2 See D.S. Hutchinson (1987), p. 375.

330

ROBERT HOWTON

I defend, DA III 12-13 are written in support of a claim on which Aristotle has premised the bulk of his investigation of soul and its capacities: that living things do whatever they do by nature for the sake of a single end, in most cases reproduction. This claim, I argue, reflects Aristotle’s view that soul is a “complex activity” of an organic body, an actuality comprising a diverse set of vital activities that are nevertheless each modes of pursuit of a single, teleologically primary activity. Aristotle expresses commitment to this view of soul in his zoological works, but its role in DA has not fully been appreciated by commentators. By recognizing its crucial role in DA, I argue that we come also to see the essential contribution of DA III 12-13 to Aristotle’s explanatory objective in that work. I begin in section 2 with an overview of DA III 12-13 and its interpretive difficulties. Section 3 criticizes a recent attempt to connect these chapters to Aristotle’s broader discussion of soul in DA. Section 4 lays the foundation for a more satisfactory interpretation by outlining Aristotle’s “complex activity model” of soul, which I trace to an argument for the priority of nutrition in DA II 4. Finally, in section 5, I present a revised interpretation of DA III 12-13 as an effort to show how the vital activities naturally available to living things contribute to the achievement of the latter’s respective teleologically primary activities.

2. The lines spanning DA III 12-13, 434a22-435b25, do not divide neatly at the conventional chapter break at 435a10. We capture more naturally the progression of thought in these lines if we divide them into roughly five contiguous passages, as follows: A [434a22-b8] Aristotle begins by discussing the necessity whereby the nutritive and perceptive capacities of soul belong to living things. Nutrition belongs necessarily to every living thing that grows, reaches maturity, and declines, since achievement of these life stages would be impossible without the ability to use food. Perception is not likewise necessary for all such living things, but belongs only to some, the animals. Nevertheless, perception belongs necessarily to animals, since “nature does nothing in vain” (a30-31). B [434b8-27] Having noted that the body of a perceptive organism must be elementally composite, Aristotle turns to the necessity of

WHY DE ANIMA NEEDS III, 12-13

331

the individual senses for animals. Contact senses, touch and taste, belong of necessity to all animals, since without these it is not possible to be an animal. Distance senses, sight, hearing, and smell, are not likewise necessary for all animals, but belong necessarily to some: the roaming animals. In framing this contrast Aristotle distinguishes two ways in which psychic capacities might belong to a psychological kind: either for the sake of being or for the sake of well-being. C [434b27-435a10] Discussion of the distance senses occasions remarks on the perceptual medium as a moved mover in episodes of distance perception. D [435a11-b3] Aristotle returns to the composition of the animal body. Since the animal must have touch, its body must be both compound and not earthen. This also explains why we do not perceive with earthy parts of our body like bones and hair, as well as why plants cannot perceive. E [435b4-25] Finally, Aristotle considers again why animals need the contact and distance senses. Animals as such must have both touch and taste, insofar as the objects of taste are tangible. By contrast, animals have the other perceptual attributes they have, not for the sake of being, but for the sake of well-being. The dominant focus of the passages is clear. A, B, and E focus on the necessity of certain psychic capacities (nutrition, perception, the senses) for different psychological kinds, including plants, stationary animals, and roaming animals; while D focuses on the elemental composition of the animal body, an issue first raised in B. Less clear is whether Aristotle’s simultaneous discussion of these themes is motivated by any unifying purpose, and if it is, whether its unifying purpose justifies its inclusion in the DA study of soul. For the discussion contained in these lines also appears haphazard in at least three respects: it seems discontinuous with the immediately preceding chapters; Aristotle’s exposition of the dominant themes seems disorganized and repeatedly loops back on itself; and C is plainly a digression. These deficiencies, coupled with the absence of any explicit mention of a unifying purpose, may invite pessimism about the prospects of interpreting III 12-13 as an integral step in the DA study of soul. Perhaps the lines known to us as III 12-13 were misplaced from their original

332

ROBERT HOWTON

position in DA3, or perhaps they are not part of the main text of DA at all4. But few commentators embrace such pessimistic conclusions. Instead, most attempt to supply III 12-13 with a unifying purpose by specifying how discussion of the chapters’ dominant themes contributes to the account of soul Aristotle has been developing in DA. The two earliest extant commentaries on III 12-13 provide examples of this sort of strategy. According to one, III 12-13 aim to make good on Aristotle’s earlier promise to determine which kinds of body are receptive to which kinds of soul, an objective finally accomplished in D5. According to the other, III 12-13 are there to clarify Aristotle’s remarks on animal locomotion in III 9-11 by indicating (in A, B, and E) how the roaming animals under consideration differ psychologically from plants, stationary animals, and superlunary movers like stars and planets6. A version of the same approach has become standard among recent commentators on III 12-137. According to it, III 12-13 address a question whose answer Aristotle postponed several times in DA II, namely: why are psychic capacities distributed “serially” among psychological kinds8? In II 2 Aristotle observed that some living things, the plants, have the capacity for nutrition but cannot perceive, whereas nothing that perceives lacks the capacity for nutrition. Among perceivers, he noticed that the contact senses are found in all animals, whereas only the roaming animals have distance senses. In II 3 he described this distribution as a hierarchy or nested series, analogous to the series of polygons in the kind figure. Just as the triangle is potentially present in the quadrangle, and the quadrangle in the pentagon, so too are the more widely distributed psychic capacities potentially present in the more narrowly distributed, the nutritive in the perceptive, for instance. Here too Aristotle postponed explanation of the serial relation among psychic capacities, and for many recent commentators, III 12-13 are where the the promised explanation is finally delivered. 3 Perhaps between II 4 and II 5; see D.S. Hutchinson (1987) and, for criticism, M.F. Burnyeat (2002), p. 30, n. 6, and R. Polansky (2007), p. 535, n. 1. 4 See D.W. Hamlyn (1993), p. 156. 5 See Ps.-Simplicius, In DA 315.21-316.12 Hayduck; cf. I 3, 407b15-17. 6 See Ps.-Philoponus, In DA 594.19-27 Hayduck. 7 See Aquinas, In DA § 300, R.D. Hicks (1907), p. 573, M. Leunissen (2010), p. 59, R. Polansky (2007), p. 534, C. Shields (2016), p. 369. 8 See II 2, 413b4-10, 413b32-414a4, and II 3, 414b33-415a1.

WHY DE ANIMA NEEDS III, 12-13

333

According to the standard approach, then, the unifying purpose of III 12-13 is to explain the serial distribution of psychic capacities among psychological kinds. This proposal offers a prima facie attractive interpretation of the connection between III 12-13 and the rest of DA. But I believe it fails to account for a crucial detail of Aristotle’s exposition in those chapters: the distinction belonging for the sake of being and belonging for the sake of well-being. The next section details this difficulty for the standard approach and points the way to a more satisfactory view of III 12-13’s unifying purpose. 3. The standard approach interprets III 12-13, especially A, B, and E, as an attempt to explain why psychic capacities are distributed serially among psychological kinds – why, for instance, every kind of perishable organism has the nutritive capacity, while only animals have the perceptive. The explanation it takes Aristotle to provide is final causal, relying on a teleological principle to which he explicitly appeals in explaining why perception belongs of necessity only to the animals ([A] 434a30-b8): Tὸ δὲ ζῷον ἀναγκαῖον αἴσθησιν ἔχειν, εἰ μηθὲν μάτην ποιεῖ ἡ φύσις. ἕνεκά του γὰρ πάντα ὑπάρχει τὰ φύσει, ἢ συμπτώματα ἔσται τῶν ἕνεκά του. εἰ οὖν πᾶν σῶμα πορευτικόν, μὴ ἔχον αἴσθησιν, φθείροιτο ἂν καὶ εἰς τέλος οὐκ ἂν ἔλθοι, ὅ ἐστι φύσεως ἔργον· πῶς γὰρ θρέψεται; τοῖς μὲν γὰρ μονίμοις ὑπάρχει τοῦτο ὅθεν πεφύκασιν. οὐχ οἷόν τε δὲ σῶμα ἔχειν μὲν ψυχὴν καὶ νοῦν κριτικόν, αἴσθησιν δὲ μὴ ἔχειν, μὴ μόνιμον ὄν, γενητὸν δέ· ἀλλὰ μὴν οὐδὲ ἀγένητον· διὰ τί γὰρ οὐχ ἕξει; ἢ γὰρ τῇ ψυχῇ βέλτιον ἢ τῷ σώματι. νῦν δ’ οὐδέτερον· ἡ μὲν γὰρ οὐ μᾶλλον νοήσει, τὸ δ’ οὐθὲν ἔσται μᾶλλον δι’ ἐκεῖνο· οὐθὲν ἄρα ἔχει ψυχὴν σῶμα μὴ μόνιμον ἄνευ αἰσθήσεως. But the animal must have perception, if nature does nothing in vain. For everything that is by nature belongs for the sake of something, otherwise it will be an attribute of things which are for the sake of something. If, then, every body were mobile but lacked perception, it would die and fail to reach the end that is the function of its nature. (For how will it nourish itself? This [scil. nourishment] is provided to stationary [bodies] from where they are naturally produced.) Nor is it possible to have soul and a discriminating intellect but not perception, if it is not stationary and generated – but not even if it is ungenerated. For why would it not have [perception]? It would have to be better for the soul or for the body, but in fact it is better for neither. For [the soul] will not think better, nor will [the body]

334

ROBERT HOWTON

be any better because of that. Therefore no non-stationary body has a soul that lacks perception.9

The principle that nature does nothing in vain, which here makes its second appearance in DA, is familiar from Aristotle’s zoological works10. In those works the principle is typically used to explain the absence, in a given subkind, of an attribute that belongs by nature to other members of the wider kind, for instance the absence of legs in snakes. Here, by contrast, it is used as a constraint on what can be said to belong by nature to a given kind: if an attribute, such as a psychic capacity, belongs by nature to a kind, it must be present for the sake of some end, either by having that end as a final cause or by belonging on account of some other, goal-directed attribute. Proponents of the standard approach see in this application of the principle a general strategy for explaining the serial distribution of psychic capacities. In general, they claim, the cause of any psychic capacity’s being “potentially present” in another, such that the latter capacity is never found without the former but the former is sometimes found without the latter, is that the psychological kinds with ends that necessitate the presence of the latter capacity also have need of the former, whereas the converse is not universally true. Thus all perishable organisms require nutrition because all need food in order to grow, reach maturity, and decline; whereas only some, the animals, have perception, since only an animal needs to perceive in order to feed itself and attain “the end that is the function of its nature”11. As the discussion proceeds, however, the principle begins to be applied in ways that do not seem to exemplify this explanatory strategy. Early in B Aristotle defends a contrast between the contact and distance senses that parallels the contrast drawn in A between nutrition and perception. The contact senses belong of necessity to all animals, touch because every animal’s survival depends on its ability to avoid some tangible bodies and apprehend others (434b14-18), and taste because it is “the sense of what is tangible and nourishing” (434b19-22); whereas 9

Text following P. Siwek (1965). See DA 432b21-26; cf. Cael. 271a33, 291b12-15, Resp. 476a11-15, PA 658a5-10, 659b17-19, 661b22-25, IA 704b12-18, 708a9-14, 711a17-19, GA 741b4-5, 744a36-38, Pol. 1253a9-18, 1256b20-22. For discussion, see D. Henry (2013), J.G. Lennox (2001), p. 205-223, and M. Leunissen (2010), p. 129-135. 11 See sect. 5. 10

WHY DE ANIMA NEEDS III, 12-13

335

the distance senses belong of necessity only to roaming animals, since only their survival depends additionally on perceiving from afar (434b2427). But then he introduces another contrast: in addition to being necessary for only some animals, the distance senses differ from the contact senses in being “for the sake of well-being” (b24). It is clear from subsequent discussion of the senses in E and the thematically continuous De Sensu (Sens.) 1 that the distinction is in fact a finer-grained application of the principle that nature does nothing in vain. Every natural attribute must belong for the sake of something, but there are two ways in which it might do so: by belonging for the sake of being (τοῦ εἶναι ἕνεκα) or by belonging for the sake of well-being (τοῦ εὖ ἕνεκα). However, it is obscure why a finer-grained application of the principle is needed to explain the serial distribution of the senses among animal kinds. Few proponents of the standard approach explicitly address this question. An exception is Mariska Leunissen, who argues that Aristotle’s remark that the distance senses are for the sake of well-being is not a further contrast with the contact senses, but an explanation of why they belong only to some animals. In her view, the contrast between belonging for the sake of being and belonging for the sake of well-being parallels a contrast drawn in the zoological works between belonging “on account of the necessary” (διὰ τὸ ἀναγκαῖον) and belonging “on account of the better” (διὰ τὸ βέλτιον)12. Aristotle uses the latter distinction to separate attributes that belong to every member of a biological kind from those that belong only to some. If an attribute belongs to all members of the relevant kind, Aristotle infers that it must belong to them necessarily; but if it belongs to only some members of that kind, he infers that it must belong “thanks to something better” (βελτίονός τινος χάριν)13. Similarly, Leunissen claims, Aristotle is using the being / well-being distinction to separate attributes that belong to every member of a psychological kind from those that belong only to some. What distinguishes attributes, like the distance senses, that belong for the sake of well-being is that they represent more complex performances of the vital functions required for the survival of every member of that kind, such as acquiring food. That such attributes belong to only some 12 13

See M. Leunissen (2010), p. 60-62. See e.g. the famous discussion of testes in GA I 4, 717a15-21.

336

ROBERT HOWTON

members of the kind indicates that these more complex performances are not necessary but merely optimal for members of that kind to achieve the relevant end(s), since other members of that kind achieve the same ends by other means, as for instance stationary animals acquire food using only the contact senses. But the text of III 12-13 does not support the parallel on which Leunissen’s interpretation relies. If the distinction between belonging for the sake of being and belonging for the sake of well-being paralleled the distinction between belonging on account of the necessary and belonging on account of the better, there would be no natural attribute that on Aristotle’s view belonged to the same psychological kind for the sake of both being and well-being. For, if there were, the attribute would, impossibly, belong both to all and to some, but not all, members of that kind. However, Aristotle recognizes at least two types of sensory attribute that belong to the same kind for both being and well-being: taste and the distance senses. The argument for taste appears in E. Aristotle has been arguing that every animal must have “touch”, here taken to include the tactile sense and taste insofar as flavor is tangible (435b12-13), since no animal can survive the destruction of its tactile organ. To this extent, he claims, “touch” is distinct among animals’ sensory attributes (435b19-25): Τὰς δ’ ἄλλας αἰσθήσεις ἔχει τὸ ζῷον, ὥσπερ εἴρηται, οὐ τοῦ εἶναι ἕνεκα ἀλλὰ τοῦ εὖ, οἷον ὄψιν, ἐπεὶ ἐν ἀέρι καὶ ὕδατι, ὅπως ὁρᾷ, ὅλως δ’ ἐπεὶ ἐν διαφανεῖ, γεῦσιν δὲ διὰ τὸ ἡδὺ καὶ λυπηρόν, ἵνα αἰσθάνηται τὸ ἐν τροφῇ καὶ ἐπιθυμῇ καὶ κινῆται, ἀκοὴν δὲ ὅπως σημαίνηταί τι αὐτῷ, γλῶτταν δὲ ὅπως σημαίνῃ τι ἑτέρῳ. But the animal has the other senses, as we said, not for the sake of being but for the sake of well-being. For instance, [it has] vision so that it may see, since it [is immersed] in air and water, and generally since it [is immersed] in transparent [stuff]; taste on account of pleasure and pain, in order to perceive and be moved by that in food and in appetite; hearing so that something may be signified to it; and a tongue so that it may signify something to another.

Recalling his claim in B (434b24-26), Aristotle again contrasts the senses in terms of their contributions to being and well-being. Here, however, taste is listed as an attribute belonging to animals both for the sake of being and well-being. This may appear problematic, since the passage seems to imply that the attributes belonging for the sake of well-being do

WHY DE ANIMA NEEDS III, 12-13

337

not belong for the sake of being14. But such a view would ignore the care with which Aristotle separates taste’s contribution to animals’ being from its contribution to their well-being. Taste is for the sake of being insofar as it is a mode of touch and so a prerequisite for the animal’s continued existence; but it is for the sake of well-being insofar as it provides for the experience of pleasure and pain and the satisfaction of appetitive desires, which do not belong to taste insofar as it is a mode of touch15. Aristotle is equally careful in distinguishing the distance senses’ contributions to the being and well-being of a subclass of roaming animals (Sens. 1, 436b12-437a3): Now, with respect to each [sense] in particular, touch and taste are consequent of necessity upon all [animals] …. But the distance senses, namely smell, hearing, and vision, [are consequent of necessity only] upon those among them that roam. They belong to all that have them for the sake of survival, so that they may pursue food and avoid what is bad and destructive by perceiving in advance; but to those who also have intelligence [they belong] for the sake of well-being. For they report many differences, from which arises intelligence about the objects of thought and of practical activity.

Distance senses contribute to the being – or “survival” (σωτηρία), as he puts it here16 – of all roaming animals. But “to those who also have intelligence” (τοῖς δὲ καὶ φρονήσεως τυγχάνουσι) they belong for the sake of well-being, since they contribute to additional cognitive ends like practical and theoretical intelligence. Aristotle appears to be isolating a subclass of cognitively advanced roaming animals and claiming of them that the distance senses contribute to ends beyond those that explain their presence in other roaming animals. But if so, then intelligent roaming animals have distance senses for both being and well-being17. 14 Cf. M. Leunissen (2010), p. 67, n. 47, who approvingly cites the suggestion of D.S. Hutchinson (1987), p. 377, n. 2, that ὄσφρησιν be read for γεῦσιν at 435b22. 15 Cf. Aquinas, In DA § 873, who refers helpfully to a parallel in the case of smell; see Sens. 443b17-444a28. 16 Cf. B, 434b26-27 17 Pace M. Leunissen (2010), p. 67-68, Aristotle probably does not regard the distance senses as a “hybrid category” that is for the being of roaming animals but the well-being only of the stationary animals that possess them. For even if he tentatively extends distance senses to some stationary animals (e.g. stationary testacea, HA 534b11-535a25), he is much more inclined to compare stationary animals to rocks and plants than to attribute to them what he here calls “intelligence” (φρόνησις); see esp. HA 588b16-21, GA 731a24b16. Cf. J.-L. Labarrière (2005b), esp. p. 415-417.

338

ROBERT HOWTON

These passages show that Aristotle’s actual usage of the being / wellbeing distinction conflicts with Leunissen’s suggestion, on behalf of the standard reading, that it is intended to separate attributes that belong to every member of a kind from those that belong only to some. Rather, it seems intended to distinguish types of end for whose sake natural attributes might belong to a psychological kind. A natural attribute belongs for the sake of being insofar as it promotes some basic or vital end(s) for members of a kind, but it belongs for the sake of well-being insofar as it promotes ends beyond those that explain its contribution to being. Thus taste belongs to animals for the sake of being because it is necessary for nutrition and survival; but it belongs for the sake of well-being because it also enables the experience of appetitive pleasure and pain. Similarly, the distance senses belong to intelligent roaming animals for the sake of being because, as roaming animals, they need to be able to spot opportunities and threats at a distance; but they belong for the sake of wellbeing to the extent that they also make available cognitive achievements like practical and theoretical intelligence. Here we might be reminded of considerations that lead Aristotle to assert elsewhere that the sentient life of an animal is intrinsically better than the insensate life of a plant (GA 731a24-b9). In his view, animals and plants share a “generating” or “reproducing” function (τὸ γεννῆσαι ἔργον) that consists in the production of generative seed (ἡ τοῦ σπέρματος γένεσις). But to the extent that animals perceive, they also share in an additional function, cognition (γνῶσις). Some animals of course enjoy a greater share than others, but even the scantest share has value that cannot be accounted for in terms of reproduction alone; even the life of a stationary animal, though paltry in comparison to the life of an intelligent animal, is supremely better than that of a plant (b1-3). Plausibly, then, Aristotle stresses the senses’ contributions to well-being in B, E and Sens. 1 in part to illustrate their contribution to ends beyond reproduction. In what follows I defend a version of this suggestion. In my view, the unifying purpose of III 12-13 is to show that the psychic capacities naturally present in most psychological kinds are there to promote the organism’s reproductive function, and Aristotle uses the being / well-being distinction in part to isolate how certain capacities contribute to reproduction. But this gets us only part of the way to the conclusion that III 12-13 is an integral step in the DA study of soul. For even if their unifying purpose is to explain how psychic capacities promote organisms’

WHY DE ANIMA NEEDS III, 12-13

339

reproductive function, as I am suggesting, we would still need to know what this explanation contributes to Aristotle’s study of the nature and essence of soul in DA. To appreciate III 12-13’s contribution to the DA account of soul, we must also take into account a model of soul that is operative in Aristotle’s zoological works, but whose significance for DA has not fully been acknowledged. 4. The model of soul I have in mind is distinguished by the thesis that the soul of a perishable organism, an animal say, is a complex activity of its organic body. It is, in other words, an actuality characterized simultaneously by: (1) a diverse set of functions (ἔργα) or activities (πράξεις) that are, nevertheless, (2) each for the sake of a single, teleologically primary function or activity. The clearest evidence of Aristotle’s commitment to this “complex activity” model of soul appears in his zoology, in a passage from Parts of Animals I 5, 645b14-20: Since every instrument is for the sake of something, and each part of the body is for the sake of something, and what each is for the sake of is a certain activity, it is apparent that the whole body too has been constituted for the sake of a certain complex activity. For sawing does not come about for the saw’s sake, but the saw for the sake of sawing, since sawing is a certain use [of the saw]. So too is the body in a way for the sake of the soul, and the parts are for the sake of the functions for which each naturally developed.

The passage recalls a pair of analogies from DA II 1, where Aristotle presented an “outline” account of soul as “the first actuality of a natural organic body”, the potential for a range of vital activities present in a body naturally suited to realize those activities (412b9-413a10). There Aristotle compared the living body endowed with such a potential to an instrument – whether artificial, like a saw, or natural, like an eye – to illustrate soul’s status as the form and essence of the living body. As a saw would be a saw in name only without the power to saw, and as an

340

ROBERT HOWTON

eye would be an eye in name only without the power of sight, so too would the living body have life only nominally if it lacked the potential for the vital activities whereby it is said to be alive. Here, by contrast, the analogies are intended to illustrate soul’s status as the end and final cause of the organic body. As a saw is for the sake of sawing in the sense that it was made to be an instrument used for sawing, and as an eye is for the sake of seeing in the sense that it naturally developed to be an instrument used for seeing, so too is the animal soul – understood here, I suggest, not as a first actuality like sight but as a second actuality like seeing – a certain use of the entire animal body, for whose sake the animal body naturally developed to be an instrument. As I read the passage, Aristotle describes the use for which the animal body is instrumental, and which characterizes the animal soul, as “a certain complex activity” (πρᾶξίς τις πολυμερής)18. The description calls attention to both aspects of the model of soul I sketched at the beginning of this section. (1) That the activity characteristic of soul is complex makes clear that it incorporates the activities for whose sake of each of the body’s instrumental parts developed, including reproduction, growth, copulation, waking, sleep, and locomotion (cf. PA I 5, 645b3335). (2) Yet it remains an activity, a single use for whose sake the body naturally developed. The diverse activities of the animal’s instrumental parts are aspects of a single (albeit complex) activity of its entire body because each is for the sake of the same teleologically primary activity. Aristotle signals the possibility that some psychic activities may be for 18 This is the reading of ms. P, accepted by D. Charles (2000), p. 330, I. Düring (1992), P. Louis (1956), S. Menn (2002), p. 109, n. 38, W. Ogle (1911), and A.L. Peck in A.L. Peck and E.S. Forster (1937). Several other commentators, following the majority of mss., read πλήρης, “complete”, in place of πολυμερής; see D.M. Balme (1992), W. Kullmann (2007), J.-L. Labarrière (2005a), p. 242, n. 6, J.G. Lennox (2002), J.G. Lennox (2010), p. 332, n. 5, and P.-M. Morel (2006), p. 135, n. 70. I think the majority reading is inferior on both textual and philosophical grounds. Textually, it ignores compelling parallels of the phrase “complex activity” at PA 646b14-15 and Poet. 1459b1; and it has plausibly been argued to be a corruption of the minority reading: see I. Düring (1992), p. 122-123. (W. Kullmann’s suggestion, on behalf of the majority reading, that P is an attempt to correct problems with the argument cannot likewise account for the corruption to πολυρους in ms. E; see his 2007, p. 356.) Philosophically, it needlessly attributes a compositional fallacy to Aristotle; see J.G. Lennox (2002), p. 176, cf. EN 1097b241098a8. Aristotle can and does establish the instrumental relation between body and soul without resorting to a compositional fallacy; see DA II 4, 415b15-21. Moreover, we will see that the complex activity model reflects an important feature of Aristotle’s account of soul in DA.

WHY DE ANIMA NEEDS III, 12-13

341

the sake of another in this way a few lines after the quoted passage, where he claims that the teleological relations among the instrumental parts of the body mirror those among their respective activities (645b2933). Some activities, he claims, are “for the sake of” others, which are therefore “prior to and end of” the former, while others belong to the animal, not for the sake of some end, but as a necessary consequence of other, goal-oriented activities; and in each case the instrumental parts of the animal’s body will relate to one another as the activities for which they are instrumental. In describing soul as a complex activity, I suggest, Aristotle is endorsing a special case of this teleological structure, one in which there is a single activity that is prior to and end of every other activity. So read, PA’s description of soul as a complex activity reflects a significant development in Aristotle’s study of soul in DA II. Aristotle claimed in DA II 1 that, in general, soul is the first actuality of a natural organic body, but in II 2 he argued that this account will not suffice for an essence-revealing definition of soul (413a11-20). A definition of soul should make clear how it operates as the cause and principle of life in ensouled things. Life, however, is said in many ways. Some ensouled things are said to live simply because they grow and reproduce, others because they perceive and roam around, and still others because they understand and reason. Such multivocity of life indicates for Aristotle that there can be no generic definition of soul that captures its role as cause and principle of every activity whereby things are said to live. Instead, we must seek definitions of soul as a cause and principle of each of the principal vital activities living things engage in – definitions of nutritive soul, of perceptive soul, and so on. But this conclusion should not be taken to imply that the soul is a mere collection of capacities for vital activities. What Aristotle takes to be the cause of psychic unity is controversial19, but an important component of his explanation seems to be that the vital activities that soul makes available to living things are structured by relations of teleological priority20. For, after concluding that the most appropriate account of soul will be an account of its principal capacities, Aristotle attributes 19 For discussion, see C. Frey (2015), T.K. Johansen (2012), p. 47-72, and J. Whiting (2002); on the related issue of Aristotle’s criterion for psychic parthood, see also K. Corcilius and P. Gregoric (2010). 20 Cf. G.B. Matthews (1992).

342

ROBERT HOWTON

special priority to nutritive soul, the capacity jointly responsible for nutrition and reproduction (DA II 4, 415a23-b7): Ἡ γὰρ θρεπτικὴ ψυχὴ καὶ τοῖς ἄλλοις ὑπάρχει, καὶ πρώτη καὶ κοινοτάτη δύναμίς ἐστι ψυχῆς, καθ’ ἣν ὑπάρχει τὸ ζῆν ἅπασιν. ἧς ἐστὶν ἔργα γεννῆσαι καὶ τροφῇ χρῆσθαι. φυσικώτατον γὰρ τῶν ἔργων τοῖς ζῶσιν, ὅσα τέλεια καὶ μὴ πηρώματα, ἢ τὴν γένεσιν αὐτομάτην ἔχει, τὸ ποιῆσαι ἕτερον οἷον αὐτό, ζῷον μὲν ζῷον, φυτὸν δὲ φυτόν, ἵνα τοῦ ἀεὶ καὶ τοῦ θείου μετέχωσιν ᾗ δύνανται· πάντα γὰρ ἐκείνου ὀρέγεται, καὶ ἐκείνου ἕνεκα πράττει ὅσα πράττει κατὰ φύσιν. τὸ δ’ οὗ ἕνεκα διττόν, τὸ μὲν οὗ, τὸ δὲ ᾧ. ἐπεὶ οὖν κοινωνεῖν ἀδυνατεῖ τοῦ ἀεὶ καὶ τοῦ θείου τῇ συνεχείᾳ, διὰ τὸ μηδὲν ἐνδέχεσθαι τῶν φθαρτῶν ταὐτὸ καὶ ἓν ἀριθμῷ διαμένειν, ᾗ δύναται μετέχειν ἕκαστον, κοινωνεῖ ταύτῃ, τὸ μὲν μᾶλλον τὸ δ’ ἧττον· καὶ διαμένει οὐκ αὐτὸ ἀλλ’ οἷον αὐτό, ἀριθμῷ μὲν οὐχ ἕν, εἴδει δ’ ἕν. For the nutritive soul belongs also to the others, and is the primary and most common capacity of soul in respect of which living belongs to all. (Its functions are reproducing and [in general] using food.) For the most natural function for living things – those that are complete and not maimed or have spontaneous generation – is for it to produce another like itself, an animal an animal, a plant a plant, so that they may share as much as possible in the eternal and divine. For all strive for that, and do whatever they do by nature for the sake of that. (And that for the sake of which is double: on the one hand that of which, on the other that for which.) So, since it cannot share in the eternal and divine in continuity, on account of the inability of any perishable thing to persist one and the same in number, each shares in it as much as it can partake, some more, some less; and not it but something like it persists, not one in number, but one in species.

That the nutritive is the primary capacity of soul is indicated by the fact that it is the most common among living things, an observation Aristotle made earlier in discussing the serial distribution of psychic capacities. Here, however, he appears to offer an explanation of the phenomenon. Nutrition is the primary and most common psychic capacity because most living things do whatever they do by nature for the sake of one of its functions, namely reproduction, since this activity is the best available means for them to share in the eternal and divine. The explanation stresses precisely the sort of teleological priority outlined in PA I 5. Reproduction is the “most natural” (φυσικώτατον) function for most kinds of living thing because it is that for whose sake most kinds of living thing do “whatever they do by nature” (ὅσα πράττει κατὰ φύσιν). What a living thing “does by nature” comprises vital activities (πράξεις) such as growing and feeding, perceiving and locomoting,

WHY DE ANIMA NEEDS III, 12-13

343

and generally any activity that corresponds to the exercise of inborn psychic capacities, since these are the activities available to it by virtue of soul, its formal nature21. In claiming that most living things do all they do by nature for the sake of reproduction, then, Aristotle means that reproduction is the end of each of their vital activities, at least insofar as reproduction is the best available means for them to share in the eternal and divine. To this extent reproduction is also their most natural function, in the sense that it corresponds to the end of their formal nature – the end, not only of nutritive soul, but of the organism as a whole22. The idea that reproduction is the end for whose sake most living things perform the vital activities naturally available to them is therefore at the heart of Aristotle’s view that nutrition is the primary and most common capacity of soul. Living things do all they do by nature for the sake of sharing as much as possible in the eternal and divine, which for most psychological kinds is limited to reproduction. As the capacity responsible for reproduction, nutrition will therefore be the most common psychic capacity among living things. For the same reason, it will also be the primary psychic capacity. Since living things do all they do by nature for the sake of the most divine activity available to them, and since for most living things this activity is reproduction, most living things will be naturally endowed with the capacity for only those nonreproductive activities that promote their ability to achieve their primary reproductive end. For the most part, then, each of the psychic capacities with which nature has endowed a living thing will be present for the sake of reproduction, though in different ways: nutritive soul will be present for the sake of reproduction because reproduction is the end and primary function of that capacity23, whereas other capacities will be present for the sake of reproduction because, and insofar as, exercises of those capacities promote reproduction. It is perhaps in order to bring out this contrast in the way that nutrition and other psychic capacities are for the sake of reproduction that Aristotle 21 Aristotle is never explicit about the identity between soul and the nature of a living thing, but cf. DA II 1, 412a19-20 (soul is form, cf. PA 641a17-18, Met. 1035b14-16) with Phys. II 1, 193b6-7 (form is nature in dominant sense, cf. PA 640b27-29, Met. 1015a1315). For discussion, see e.g. C. Frey (2015), T.K. Johansen (2012), p. 85-89, R. Polansky (2007), p. 7. 22 Contrast T.K. Johansen (2012), p. 119, R. Polansky (2007), p. 205. On a thing’s end as its nature, see H. Bonitz (1870), 836a51-b28. 23 See DA II 4, 416b23-25.

344

ROBERT HOWTON

reminds us in this context that ‘that for the sake of which’ is said in at least two ways, as that of which and as that for which. According to many commentators, the distinction highlights two ways of being the end of some process or activity: on the one hand being the good achieved by the fulfillment of a process or activity (the “end of which”), and on the other being what benefits from the achievement of that good (the “end for which”). So, for instance, the end of which of a psychic capacity is its function, the good activity or product characterizing its exercise, whereas its end for which is who or what benefits from the achievement of that good24. Aristotle’s appeal to the distinction in this context has often been read as an attempt to discourage a potential implication of the claim that living things do what they do by nature for the sake of reproduction, namely that there is some beneficiary of an organism’s reproductive activity apart from the organism itself, such as god or the biological kind to which it belongs25. On the present interpretation, however, there is a more immediate concern for the distinction to address. Aristotle has just argued that, for the most part, each of the vital activities of which a living thing is naturally capable is for the sake of reproduction, and he has inferred from this that the capacities for these activities must belong to the psychological kinds that have them for the sake of reproduction. But, while it may be obvious that living things have the ability to produce generative seed for the sake of reproduction, it is far from clear that reproduction also explains why they are capable of such diverse nonreproductive activities as growth, self-nutrition, and perception. To show how reproduction could be the end for whose sake living things for the most part have every psychic capacity naturally present in them, Aristotle clarifies that we must think of reproduction as “that for the sake of which” in both of its senses. Nutritive soul in its reproductive capacity belongs for the sake of reproduction as the end of which, the function characterizing the fulfillment of that capacity. By contrast, living things’ non-reproductive capacities belong for the sake of reproduction as the end 24 See J.K. Johansen (2015), M.R. Johnson (2005), p. 65-69, W. Kullmann (1985), M. Leunissen (2010), p. 56, and S. Menn (2002), p. 113. This reading takes the genitive expressed by τὸ οὗ to be what H.W. Smyth (1956), § 1349, calls the “genitive of the end desired”, similar to the genitive verbs like ὀρέγεται and στοχάζεται take as object; whereas the dative expressed by τὸ ᾧ is what H.W. Smyth (§ 1474, cf. § 1461) calls the “dative of interest” or benefit; cf. D. Sedley (1991), p. 180, n. 3. 25 Cf. GA 731b24-732a1. The only exceptions I know of are M.R. Johnson (2005) and T.K. Johansen (2015).

WHY DE ANIMA NEEDS III, 12-13

345

for which, since for the most part it is as reproducers that they benefit from the exercise of those capacities26. But even if Aristotle distinguishes these two ways of being an end to clarify how reproduction can be the end for whose sake a living thing has each of the psychic capacities naturally belonging to it, substantial gaps remain in his exposition. If his view is that most living things possess only those non-reproductive capacities that benefit them as reproducers, it remains for him to state how the diverse capacities present in a psychological kind contribute to its members’ primary reproductive end – how, for instance, plants and animals are benefited as reproducers by feeding and growing, while only animals are benefited as reproducers by perceiving. It also remains for him to consider the ends that explain the presence of psychic capacities in living things for whom reproduction is not the most divine activity. In arguing for the priority of nutritive soul for most psychological kinds, Aristotle implicitly acknowledged two types of living thing to which this generalization does not apply. One type comprises rational living things in possession of theoretical intellect, such as human beings and god27. Nutrition is not primary for these living things, for though some undoubtedly strive for reproduction28, it is not reproduction but contemplation, the exercise of theoretical intellect, that represents their best available means of sharing in the eternal and divine29. Another type comprises living things for whom reproduction is not an attainable end, either because they are congenitally sterile, like mules or spontaneously generated animals, or because they have suffered damage to their reproductive organs30. Since, for 26 Cf. Phys. 194a25-26: the function of helm-making is a helm (or the production of one), but it is the navigator who benefits as such from the achievement of that function. In contrast to W. Kullmann (1985), then, my interpretation aligns this passage with those, like DA II 4, 415b20-21 and the Phys. passage just cited, in which Aristotle seems to be spelling out the οὗ ἕνεκά τινι of some process or activity, rather than those in which he seems to be denying that some process or activity has a οὗ ἕνεκά τινι, as at Met. 1072b1-2 and EE 1249b15-16. 27 See DA II 3, 415a11-12. 28 See Pol. 1252a26-30. 29 See EN 1177a12-18, 1178b27-32. 30 I take these to be the cases Aristotle has in mind as the incomplete, maimed, or spontaneously generated organisms mentioned at 415a27-28. Significantly, I deny that the “incomplete” organisms mentioned there include adolescents of fertile kinds, since, as we will see, Aristotle takes growth and maturation to benefit fertile organisms as reproducers; contrast Aquinas, In DA § 313, Philoponus, In DA 267.24-25 Hayduck, and perhaps C. Shields (2016), p. 201.

346

ROBERT HOWTON

different reasons, reproduction is not the most divine activity available to these living things, it can neither be the teleologically primary activity for whose sake they have the psychic capacities with which nature has endowed them. If, then, there is some other teleologically primary activity that explains the presence of the psychic capacities belonging to these living things, Aristotle needs to specify what the relevant activities are, and how they are promoted by the presence of the capacities belonging by nature to the relevant kinds. These are questions left open by Aristotle’s discussion of the priority of nutrition in II 4, and what I suggest III 12-13 are meant to address. Having concluded in III 11 the examination of the principal capacities of soul inaugurated in II 4, and having determined the function definitive of each, Aristotle returns in III 12-13 to their final cause, the ends promoted by the achievement of their respective functions31. His aim, as I’ve suggested, is to determine the contribution of the capacities present in the broadest psychological kinds he has considered – plants, stationary and roaming animals, and humans32 – to their teleologically primary activity, whether reproduction, contemplation, or something else. To do so he appeals to the principle that nature does nothing in vain, which he here applies as a constraint on what can be said to belong by nature. Since psychic capacities number among living things’ natural attributes, and since Aristotle has argued that each of the psychic capacities present in the living thing belongs for the sake of its teleologically primary activity, he can use the principle to identify the contribution of a psychic capacity by reasoning counterfactually from its absence to the organism’s consequent inability to achieve its primary activity. I devote the final section to spelling out how this strategy unfolds in III 12-13, and I conclude by highlighting some lingering interpretive questions. 5. Aristotle begins with the capacities for nutrition and perception. All perishable living things must have nutritive soul, here conceived in its 31 Hence I read the transitional μὲν οὖν at 434a22 as having a resumptive force. Contrast T.K. Johansen (2012), p. 277, who also reads III 12-13 as addressing the “beneficiary” of the principal capacities of soul but does not see the connection with II 4 I am defending. 32 Cf. DA II 3, 414b32-33

WHY DE ANIMA NEEDS III, 12-13

347

threptic capacity as the power to maintain an organism as such33, since none could grow, reach maturity, and decline without food ([A] 434a2226). Though it may not initially be clear, reproduction has an important role to play in this explanation. In stressing the need for perishable living things to reach “maturity” (ἀκμή), Aristotle seems to have in mind specifically reproductive maturity, since, as recently as III 9, he has characterized as “complete” living things that “reproduce and have maturity and decline” (432b24-25). Hence, part of Aristotle’s claim seems to be that nutritive soul in its threptic capacity benefits perishable living things as reproducers by enabling development into reproductive maturity. He may also intend for the account to apply to incomplete living things, like mules and barnacles, for whom sharing in the eternal and divine seems limited to self-maintenance and individual existence. For the account also makes clear that these ends are not attainable without nutritive soul in its threptic capacity. Reproduction also has an important place in perception’s contribution to animal life ([A] 434a27-b8). Perception, in Aristotle’s view, is a discriminative capacity: its function is to discriminate objects’ sensible attributes by receiving sensible form without the matter34. Without the ability to discriminate sensible attributes, no animal could feed itself, and hence none would be able to achieve “the end that is the function of its nature” (τέλος… ὅ ἐστι φύσεως ἔργον)35. It is not initially clear what Aristotle is thinking of as the function of an animal’s nature, but the reference comes into sharp relief against the background of another context in which he stresses the teleological priority of reproduction36. It is a basic principle of Aristotle’s developmental biology that reproduction is the “natural function” (ἔργον φύσει) of every complete animal and plant (GA II 1, 735a18-19). It explains, for instance, why the first part of the animal to develop is its primary nutritive organ, since this is the part “possessing the source and end of its entire nature” (τὸ πάσης ἔχον τῆς φύσεως ἀρχὴν καὶ τέλος), namely the power to reproduce, for 33

See DA II 4, 416b17-19. See DA III 9, 432a16 with II 12, 424a17-19. 35 Although it is clear that Aristotle thinks all animals need at least contact senses, he oddly restricts his argument here to roaming animals. Perhaps he thought roaming animals would better emphasize perception’s contribution to nutrition and reproduction, since in his view stationary animals live pretty much like plants. 36 Cf. R.D. Hicks (1907), p. 576-577. 34

348

ROBERT HOWTON

whose sake the animal’s other instrumental parts are present and naturally develop37. The point made here about animals’ capacity to perceive is similar: because perception belongs to most animals for the sake of reproduction, it benefits them as reproducers, since without perception they could not nourish themselves and reach reproductive maturity38. Perception also benefits some animals in ways that go beyond nutrition and reproduction. Notably, it promotes thinking in rational animals, since without perception it would be impossible to acquire “discriminative intellect” (νοῦς κριτικός, [A] 434b3). Aristotle’s remark is telegraphic, but given that contemplation is the most divine activity available to rational animals, his point may be that perception benefits rational animals as contemplators as well as reproducers. But the cognitive benefits of perception are not limited to rational animals. In B, E, and Sens. 1 Aristotle considers the contributions of the individual senses, each of which functions to discriminate the sensible qualities it uniquely perceives. While the contact senses benefit all animals as self-nourishers and reproducers, since the qualities they discriminate are what nourish and grow animals, the distance senses benefit only roaming animals, since discrimination of their special objects enables remote perception of opportunities and threats. In some roaming animals, however, the distance senses also promote cognitive ends like practical and theoretical intelligence. In this respect they compare with taste, which Aristotle takes to benefit animals cognitively in addition to promoting nutrition and reproduction. From the point of view of Aristotle’s project in III 12-13, there are at least two reasons to highlight these contributions to animal “wellbeing”. First, by distinguishing a capacity’s contribution to being from its contribution to well-being, Aristotle can isolate its contribution to its possessor’s teleologically primary activity from other goods promoted through its exercise. In doing so, however, Aristotle also indicates that possession of a psychic capacity can make available goods whose value 37 See GA II 6, 742a27-33, b1-2. Unlike the nutritive organ, however, these parts are for the sake of reproduction, not as the dynamic source of the organism’s reproductive activity, but as instruments for the organism’s pursuit of its reproductive end. See J. Gelber (2018) for a compelling interpretation of the τὸ οὗ/τὸ ᾧ distinction based on the distinction between these two ways of being for the sake of something; cf. J. Rosen (2014), p. 98-101. 38 Hence animals’ natural function should be distinguished from what Aristotle at EN 1097b33-1098a3 calls their peculiar function, namely perception. I’m grateful to Jennifer Whiting for pressing me on this point.

WHY DE ANIMA NEEDS III, 12-13

349

goes beyond the ends that explain its present in the relevant kind. For even if their benefit to animals as self-nourishers and reproducers explains the presence of taste and the distance senses in the animals that have them, they may nevertheless benefit their possessors by promoting cognitive ends whose value is not exhausted by any contribution to nutrition and reproduction. This interpretation of III 12-13 leaves open several interpretive questions. I conclude by addressing three of them to the extent the scope of the present study will allow. First, there is the question of what do to with C and D. C, I claimed, is a digression on the conditions for distance perception occasioned by Aristotle’s observation that roaming animals must be able to perceive remotely if they are to survive. But D, which argues in detail for the elemental compositeness of the animal body, appears more central to the discussion, since Aristotle has primed us for this topic earlier, in B (434b8-11). I suggest that Aristotle’s remarks in D are a further application of the non-futility principle that has guided his examination of the final causes of psychic capacities. Animals could not nourish themselves and reproduce without contact perception, and since contact perception requires a body that is elementally composite ([D] 435a21-b4), for the same reason animals could not survive and reproduce without an elementally composite body. Second, if the unifying purpose of III 12-13 is as I’ve suggested, one might wonder whether the chapters are not woefully inadequate to the task. Aristotle may have shown how nutrition, perception, and the senses contribute to the teleologically primary activity appropriate to various psychological kinds. But is this enough to show that every capacity belonging by nature to a living thing, including the capacities for locomotion, sleep, respiration, and other vital activities, is also for the sake of that activity? I suspect this worry underestimates what Aristotle takes himself to have accomplished in III 12-13. In establishing, for instance, that perception belongs to animals for the sake of reproduction, Aristotle plausibly takes himself to have also shown that the animal is capable of respiration, sleep, and death for the same end, since in his view these attributes belong to animals on account of their ability to perceive39. And in general, by explaining the presence of the principal capacities of soul in a psychological kind, Aristotle can plausibly take himself also to be 39

See Sens. 436b3-6.

350

ROBERT HOWTON

explaining the presence of any capacity belonging to that kind on account of one or more of those principal capacities40. Finally, this interpretation of III 12-13 entails a controversial view of the teleological relations among the capacities present in a psychological kind. On my view, neither are higher (non-rational) psychic capacities like perception all present in a psychological kind for the sake of lower functions like nutrition and reproduction, nor are lower capacities all present for the sake of higher functions like cognition41. Rather, Aristotle distinguishes two fundamentally different types of teleological relation, which enables him to say that the same capacity can be for the sake of lower functions insofar as it contributes to an organism’s being but for the sake of higher functions insofar as it contributes to its well-being. I have argued that, of these, what explains a capacity’s presence in a psychological kind is its contribution to being. And while it is undeniable, on the present reading, that Aristotle takes a capacity’s contributions to well-being to increase the objective value of an organism’s life, it is unclear whether he regards these goods as strictly incidental to an organism’s pursuit of its teleologically primary activity, or whether he envisions a more robust role for them in explaining the mode of living proper to that organism42. III 12-13 do not help us to settle this important question, so full treatment lies outside the scope of the present study. But if the interpretation I have defended is correct, any explanatory role for a capacity’s contributions to wellbeing will be distinct from an account of its presence in a psychological kind, which for Aristotle must be explained by its contribution to the teleologically primary activity appropriate to that kind43. 40 See also M. Leunissen (2010), p. 70-74, for discussion of the ends promoted by locomotion. 41 For the first view, see e.g. M. Leunissen (2010), p. 59; for the second, see e.g. M.R. Johnson (2005), p. 5, and perhaps S. Menn (2002), p. 121. Contrast T.K. Johansen (2012), p. 280-281, and R. Polansky (2007), p. 541-542. 42 This may be an important point of contrast between rational and non-rational animals, but, at least in the context of III 12-13, Aristotle does not describe perception’s contribution to thinking as a contribution to well-being. Though cf. his remarks on hearing’s contribution to intellect at Sens. 437a5, 10-15. 43 I am grateful to participants in the De Anima III Workshop in Honor of Michel Crubellier at Université de Lille and a spring 2017 graduate seminar on De Anima at the University of Pittsburgh for questions and comments; to Jessica Gelber and Jim Lennox for discussion of various aspects of this chapter; and especially to David Charles and Jennifer Whiting for written comments on earlier drafts.

BIBLIOGRAPHIE AUTEURS ANCIENS ET MÉDIÉVAUX Alexandre d’Aphrodise De Anima BRUNS, I. (ed.), 1887-1892, Alexandri Aphrodisiensis praeter commentaria, scripta minora  : De Anima liber cum Mantissa, Berlin, Reimer. Quaestiones BRUNS, I. (ed.), 1892, Alexandri Aphrodisiensis Praeter Commentaria scripta minora  : Quaestiones, De Fato, De Mixtione, in Supplementum Aristotelicum, II.2, Berlin.

Aristote Œuvres complètes BARNES, J. (éd.), 1984, The Complete Works of Aristotle. The Revised Oxford Translation, vol. I-II, Princeton, Princeton University Press. PELLEGRIN, P. (dir.), 2014, Aristote. Œuvres complètes, Paris, Flammarion. Analytica Priora et Posteriora (A.Pr. ; A.Po.) CRUBELLIER, M., 2014, Aristote. Premiers Analytiques, Paris, GF-Flammarion. BARNES, J., 1975, 1993², Aristotle’s Posterior Analytics, Oxford, Clarendon Press. PELLEGRIN, P., 2005, Seconds Analytiques, Paris, GF. ROSS, W.D., 1949, Prior and Posterior analytics, Oxford, Clarendon Press. De Anima (DA) BODÉÜS, R., 1993, Aristote. De l’âme, Paris, GF-Flammarion. CORCILIUS, K., 2017, Aristoteles. Über die Seele/De anima, Hamburg, Felix Meiner Verlag. FÖRSTER, A., 1912, Aristoteles: De anima libri III, Budapest, Academiae Litterarum Hungaricae. HAMLYN, D.W., 1968, Aristotle. De Anima, Books II and III with Passages from Book I, Oxford, Clarendon Press.

352

BIBLIOGRAPHIE

— 1993, Aristotle. De Anima, Books II and III with Passages from Book I, and a report on recent work and a revised bibliography by C. Shields, Oxford, Clarendon Press. HETT, W.S., 1936, Aristotle. On the Soul, Parva Naturalia, On Breath, Cambridge, MA, London, Heinemann, Harvard University Press. HICKS, R.D., 1907, Aristotle. De Anima, with Translation, Introduction and Notes, Cambridge, Cambridge University Press. JANNONE, A., BARBOTIN E., 1966, Aristote. De l’âme, Texte établi par A. Jannone et traduit par E. Barbotin, Paris, Les Belles Lettres. MOVIA, G., 19912, Aristotele, L’anima, Loffredo, Napoli. RODIER, G., 1900 (19852), Aristote, Traité de l’âme (Texte, traduction et notes par G. Rodier), Paris, E. Leroux. ROSS, W.D., 1956, Aristotelis De Anima, Oxford, Oxford Classical Texts (« editio minor »). — 1961, Aristotle. De anima, Oxford, Clarendon Press (« editio maior »). SEIDL, H., 1995, Aristoteles. Über die Seele (with the Greek text of Wilhelm Biehl and Otto Apelt), Hamburg, Felix Meiner Verlag. SHIELDS, C., 2016, Aristotle: De Anima, Translated with Introduction and Commentary, Oxford, Clarendon Press. SIWEK, P., 1965, Aristotelis Tractatus de Anima: Graece et Latine, Rome, Desclée & Ci. Editori Pontifici. SMITH, J.A., 1931, Aristotle De Anima, Oxford, Clarendon Press (repr. in J. Barnes (ed.) The Complete Works of Aristotle, Princeton 1995). THILLET, P., 2005, Aristote  : De l’âme, Traduction, présentation et notes, Paris, Gallimard. TORSTRIK, A., 1862, Aristoteles De anima libri tres, Berlin, Weidmann. TRENDELENBURG, F.A., 1833, 1877, Aristotelis. De Anima libri tres, Iena, Walzer. TRICOT, J., 1934, Aristote  : De l’âme, Paris, Vrin. WALLACE, E.,1882, ΑΡΙΣΤΟΤΕΛΗΣ ΠΕΡΙ ΨΥΧΗΣ, Aristotle’s Psychology, Cambridge, Cambridge University Press. De Generatione Animalium (GA) LEFEBVRE, D., 2014, Aristote. La Génération des animaux, in P. Pellegrin (dir.), Aristote. Œuvres complètes, Paris, Flammarion, p. 1575-1730. Metaphysica (Met.) ALEXANDRU, S., 2014, Aristotle’s Metaphysics Lambda, Leiden, Brill. ROSS, W.D., 1924, Aristotle’s Metaphysics, A Revised Text with Introduction and Commentary, Oxford, Clarendon Press, 2 vol. De Motu Animalium (DMA) NUSSBAUM, M.C., 1978, Aristotle’s De Motu Animalium, Text with Translation, Commentary and Interpretive Essays, Princeton, Princeton University Press.

BIBLIOGRAPHIE

353

De Partibus Animalium (PA) BALME, D.M., 1972, Aristotle. De Partibus Animalium I and De Generatione Animalium I (with passages from II. 1-3), Translated with notes, Oxford, Clarendon Press. — 1992, Aristotle. De Partibus Animalium I and De Generatione Animalium I, (with passages from II, 1-3), Translated with Notes, with a Report on Recent Work and an Additional Bibliography by A. Gotthelf, Oxford, Clarendon Press. DÜRING, I., 1992, Aristotle’s de Partibus Animalium: Critical and Literary Comments, New York – London, Garland Publishing. KULLMANN, W., 2007, Aristoteles. Über Die Teile Der Lebewesen, Berlin, Akademie Verlag. LENNOX, J.G., 2002, Aristotle. On the Parts of Animals I-IV, Oxford, Oxford University Press. LOUIS, P., 1956, Aristote. Parties des Animaux, Texte établi et traduit, Paris, Les Belles Lettres. OGLE, W., 1911, Aristotle: De Partibus Animalium, Oxford, Clarendon Press. PECK, A.L., FORSTER, E.S., 1937, Aristotle: Parts of Animals, Movement of Animals, Progression of Animals, Cambridge (MA), Harvard University Press. Parva naturalia (PN) ROSS, D., 1955, Aristotle. Parva Naturalia. A revised text with introduction and commentary, Oxford, Clarendon Press. MOREL, P.-M., 2000, Aristote. Petits traités d’histoire naturelle, Traduction et présentation, Paris, GF-Flammarion. Physica (Phys.) ROSS, W.D., 1950, 1966, Aristotelis Physica, recognovit brevique adnotatione critica instruxit W.D. Ross, Oxford, Clarendon Press. Averroès STUART CRAWFORD, F., 1953, Averrois Cordubensis Commentarium magnum in Aristotelis de Anima libros, Cambridge (MA), Mediaeval Academy of America. Philopon HAYDUCK, M., 1897, In Aristotelis ‘De Anima’ libros commentaria, Berlin, Reimer. Simplicius HAYDUCK, M., 1882, In libros Aristotelis De Anima Commentaria, Berlin, Reimer.

354

BIBLIOGRAPHIE

Thémistius HEINZE, R., 1899, In libros Aristotelis De Anima paraphrasis, Berlin, Reimer. Théophraste FORTENBAUGH, W.W., HUBY, P.M., SHARPLES, R.W., GUTAS, D., Theophrastus of Eresus, Sources for his Life, Writings, Thought and Influence, Part Two, Leiden, Brill, 1992. Thomas d’Aquin PIROTTA, A.M., 1925, In Aristotelis librum de Anima commentarium, Torino, Marietti. VERNIER, J.-M., 1999, Thomas d’Aquin, Commentaire du traité De l’âme d’Aristote, Paris, Vrin. Thucydide JONES, H.S., 1942, Thucydidis Historiae, 2 vols., Oxford, Oxford University Press. Zabarella ZABARELLA, G., 1605, In tres libros Aristotelis De anima commentarii, Venezia. ÉTUDES CONTEMPORAINES ANAGNOSTOPOULOS, A., 2017, « Change, Agency and the Incomplete in Aristotle », Phronesis, 62, p. 170-209. BERRYMAN, S., 2002, « Aristotle on Pneuma and Animal Self-Motion », Oxford Studies in Ancient Philosophy, 23, p. 85-97. BERTI, E., 1978, « The intellection of ‘indivisibles’ according to Aristotle, De Anima III 6 », dans G.E.R. Lloyd, G.E.L. Owen (éds.), Aristotle on Mind and the Senses. Proceedings of the Seventh Symposium Aristotelicum, Cambridge, Cambridge University Press, p. 141-163. — 1996, « Reconsidérations sur l’intellection des ‘indivisibles’ selon Aristote, De Anima, III, 6 », dans G. Romeyer Dherbey, C. Viano (dir.), Corps et âme, Sur le De Anima d’Aristote, Paris, Vrin, p. 391-404. BRENTANO, F., 1867, Die Psychologie des Aristoteles, insbesondere seine Lehre vom νοῦς ποιητικός, Mainz, Franz Kirchheim. — 1992 [1867], « Nous poiêtikos: survey of earlier interpretations » dans M.C. Nussbaum, A.O. Rorty (éds.), Essays on Aristotle’s De Anima, Oxford, Clarendon Press, p. 313-342. BRUNSCHWIG, J., 1991, « Les multiples chemins aristotéliciens de la sensation commune », Revue de Métaphysique et de Morale, 4, p. 455-474.

BIBLIOGRAPHIE

355

— 1996, « En quel sens le sens commun est-il commun ? » dans G. Romeyer Dherbey, C. Viano (dir.), Corps et âme, Sur le De Anima d’Aristote, Paris, Vrin, p. 189-218. BURNYEAT, M.F., 1990, The Theaetetus of Plato, Indianapolis, Hackett. — 2001, A Map of Metaphysics Zeta, Mathesis, Pittsburgh. — 2002, « De Anima II 5 », Phronesis, 47, p. 28-90. BYWATER, I., 1888, « Aristotelia », Journal of Philology, 17, p. 53-74. CAMPBELL, J., 1987, « Is sense transparent? », Proceedings of the Aristotelian Society, 88, p. 273-292. CAMPBELL, J., CASSAM, Q., 2014, What does experience teach us?, Oxford, Oxford University Press. CARBONE, A.L., 2011, Aristote illustré. Représentations du corps et schématisation dans la biologie aristotélicienne, Paris, Classiques Garnier. CASTON, V., 1996, « Why Aristotle Needs Imagination », Phronesis, 41, p. 20-55. — 1999, « Aristotle’s Two Intellects: a Modest Proposal », Phronesis, 44, p. 199-227. CHARLES, D., 2000, Aristotle on Meaning and Essence, Oxford, Clarendon Press. COHOE, C., 2013, « Why the Intellect Cannot have a Bodily Organ: DA III. 4 », Phronesis, 58, p. 347-377. CORCILIUS, K., 2008, Streben und Bewegen: Aristoteles’ Theorie der animalischen Ortsbewegung, Berlin, De Gruyter. — 2011, « Aristotle’s Notion of Non-Rational Desire », dans J. Miller (éd.), Aristotle’s Nicomachean Ethics. A Critical Guide, Cambridge, Cambridge University Press, p. 117-143. — 2013, « Phantasia und Phantasie bei Aristoteles », dans P. Brüllmann, U. Rombach, C. Wilde (éds.), Imagination, Transformation und die Entstehen des Neuen. Beiträge zur Entstehung des Neuen, Berlin, De Gruyter, p. 71-87. — 2013, « Aristotle’s Model of Animal Motion », Phronesis, 58, p. 52-97. CORCILIUS, K., GREGORIC, P., 2010, « Separability vs. Difference: Parts and Capacities of the Soul in Aristotle », Oxford Studies in Ancient Philosophy, 39, p. 81-120. CRUBELLIER, M., 2004, « Le syllogisme pratique, ou comment la pensée meut le corps », dans A. Laks et M. Rashed (éds.), Aristote et le mouvement des animaux  : Dix études sur le De motu animalium, Villeneuve d’Ascq, Presses Universitaires du Septentrion, p. 9-26. DE KONINCK, T., 1990, « La noêsis et l’indivisible selon Aristote », dans J.-F. Mattéi (éd.), La Naissance de la raison en Grèce, Paris, PUF, p. 215-228. DELCOMMINETTE, S., 2014, « Métaphysique Η 6 : Unité de l’ousia, unité de l’eidos », Elenchos, 35, p. 89-125. — 2019, « Physics I 5 », dans K. Ierodiakonou, P. Kalligas, V. Karasmanis (éds.), Aristotle’s Physics Alpha. Proceedings of the XXth Symposium Aristotelicum, Oxford, Oxford University Press, p. 156-189. DOW, J., 2014, « Feeling Fantastic Again – Passions, Appearances and Beliefs in Aristotle », Oxford Studies in Ancient Philosophy, 46, p. 213-251.

356

BIBLIOGRAPHIE

FREDE, D., 1992, « The Cognitive Power of phantasia », dans M.C. Nussbaum, A.O. Rorty (éds.) Essays on Aristotle’s De Anima, Oxford, Clarendon Press, p. 279-295. — 2001, « Aquinas on phantasia », dans D. Perler (éd.) Ancient and Medieval Theories of Intentionality, Leiden, Brill, p. 155-184. FREY, C., 2015, « Two Conceptions of Soul in Aristotle », dans D. Ebrey (éd.), Theory and Practice in Aristotle’s Natural Science, Cambridge, Cambridge University Press, p. 137-160. GELBER, J., 2018, « Two Ways of Being for an End », Phronesis, 63 (1), p. 64-86. GERSON, L.P., 2005, Aristotle and Other Platonists, Ithaca, Cornell University Press. GREGORIC, P., 2007, Aristotle on the Common Sense, Oxford, Oxford University Press. — 2020, « The Origin and the Instrument of Animal Motion: MA 9-10 », dans C. Rapp, O. Primavesi (éds.), Aristotle’s De Motu Animalium, Proceedings of the Symposium Aristotelicum, Oxford, Oxford University Press, p. 416-444. HARVEY, J., 1978, « Aristotle on Truth and Falsity in De Anima III 6 », Journal of the History of Philosophy, 16, p. 219-220. HENRY, D., 2013, « Optimality Reasoning in Aristotle’s Natural Teleology », Oxford Studies in Ancient Philosophy, 45, p. 225-263. HUBY, P., 2015, « When was Theophrastus’ Metaphysics written? », dans A. Jaulin, D. Lefebvre (éds.), La Métaphysique de Théophraste. Principes et apories, Louvain, Peeters, p. 5-15. HUTCHINSON, D. S., 1987, « Restoring the Order of Aristotle’s de Anima », The Classical Quarterly, 37, p. 373-381. — 1990, « Aristotle and the Spheres of Motivation: De Anima III.11 », Dialogue, 29, p. 7-20. JAEGER, W., 1913, « Das Pneuma im Lykeion », Hermes, 48, p. 29-74. JOHANSEN, T.K., 2012, The Powers of Aristotle’s Soul, Oxford, Oxford University Press. — 2015, « The Two Kinds of End in Aristotle: The View from the de Anima  », dans D. Ebrey (éd.), Theory and Practice in Aristotle’s Natural Science, Cambridge, Cambridge University Press, p. 119-136. JOHNSON, M.R., 2005, Aristotle on Teleology, Oxford, Clarendon Press. KULLMANN, W., 1985, « Different Concepts of the Final Cause in Aristotle: Philosophical and Historical Studies », dans A. Gotthelf (éd.), Aristotle on Nature and Living Things, Pittsburgh, PA, Mathesis Publications, p. 169-75. LABARRIÈRE, J.-L., 2005a, « Aristote et l’éthologie », La Condition Animale, Louvain, Peeters, p. 239-258. — 2005b. « De la phronêsis animale chez Aristote », La Condition Animale, Louvain, Peeters, p. 121-148. LAKS, A., RASHED, M., (éds.), 2004, Aristote et le mouvement des animaux  : Dix études sur le De motu animalium, Villeneuve d’Ascq, Presses Universitaires du Septentrion.

BIBLIOGRAPHIE

357

LAKS, A., 2020, « Articulating the De Motu Animalium. The Place of the Treatise within the Corpus Aristotelicum », dans C. Rapp, O. Primavesi (éds.), Aristotle’s De Motu Animalium, Proceedings of the Symposium Aristotelicum, Oxford, Oxford University Press, p. 472-496. LENNOX, J.G., 2001, Aristotle’s Philosophy of Biology: Studies in the Origins of Life Science, Cambridge, Cambridge University Press. — 2010, « Bios and Explanatory Unity in Aristotle’s Biology », dans D. Charles (éd.), Definition in Greek Philosophy, Oxford, Oxford University Press, p. 329-353. LEUNISSEN, M., 2010, Explanation and Teleology in Aristotle’s Science of Nature, Cambridge, Cambridge University Press. LEWIS, F.A., 2003, « Is There Room for Anaxagoras in an Aristotelian Theory of Mind? » Oxford Studies in Ancient Philosophy, 25, p. 89-129. LLOYD, G.E.R., 1996, « The theories and practices of demonstration », dans G.E.R. Lloyd, Aristotelian Explorations, Cambridge, Cambridge University Press, p. 7-37. LORENZ, H., 2006, The Brute Within. Appetitive Desire in Plato and Aristotle, Oxford, Oxford University Press. LOUGUET, C., 2013, « Anaxagore : analogie, proportion, identité », Philosophie antique, 13, p. 117-145. LOWE, M., 1993, « Aristotle on kinds of thinking », dans M. Durrant (éd.), Aristotle’s De Anima in Focus, London and New York, Routledge, 1993, p. 110-127, [repris de « Aristotle on kinds of thinking », Phronesis, 28, 1983, 17-30]. MCDOWELL, J., 1998, « Some Issues in Aristotle’s Moral Psychology », dans S. Everson (éd.), Companions to Ancient Thought 4: Ethics, Cambridge, Cambridge University Press, p. 107-128. Réimprimé dans Mind, Value, and Reality, Cambridge, Harvard University Press, p. 23-49. MARMODORO, A., 2014, Aristotle on Perceiving Objects, Oxford, Oxford University Press. MATTHEWS, G.B., 1992, « De Anima 2. 2-4 and the Meaning of Life », dans M.C. Nussbaum, A.O. Rorty (éds.), Essays on Aristotle’s de Anima, Oxford, Clarendon Press, p. 185-194. MENN, S., 1992, « Aristotle and Plato on God as Noûs and as the Good », Review of Metaphysics, 45, p. 543-573. — 1994, « The Origins of Aristotle’s Concept of ̓Ενέργεια: ̓Ενέργεια and Δύναμις », Ancient Philosophy, 14, p. 73-114. — 1995a, Plato on God as Noûs, Journal of the History of Philosophy Monograph Series, Southern Illinois University Press, réimprimé par St. Augustine’s Press, 2002. — 1995b, « Physics as a Virtue », Proceedings of the Boston Area Colloquium in Ancient Philosophy, 11, p. 1-34. — 2001, « Plotinus on the Identity of Knowledge with its Object », Apeiron, 34, p. 233-246. — 2002, « Aristotle’s Definition of Soul and the Programme of the de Anima », Oxford Studies in Ancient Philosophy, 22, p. 83-139.

358

BIBLIOGRAPHIE

— 2012, « Aristotle’s Theology », dans C. Shields (éd.), The Oxford Handbook of Aristotle, Oxford, Oxford University Press, p. 422-464. — 2019, « Physics I 1: The Path to the Principles », dans K. Ierodiakonou, P. Kalligas, V. Karasmanis (éds.), Aristotle’s Physics Alpha, Proceedings of the XXth Symposium Aristotelicum, Oxford, Oxford University Press, p. 19-52. MOREL, P.-M., 2006, « ‘Common to Soul and Body’ in the Parva Naturalia », dans R.A.H. King (éd.), Common to Body and Soul: Philosophical Approaches to Explaining Behavior in Greco-Roman Antiquity, Berlin, De Gruyter, p. 121-139. — 2007, De la matière à l’action. Aristote et le problème du vivant, Paris, Vrin. — 2016a, « La physiologie des passions dans le De motu animalium d’Aristote », dans I. Boehm, J.-L. Ferrary, S. Franchet d’Espèrey (éds.), L’Homme et ses Passions. Actes du XVIIe Congrès International de l’Association Guillaume Budé, organisé à Lyon du 26 au 29 août 2013, Paris, Les Belles Lettres, p. 291-301. — 2016b, « ‘La nature ne fait rien en vain’. Sur la causalité finale dans la Locomotion des animaux d’Aristote », Philosophie antique, 16, p. 9-30. OSBORNE, C., 1998, « Perceiving white and sweet (again): Aristotle, De Anima 3.7, 431a20-b1 », Classical Quarterly, 48, p. 433-446. OWENS, J., 1982, « Aristotle on Common Sensibles and Incidental Perception », Phoenix, 36, p. 215-236. PELLEGRIN, P., 1990, « De l’explication causale dans la biologie d’Aristote », Revue de Métaphysique et de Morale, 95, p. 197-219. PICKAVÉ, M., WHITING, J., 2008, « Nicomachean Ethics 7.3 on Akratic Ignorance », Oxford Studies in Ancient Philosophy, 34, 323-71. POLANSKY, R., 2007, Aristotle’s De anima. A Critical Commentary, Cambridge, Cambridge University Press. POLITIS, V., 2001, « Aristotle’s Account of the Intellect as Pure Capacity », Ancient Philosophy, 21, p. 375–402. RAPP, C., 2001, « Intentionalität und phantasia bei Aristoteles », dans D. Perler (éd.), Ancient and Medieval Theories of Intentionality, Leiden, Brill, p. 63-96. — 2020, « Introduction I: The Argument of De Motu Animalium  », dans C. Rapp, O. Primavesi (éds.), Aristotle’s De Motu Animalium, Proceedings of the Symposium Aristotelicum, Oxford, Oxford University Press, p. 1-66. RAPP, C., PRIMAVESI O. (éds.), 2020, Aristotle’s De Motu Animalium, Proceedings of the Symposium Aristotelicum, Oxford, Oxford University Press. RASHED, M., 2004, « Agrégat de parties ou vinculum substantiale  ? Sur une hésitation conceptuelle et textuelle du corpus aristotélicien », dans A. Laks, M. Rashed (éds.), Aristote et le mouvement des animaux  : Dix études sur le De motu animalium, Villeneuve d’Ascq, Presses Universitaires du Septentrion, p. 185-202. ROMEYER DHERBEY, G., VIANO, C. (dir.), 1996, Corps et Âme. Sur le De Anima d’Aristote, Paris, Vrin.

BIBLIOGRAPHIE

359

ROREITNER, R., 2018, The Unmoved Causes of Receptivity: Sense-Perception and Thinking as Passive Activities in Aristotle’s De Anima, Doctoral Dissertation, Humboldt-Universität Berlin. ROSE, V., 1854, De Aristotelis librorum ordine et auctoritate commentatio, Berlin, G. Reimer. ROSEN, J., 2014, « Essence and End in Aristotle », Oxford Studies in Ancient Philosophy, 46, p. 73-107. ROSS, W.D., 1923, 19595, Aristotle: A Complete Exposition of His Works and Thought, London, Methuen ; New York, Meridian Books (rééd.). SCHOFIELD, M., 1992, « Aristotle on the Imagination », dans M. Nussbaum, A. Rorty (éds.), Essays on Aristotle’s De Anima, Oxford, Clarendon Press, p. 249-278. SEDLEY, D., 1991, « Is Aristotle’s Teleology Anthropocentric? », Phronesis, 36 (2), p. 179-196. STEVENS, A., 2009 « L’apparition de la conscience dans le De Anima et d’autres œuvres d’Aristote », dans G. van Riel, P. Destrée (éds.), Ancient Perspectives on Aristotle’s De Anima, Leuven, Leuven University Press, p. 35-48. STRAWSON, P.F., 1970, « Imagination and Perception », dans L. Foster, J. W. Swanson (éds.), Experience and Theory, Amherst, University of Massachusetts Press, p. 31-54. TEICHMÜLLER, G., 1874, Studien zur Geschichte der Begriffe, Berlin, Weidmann. TRENTINI, S., 2016, « The internal coherence of De Anima III 6 », dans G. Sillitti, F. Stella, R. Fronterotta (éds.), Il Noûs di Aristotele, Sankt Augustin, Academia Verlag, p. 175-213. WEDIN, M., 1988, Mind and Imagination in Aristotle, New Haven, Yale University Press. WEIDEMANN, H., 2001, « War Aristoteles ein Repräsentationalist? », dans D. Perler (éd.), Ancient and Medieval Theories of Intentionality, Leiden, Brill, p. 97-104. WHITING, J., 2002, « Locomotive Soul: The Parts of Soul in Aristotle’s Scientific Works », Oxford Studies in Ancient Philosophy, 22, p. 141-200. — 2019, « Hylomorphic Virtue : cosmology, embryology, and moral development in Aristotle », Philosophical Explorations, 22 (2), p. 1-21. OUTILS BONITZ, H., 1870, Index Aristotelicus, Berlin, Georg Reimer. DENNISTON, J.D., 1934, 1954², The Greek Particles, Oxford, Clarendon Press. KÜHNER, R., GERTH, B., 1904, Ausführliche Grammatik der Griechischen Sprach, Darmstadt, Wissenschaftliche Buchgesellschaft. LIDDELL, H.G. & SCOTT, R. (rev. by H.J. Jones), 1996, Greek-English Lexicon, Oxford, Clarendon Press. MUGLER, C., 1958-1959, Dictionnaire historique de la terminologie géométrique des Grecs, Paris, Gauthier-Villars et C. Klincksieck, 2 vol. SMYTH, H.W., 1956, Greek Grammar, Cambridge, MA, Harvard University Press.

INDEX DES NOMS ANCIENS ET MÉDIÉVAUX ALEXANDRE D’APHRODISE : 7, n. 3 ; 79, n. 13 ; 85 ; 88, n. 45 ; 104, n. 13 ; 113 ; 114 ; 120 ; 121 ; 121, n. 30 ; 122, n. 31 ; 137 ; 148 ; 167, n. 27 ANAXAGORE : 15 ; 78 ; 83 ; 83, n. 28 ; 84 ; 85 ; 87 ; 93 ; 93, n. 57 ; 94 ; 96 ; 102 ; 104 ; 105 ; 106 ; 106, n. 16 ; 107 ; 132 DÉMOCRITE : 7 ; 47 ; 47, n. 8 PHILOPON : VIII, n. 1 ; 27, n. 9 ; 79 ; 158, n. 3 ; 161 ; 161, n. 15 ; 167 ; 167, n. 29 ; 173 ; 173, n. 42 ; 174 ; 174, n. 44 ; 179 ; 179, n. 54 et 56 ; 180 ; 180, n. 57 ; 190, n. 10 ; 242, n. 45 ; 332, n. 6 ; 345, n. 30 PLATON : 6 ; 18 ; 45 ; 47, n. 9 ; 57 ; 63 ; 63, n. 18 ; 71 ; 76, n. 7 ; 77 ; 82 ; 83, n. 28 ; 84 ; 87 ; 87, n. 43 ; 88 ; 92 ; 94 ; 95 ; 96 ; 102 ; 114 ; 115 ; 124 ; 126 ; 127 ; 128 ; 129 ; 129, n. 40 ; 130 ; 130, n. 41 ; 131 ; 135, n. 49 ; 140 ; 141 ; 144 ; 145 ; 151 ; 153 ; 154, n. 64 ; 174 ; 189, n. 8 ; 213 ; 214 ; 225 ; 227, n. 13 ; 240 ; 241 ; 241, n. 41 ; 243 ; 245, n. 53 ; 256 ; 256, n. 3 ; 267, n. 32 ; 322, n. 30

PLOTIN : 45 ; 130, n. 41 ; 145 ; 146 PROTAGORAS : 46 ; 76 SIMPLICIUS : VIII, n. 1 ; 10 ; 27, n. 9 ; 79 ; 80 ; 157 ; 158, n. 3 et 4 ; 171, n. 37 ; 174, n. 44 ; 178 ; 178, n. 52 ; 180 ; 180, n. 57 ; 190, n. 10 ; 242, n. 45 ; 320 ; 322 ; 322, n. 30 ; 332, n. 5 THÉMISTIUS : 27, n. 9 ; 78 ; 80 ; 104, n. 13 ; 127 ; 128, n. 36 ; 148 ; 149 ; 162 ; 162, n. 18 ; 165 ; 165, n. 22 ; 173 ; 173, n. 43 ; 174, n. 44 ; 175, n. 47 ; 180 ; 181, n. 59 ; 183, n. 65 ; 266, n. 31 ; 321, n. 30 THÉOPHRASTE : VIII, n. 4 ; 89 ; 102, n. 11 ; 128 THOMAS D’AQUIN : 48 ; 51 ; 74, n. 30 ; 104, n. 13 ; 114, n. 23 ; 122, n. 30 ; 127 ; 128, n. 36 ; 148 ; 149 ; 158, n. 3 ; 169, n. 33 ; 332, n. 7 ; 337, n. 15 ; 245, n. 30 THUCYDIDE : 208, n. 35 ZABARELLA : 114 ; 115 ; 116 ; 121, n. 30 ; 122 ; 122, n. 31 ; 123, n. 32 ; 149, n. 61

INDEX DES NOMS MODERNES ET CONTEMPORAINS

ALEXANDRU S. : 96, n. 3 ANAGNOSTOPOULOS A. : 196, n. 16 BALME D.M. : 117, n. 25 ; 340, n. 18 BARBOTIN E. : 41 ; 48 ; 75, n. 1 ; 78, n. 12 ; 159, n. 9 ; 160, n. 13 ; 166, n. 24 ; 173, n. 43 ; 183, n. 67 ; 187, n. 6 ; 257, n. 6 ; 315, n. 19 BARNES J. : 280, n. 5 BERRYMAN S. : 260, n. 20 BERTI E. : 101, n. 10 ; 157, n. 2 ; 161, n. 16 ; 167, n. 29 ; 168 ; 168, n. 30 ; 174 ; 174, n. 45 ; 175 ; 181 ; 181, n. 60 ; 182 ; 182, n. 64 BODÉÜS R. : 41 ; 48 ; 51 ; 78, n. 12 ; 80 ; 80, n. 17 ; 164, n. 21 ; 166, n. 24 et 26 ; 167, n. 29 ; 173, n. 43 ; 241, n. 41 ; 257 ; 257, n. 6 ; 261, n. 21 BONITZ H. : 135, n. 49 ; 182, n. 61 ; 198, n. 19 ; 200, n. 21 ; 343, n. 22 BRENTANO F. : 89, n. 50 ; 127 ; 128, n. 36 BRUNSCHWIG J. : 12 ; 12, n. 8 BURNYEAT M. : 84, n. 37 ; 186, n. 1 ; 214, n. 43 ; 332, n. 3 BYWATER I. : 107, n. 18 ; 171 ; 180 ; 315 ; 320 ; 320, n. 28 CAMPBELL J. : 19, n. 1 ; 31, n. 15 CARBONE A.L. : 250, n. 62 CASTON V. : 69, n. 27 ; 95 ; 95, n. 1 ; 99 ; 99, n. 7 ; 100 ; 101 ; 125 ; 126 ; 128 ; 135, n. 49 et 50 ; 138, n. 52 ; 141 ; 142 ; 142, n. 55 ; 143, n. 56 et 59 ; 144 ; 144, n. 59 ; 146 ; 153

CHARLES D. : 34, n. 19 ; 39, n. 1 ; 157, n. 1 ; 271, n. 37 ; 328, n. 39 ; 340, n. 18 ; 350, n. 43 COHOE C. : 77, n. 11 ; 84, n. 37 CORCILIUS K. : 38, n. 25 ; 59, n. 9 ; 63, n. 18 ; 95 ; 98, n. 6 ; 100, n. 9 ; 101, n. 10 ; 104, n. 12 ; 105, n. 15 ; 107, n. 18 ; 125, n. 33 et 35 ; 131, n. 42, 43 et 44 ; 133, n. 45 ; 141, n. 53 ; 157, n. 1 ; 189, n. 9 ; 201, n. 23 ; 225, n. 7 ; 240, n. 39 ; 258, n. 11 ; 259, n. 13 ; 314, n. 19 ; 315 ; 341, n. 19 CORNFORD F.M. : 320 CRUBELLIER M. : VIII, n. 2 ; 38, n. 25 ; 95 ; 95, n. 1 ; 157 ; 157, n. 1 ; 328, n. 39 ; 350, n. 43 DE KONINCK T. : 167, n. 29 DELCOMMINETTE S. : 167, n. 27 ; 174, n. 46 ; 182, n. 62 DENNISTON J.D. : 77 ; 77, n. 9 et 10 ; 197 ; 198, n. 18 et 19 DILLON J. : 84, n. 37 DOW J. : 59, n. 9 DÜRING I. : 340, n. 18 FÖRSTER A. : 98, n. 6 ; 100, n. 9 ; 105, n. 15 ; 107, n. 18 ; 125, n. 33 et 35 ; 131, n. 42, 43 et 44 ; 186 ; 189 ; 206, n. 32 ; 222, n. 2 ; 232 ; 22 ; 238, n. 34 FORSTER E.S. : 340, n. 18 FREDE D. : 59, n. 9 ; 74, n. 30 FREDE M. : 141 FREY C. : 341, n. 19 ; 343, n. 21 GELBER J. : 348, n. 37 ; 350, n. 43 GERSON L.P. : 177, n. 48

364

INDEX DES NOMS MODERNES ET CONTEMPORAINS

GREGORIC P. : 37, n. 24 ; 201, n. 23 ; 301, n. 52 ; 341, n. 19 HAMLYN D.W. : 59, n. 9 ; 78, n. 12 ; 159 ; 159, n. 7 ; 164, n. 21 ; 166, n. 24 et 25 ; 170, n. 33 ; 171 ; 173, n. 42 ; 256, n. 3 ; 257, n. 7 ; 332, n. 4 HARVEY J. : 161, n. 14 HAYDUCK M. : 232, n. 22 HENRY D. : 264, n. 28 ; 334, n. 10 HETT W.S. : 240, n. 40 ; 261, n. 21 HICKS R.D. : 28, n. 9 ; 30, n. 13 ; 31, n. 16 ; 34, n. 20 ; 35, n. 21 ; 79 ; 80, n. 17 ; 84 ; 84, n. 36 ; 105, n. 15 ; 112 ; 121 ; 149 ; 150 ; 187, n. 6 ; 240, n. 40 ; 241, n. 41 ; 253 ; 315, n. 21 ; 320 ; 321, n. 30 ; 347, n. 36 HUBY P. : IX, n. 4 ; 102, n. 11 HUTCHINSON D.S. : 322, n. 30 ; 329, n. 2 ; 332, n. 3 ; 337, n. 14 JAEGER W. : 293 ; 293, n. 26 JANNONE A. : 75, n. 1 ; 159 ; 159, n. 9 ; 160, n. 12 et 13 ; 166, n. 25 ; 171 ; 173, n. 43 ; 180 ; 261, n. 21 JOHANSEN T.K. : 37, n. 24 ; 55, n. 3 ; 59, n. 9 ; 64, n. 20 ; 67, n. 24 ; 256, n. 3 ; 265, n. 29 ; 341, n. 19 ; 343, n. 21 et 22 ; 344, n. 24 et 25 ; 346, n. 31 ; 350, n. 41 JOHNSON M.R. : 344, n. 24 et 25 ; 350, n. 41 KULLMANN W. : 340, n. 18 ; 344, n. 24 ; 345, n. 26 LABARRIÈRE J.-L. : 38, n. 25 ; 337, n. 17 ; 340, n. 18 LAKS A. : 95, n. 1 ; 290, n. 22 ; 293, n. 29 LEAR G.R. : 210, n. 37 ; 219, n. 52 LEFEBVRE D. : 16 LENNOX J.G. : 334, n. 10 ; 340, n. 18 ; 350, n. 43 ; 117, n. 25 et 26 LEUNISSEN M. : 332, n. 7 ; 334, n. 10 ; 335 ; 335, n. 12 ; 336 ; 337, n. 14 et 17 ; 338 ; 344, n. 24 ; 350, n. 40 et 41

LEWIS F.A. : 84, n. 36 LLOYD G.E.R. : 117, n. 25 LORENZ H. : 69, n. 26 LOUGUET C. : 85, n. 39 LOUIS P. : 340, n. 18 LOWE M. : 77, n. 8 ; 91, n. 56 MADVIG J.N. : 3, n. 1 ; 56, n. 5 MARMODORO A. : 20 ; 20, n. 2 MATTHEWS G.B. : 341, n. 20 MCDOWELL J. : 325 ; 325, n. 33 MENN S. : 95, n. 2 ; 96, n. 4 ; 97, n. 5 ; 118, n. 26 ; 141, n. 54 ; 143, n. 57 ; 145, n. 60 ; 154, n. 64 ; 192, n. 13 ; 196, n. 17 ; 207, n. 33 ; 219, n. 52 ; 340, n. 18 ; 344, n. 24 ; 350, n. 41 MOREL P.-M. : 16 ; 258, n. 10 ; 264, n. 28 ; 265, n. 29 ; 268, n. 33 ; 340, n. 18 MORISON B. : 295, n. 34 MOVIA G. : 79 ; 80, n. 17 MUGLER C. : 243, n. 48 MURGIER C. : 157, n. 1 ; 168, n. 31 ; 271, n. 37 NUSSBAUM M.C. : 89, n. 50 ; 285, n. 12 ; 293, n. 28 OGLE W. : 340, n. 18 OSBORNE C. : 186 ; 186, n. 4 ; 187 ; 188 OWENS J. : 28, n. 9 PECK A.L. : 340, n. 18 PELLEGRIN P. : 250, n. 61 ; 263, n. 27 PICKAVÉ M. : 326, n. 34 ; 327, n. 36 POLANSKY R. : 57, n. 6 ; 75 ; 75, n. 4 ; 79 ; 79, n. 13 ; 83 ; 83, n. 31 ; 91, n. 55 ; 158 ; 158, n. 5 ; 167, n. 28 ; 170, n. 34 ; 171 ; 172, n. 38 ; 173, n. 42 ; 181, n. 60 ; 186 ; 186, n. 5 ; 187 ; 188 ; 240, n. 40 ; 257, n. 8 ; 259, n. 12 ; 260, n. 16 ; 261, n. 21 ; 269, n. 36 ; 271, n. 37 ; 281, n. 9 ; 332, n. 3 et 7 ; 343, n. 21 et 22 ; 350, n. 41 POLITIS V. : 84, n. 37 PRIMAVESI O. : 284, n. 11 ; 295, n. 37 RAPP C. : 59, n. 9 ; 284, n. 11 ; 293, n. 28 ; 295, n. 34 ; 297, n. 43

INDEX DES NOMS MODERNES ET CONTEMPORAINS

RASHED M. : 290, n. 22 RODIER G. : 2 ; 3, n. 1 ; 41 ; 79 ; 80 ; 80, n. 16 ; 84 ; 84, n. 33 ; 91 ; 91, n. 54 ; 158, n. 3 ; 162 ; 162, n. 17 ; 169 ; 169, n. 32 ; 172 ; 172, n. 39 et 41 ; 173, n. 43 ; 175 ; 175, n. 47 ; 181, n. 59 ; 186, n. 1 ; 196, n. 15 ; 241, n. 41 ; 257, n. 5 ; 263, n. 27 ; 266, n. 31 ; 268, n. 35 ROREITNER R. : 95 ; 101, n. 10 ; 105, n. 15 ; 152, n. 63 RORTY A.O. : 89, n. 50 ROSE V. : 293 ; 293, n. 27 ROSEN J. : 348, n. 37 ROSS W.D. : 30, n. 13 ; 31, n. 16 ; 41 ; 44, n. 5 ; 45, n. 6 ; 48 ; 49 ; 50 ; 51 ; 69, n. 27 ; 75, n. 1 ; 84, n. 37 ; 98, n. 6 ; 100, n. 9 ; 104 ; 104, n. 12 ; 105, n. 15 ; 107, n. 18 ; 111, n. 20 ; 112 ; 112, n. 21 ; 120 ; 121 ; 121, n. 29 ; 125, n. 33 et 35 ; 127 ; 128 ; 128, n. 36, 37, 38 et 39 ; 131, n. 42, 43 et 44 ; 132 ; 133 ; 133, n. 45 et 46 ; 139 ; 149 ; 150 ; 159 ; 159, n. 8 ; 160, n. 12 ; 165, n. 23 ; 166, n. 24 et 25 ; 171 ; 171, n. 36 ; 172 ; 172, n. 40 ; 173, n. 41 et 42 ; 178, n. 49 et 51 ; 179 ; 179, n. 55 ; 180 ; 182, n. 62 et 63 ; 183, n. 66 et 68 ; 186 ; 187 ; 193, n. 14 ; 196, n. 17 ; 211, n. 38 ; 232, n. 22 ; 235, n. 29 ; 261 ; 261, n. 22 ; 276, n. 1 ; 280, n. 5 SCHOFIELD M. : 59, n. 9 ; 62, n. 15 ; 66, n. 22 SEDLEY D. : 344, n. 24 SEIDL H. : 79 ; 80, n. 17 ; 240, n. 39 ; 314, n. 19 ; 320, n. 28

365

SHIELDS C. : 27, n. 9 ; 53, n. 1 ; 56, n. 5 ; 59, n. 9 et 11 ; 60, n. 13 ; 120, n. 28 ; 159, n. 6 ; 171 ; 173, n. 42 ; 186 ; 187 ; 189 ; 211, n. 38 ; 212, n. 41 et 42 ; 215, n. 45 ; 216, n. 47 ; 256, n. 3 ; 258, n. 11 ; 259, n. 12 ; 277, n. 2 ; 280 ; 280, n. 6 et 7 ; 281 ; 281, n. 8 ; 282 ; 288, n. 19 ; 191, n. 24 ; 298 ; 298, n. 45 ; 299 ; 314, n. 19 ; 315 ; 332, n. 7 ; 345, n. 30 SIWEK P. : 79 ; 160, n. 13 ; 169, n. 32 ; 172, n. 41 ; 186 ; 276, n. 1 ; 334, n. 9 SMYTH H.W. : 344, n. 24 STRAWSON P.F. : 55, n. 4 TEICHMÜLLER G. : 84 ; 84, n. 34 THEILER W. : 186 ; 314, n. 19 THILLET P. : 41, n. 3 ; 48 ; 52 ; 78, n. 12 ; 160, n. 13 ; 166, n. 24 ; 173, n. 43 ; 180 ; 180, n. 58 ; 182, n. 63 ; 183, n. 67 ; 241, n. 41 TORSTRIK A. : 27, n. 9 ; 61, n. 14 ; 160, n. 12 ; 165, n. 23 ; 172, n. 41 ; 182, n. 63 ; 183, n. 66 ; 185 ; 185, n. 1 ; 186 ; 186, n. 3 ; 187 ; 187, n. 6 ; 197 ; 198, n. 19 ; 276, n. 1 TRENDELENBURG F.A. : 48 ; 186, n. 1 ; 319 ; 321 TRENTINI S. : 157, n. 2 TRICOT J. : 41 ; 48 ; 173, n. 43 ; 183, n. 67 ; 240, n. 40 WALLACE E. : 11 WEIDEMANN H. : 72, n. 29 WHITING J. : 38, n. 25 ; 311, n. 14 ; 325, n. 33 ; 326, n. 34 ; 327, n. 36 ; 341, n. 19 ; 348, n. 38 ; 350, n. 43

INDEX DES LIEUX

Ne sont indiqués que les passages mentionnés en corps de texte. ALEXANDRE D’APHRODISE De anima (Bruns) 84, 15 : 85 87, 4-23 : 148 87, 24-28 : 113 87, 24-88, 16 : 114 88, 2-3 : 114 ANAXAGORE (D.-K.) 59B12 : 83 ARISTOTE Seconds Analytiques (A.Po.) I 13, 79a7-10 : 215 I 18, 81a38-39 : 1 I 31, 87b37-39 : 25 I 31, 87b40 : 85 II 1, 89b31-35 : 161 II 2, 90a15 : 85 II 19, 99b15-100b5 : 174 II 19, 99b25-26 : 86 II 19, 99b26-27 : 99 II 19, 99b35 : 86 II 19, 100a6-8 : 248 II 19, 100b5-17 : 174 De Anima (DA) I 1, 402a7-8 : 329 I 1, 402b14-16 : 94 I 1, 403a7-10 : 68 I 1, 403a8-10 : 80, 150, 203 I 1, 403b14-15 : 215 I 2, 404b1-3 : 83

I 2, 405a16-17 : 83 I 2, 405b19-21 : 83 I 4, 408b14-18 : 90 I 4, 408b18-30 : 89, 138 I 4, 408b25-30 : 138 I 5, 410a23-26 : 81 II 1, 412b9-413a10 : 339 II 2, 413a11-20 : 341 II 2, 413b11-16 : 79 II 2, 413b13 : 80 II 3, 414b1-3 : 104 II 4, 415a14-22 : 158 II 4, 415a20 : 38 II 4, 415a23-b7 : 342 II 5, 416b33 : 38 II 5, 417a4 : 9 II 5, 417a4-5 : 9 II 5, 417a6 : 9 II 5, 417a15-16 : 81 II 5, 417a18-21 : 81 II 5, 417a22-24 : 81 II 5, 417a26-29 : 86 II 5, 417a30-b2 : 86 II 5, 417b2-16 : 80 II 5, 417b2-27 : 107 II 5, 417b6-7 : 193, 194 II 5, 417b7 : 194 II 5, 417b16 : 86 II 5, 417b16-19 : 86 II 5, 417b18 : 81 II 5, 417b18-20 : 80 II 5, 417b20-26 : 45 II 5, 417b22-28 : 25

368 II 5, 417b29-418a3 : 81 II 6, 418a11 : 22 II 6, 418a15 : 23 II 6, 418a15-16 : 24 II 6, 418a17-18 : 228 II 6, 418a18-20 : 22 II 6, 418a19-20 : 23, 25 II 6, 418a24-25 : 21, 23 II 7, 418a26-31 : 24 II 7, 418a31 : 7 II 7, 418b18-24 : 107 II 7, 419a15 : 7 II 8, 419b18 : 6 II 10, 422a8 : 24 II 10, 422a20-22 : 43 II 10, 422a31-34 : 24 II 10, 422b3 : 192 II 11, 422b23 : 16 II 11, 423b1-4 : 8 II 11, 423b17-19 : 8 II 11, 424a1-6 : 89 II 12, 424a17-20 : 104 II 12, 424a17-25 : 70 II 12, 424a17-28 : 82 II 12, 424a17-b18 : 43 II 12, 424a18-19 : 238, 251 II 12, 424a21-24 : 34 II 12, 424a25-b18 : 50 III 1, 424b22-23 : 11 III 1, 424b24 : 2 III 1, 424b24-27 : 2 III 1, 424b25 : 3 III 1, 424b26 : 1, 3 III 1, 424b29-30 : 6, 8, 11 III 1, 424b31-34 : 4 III 1, 424b34-425a3 : 5 III 1, 425a4 : 5 III 1, 425a5 : 8 III 1, 425a5-6 : 8 III 1, 425a13 : 32 III 1, 425a13-20 : 27 III 1, 425a15-19 : 228 III 1, 425a15-21 : 35 III 1, 425a16-17 : 34 III 1, 425a17 : 28 III 1, 425a19 : 27, 28

INDEX DES LIEUX

III 1, 425a20-30 : 29 III 1, 425a30-b2 : 30 III 1, 425a30-b3 : 33 III 1, 425a31-b2 : 31 III 1, 425b2-3 : 31 III 1, 425b2-4 : 33 III 1, 425b3-4 : 32, 33 III 2, 425b11-15 : 94 III 2, 425b12-17 : 40 III 2, 425b13-14 : 41 III 2, 425b17-25 : 42 III 2, 425b18-19 : 24 III 2, 425b21-22 : 43 III 2, 425b22-25 : 43 III 2, 425b24 : 70 III 2, 425b26-426a19 : 65 III 2, 425b26-426a26 : 45, 232 III 2, 426a10 : 65 III 2, 426a15-25 : 233 III 2, 426a27-30 : 48 III 2, 426a30-b12 : 49 III 2, 426b3-6 : 48 III 2, 426b7 : 49, 50 III 2, 426b7-8 : 89 III 2, 426b12-23 : 51 III 2, 426b17 : 33 III 2, 426b22 : 51 III 2, 426b23-427a16 : 52 III 2, 426b30 : 35 III 2, 426b31-427a1 : 192 III 2, 427a1 : 51 III 3, 427a17-19 : VII, 75 III 3, 427a17-b14 : 55 III 3, 427a17-b16 : 53 III 3, 427a18-20 : 75 III 3, 427a19-20 : 76, 80 III 3, 427a19-21 : 224 III 3, 427a19-22 : 76 III 3, 427a20-21 : 79 III 3, 427a21-22 : 76 III 3, 427b6-7 : 76 III 3, 427b8-11 : 58 III 3, 427b14-16 : 53, 54 III 3, 427b14-26 : 55-58 III 3, 427b16 : 68, 70 III 3, 427b17-24 : 56

INDEX DES LIEUX

III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III III

3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,

427b24-26 : 58 427b26-429a9 : 54 427b27-428b9 : 54 428a1-4 : 251 428a1-5 : 58 428a1-b9 : 58-64 428a1-429a2 : 258 428a5-13 : 60 428a5-16 : 62 428a13 : 61 428a16-18 : 62 428a18-24 : 62 428a24-b9 : 63 428a25-26 : 258 428b2-4 : 252, 314 428b3 : 61 428b10-17 : 64 428b10-429a9 : 54, 64-66 428b17-30 : 65 428b18-25 : 61 429a4 : 66, 69 429a4-9 : 66 429a10-11 : 90, 98 429a10-13 : 77, 78-80 429a10-29 : 80-88 429a10-b22 : 103 429a13 : 38 429a13-14 : 76 429a13-15 : 106 429a13-18 : 104 429a13-29 : 77 429a14 : 81 429a14-15 : 77, 98 429a15-16 : 106, 147 429a15-18 : 25, 82 429a15-27 : 227 429a16 : 82 429a18-22 : 104 429a20 : 84 429a21 : 78 429a21-22 : 100 429a21-24 : 102, 108 429a22 : 87, 128 429a22-24 : 94, 100, 126 429a24 : 85 429a24-25 : 86

369

III 4, 429a27 : 225 III 4, 429a27-28 : 87 III 4, 429a27-29 : 85 III 4, 429a28-29 : 87 III 4, 429a29 : 76, 88 III 4, 429a29-30 : 82, 107 III 4, 429a29-b5 : 88, 107, 193 III 4, 429a29-b22 : 78, 88-93 III 4, 429b3 : 90 III 4, 429b4-5 : 107, 146, 150, 152 III 4, 429b5-9 : 107, 236 III 4, 429b5-10 : 89 III 4, 429b7 : 90 III 4, 429b8-9 : 90 III 4, 429b9 : 94 III 4, 429b9-10 : 94 III 4, 429b10 : 91 III 4, 429b10-18 : 92 III 4, 429b10-22 : 89, 94, 107, 147, 152, 216 III 4, 429b12-13 : 91 III 4, 429b13 : 91 III 4, 429b15 : 91 III 4, 429b16 : 91 III 4, 429b16-17 : 91 III 4, 429b17 : 91 III 4, 429b18-20 : 92, 176 III 4, 429b19-20 : 92 III 4, 429b20-21 : 177 III 4, 429b21 : 91 III 4, 429b21-22 : 91, 93, 153, 177 III 4, 429b22 : 91, 103, 124 III 4, 429b22-23 : 106 III 4, 429b22-24 : 93 III 4, 429b22-25 : 108 III 4, 49b22-26 : 93 III 4, 429b22-29 : 107, 124 III 4, 429b22-430a9 : 78, 83, 93-94, 103, 141 III 4, 429b22-430a25 : 103 III 4, 429b25-26 : 108 III 4, 429b26 : 93 III 4, 429b26-29 : 94, 108, 114 III 4, 429b26-430a9 : 225 III 4, 429b27-29 : 112 III 4, 429b29-430a2 : 111

370

INDEX DES LIEUX

III 4, 429b29-430a9 : 124 III 4, 429b30-31 : 108 III 4, 429b31-430a2 : 108 III 4, 430a1-2 : 106 III 4, 430a2 : 89 III 4, 430a2-9 : 94, 111, 112, 114 III 4, 430a3 : 112, 128 III 4, 430a3-4 : 115, 167 III 4, 430a3-5 : 113, 116, 194 III 4, 430a3-6 : 114, 115 III 4, 430a4-5 : 115 III 4, 430a6 : 112, 128 III 4, 430a6-7 : 113, 121 III 4, 430a6-9 : 114, 116, 119 III 4, 430a7 : 113 III 4, 430a7-8 : 113 III 5, 430a10-11 : 128 III 5, 430a10-14 : 110 III 5, 430a10-17 : 125 III 5, 430a13-14 : 98 III 5, 430a14-15 : 99 III 5, 430a15-17 : 249 III 5, 430a16-17 : 153 III 5, 430a17-18 : 98 III 5, 430a17-25 : 132 III 5, 430a18 : 98, 99 III 5, 430a19 : 162 III 5, 430a19-20 : 98 III 5, 430a19-21 : 194, 226 III 5, 430a19-22 : 132, 133 III 5, 430a20-21 : 191 III 5, 430a22 : 98, 99 III 5, 430a23 : 98 III 5, 430a23-25 : 99, 137, 138, 140 III 5, 430a24 : 88, 98 III 5, 430a24-25 : 126 III 5, 430a25 : 103, 124 III 6, 430a26-b6 : 160-165 III 6, 430a27 : 160 III 6, 430a27-b6 : 252 III 6, 430a28 : 159 III 6, 430a30 : 163 III 6, 430b3-4 : 164 III 6, 430b6 : 162 III 6, 430b6-14 : 165-171 III 6, 430b10-11 : 169

III 6, 430b11-13 : 169 III 6, 430b13-14 : 169 III 6, 430b14 : 166 III 6, 430b14-20 : 171-177 III 6, 430b17-18 : 176 III 6, 430b18 : 176 III 6, 430b20-26 : 177-183 III 6, 430b23 : 179 III 6, 430b23-24 : 180 III 6, 430b24-26 : 101, 181 III 6, 430b26-31 : 161, 183-184 III 6, 430b27-31 : 161 III 6, 430b28 : 170, 172 III 6, 430b30-31 : 170 III 7, 431a1-3 : 132, 133, 226 III 7, 431a1-4 : 187, 189, 195 III 7, 431a1-8 : 198 III 7, 431a4 : 186, 197, 198 III 7, 431a4-5 : 187 III 7, 431a4-6 : 187 III 7, 431a4-7 : 81, 187, 190, 195 III 7, 431a6 : 192 III 7, 431a7-14 : 187 III 7, 431a8 : 197, 198 III 7, 431a8-14 : 187, 195, 199, 250 III 7, 431a8-16 : 187 III 7, 431a8-17 : 187 III 7, 431a8-b1 : 187 III 7, 431a9-10 : 202 III 7, 431a11 : 205 III 7, 431a14 : 69 III 7, 431a14-16 : 187, 195, 201 III 7, 431a14-17 : 187 III 7, 431a14-20 : 187 III 7, 431a15-16 : 266 III 7, 431a16-17 : 151, 187, 195, 203, 250 III 7, 431a17-20 : 33, 186, 187 III 7, 431a17-b1 : 187, 195, 204-207 III 7, 431a19 : 205 III 7, 431a20-b1 : 187 III 7, 431a22 : 205 III 7, 431a22-23 : 33 III 7, 431a29-b2 : 33 III 7, 431b2 : 151 III 7, 431b2-5 : 266

INDEX DES LIEUX

III 7, 431b2-10 : 69, 187, 195, 207212 III 7, 431b2-12 : 187 III 7, 431b2-16 : 187 III 7, 431b2-17 : 187 III 7, 431b3 : 210 III 7, 431b7 : 211 III 7, 431b8 : 213 III 7, 431b8-9 : 210 III 7, 431b10 : 211 III 7, 431b10-12 : 187, 195, 212 III 7, 431b12-19 : 187, 195, 214 III 7, 431b16 : 221, 225 III 7, 431b16-19 : 187 III 7, 431b16 – III 8, 432a14 : 221223 III 7, 431b17 : 227 III 7, 431b17-19 : 80, 187, 225 III 7, 431b18-19 : 228 III 7, 431b19 : 80 III 8, 431b21 : 223, 230 III 8, 431b21-23 : 230 III 8, 431b22-23 : 225 III 8, 431b24-26 : 231 III 8, 431b24-28 : 236 III 8, 431b27 : 233 III 8, 431b28-29 : 233 III 8, 431b28-432a3 : 236, 247 III 8, 432a2-3 : 247 III 8, 432a3 : 223, 230, 238 III 8, 432a3-4 : 240, 242 III 8, 432a3-5 : 245 III 8, 432a3-9 : 151 III 8, 432a3-10 : 69 III 8, 432a4 : 240 III 8, 432a4-5 : 223, 240, 242 III 8, 432a6 : 245 III 8, 432a7-10 : 246 III 8, 432a7-14 : 249 III 8, 432a8-9 : 250 III 8, 432a9-10 : 251 III 8, 432a12-13 : 203 III 8, 432a12-14 : 151, 253 III 9, 432a15-17 : 255 III 9, 432a15-18 : 223 III 9, 432a15-19 : 75

371

III 9, 432a17-22 : 255 III 9, 432a20 : 79 III 9, 432a22-23 : 256 III 9, 432a23-b7 : 256 III 9, 432b7 : 257 III 9, 432b7-14 : 259-260 III 9, 432b7-433a8 : 255, 257 III 9, 432b13-14 : 260 III 9, 432b14-15 : 261 III 9, 432b14-16 : 308 III 9, 432b14-19 : 261-262 III 9, 432b16 : 261, 262 III 9, 432b17-18 : 262 III 9, 432b19-26 : 262-265 III 9, 432b21-22 : 263 III 9, 432b24-25 : 347 III 9, 432b26-433a6 : 264-269 III 9, 432b27-29 : 266 III 9, 432b27-433a1 : 309 III 9, 432b31-433a1 : 267 III 9, 433a1-2 : 268 III 9, 433a1-3 : 268 III 9, 433a1-6 : 309 III 9, 433a4-6 : 269 III 9, 433a6-8 : 269-271, 310 III 10, 433a9 : 261, 270 III 10, 433a9-12 : 310 III 10, 433a13-22 : 312 III 10, 433a14-16 : 210 III 10, 433a23-25 : 270 III 10, 433b1 : 256 III 10, 433b10-13 : 210 III 10, 433b13-18 : 201, 276-284 III 10, 433b19 : 282, 283, 292 III 10, 433b19-21 : 279, 288-291 III 10, 433b19-22 : 267 III 10, 433b21-27 : 291-297, 299 III 10, 433b23-25 : 79 III 10, 433b27-30 : 303 III 11, 433b31-434a6 : VIII, 304 III 11, 434a4 : 61 III 11, 434a6-10 : 304 III 11, 434a10-12 : 306 III 11, 434a13-16 : 306 III 11, 434a16-21 : 307 III 12, 434a22-26 : 347

372 III III III III III III III III III III III III III III III III III III III III

INDEX DES LIEUX

12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 13, 13, 13, 13, 13,

434a22-b8 : 330 434a22-b25 : 330-331 434a27-b8 : 347 434a30-b8 : 333 434b3 : 348 434b8-11 : 349 434b8-27 : 330 434b14-18 : 334 434b19-22 : 334 434b24-26 : 336 434b24-27 : 335 434b26-435a5 : 191 434b27-435a10 : 331 435a8-10 : 192 435a10 : 330 435a11-b3 : 331 435a21-b4 : 349 435b4-25 : 331 435b12-13 : 336 435b19-25 : 336

Categoriae (Cat.) 12, 14b21-22 : 233 De Caelo (Cael.) II 13, 293b25 : 85 Ethica Nicomachea (EN) I 6, 1097b34 : 258 III 5, 1112b15-20 : 210 III 9, 1115a14-24 : 59 VII 7, 1149a23 : 59 VII 8, 1150b21 : 323 De Generatione Animalium (GA) I 23, 731a24-b9 : 338 II 1, 735a18-19 : 347 V 2, 781b17-22 : 16 De Generatione et Corruptione (GC) I 7, 324a24-b6 : 110 II 2, 329b34 : 10 Historia Animalium (HA) IV 8, 534a13 : 5 IV 9, 535b26 : 6

De Incessu Animalium (IA) 2, 704b12-18 : 264 4, 706a19 : 17 De Interpretatione (Int.) 1, 16a9-18 : 161 1, 16a16 : 161 3, 16b23-25 : 163 3, 16b6-9 : 165 3, 16b16-18 : 165 12, 21b9-10 : 165 Metaphysica (Met.) A 8, 989b15-16 : 83 A 8, 989b16-21 : 85 B 2, 997b5-12 : 155 Γ 2, 1004b27 : 179 Γ 5, 1010b30-1011a2 : 46 Γ 6, 1011b18-20 : 179 Δ 6, 1016b1-3 : 172, 173 Δ 6, 1016b24-36 : 178 Δ 6, 1016b25-26 : 178 Δ 6, 1016b30-31 : 178 Δ 7, 1017a27-30 : 165 Δ 15, 1021a29-b3 : 46 Δ 23, 1023a23-25 : 118 Δ 24, 1023a29 : 193 E 1, 1025b21 : 117 E 1, 1025b24 : 117 E 1, 1025b25-28 : 117 E 1, 1026a4-6 : 118 Z 6, 1031a15-16 : 92 Z 6, 1031b18-20 : 93 Ζ 7, 1032b2-5 : 180 Z 8, 1033b19-26 : 93 Z 11, 1036b7-20 : 92 Z 11, 1037a1-5 : 93 Z 11, 1037a21-b5 : 93 Ζ 17, 1041b11-33 : 176 H 1, 1042a26-31 : 79 Θ 2, 1046b8-9 : 180 Θ 9, 1051a17-19 : 181 Θ 10, 1051b9-17 : 163 Θ 10, 1051b11-13 : 163, 164 Θ 10, 1051b17-1052a4 : 161 Θ 10, 1051b20-21 : 163

INDEX DES LIEUX

Θ 10, 1051b26 : 161, 172 Θ 10, 1051b27 : 172 Θ 10, 1051b29 : 172 Θ 10, 1051b32 : 161 Ι 1, 1052a29-31 : 173, 175 Ι 1, 1052a29-34 : 172 Ι 2, 1053b30-31 : 179 Ι 3, 1054a26-27 : 178 Ι 4, 1055a33 : 178 Ι 4, 1055b17-29 : 179 I 6, 1056b34 : 46 Κ 3, 1061a19-20 : 179 K 6, 1063b17-19 : 179 Λ 5, 1071a8-11 : 181 Λ 6, 1071b20 : 96 Λ 7, 1072a26 : 97 Λ 7, 1072a29-30 : 143 Λ 7, 1072b1-3 : 143 Λ 7, 1072b22-24 : 97 Λ 9, 1074b23-27 : 183 Λ 9, 1074b38-1075a5 : 94, 97 Λ 9, 1075a3-5 : 97 Λ 9, 1075a5-7 : 146 Λ 10, 1075a11-15 : 130 Λ 10, 1075b21-24 : 182 M 10, 1087a15-21 : 239 De Motu Animalium (DMA) 1, 698a8-11 : 284 4, 700a18-20 : 295 6, 700b17-19 : 210 6, 700b23-25 : 285 6, 700b35-701a2 : 285 7, 701b16-32 : 71 7, 701b33-702a7 : 210 8, 702a21 : 301 8, 702a24-27 : 295 10, 703a4-6 : 286 10, 703a12-14 : 296 10, 703a28-29 : 287 11, 703b5-8 : 267 11, 703b21 : 267 11, 704a3-b1 : 290 De Partibus Animalium (PA) : I 1, 639b30-640a9 : 116

373

I 1, 641a32-b4 : 87 I 5, 645b14-20 : 339 I 5, 645b29-33 : 341 I 5, 645b33-35 : 340 II 1, 647a16 : 16 II 2, 648a11-16 : 14 IV 10, 687a19-21 : 237 IV 10, 687b3-5 : 237 Physica (Phys.) I 7, 191a6-7 : 179 II 2, 194a14-15 : 118 III 3, 202a21-202b5 : 65 III 7, 207b35-208a1 : 85 IV 11, 218b21-219a1 : 176 IV 11, 220a19 : 174 IV 11, 220a21 : 174 VIII 5, 256b14-20 : 280 Parva Naturalia (PN) De Memoria (Mem.) 449b28-30 : 165 450a19 : 165 450a21-23 : 70 450a29-451a18 : 72 450a31-b6 : 71 450b11-451a18 : 72 450b24-27 : 72 451a8-12 : 72 451a16-17 : 165 De Sensu (Sens.) 436b12-437a3 : 16, 337 437a2-3 : 16 438b30-31 : 9 439a23 : 7 439a27 : 7 De Somno et Vigilentia (Somno) 455a16-26 : 50 De Insomniis (Insomn.) 459b1-6 : 71 461a24-30 : 71 461b23 : 71 461b25-462a15 : 71

374 PSEUDO-ARISTOTE Problemata XXIII 9, 932b22-24 : 105 DÉMOCRITE (D.-K.) 68B125 : 47 EMPÉDOCLE (D.-K.) 31B17.21 : 162 PHILOPON In De Anima (Hayduck) 520, 25 : 79 PLATON Lois X, 897b1-4 : 128 République (Rep.) VI, 508a5-6 : 129 VI, 508b13-c2 : 129 X, 606c : 57 Sophiste 235b-239e : 57 260d : 57 Théétète 197c : 90

INDEX DES LIEUX

Timée 48e-50c : 82 50c2 : 106 50d : 84 50d5-e5 : 84 50d7-e4 : 105 50e8-51a1 : 106 SIMPLICIUS In De Anima (Hayduck) 118, 26 : 10 222, 5 : 79 THÉMISTIUS In De Anima (Heinze) 96, 25 : 148 172, 23 : 79 THÉOPHRASTE 307B (FHS&G) : 128 THOMAS D’AQUIN In De Anima : 713 : 148 ZABARELLA In De Anima 55v, col. 1-2 : 116 56r, col. 2 : 115 57r, col. 1 : 123

PRINTED ON PERMANENT PAPER

• IMPRIME

SUR PAPIER PERMANENT

N.V. PEETERS S.A., WAROTSTRAAT

• GEDRUKT

OP DUURZAAM PAPIER

50, B-3020 HERENT

- ISO 9706