619 29 5MB
English Pages 358 [351] Year 2017
Classification from Antiquity to Modern Times
Classification from Antiquity to Modern Times Sources, Methods, and Theories from an Interdisciplinary Perspective Edited by Tanja Pommerening and Walter Bisang
ISBN 978-3-11-053612-6 e-ISBN (PDF) 978-3-11-053877-9 e-ISBN (EPUB) 978-3-11-053727-7 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar. © 2017 Walter de Gruyter GmbH, Berlin/Boston Einbandabbildung: Relief mit Kranichen und anderen Vögeln, Replikat der Darstellung aus dem Grab des Ptahhotep, Saqqara (um 2400 v. Chr.) (Foto: ADD der Universitätsbibliothek Mainz) Druck und Bindung: CPI books GmbH, Leck ♾ Gedruckt auf säurefreiem Papier Printed in Germany www.degruyter.com
Preface The present volume is based on a workshop held in 2014 within the framework of the Research Training Group Early Concepts of Man and Nature: Universal, Local, Borrowed, a Research Training Group installed by the German Research Foundation at the University of Mainz since 2013. The Mainz Research Training Group aims at systematically recording concepts of natural objects in a selection of topics from the domain of man and nature and at investigating their (trans)cultural changes in space and time from a wide range of perspectives that also include the existence of universal concepts, i.e. concepts that may arise independently under temporally and spatially similar conditions. The doctoral students of the Research Training Group and their supervisors jointly discussed these issues with national and international guests on November 21 and 22, 2014 under the title From Classification to Concepts: Interdisciplinary Heuristics on the Conceptualization of Flora, Fauna, Man and Landscape. They exchanged ideas on how to recover and reconstruct concepts from historical sources with researchers from Linguistics, Anthropology, History of Science, Biology, Sociology, Egyptology, Greek Studies and Computational Linguistics. In the view of the editors of this volume and the organizers of the workshop, ways of classification and categorization as they were transmitted through time provide important keys for resolving these questions. The extensive discussion of pre-defined key questions during the presentations clearly showed the innovative potential that lies in an interdisciplinary synopsis of different approaches from different disciplines. As a consequence, the present volume seeks to address a wider public with the aim of stimulating further research on classification and conceptualisation that leads to the reflection and maybe rethinking of traditional views and habits by crossing traditional disciplinary boundaries. Where and how does classification take place and what determines classification systems? To what extent are there universal patterns and where do we find specific patterns and patterns that are not universal but stable enough to be transmitted through space and time? Such questions need an integrative basis that looks at how conceptualization operates at the level of the individual and how it gets conventionalized at the level of social groups and societies as a whole. Moreover, classification manifests itself in language, writing systems, pictures and frames. This leads to the question of differences and similarities of classification in different modalities of representation. The solution of all these questions will crucially depend on cross-fertilization between the humanities (ancient studies, linguistics), social sciences and natural sciences. We would like to thank all the presenters who took the effort to go through their papers again and to adapt them to the specific requirements of our volume. We would also like to thank all participants for their constructive questions and discussions. The doctoral students of the Research Training Group contributed to the success DOI 10.1515/9783110538779-202
VI
Preface
of the workshop in various ways – a heartfelt thank you to all of them, and to the German Research Foundation who has supported the Research Group and the symposium. Special thanks go to the coordinator of the Research Training Group, Silke Bechler, who did not only an excellent job in coordinating the workshop but also helped during the printing of this volume. We are also very grateful to Peggy Zogbaum for preparing the proofs and the index, and to Douglas Fear for his corrections of the English texts and for translating some of the papers and certain passages within individual contributions from German to English. Finally, we would like to express our gratitude to the publisher, Walter de Gruyter, first and foremost to Mirko Vonderstein, Antonia Schrader, and Katrin Hofmann for their excellent support throughout the whole time it took for producing this volume.
Mainz, July 2017 Tanja Pommerening & Walter Bisang
Contents Tanja Pommerening and Walter Bisang Classification and Categorization through Time
1
Part I: Classification in Ancient Times Jochen Althoff Categorization and Explanation of the World in Hesiod’s Theogony
21
Orly Goldwasser What Is a Horse? Lexical Acculturation and Classification in Egyptian, Sumerian, and Nahuatl 45 Sonja Gerke All Creatures Great and Small – The Ancient Egyptian View of 67 the Animal World Iolanda Ventura Classification Systems and Pharmacological Theory in Medieval Collections of Materia Medica: A Short History from the Antiquity to the 101 End of the 12th Century Tanja Pommerening Classification in Ancient Egyptian Medical Formulae and its Role in Re-Discovering 167 Comprehensive and Specific Concepts of Drugs and Effects
Part II: Classification in Modern Times Walter Bisang Classification between Grammar and Culture – a Cross-Linguistic Perspective Joachim W. Kadereit Classification and Naming of Living Objects – a Biologist’s Perspective 231 (Extended Abstract) Roy Ellen Tools, Agency and the Category of Living Things
239
199
VIII
Contents
Roy Ellen Categorizing Natural Objects: Some Issues Arising from Recent Work in Cognitive 263 Anthropology and Ethnobiological Classification
Part III: Tools Dietrich Busse Frames as a Model for the Analysis and Description of Concepts, Conceptual 281 Structures, Conceptual Change and Concept Hierarchies Frederik Elwert (Bochum) and Simone Gerhards Tracing Concepts – Semantic Network Analysis as a Heuristic Device 311 for Classification Index
339
List of Contributors
349
Tanja Pommerening and Walter Bisang
Classification and Categorization through Time The process of ordering objects of their environment into groups of objects according to certain criteria that may be related to their form or their usefulness for certain purposes seems to be a universal (Reichelt 2001) ability of humans. As Berlin (1992) has shown in his book on Ethnobiological Classification: Principles of Categorization of Plants and Animals in Traditional Societies, all the societies he analyzed give names to animals and plants and they seem to follow certain criteria when they do so and when they are asked to assign them to hierarchies. If the ability to classify is universal, one may wonder what can be said about its historical dimension. The present volume intends to go new ways in the research of classification by asking exactly that question. Given that this is a huge task, it will not be able to come up with a comprehensive analysis. Its intention is rather to present relevant examples from classification as it can be observed in ancient times and in modern times and to integrate them into the current theoretical discussion in such a way that it stimulates and cross-fertilizes research across a wide range of disciplines. Research of this type also includes new and promising methods from frame semantics and automated network analysis in computational sciences. As it will turn out, insights from prototype theory and insights on the flexibility and the context dependence of classification as they have been developed on the basis of observation from modern times also work in ancient cultures. The aim of this chapter is to provide the framework for stimulating further research on classification at the crossroads of the present and the past. For that purpose, it is divided into three sections. The first section takes the theoretical perspective as it is based on observations from currently accessible cultures. It discusses the notions of concepts, categories and classification from the perspectives of philosophy, cognitive sciences, semiotics, linguistics, cultural and social anthropology and biology. The second section starts out from the methodological question of how to reconstruct the mental world of past cultures from the available historical sources and then goes on to introduce certain aspects of classification in ancient cultures. Finally, the third section outlines the structure of the volume with its individual chapters.
DOI 10.1515/9783110538779-001
2
Tanja Pommerening and Walter Bisang
1 Concepts, categories and classification from the current theoretical perspective The notions of concepts, categories and classification and the question of how they are related are subject to rather divergent views, depending on the disciplines involved and the various approaches within them. Starting out from some philosophical deliberations, this section will offer a short survey of different approaches in cognitive sciences, semiotics, linguistics, cultural and social anthropology and biology. At the end, we will point out the aspects that will be relevant for our volume. The classical philosophical approach to classification goes back at least to Aristotle. It defines a category in terms of a set of necessary and sufficient membership criteria (conditions, features). Thus, a concept X can be included in the category girl if it fulfils the criteria of ‘human’, ‘female’ and ‘young’ (cf. e.g. the discussion in Cruse 2000: 130). In the course of time, this definition has been subject to many specifications. Wolters (2004: 409 f.) offers a set-based logical definition of classification that adds the criterion of binarity (pairwise subsets). Moreover, he rightly points out that the notion of ‘classification’ may refer to an operation or a result: [Classification is a] logical and methodological term. It is frequently used for referring to structured terminological systems in individual disciplines. From a logical perspective, classification stands for a division into classes, i.e. the complete decomposition of a non-empty set in pairwise disjunct subsets ... Thereby, classification denotes the operation of decomposition as well as its result. (Wolters 2004: 409.)1
In the view of this definition, binarity is an important criterion for distinguishing classification from categorization. Thus, in contrast to classification, categorization does not presuppose pairwise disjunction (Wolters 2004: 410). From a more general perspective that is underlying cognitive and linguistic research, complex organisms cannot survive without the ability to associate certain aspects of experience to reliable concepts. Without such concepts they would not be able to recognize them as having happened before and to associate them with previous successful or unsuccessful reactions. Moreover, they would not be able to communicate without at least partially shared concepts or categories. In models of the semiotic triangle, concepts can be seen as thoughts or references that are represented in our mind and that are associated with symbols (often
1 [Klassifikation ist ein] ‘logischer und methodologischer Terminus; [er wird] in den Einzelwissenschaften vielfach zur Bezeichnung gliedernder terminologischer Systeme (und der entsprechend gegliederten Gegenstandsbereiche) verwendet. Logisch gesehen stellt eine K. eine Klasseneinteilung, d.h. die vollständige Zerlegung einer nicht-leeren Menge in paarweise disjunkte Teilmengen (Einteilungsglieder) dar. Klassifikation bezeichnet dabei sowohl die Operation der Zerlegung als auch ihr Resultat. (...)’. (S. 409).
Classification and Categorization through Time
3
linguistic signs) that, in turn, stand for referents as real objects in the outside world (Ogden and Richards 1923)2. The terminology that is used for the three elements of the triangle varies across different approaches (cf. Eco 1973: 26 for an overview). In Peirce’s (1931–1935) theory, ‘interpretant’ corresponds to ‘thought/referent’, while ‘sign’ stands for ‘symbol’ and ‘object’ for ‘referent’. In general, a linguistic symbol like tree represents a mental concept of [tree] that is linked to real trees (oaks, elms, birches, etc.). Semiotic approaches of this type are limited to the notion of concepts and do not address the issue of categories. In this volume, categories are seen as entities that can be assigned to a certain concept. This view on concepts and categories follows a general agreement among cognitive scientists ‘that a concept is a mental representation that picks out a set of entities, or a category. That is, concepts refer, and what they refer to are categories’ (Rips et al. 2013: 177). In terms of Goldstone and Kersten’s (2003) very explicit definition, a concept is a mental representation of a class and a category is a set of entities that can be appropriately subsumed under that concept. In linguistics, the notion of concepts as they are assigned to categories is directly linked to concrete articulation as it is reflected in linguistic expression formats like words or more complex linguistic structures. This can be seen from the following quotation from Cruse (2000): They [= concepts, T.P and W.B] are organized bundles of stored knowledge representing an articulation of events, entities and situations, and so on in our experience. If we were not able to assign aspects of our experience to stable categories, it would remain disorganized chaos. (Cruse 2000: 127)
The process or activity that determines the assignment of entities to a category as it is associated with a given concept as a mental representation of a class will be called classification in this volume. This definition does not include the condition of binarity as introduced in the above philosophical definition of classification because our research is focused on what determines humans when they divide their experiences into different domains. In such a context, binarity seems to be irrelevant. The definitions of concepts and categories are associated with various problems which affect the notion of classification (for a good survey, cf. Ramscar and Port 2015). We will briefly introduce three of them. The first problem has to do with the relation between concepts and categories. As Goldstone and Kersten (2009) point out, their definition of the two terms leads to a stalemate in which either concepts determine categories or, vice versa, categories determine concepts. Another problem is concerned with what makes a category. As it
2 We cannot discuss the abundance of literature on this topic in our short introduction. Eco (1973) offers still a good survey of the debates in semiotics up to the time of its publication.
4
Tanja Pommerening and Walter Bisang
turns out in practice, it is hard to define a concept without at least implicitly labeling it by means of elements from natural language (also cf. Busse, this volume). Thus, a given object is a tree if the speakers of English refer to it as a ‘tree’. The problem of the reciprocal relation between concept and category and the dependency of concepts from concrete linguistic items led Ramscar and Port (2015) to the following practical solution: [I]n the rest of this review the word concept will be used to describe a specific relationship between a group of items: that they share a label, typically a word (or phrase, discriminable symbol, etc.). ... Category will be used to describe a set of items with a common label (including new items that could be considered to be members of that category in an appropriate context), etc. (Ramscar and Port 2015: 78)
The third problem is based on the question of how to define concepts and the criteria for assigning them to categories — a topic that has been discussed extensively. The classical approach described above in terms of Aristotle is problematic for various reasons, first and foremost because there is only a relatively small number of concepts/words that can be defined by means of a set of necessary and sufficient criteria (cf. girl above). Another approach is the prototype approach as introduced by Rosch and Mervis (1975). In this view, concepts are not defined on the basis of necessary and sufficient criteria but rather on the basis of similarities between members of a category. Individual examples of a concept can share more or less properties associated with that concept. Thus, penguins are seen as worse representatives of the concept [bird] than canaries because they lack the property of being able to fly. Yet another approach comes from psychology and category learning. It is based on exemplars (Nosofsky 1991; 1992) as instances of a category that is already stored in memory and the idea that humans assign new stimuli to categories by comparing their properties to the properties of previously learned exemplars. Since the assignment of a new stimulus to an exemplar typically depends on the number of shared properties, exemplar models are argued to explain the property configurations found in prototypicality approaches. Each of the above definitions starts out from the assumption that concepts are discrete constructs that are associated with invariant context-independent properties. That this can hardly be the case has already been pointed out by Wittgenstein (1953) who argues that concepts and words only get their meaning by their use within specific systems. Bisang (2004) discussed the same problem for traditional semiotic approaches (e.g. Eco 1973) which cannot model the creation of meaning in context because they define symbols and their relations to concepts in terms of fixed conventionalizations that exclude pragmatic inference. This criticism is further supported by experimental research from cognitive science on category learning.
Classification and Categorization through Time
5
Thus, Ramscar and Port (2015: 88–89) point out findings like the following in their summary of recent cognitive research: 1. Category assignments vary with context. 2. Even when people learn concepts with clear and consistent definitions, the representations they acquire diverge from these definitions. 3. The tasks people perform when learning to apply concepts has a strong effect on the representations they acquire. Moreover, the brain regions activated in categorization experiments seem to vary depending on the task and the context. The question of how these findings can be adequately modeled certainly needs more research. Potential candidates for integrating the effects of use and context dependence are approaches based on frames. Even if some approaches to frames may still more or less implicity be based on discrete notions of concepts (e.g. Fillmore 1982, also cf. Ramscar and Port 2015: 82), approaches like the one by Barsalou (1992; 1993) and others described by Busse (this volume) integrate frames or frame elements into larger frame systems or frame networks which may offer tools for capturing context-dependent categorizations and for modeling differences in conceptualization between individuals within a social group or a society. The flexibility of the assignment of entities to categories and the concepts associated with them is further supported by evidence from social and cultural anthropology. Classification is not only a cognitive but also a social phenomenon. As pointed out by Ellen (1996: 103), it ‘refers to the cognitive and cultural mechanisms’ by which the assignment of entities to concepts is ‘achieved’. Ellen (1996) also explicitly distinguishes the process of classification from its results. This distinction is of particular importance for the historical sciences, since they can only see the results of classification from which they have to reconstruct the initial intentions and their mental and cultural foundations from which they developed. No matter whether the cognitive or cultural motivations are accessible, approaches to classification that disregard flexibility and focus on conventionalised relations between entities and categories may easily fall for what Ellen (1996: 103) calls ‘the classificatory fallacy’: This reifies schemes as permanent cultural artefacts or mentally stored old knowledge, when they are more properly understood as the spontaneous and often transient end-product of underlying processes in an individual classifying act. We might call such a misinterpretation ‘the classificatory fallacy’, and there is every reason to believe that it is potentially evident whenever ethnographers try to make sense of their data, whether these be tables of symbolic oppositions, animal taxonomies or relationship terminologies.
6
Tanja Pommerening and Walter Bisang
In biology, classification is either geared to taxonomy or to evolutionary relationship.3 While the former is based on criteria of similarity for determining categories, the latter reflects the theory and practice of the phylogenetic classification of organisms based on molecular genetic information (cf. Kadereit, this volume). The extent to which the two perspectives overlap is unclear, since similarity-based criteria and genetic criteria do not necessarily lead to the same classification of individual organisms. Moreover, and this is remarkable from the perspective of the flexibility of classification, both approaches lead to large numbers of different classification systems. In the case of similarity, the result largely depends on the properties that are chosen for the classification, a fact that shows clear parallels with the absence of context-independent criteria of classification discussed above. In the case of phylogenetic classification, the increasing quantity of data plus the increasing quality of analytic tools has led to a large number of phylogenetic trees with an increasing number of hierarchical levels whose analysis and interpretation is beyond the topic of this volume. As the above overview has shown, classification can be seen from many perspectives and has been the subject of extensive discussions. To conclude this section, the following three aspects are of basic importance for this volume: –– The terms ‘classification’ and ‘categorization’ are used synonimously in this volume. –– Classification includes conscious processes for the systematic ordering of objects, notions, knowledge, etc. as well as unconscious processes that are part of a social context. Both types of processes are reflected in historical sources and it is not always possible to determine which of them applies. –– Concepts are often classified. For that reason, it is possible to derive information on conceptual content from different ways of classification (cf. section 2).
2 The view from the historical sciences The various approaches and analyses of the previous section are the result of a long tradition of theoretical reasoning combined with cognitive experiments with currently living informants, field work studies in currently observable living cultures and a rich tool kit of methods for biological classification with an accessible scientific history of the different schools and researchers involved. Most of this information is inaccessible to researchers specialised in pre-modern cultures. The available sources for reconstructing the history and the mental world of past cultures call for a considerable number of different methods that are not necessary in disciplines working
3 Since the term ‘class’ itself already refers to a hierarchical level in the system of classification, the systematic groups distinguished in biology are called ‘categories’.
Classification and Categorization through Time
7
on contemporary cultures. Thus, disciplines like Egyptology, Ancient Near Eastern Studies, Archeology, Classical Philology or Medieval Studies with their isolated sources in the form of texts, pictures and artefacts see themselves in the situation of puzzle players who have to complete incomplete puzzles. Often it is not even possible to trace back the origin of their research objects. Therefore, a scientific analysis of a historical source is not limited to the detailed assessment of its content (description, translation or typology, depending on the medium), it first and foremost also includes the reconstruction of its original context with indications of its temporal setting, its geographic origin, its usage or its producers and its potential users with their social position. From the perspective of the semiotic model discussed in section 1, words, pictures and artefacts in historical and archeological sources can be seen as signs (Peirce 1931–1935) or symbols (Ogden and Richards 1923). The access to the underlying concepts (references, interpretants) and the objects they refer to are blocked and there is no way to ask informants or to use other methods that are available to disciplines dealing with contemporary cultures. Disciplines of pre-modern cultures mainly use semiotic models for the interpretation of signs in a narrow sense. The following quotation from Loprieno (1998) is a good example: The present article interprets the notion of ‘sign’ in explicit recognition of the semiotic character (i.e. ‘A interprets B as a sign for C’) of ancient oriental cultural phenomena and their integration into an empirical or scientific framework, i.e. in terms of intracultural awareness. The simultaneous consideration of implicit semantic relations (cf. the discussion in Eco 1983, 79ff. ...), in particular the consideration of intracultural aspects, would imply a comprehensive cultural history of the world of the ancient Near East.4 (Loprieno 1998: 1785)
In actual fact, semiotic models are ideally suited for describing the problems that the historical sciences have to face (also cf. Theuerkauf 2003). Historians are thrown back on symbols of the world from which they have to reconstruct the underlying ‘real’ world or at least the view of their originator(s). In such a model, the content of a symbol occurring in the sources may only be derived directly from the source if it is fully congruent with its concept. Given the observation that a sign/symbol always needs pragmatic enrichment for being fully interpretable in a given context, such a
4 The German original of this quotation runs as follows: ‘In vorliegendem Artikel wird der Begriff ‘Zeichenkonzeption’ im Sinne der expliziten Anerkennung des semiotischen Charakters (d.h. ‘A interpretiert B als Zeichen für C’ ...) eines altorientalischen Kulturphänomens und dessen Einbindung in einen empirischen oder wissenschaftlichen Rahmen, also im Sinne eines intrakulturellen Bewußtseins interpretiert; denn die gleichzeitige Behandlung impliziter semantischer Beziehungen (vgl. die Diskussion in Eco 1983, 79ff. ...), insbesondere unter Heranziehung interkultureller Erwägungen, würde einer Gesamtkulturgeschichte der altorientalischen Welt gleichkommen.’ (Loprieno 1998: 1785).
8
Tanja Pommerening and Walter Bisang
situation is extremely unlikely if not impossible. In general, a sign/symbol can only reflect certain aspects of the originator’s conceptual intention and, needless to say, it cannot be expected that its interpreter (including the researcher) will fully retrieve that concept. Since there is no possibility to have direct exchange with the members of an ancient culture for asking questions, elicitation or even experiments, modern interpreters need to acquire a maximally broad knowledge about a culture, its concepts and their objects of reference for reconstructing the mental world of its members. This needs constant checking with already known signs/symbols and their contexts to avoid projections of our own associations from our conceptualization of the world into the signs/symbols of a past culture. What is needed is a maximal approximation to an emic (i.e, an intercultural) view – a goal that can only be reached to a limited extent (Trigger 2003; Harris 1976). In the context of translation, this implies that it is necessary for each word to look at its overall meaning in terms of the concept as a comprehensive whole and its concrete meaning in terms of the relevant subset of that whole. In practice, this means that it is inevitable to add a comment which clarifies the relation between that word as a sign/symbol and the concept with its extralinguistic referent for which it stands (for more information, cf. Pommerening 2010, Imhausen and Pommerening 2016). Theoretical considerations on the relation between concepts, categories and classification and the models derived from them (cf. section 1 on prototype semantics and the more complex context-dependent perspectives) offer new options for a better understanding of the concepts behind the signs/symbols through interdependencies between their properties. Starting out from concepts as combinations of properties (e.g. in terms of prototype theory), the class under which they are subsumed also contains the properties that determine those concepts. Taking the example of a sign/symbol of a conceptually unknown species, some properties of that species can be reconstructed if it is assigned to a certain class of species whose properties are known. The success of such reconstructions crucially depends on the extent to which information on the criteria of classification and the practice of classification in a given culture is available. In addition, it is necessary to know if there are clearly defined class boundaries or if classification is more flexible and context-dependent. With these questions, we are at the center of the topics discussed in part II of the present volume. Some aspects of these questions will be first illustrated with three examples: The first example starts out from the description of the word ‘sycamore’ in the Oxford English Dictionary (OED). The description of this plant is divided into several sublemmata or semantic subcategories. The definition given in the first position will be enough for the purpose of this chapter. It runs as follows:
Classification and Categorization through Time
9
‘A species of fig-tree, Ficus Sycomorus, common in Egypt, Syria, and other countries, and having leaves somewhat resembling those of the mulberry.’5 In Allaby’s (2013, ‘figs’) A Dictionary of Plant Sciences we find the following additional division: 1. 2. 3.
(UK) See Acer. (Near East) See Ficus. (N. America) See Platanus.
The description of ‘Ficus’ under point 2 runs as follows: Ficus (figs) (Family Moracaea): A big genus of trees, stranglers, root-climbers, and epiphytes which produce watery latex. The leaves usually have conspicuous stipules enveloping the bud. The flowers are minute, unisexual, and inserted on a concave receptacle forming a closed sphere. They are pollinated by specialized wasps. Many are cauliflorous. The fruits are tiny drupes enclosed in a receptacle, which is often brightly coloured and is eaten by many birds and mammals, which disperse the seeds. The receptacle of F. carica is the fig eaten by humans. F. religiosa is the pipal-tree or bodh-tree. There are about 800 species, most of which are tropical. (Allaby 2013, ‘sycamore’)
In both definitions, a species is assigned to its genus. The Oxford Dictionary of Plant Science provides additional information on the properties of the genus. Certain properties are marked as specific (latex, minute, unisexual), while the properties of shape are subject to variation. The sources from past cultures do not provide any information on what individuals understood under a word like ‘tree’. Nor do they address questions like ‘What do you associate with a sycamore?’; ‘To what category do you assign sycamores and what are its properties?’ or ‘Would you assign palm trees to the same category?’. What we find in various traditions of different cultures, however, are more or less comprehensive descriptions of living things like trees or plants. This can be illustrated by a second and a third example that come from Egyptology and Classical Studies, which both contributed to this volume. Two Egyptian papyri from about 1.550 BC, compare the habitus of a plant plus its flower and its leaf structure with other plants that are better known: A herb, ‘creeping herb’ is its name. It grows on its belly like the snake herb. Its flower is generally like [that of] the blue lotus. Its leaves are found to be like those of the ‘White-Wood Plant’. [Eb 294 (51,15–17) = H 35 (3,4 f.)]6.
All expressions referring to plants are assigned to the class of ‘plants’ by their corresponding determiner or classifier in their written representation. Moreover, certain
5 http://www.oed.com/view/Entry/196090?redirectedFrom=sycamore#eid (Accessed: 7/2017). 6 The abbreviations follow Westendorf (1999: 808–831).
10
Tanja Pommerening and Walter Bisang
typical representatives of the class of plants are pointed out for a closer characterization of certain of its components. Examples like these may recall prototype semantics. Thus, one can see them as a confirmation of the universality of prototype semantics by data from ancient Egypt. Similar conclusions may be drawn from examples from ancient Greece. The following passage from Theophrast’s Natural History of Plants (3rd century BC) shows similar procedures of comparing components of plants for highlighting certain specifics of a tree like the sycamore: Thus in Egypt there are a number of trees which are peculiar to that country, the sycamore the tree called persea the balanos the acacia and some others. Now the sycamore to a certain extent resembles the tree which bears that name in our country; its leaf is similar, its size, and its general appearance; but it bears its fruit in a quite peculiar manner, as was said at the very outset; it is borne not on the shoots or branches, but on the stem; in size it is as large as a fig, which it resembles also in appearance, but in flavour and sweetness it is like the immature figs, except that it is much sweeter and contains absolutely no seeds, and it is produced in large numbers. It cannot ripen unless it is scraped; but they scrape it with iron ‘claws’; the fruits thus scraped ripen in four days. If these are removed, others and others again grow from exactly the same point, and this some say occurs three times over, others say it can happen more times than that. Again the tree is very full of sap, and its wood is useful for many purposes. There is another peculiar property which it appears to possess; when it is cut, it is at first green, but it dries in deep water; they put it at once in a hole or in pools and so season it; and it becomes dry by being soaked in the deep water, and when it is completely dry, it is fetched up and floats and is then thought to be duly seasoned; for it is now light and porous. Such are the peculiarities of the sycamore. (Theophrastus, translated by Arthur Hort. 1916. Book IV, ch. II. Loeb Classical Library 70: 292 f.)
Both texts refer to the superordinate class of ‘tree’ or ‘plant’ for classifying the above organisms without adding any further hierarchical distinctions. Instead, certain similarity-based properties that are characteristic of a given plant or tree are picked out and compared to plants or trees which are well-known for having them. The above discussion and the contributions by Goldwasser and Pommerening (both this volume) show that the prototype model is very useful for interpreting sources from ancient cultures. In addition, our modern perspective which tries to understand early concepts and categorizations crucially depends on context. This becomes obvious if one investigates the sources for certain phenomena of classification from the perspectives of their temporal and geographic setting, their text type, the medium in which they occur, their originator and their use. The sources provide ample evidence for context dependence in various domains and they also show that there were no uniform criteria of classification (Pommerening 2015; also cf. the contributions to this volume by Gerke, Pommerening and Ventura). For that reason, semiotic models must be used in a semantically open form that allows for cultural pecularities – a point that has already been made with regard to historical sources by Aleida Assmann (1998):
Classification and Categorization through Time
11
Semiotics is constituted as a modern science by its interest in signs as such, which it investigates from a universalistic generalized perspective. The present article is based on the hypothesis that semiotics always has to be linked back to culture-specific conditions for avoiding the danger of getting lost in abstract typologies. Its strength lies exactly in providing an analytic tool for cross-cultural comparison that can contribute to a more differentiated understanding of cultural peculiarities. (Assmann 1998: 727)7
The classification criteria that were used in ancient cultures can be clearly derived from the historical sources and they form an excellent basis for stimulating effects on theoretical semiotic discussions in the light of A. Assmann’s (1998) above statement that generalized conceptualizations need backing up from cross-cultural variation, irrespective of whether this comes from ancient or modern cultures. Moreover, the findings on the context dependence of classification in ancient times tie in very neatly with findings from the studies of classification in modern times as they are presented in this volume, in particular in the two contributions by Ellen from Social and Cultural Anthropology. As in studies on modern times, the perspective of context dependence is not limited to the monolithic level of individual cultures as in A. Assmann’s (1998) quotation, it also applies to lower levels down to individual actors and the particular angle from which they look at a particular object. To conclude this section, it is important to point out that flexibility and context dependence do not exclude stability. The transmission of certain domains of knowledge from one generation to the next often took place with no change over long periods of time. In the case of ancient Egypt, temple libraries and archives can be seen as repositories of its cultural memory (Assmann 1992).
3 The structure of this volume The structure of this volume combines the current theoretical perspective discussed in section 1 with the view from the historical sciences in section 2. Inverting the order of these two sections, the chapters that discuss classification in historical periods are presented in Part I on Classification in Ancient Times and the chapters that address classification as it is can be currently observed are gathered under the title of Classification in Modern Times in Part II. Since two chapters explicitly address new methods
7 The original German text runs as follows: ‘Die Semiotik konstituiert sich als moderne Wissenschaft mit ihrem Interesse an Zeichen als solchen, die sie in einer universalistisch generalisierten Perspektive erforscht. Dem vorliegenden Artikel liegt die These zugrunde, daß die Semiotik von dieser allgemeinen Ebene immer wieder auf die Stufe der kulturspezifischen Rahmenbedingungen zurückgeführt werden muß, will sie der Gefahr entgehen, sich in abstrakten Typologien zu verlieren. Ihre Stärke liegt gerade darin, daß sie der vergleichenden Kulturbetrachtung ein analytisches Instrumentarium anbietet und so gerade auch zu einer differenzierteren Erkenntnis kultureller Besonderheiten beitragen kann.’ (Assmann 1998: 727)
12
Tanja Pommerening and Walter Bisang
of approaching and representing classification, two chapters are presented separately in Part III on Tools. Since one of these chapters deals with frames in current times and the other uncovers concepts and classification with the computational method of semantic network analysis in Ancient Egypt, this last chapter combines again the modern perspective with the historical view. The arrangement of the chapters in Parts I and II roughly follows the criterion of the extent to which context dependence is necessary for describing classification. It thus takes up the discussion at the end of section 2 which showed that the perspective from which a given object is looked at is crucial for analyzing classification in ancient cultures as well as in modern cultures. Chapters with comparatively low context independence are followed by chapters in which context plays a more important role in describing and modelling classification. Thus, the order of the chapters roughly reflects the increasing complexity that is needed for adequately describing the concepts and processes of classification under study. At the end of these two parts, part III discusses two promising tools for adequately detecting and modelling the overall complexity of classification no matter whether it manifests itself in ancient or in modern cultures. Part I: Classification in Ancient Times This part starts out from a context independent genealogical model of classification in ancient Greece, a model that was abandoned later until it became relevant again in modern genetics, and goes on to the example of the horse as it was introduced and classified in Egyptian, Sumerian and Nahuatl. The last three contributions discuss the absence of uniform classification criteria and the relevance of context. One of them looks at the animal world in ancient Egypt and the other two investigate medical classification in medieval Europe and in ancient Egypt, respectively. The chapter Categorization and Explanation of the World in Hesiod’s Theogony by Jochen Althoff shows how Hesiod, who is assumed to be the first poet of Europe, structures the world as it is seen from the human perspective on the basis of a genealogical model – a model that is context-independent and intends to explain nothing less than ‘the inner relations of things, attributes and forces’ (Althoff, this volume). Even though this model was clearly motivated by the aristocratic audience of Hesiod and the social importance of family lineages at his time it is quite remarkable that this approach was displaced by similarity-based models since Aristotle and that it is only since Darwin that genealogy has gained significant momentum again in modern biology. The chapter What is a Horse? – Lexical Acculturation and Classification in Egyptian, Sumerian and Nahuatl by Orly Goldwasser takes the example of the horse and shows how new objects are conceptually and linguistically integrated when they appear in a culture. As it turns out from the words that were created for the horse, the conceptualization of that animal started out from rather different perspectives in each of the
Classification and Categorization through Time
13
three cultures under scrutiny before it eventually became lexically more acculturated at a second stage (Brown 1999). In the case of Egyptian and Sumerian, classification is also reflected at the level of the writing system. In both cases, the criteria of classification do not primarily depend on context, even though the criteria that are used differ in each of the two cultures. The chapter All Creatures Great and Small – The Ancient Egyptian View of the Animal World by Sonja Gerke shows that there was no general ancient Egyptian view of the animal world nor any standardised nomenclature or hierarchisation. In spite of this, there were examples of classification and classification models in individual textual and iconographic sources but they were all from different contexts and did not focus on concrete knowledge about animals but rather on the broader conceptual domains of theology and religion. Thus, each case must be seen as a special case and must be analyzed within its own context and purpose. The chapter Classification Systems and Pharmacological Theory in Medieval Collections of materia medica: A Short History from the Antiquity to the End of the 12th Century by Iolanda Ventura shows how the arrangement of drugs (parts of plants, animals and minerals) in the pharmacological literature depends on their usage and on the theoretical pharmacological concepts that underly them. There are the following three main principles by which this arrangement is determined: (i) the naturalistic principle (Dioscorides: plants, animals, minerals; depending on family similarity), (ii) the Galenic principle (primary qualities: cold, warm, humid, dry; secondary and tertiary qualities), and (iii) the alphabetical principle (initial letter(s)). The motivation that drives this classification is information overload and the need to manage it. As the author points out, the more abstract principles (ii) and (iii) prevailed over morphologically oriented principles (i) in the medieval period. The chapter From Classification to Concepts: Classification in Ancient Egyptian Medical Formulae and its Role in Re-discovering Compehensive and Specific Concepts of Drugs and Effects by Tanja Pommerening aims at gaining information on the concepts of drugs and their healing effects in ancient Egypt through the ways in which they are classified. For that purpose, the following four levels of classification are analyzed: (i) classifiers in the Egyptian writing system (also cf. Goldwasser, this volume), (ii) nominal compounds, (iii) the internal structure of drug prescriptions and (iv) the arrangement of these prescriptions within the entire papyrus. This method does not only offer deep insights into the structure of individual drug concepts, it also elucidates the criteria used for classifying drugs. While level (i) is clearly based on the semantically most fine-grained criteria, level (ii) can be mostly accounted for in terms of prototype-based classification. The last two levels clarly depend on the context of the practical application and the efficacy of a drug. The overal motivation of classification is precision for capturing the healing effects of the drugs. It is this overall motivation that makes classification a valuable tool for recovering the concepts of Egyptian drugs.
14
Tanja Pommerening and Walter Bisang
Part II: Classification in Modern Times The first chapter presented in this part is about linguistics. It shows that classification is context independent in the grammar of some languages and context dependent in the grammar of others. The second chapter on classification in biology illustrates the intricate relationship between taxonomic and molecular genetic approaches. The last two chapters are about context dependence in Social and Cultural Anthropology. One of them presents the classification of living things, the other one raises important theoretical issues that follow from context dependence. The chapter Classification between Grammar and Culture – a Cross-linguistic Perspective by Walter Bisang presents an overview of how classification manifests itself in the grammar of the world’s languages. His typological analysis shows that depending on the individual language classification can follow rigid context-independent rules as well as very flexible context-dependent rules that even contribute to conceptualization, i.e. the specification of the concept as it matters in a given context. Moreover, some criteria of classification are recurrent in many languages from different families across the globe. These criteria are good candidates for universals of classification. The chapter Classification and Naming of Living Objects – a Biologist’s Perspective by Joachim W. Kadereit introduces the similarity-based taxonomic approach and the molecular genetic approach to classification in biology and illustrates how these two approaches produce different options for naming individual species on the basis of a concrete example. As was pointed out above, the taxonomic approach does not allow for a neutral perspective and is a good example for the context dependence of classification. The chapter Tools, Agency and the Category of Living Things by Roy Ellen offers an excellent example of the context dependence of classification by describing how the Nuaulu people in eastern Indonesia assign properties of living things to tools and other cultural objects – a general tendency of humans that is based on ‘the use of social intelligence to make sense of the natural world’ (Ellen, section 5, this volume). A look at meat-skewers, outboard motors, coconut graters and sago-processing devices from the perspective of the conditions that determine the recognition of life provides ample evidence of the context dependence of classification and shows that certain objects are more likely to be assigned to the category of living things than others. Methodologically, context dependence is reflected in the important observation that different kinds of ethnographic contexts and different elicitation procedures yield different results – a fact that needs to be considered in theoretical discussions on the validity of existing anthropological classifications of living things. The chapter Categorising Natural Objects: Some Issues Arising from Recent Work in Cognitive Anthropology and Ethnobiological Classification by Roy Ellen provides an overview on classification as a culturally and contextually dependent phenomenon by addressing the following issues that are widely discussed in the literature: classification as a situated process, the relationship between words and categories, hierar-
Classification and Categorization through Time
15
chies and taxonomy, cultural universals, domain specificity, the position of humans within folk classifications, and the impact of writing and depiction technologies. It is needless to say that such an overview automatically goes with certain simplifications but it is exactly these simplifactions which may give the reader a certain impression of how complex the task of developing an integrative model of classification is. Part III: Tools Analyzing and understanding the full complexity of classification in its interaction with cognition, context and historical change is an enormous task that calls for new tools. This volume introduces frames (Busse) and automated network analysis (Elwert and Gerhards) as two very promising approaches to this problem. The chapter Frames as a Model for the Analysis and Description of Concepts, Conceptual Structures, Conceptual Change and Concept Hierarchies by Dietrich Busse explores the potential of frames for classification and complex conceptual structures in general. Frames can be roughly seen as concepts or structures of knowledge with their core and their slots (or attributes, terminals) as the basic elements of knowledge which are interconnected. Since all components of a frame are also frames, frame theory is a recursive model that can be infinitely refined. Moreover, frames are organised in larger structures or networks that can be described in terms of ontologies or taxonomies as well as in terms of proximity, analogy and frame inheritance. As the author points out, frame theory should not be seen as a general solution to all the problems associated with classification but it is a promising approach for analyzing and understanding concepts whose complexity has been underestimated in other approaches. The chapter Tracing Concepts – the Classification of the Use of Ancient Egyptian Terms with Semantic Network Analysis by Frederik Elwert and Simone Gerhards presents semantic network analysis as a heuristic computational tool for tracing concepts and their relationships from large text corpora. The authors illustrate the detection of the conceptual properties of Egyptian expressions for day by the computational analysis of their co-occurrence with other concepts in a corpus of about one million lexical units. With this method, it is possible to see indicators of different conceptualizations (day connected to feasts, wars, etc.). As the authors point out, such computational methods cannot replace close reading. They uncover potential conceptual properties of a lexical unit that will need further corroboration from close reading, i.e. the detailed analysis of how these expressions are used in individual texts.
16
Tanja Pommerening and Walter Bisang
References Allaby, Michael 32013. A Dictionary of Plant Sciences. Oxford: Oxford University Press. Assmann, Aleida. 1998. “Probleme der Erfassung von Zeichenkonzeptionen im Abendland.” In Semiotik. Ein Handbuch zu den zeichentheoretischen Grundlagen von Natur und Kultur, vol. 1, edited by Roland Posner et al., 710–729. Berlin: de Gruyter. Assmann, Jan 1992. Das kulturelle Gedächtnis. Schrift, Erinnerung und politische Identität in frühen Hochkulturen. München: Beck. Barsalou, Lawrence. 1992. “Frames, concepts, and conceptual fields.” In Frames, Fields and Contrasts, edited by Adrienne Lehrer and Eva F. Kittay, 21–71. Hillsdale, N.J.: Lawrence Erlbaum. Barsalou, Lawrence. 1993. “Flexibility, structure and linguistic vagary in concepts: manifestations of a compositional system of perceptual symbols.” In Theories of Memory, edited by Alan F. Collins, Susan E. Gathercole, Martin A. Conway and Peter E. Morris, 29–101. Hove, UK/Hillsdale, N.J.: Lawrence Erlbaum. Berlin, Brent. 1992. Ethnobiological Classification: Principles of Categorization of Plants and Animals in Traditional Societies. Princeton/New Jersey: Princeton University Press. Bisang, Walter. 2004. “Kultur und Sprache aus der Perspektive des Kontaktes.” In Kultur, Sprache, Kontakt, edited by Walter Bisang, Thomas Bierschenk, Detlef Kreikenbom and Ursula Verhoeven, 1–52. Würzburg: Ergon. Brown, Cecil H. 1994. “Lexical acculturation in Native American languages.” Current Anthropology 35: 95–117. Cruse, Alan. 2000. Meaning in Language. An Introduction to Semantics and Pragmatics. Oxford: Oxford University Press. Eco, Umberto. 1973. Il segno. Milano: Arnaldo Mondadori. Ellen, Roy, 1996. “Classification.” In Encyclopedia of Social and Cultural Anthropology, edited by Alan Barnard and Jonathan Spencer, 103–106. London: Routledge. Ellen, Roy. 2006. The Categorical Impulse: Essays in the Anthropology of Classifying Behaviour. New York/Oxford: Berghahn. Fillmore, Charles John. 1982. “Frame semantics.” In Linguistics in the Morning Clalm, edited by Linguistic Society of Korea, 111–137. Seoul: Hanshin. Goldstone, Robert and Alan Kersten. 2003. “Concepts and categorization.” In: Comprehensive Handbook of Psychology. Vol. 4: Experimental Psychology, edited by Alice F. Healy, and Robert W. Proctor, 591–621. New York: Wiley. Harris, Marvin. 1976. “History and Significance of the Emic/Etic Distinction.” Annual Review of Anthropology 5: 329–350. Imhausen, Annette and Tanja Pommerening (ed.). 2016. Translating Writings of Early Scholars in the Ancient Near East, Egypt, Greece and Rome. Methodological Aspects with Examples. (Beiträge zur Altertumskunde; , 344) Berlin/New York: de Gruyter. Lakoff, George. 1987. Women, Fire, and Dangerous Things. Chicago: University of Chicago Press. Loprieno, Antonio. 1998. “Zeichenkonzeptionen im alten Orient.” In Semiotik. Ein Handbuch zu den zeichentheoretischen Grundlagen von Natur und Kultur, vol. 2, edited by Roland Posner et al., 1785–1799. Berlin: de Gruyter. Murphy, Gregory L. 2002. The Big Book of Concepts. Cambridge, Mass.: The MIT Press. Nosofsky, Robert. 1991. “Tests of an exemplar model for relating perceptual classification and recognition memory.” Journal of Experimental Psychology: Human Perception and Performance 17: 3–27. Nosofsky, Robert. 1992. “Similarity scaling and cognitive process models.” Annual Review of Psychology 43.1: 25–53.
Classification and Categorization through Time
17
Ogden, Charles Kay and I. A. Richards. 1923. The Meaning of Meaning. New York: Harcourt, Brace & World, Inc. Peirce, Charles Sander. 1931–1935. Collected Papers. Cambridge, Mass.: Harvard University Press. Pommerening, Tanja. 2010. “Von Impotenz und Migräne – eine kritische Auseinandersetzung mit Übersetzungen des Papyrus Ebers.” In Writings of Early Scholars in the Ancient Near East, Egypt, Rome and Greece. Translating Ancient Scientific Texts, edited by Annette Imhausen and Tanja Pommerening, 153–174. (Beiträge zur Altertumskunde, 286) Berlin/New York: de Gruyter. Pommerening, Tanja. 2015. “Bäume, Sträucher und Früchte in altägyptischen Listen – eine Betrachtung zur Kategorisierung und Ordnung.” In Die Liste. Ordnungen von Dingen und Menschen in Ägypten, edited by Susanne Deicher and Erik Maroko, 125–166. Berlin: Kulturverlag Kadmos. Ramscar, Michael, and Robert Port. 2015. “Categorization (without categories).” In Handbook of Cognitive Linguistics, edited by E. Dawbroska and D. Divjak, 75–99. Berlin: Mouton de Gruyter. Reichelt, Gregor 2001. Universalien. (Zusammenstellung aus dem Sonderforschungsbereich “Literatur und Anthropologie” an der Universität Konstanz.) http://www.uni-konstanz.de/FuF/ ueberfak/sfb511/publikationen/universalien.html [Accessed 5/2016] Rips, Lance, Edward Smith and Alan Wagner. 2013. “A theory of Pavlonian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement.” In Classical Conditioning II: Current Research and Theory, edited by A. H. Black and W. F. Prokasy, 64–99. New York: Crofts. Rosch, Eleonor, and Carolyn Mervin. 1975. “Family resemblances: Studies in the internal structure of categories.” Cognitive Psychology 7.4: 573–605. Theophrastus, translated by Arthur F. Hort. 1916. Enquiry into Plants, vol I: Books 1–5. (Loeb Classical Library, 70) Cambridge: Harvard University Press. Theuerkauf, Gerhard. 2003. “Semiotische Aspekte der Geschichtswissenschaften: Geschichtssemiotik.” In Semiotik. Ein Handbuch zu den zeichentheoretischen Grundlagen von Natur und Kultur, vol. 3, edited by Roland Posner et al., 2937–2976. Berlin: de Gruyter. Trigger, Bruce G. 2003. Understanding Early Civilizations: A Comparative Study. New York: Cambridge University Press. Westendorf, Wolfhart. 1999 . Handbuch der altägyptischen Medizin. 2 vols. (Handbuch der Orientalistik, Abt. I, 36) Leiden et al.: Brill. Wittgenstein, Ludwig. 1953. Philosophical Investigations. London: Blackwell. Wolters, Gereon. 2004. “Klassifikation.” In Enzyklopädie Philosophie und Wissenschaftstheorie, vol. 2, edited by Jürgen Mittelstraß, 409 f. Unveränderte Sonderausgabe, Stuttgart/Weimar: Metzler.
Jochen Althoff
Categorization and Explanation of the World in Hesiod’s Theogony Summary: Hesiod (ca. 730 BC) in his hexametric poem Theogony describes his known world in terms of a widespread genealogy, starting from Chaos in the beginning, unfolding in a series of sexual encounters and births and ending with the god Zeus and his kingship in heaven. Natural structures (earth, mountains, rivers etc.) are covered as well as basic human emotions (love, strife, anger etc.) and actions (war, deceit etc.) and all kinds of gods (partly identical with geological structures, like Earth and Sky). By grouping these phenomena as parents and offspring, brothers and sisters etc. he categorizes them by allocating groups to certain parents and tells us a lot about their characters, values and relations to each other. This mode of structuring works for concrete gods as well as for more abstract phenomena who thus all appear as anthropomorphic figures. The genealogical model is transferred from heroic poetry, where it serves a different purpose, to the fundamental entities of the world. It always needs interpretation, because common features and characteristics are hardly ever spelled out explicitly and the meanings of genealogical descent can be manifold. The term genos (‘race, stock, kin’) and its cognates (genealogia and the like) have later been used in the classificatory systems of Aristotle’s biology and thus found their way into modern classificatory efforts (Linnaeus and in a different sense Darwin). Hesiod seems to have prepared the way for a powerful explanatory tool of natural science. Most modern handbooks on Ancient Greek Literature list Homer as the first poet of Europe and Hesiod as a younger contemporary, and this has long been the standard dating.1 Some time ago M. L. West2 and recently W. Blümer have checked the few and complicated arguments and have come to the opposite conclusion. According to them, Hesiod probably is the earliest Greek (and European) poet and Homer wrote about 50 years or even more after him.3 I am convinced by Blümer’s arguments and
1 Cf. Paulsen 2004: 15: Iliad ca. 730 BC, Odyssee ca. 700 BC. Id. 41: ‘… Hesiod from Ascra in Boeotia who lived around 700 BC.’ 2 West 1966: 46: ‘The Theogony may well be the oldest Greek poem we have.’; 46f.: ‘That Hesiod is earlier than Homer is no revolutionary view; but as the reverse is taken as axiomatic by most writers, it may be worth recalling that until the latter part of the fourth century B.C., Hesiod’s priority was widely accepted.’ 3 Cf. Reichel 2011: 28 f.: Iliad ca. 660–650 BC, cf. especially 29: ‘Daß nicht Homer, sondern Hesiod der älteste Dichter des Abendlandes ist, muß demnach als Möglichkeit erwogen werden.’ – for Hesiod cf. ibid. 81. I cite Blümer 2001, because he is unfortunately often omitted, 107–260. DOI 10.1515/9783110538779-002
22
Jochen Althoff
follow his tentative dating of the Theogony to the last three decades of the 8th century BC.4 We are, then, dealing here with the oldest poet of European culture. Since ancient times Hesiod has been taken to be the first natural philosopher who, in many ways, influenced the thoughts and writings of the early Presocratics.5 Aristotle names him in Metaphysics A 4. 984 b 23–25 together with Parmenides as the inventor of a kind of ‘principle of movement’ (ἀρχὴ κινήσεως), which is a prominent factor in his set of the four causes.6 Hesiod assumes the god Eros, who comes into being very early in his Theogony (line 119, see shortly), as his basic moving cause. Even the literary form of the epic hexameter, which Hesiod uses, was adopted by such famous Presocratics as Parmenides and Empedocles.7 Although many aspects of the close relation between the Presocratics and Hesiod have been detected and well described, little has been said about the effort of systematization and categorization which Hesiod undertakes by using a genealogical model.8 This will therefore be the focus of the following pages. After a long and famous proem, Hesiod begins the main part of his Theogony in line 116 with the very first elements of the world. He writes: ἤτοι μὲν πρώτιστα Χάος γένετ’· αὐτὰρ ἔπειτα Γαῖ’ εὐρύστερνος, πάντων ἕδος ἀσφαλὲς αἰεὶ ἀθανάτων οἳ ἔχουσι κάρη νιφόεντος Ὀλύμπου, Τάρταρά τ’ ἠερόεντα μυχῷ χθονὸς εὐρυοδείης, ἠδ’ Ἔρος, ὃς κάλλιστος ἐν ἀθανάτοισι θεοῖσι, λυσιμελής, πάντων τε θεῶν πάντων τ’ ἀνθρώπων δάμναται ἐν στήθεσσιν νόον καὶ ἐπίφρονα βουλήν.
4 Blümer himself is very careful, cf. e.g. 2001: 210: ‘Natürlich wüsste man gerne, wann die Gedichte verfasst wurden. Wenn man daher die Ergastelle in die Zeit vor 700, die Iliasstelle in die Zeit nach 700 datiert, mag dieser Ansatz dem gegenwärtigen Stand der Forschung Rechnung tragen.’ 5 A good recent overview is given by Koning 2010: 189–235, with references to older works 190 f., notes 2–4. Cf. also Bremer, in: Flashar/Bremer/Rechenauer 2013: 61: ‘So ist die philosophische WoherFrage, wie sie Aristoteles entwickelt und den frühen Denkern von Thales an zuspricht, historisch fassbar als genealogische Denkfigur in den vielfältigen Formen des Hervorgehens, in denen Hesiod in seiner ‘Theogonie’ eines aus dem anderen entstehen lässt.’ Important are the older German contributions by Diller 1946 and Snell 1986: 45–55. Cf. also Boys-Stones 2010: 31–51. 6 Cf. e.g. Aristotle, Physica II 3. 194 b 23–195 a 3: 1. τὸ ἐξ οὗ γίγνεταί τι ἐνυπάρχοντος (material cause), 2. τὸ εἶδος καὶ τὸ παράδειγμα (formal cause), 3. ὅθεν ἡ ἀρχὴ τῆς μεταβολῆς ἡ πρώτη ἢ τῆς ἠρημήσεως (moving cause), 4. τὸ τέλος· τοῦτο δ’ ἐστὶν τὸ οὗ ἕνεκα (final cause). 7 Cf. Patzer, in: Flashar/Bremer/Rechenauer 2013: 129–135. 8 Unfortunately the new Brill’s Companion to Hesiod (Montanari, Rengakos and Tsagalis 2009) does not contain a separate chapter on Hesiod as a philosopher, although a few remarks can be found in F. Montanari‘s contribution (in: Montanari, Rengakos and Tsagalis 2009: 313–342). Cf. Philippson 1966.
Categorization in Hesiod’s Theogony
23
And really as the very first thing Chaos came to be, and then Gaia, broad-breasted, the ever secure seat of all the immortals who dwell on the peaks of snowy Olympus, and the realms of dark Tartarus in a corner deep in the broad-pathed earth and Eros who is the most beautiful among the immortal gods, slackening (their) limbs, and the mind of all the gods and all the mortals he overpowers and their thoughtful counsel as well.9 The absolute starting point of the world is Chaos. Quite amusingly, Hesiod uses the superlative of the (already superlative) adverb πρῶτος ‘at first’: πρώτιστα ‘at the very first time’. Hesiod does not say where Chaos came from or from what it developed. It was just the very first thing to come to be. Nor does he say what Chaos is, and it is important not to confuse this first entity with what we nowadays understand as chaos. To us, chaos means a disordered, unorganized jumble of matter, but this is only a later development of the word’s meaning.10 In Hesiod’s cosmogony, however, Chaos has to be understood as the animal metaphor it originally was. The word is connected with the verb chasko (χάσκω), which means ‘to open one’s mouth wide while yawning or looking at something very strange or awful’. The basic idea seems to be one of a cosmogonic animal that opens its mouth to produce a space into which things can be placed. Nevertheless, Hesiodic Chaos already contains features of the dark, murky, undifferentiated and unordered state of the primeval world which the later use connotes, as becomes clear from Chaos’s immediate offspring. It follows quite naturally that, after empty space, there have to come into being some firm and steady foundations for further cosmogonic activities.11 These are provided by the earth, Gaia, who is described as ‘broad-breasted’. Quite concretely, Hesiod only speaks of all the immortal gods, who find their place on earth.12 But obviously
9 Translations are my own, if not otherwise noted, but I gratefully acknowledge the English version of Most which I compared throughout: Most 2006. 10 Best known from the beginning of Ovid’s Metamorphoses I. 6–9: unus erat toto naturae vultus in orbe, / quem dixere Chaos: rudis indigestaque moles, / nec quicquam nisi pondus iners congestaque eodem / non bene iunctarum discordia semina rerum. Anaxagoras seems to be the first to propose a primordial matter as a total mixture of things and qualities, although he does not use the word ‘chaos’ for it: 59 B 4 Diels/Kranz: p. 34, line 17–35, 5. 11 Cf. West 1966: 192 on 116–53: ‘The universe is, naturally, built from the bottom upward, starting with the foundations (Chaos …). Then come the floor and the walls (Earth with her subsidiary parts Tartarus, mountains and sea) and the roof (Uranos). When the house is ready, the inhabitants move in; …’ 12 Cf. West 1966: 194 on 118f.: ‘It is clearly possible to describe the earth as the ‘seat’ of the gods, who dwell in Olympus; … The difficulty is to reconcile it with 128, where it is the sky that is the gods’ ἕδος ἀσφαλές. … The answer to it is perhaps that, after all, the gods do have homes both on earth and in heaven, and their Olympian city may be thought of as in either realm.’
24
Jochen Althoff
he takes the perspective of human beings for whom the most solid bottom and basis of everything is the earth on which we stand. It is only because humans have not yet been created that he at first focusses on immortal gods. And, of course, all the next entities to come into being are conceived of as immortal gods. The earth is ‘broadbreasted’ because it stretches far into every direction. Earth is ‘broad-breasted’ and not ‘broad-backed’ or the like, because she is also – as mother earth – the nurturing ground of all the living beings. It becomes clear at once, that Earth, like Chaos, is conceived of as an animal or human being, because only animals and humans have a breast. Once again Hesiod does not tell us where Earth comes from. It is just the second thing to spring into existence after Chaos. More by intuition than by logical reasoning, Hesiod has understood that a very first starting point of the cosmos cannot itself have another cause for its coming into being. Otherwise the reason for this starting point would be the very first starting point and so on, until you finally reach a very, very first starting point. An absolute beginning cannot have a beginning itself; we have to reach an end of the backwards extending chain of principles. Later philosophers coined the term of the ‘unconditional beginning’ (ἀνυπόθετος ἀρχή);13 Hesiod’s Chaos and Earth are no doubt his ‘unconditional beginnings’ of the world. The ‘realms of Tartaros’, which are mentioned as a third basic entity, pose certain difficulties.14 I will leave them aside here. After that, Eros is another basic element of the cosmos at its very beginning. Eros is supposed to be a male god (ὅς, 120) overpowering gods and men alike. He weakens the normal powers of reason and rational planning and relaxes bodily tension. These consequences are aspects of sexual activity and a foreshadowing of the poem’s central topic: the coming to be of the following entities as a series of sexual encounters and births. Sexual encounter presupposes love and erotic attraction, which are the moving forces behind copulation and begetting. The person that represents these forces is Eros, who is the most beautiful god among the immortals. Whereas we moderns tend to think about love as an immaterial force, Hesiod clearly remains within the perspective of personification. We cannot be sure, therefore, if he really distinguishes between forces and matter, which are later concepts.15
13 Plato, State VI. 510 B 7; Aristotle, Metaphysics Γ 3. 1005 b 14. 14 The lines 118 f. are missing in two ancient quotations by Plato, Symposium 178 B 5–7 and Aristotle, Metaphysics A 4. 984 b 27–29. Cf. West 1966: 194 (on 119): ‘It is possible that Hesiod originally began with the trio Chaos, Earth, Eros, and inserted Tartarus later, when he came to the Titanomachy and realized that an important part of the universe had been omitted from the cosmogony.’ 15 They are most clearly developed by Empedocles, who reduces the world to four material elements (fire, air, water, earth) and two immaterial forces (love and strife).
Categorization in Hesiod’s Theogony
25
The next lines (123–125) describe the beginning of the series of generations: ἐκ Χάεος δ’ Ἔρεβός τε μέλαινά τε Νὺξ ἐγένοντο· Νυκτὸς δ’ αὖτ’ Αἰθήρ τε καὶ Ἡμέρη ἐξεγένοντο, οὓς τέκε κυσαμένη Ἐρέβει φιλότητι μιγεῖσα. From Chaos black Erebos and black Night came to be. From Night again Aither and Day came forth, both of which Night conceived and bore for Erebos after mingling (with him) in love. Whereas the verb γίγνεσθαι ... ἐξ ‘to come to be ... from’ does not necessarily express a sexual reproduction, this is quite clearly the case with the verbs τέκε κυσαμένη (‘conceived and bore’), especially with the added standard phrase φιλότητι μιγεῖσα (‘having mingled in love’). Philotes obviously represents the force of Eros, but the word is not confined to sexual desire in other contexts.16 From Chaos, being at least partly an animal-like individual, Erebos and Night come forth. No second parent is mentioned, so that the obvious conclusion is that Chaos is female. 17 Erebos is qualified as ‘black’, i.e. ‘dark’; it is a place below the earth, mostly the same as Tartaros (who had a similar epithet in line 119: ἠερόεις ‘murky’). But the word Erebos itself stresses the quality of darkness (which is its basic meaning). This is the reason why Erebos is closely connected with Night, which is the time of darkness on earth. The difference between the two is that Erebos is perpetual darkness under or in the earth, whereas Night is temporary darkness on the earth. Their similarity lies in the quality of being dark. An important question is: why do Erebos and Night come forth from Chaos and not from Earth or Eros? Eros is least likely to have offspring, because from the beginning he is primarily an active force, not so much a parent. Except for one other place, he is never mentioned again in the Theogony and has no offspring of his own.18 The idea must be somewhat like the following: Chaos is a sinister, negatively connoted entity. It is something like mere space, dark, undefined, without clear structures. Earth, however, is defined as a hard, stable surface, a real, recognizable structure
16 Philotes is named in 224 as a child of the Night, so – strictly speaking – it does not yet exist at this early time. Therefore, it seems to be a synonym for Eros here. The verse-end φιλότητι μεγεῖσα is formulaic in epic language. Cf. West 1966: 196: ‘if he (sc. Eros) is not named as such (sc. as a force of generation and reproduction), it is only because the formulae describing sexual union refer to φιλότης and Ἀφροδίτη instead.’ But the question of why Hesiod names philotes separately in 224 remains. 17 Cf. West 1966: 193 on 116: ‘Note that although grammatically neuter, Chaos is treated as female (123, Aristophanes, Aves 698).’ 18 In 201 Eros and Himeros (‘longing’) are listed as companions of Aphrodite, who is a third representative of the force of love.
26
Jochen Althoff
with positive connotations of stability. Darkness, on the other hand, is also a negative quality: you cannot see anything, everything looks as black as everything else, you are disoriented and become frightened. So the connection between Chaos and Darkness lies in the negative, undefined, unstructured and even horrifying qualities of both. The relation between parent and children in this first case of genesis is then that of similarity. And this is, indeed, often observed in the real world. Children resemble their parents in most respects, first of all by belonging to the same species, but also sharing more specific common features (like colour of the eyes or hair, form of the nose etc.) in a concrete, as well as a more abstract sense: sometimes children have the same intelligence, the same habits, the same character as their parents. Hesiod has transferred these observations to his description of the first entities of the cosmos, and it even holds in cases like darkness, where not even a complete sexual generation is described. From line 124 onwards there is finally enough staff, enough ‘persons’, to move on to a first sexual encounter between Night and Darkness. This couple of similar character ‘mingles in love’, conceives and bears two children: Aither and Day, quite obviously a pair like Erebos and Night. But this time the relation between parents and children seems to be different. Aither is a nomen agentis from the verb aitho (αἴθω), ‘to burn, to kindle a fire’, and therefore here represents light (‘the burner’). Compared with day, location and durability once again make the difference: Aither seems to be the eternal heavenly light in an upper sphere, day is the temporary light on earth, alternating with night. So this time the children are the logical opposites of their parents. This relation is not quite as obvious as that of similarity, but it can just as well be observed in human reality: every now and then children show different habits and features from their parents. Moreover, ancient Greeks tended to think in terms of analogy or polarity.19 Therefore, when the model of an analogical character of the offspring does not work, they tend to switch to the polar model, and that means, as in the case discussed here, that children are not just slightly dissimilar to their parents, they become their opposites. The children of darkness and night become light and day.20 With the second generation of Aither and Hemera, we have reached the beginning of a genealogical description of the entities of the world. From the very beginning, Chaos, come forth Darkness and Night, from the sexual coupling of Night and Darkness, Light and Day are produced. The first pair is similar in character to its
19 Cf. Lloyd 1992. 20 It has long been noted that Light and Darkness are not just opposites, they are complementary opposites. It is always either dark or bright, no third quality exists (not even when one admits intermediate states like dawn, which can be explained as mixtures of the two basic states). Cf. West 1966: 197 on line 124.
Categorization in Hesiod’s Theogony
27
single parent, the second pair, though stemming from parents of similar character, has the opposite character traits. But they also have features in common. Parents and children are states of light in different places and with a comparable durability. As grandchildren of Chaos, they still have something in common with Chaos. Hesiod’s idea seems to be that brightness and distinct features necessarily belong to the indistinct and dark chaos – like existence belongs to non-existence or the mere possibility of existence. But, of course, such very abstract sentences cannot be found in Hesiod’s poem. These insights have to be gained by interpreting the genealogical model. This is a severe drawback of such models (and of poetry as a whole): as a reader and interpreter you can never be sure to which conclusions and insights Hesiod had already come. It is, nevertheless, obvious that the genealogical scheme is intended to disclose (or at least hint at) relations and defining features of the cosmic entities described. In this very first phase of cosmogony, we already find a broad variety of things. There are, in sequence, quite undefined chaos (unstructured space in modern words), solid earth as the firm seat of everything else, and very abstract astronomical phenomena such as night and day, which are combined with even more abstract terms such as darkness and light. For Hesiod, these are all animal-like or man-like individuals.21 Every now and then one gains the impression that some entities are thought of as less personal than others, e. g. Eros, because nothing is said about his bodily features and he never has children of his own. But that is never expressly stated. Having followed the offspring of Chaos until the birth of Light and Day, Hesiod in line 126 switches back to the offspring of Gaia (Earth): Γαῖα δέ τοι πρῶτον μὲν ἐγείνατο ἶσον ἑωυτῇ Οὐρανὸν ἀστερόενθ’, ἵνα μιν περὶ πᾶσαν ἐέργοι, ὄφρ’ εἴη μακάρεσσι θεοῖς ἕδος ἀσφαλὲς αἰεί, γείνατο δ’ οὔρεα μακρά, θεᾶν χαρίεντας ἐναύλους Νυμφέων, αἳ ναίουσιν ἀν’ οὔρεα βησσήεντα, ἣ δὲ καὶ ἀτρύγετον πέλαγος τέκεν οἴδματι θυῖον, Πόντον, ἄτερ φιλότητος ἐφιμέρου· αὐτὰρ ἔπειτα Οὐρανῷ εὐνηθεῖσα τέκ’ Ὠκεανὸν βαθυδίνην Κοῖόν τε Κρεῖόν θ’ Ὑπερίονά τ’ Ἰαπετόν τε Θείαν τε Ῥείαν τε Θέμιν τε Μνημοσύνην τε Φοίβην τε χρυσοστέφανον Τηθύν τ’ ἐρατεινήν.
21 According to modern spelling habits, this poses the problem of whether to write the names of the phenomena with a capital letter or not; if Hesiod already had that spelling habit (he probably wrote everything in capitals without word-boundaries and boustrophedon, cf. West 1978: 60), he would perhaps always use the capital. I have changed the spelling as it seems adequate. Cf. Philippson 1966: 662 f.
28
Jochen Althoff
τοὺς δὲ μέθ’ ὁπλότατος γένετο Κρόνος ἀγκυλομήτης, δεινότατος παίδων, θαλερὸν δ’ ἤχθηρε τοκῆα. Gaia, you see, at first brought forth as equal to herself starry Sky, that he may cover her all around her, and that he would be an always secure seat of the blessed gods. And she bore high mountains, the graceful dwellings of the Nymph-goddesses, who live scattered through the mountains with their wooded glens. She also gave birth to the barren (?) sea, seething with its swell, Pontos, without love full of desire. And then, having slept with Sky, she bore Okeanos with deep swirls, Koios, Kreios, Hyperion, Iapetos, Theia, Rheia, Themis, Mnemosyne, Phoibe with a golden crown and delightful Tethys. After those as the youngest Kronos came to be, crooked-counseled, the most terrible of her children: he hated his stout father. Gaia’s offspring fill a long list. First she gives birth to the Sky with his stars, obviously without a father. This production is motivated by two intentions (ἵνα ... ὄφρ’, 127f.), which again make it clear that Earth is a person with plans and purposes. Firstly, Sky is born in order to enclose earth all around. This sounds like Sky is needed as a kind of border or boundary of the earth. Earth, as we have already heard, stretches far into all directions, but not endlessly. There has to be a border against Chaos, the undefined space. It is not easy to understand what form of our world Hesiod implies by these lines. The picture seems to be that of a sandwich, the sky lying flat on the surface of the earth like a slice of cheese on bread. But this is only the first stage. Later sky will be removed to a much greater distance, probably roughly parallel to the earth.22 In lines 154–206 we hear of a typical myth of separation of both partners which initiates this second stage. If we ask about the relation between Sky and Earth as son and mother, the answer is that they are of the same size; this is expressly said in line 126 (ἶσον ἑωυτῇ). This same size is necessary in order for the earth to be covered completely. The second aim of giving birth to Sky is to produce a firm seat for the gods. This line conflicts a little with lines 117/18, where Earth has already been described as
22 In 517–20 we read that Atlas supports broad Ouranos at the border of the earth, near the Hesperides (that is: in the far west), with his head and hands. This sounds like heaven was conceived as a parallel level above the earth, kept aloft by Atlas. Cf. West 1966: 311 on line 516, who cites Odyssey 1. 52–54, where Atlas lives in the sea and supports long columns ‘which keep heaven and earth apart’ (1. 54).
Categorization in Hesiod’s Theogony
29
the seat of the Olympic gods.23 Perhaps the close relation between Earth and Sky expressed here demonstrates that it does not make too much difference whether the sky or the earth is called the seat of the gods. This is all the more true because the sky is a fatherless child, i.e. a secondary phenomenon or different aspect of the earth. Sky later becomes a prolific sexual partner of the Earth (132–153) and plays a major role in the first scene of the so-called succession myth. But before Hesiod lists the children of Earth and Sky, he names two major structures of the earth to which she also gives birth: the mountains and the sea. Like earth and sky, the mountains predominantly figure as homes of the Nymphs, lesser goddesses who are, in Hesiod’s time, closely connected with trees and springs.24 It is natural for a Greek living in an area very much characterized by mountains and valleys to believe that mountains appear in a very early cosmogonic stage. Much the same holds for the sea, which has always been a vital part of Greek life.25 Mountains and the sea give a distinct structure to the earth and are therefore rightly listed among the first offspring. Also Okeanos is briefly mentioned (133), the circular freshwater river surrounding the earth at its outer boundary. This concept and the name seem to have been taken over from the East.26 Like the sky itself these geological phenomena can be understood as aspects of the earth. This is underlined by the fact that all of these entities (except for Okeanos27) are produced without a father. So far in this cosmogony, there seems to be a tendency to move from very indistinct and murky first beginnings (Chaos) towards more defined structures. From line 134 onwards, there follows a list of various children who are the result of a sexual encounter between Earth and Sky. The connection has been prepared for by Okeanos, who is the first common child of Earth and Ouranos. I mention these offspring only in passing, because they are mostly not physical entities (like earth, sea, heaven etc.), but early personalized ancestors of the anthropomorphic gods whom their father Ouranos summarily calls ‘Titans’ (Τιτῆνες, 20728). Most of these names occur only rarely in later Greek mythology or religion and do not have cults.29 Hesiod
23 Cf. note 12 above. 24 They are named in a long list as the daughters of Okeanos and Tethys in 346–366, although there the word ‘Nymph (νύμφη)’ does not occur. 25 Cf. Lesky 1947; Schulz 2005. 26 Cf. West 1966: 201. Herodotus already criticizes the poets (‘Homer or someone of the earlier poets’) for having invented such a river, which does not really exist: II 23. In IV 8 Herodotus shows that the sea outside the columns of Heracles originally was thought to be this Okeanos, which therefore later became the name of the (salty) Atlantic ‘Ocean’. 27 It is not clear why Okeanos has a father. Probably this has to do with Ouranos’ character as a boundary of the earth, which also applies to Okeanos. 28 With a (probably wrong) etymology added (209 f.): ‘He said that they performed a great deed stretching out (τιταίνοντας) in wickedness:’ The true etymology is obscure, cf. West 1966: 200. 29 Cf. West 1966: 200 f. on 133. He assumes them to be taken over from the Orient as ‘the former gods’, who will be defeated by Zeus and condemned into Tartarus, cf. 664–735.
30
Jochen Althoff
seems to have introduced these figures here in order to have a generation of very early gods, from whom the more common gods of later Greek religion come forth. Especially notable is Rheia, who will (453–458) give birth to Zeus after having married Kronos, her youngest and most important brother. The installation of the kingdom of Zeus and his justice is the final stage of Hesiod’s poem and actually the state of the world at the time of Hesiod. This throws a light on the twofold character of Hesiod’s Theogony. On the one hand, it describes the coming to be of the important structures and abstract forces that constitute the world. On the other hand, it narrates a lively story of the successive generations of anthropomorphic gods, who they are and why they fight with each other. This second feature is much nearer to the narrative epos as we have it in Homer’s Iliad. It is crucial to note that both strands are united by means of the genealogical model. Most plausibly, the genealogical model originated in narrative epic as a tool to identify persons and to explain their heroic status by connecting it to the heroic status of their ancestors. Hesiod’s innovation lies in the transfer of this model to more abstract things and forces that constitute the world. Kronos plays the most important role in the first stage of the succession myth and is named as the last and youngest child in this list (137). From 139 until 210 he figures prominently in the first phase of the succession myth, using a newly invented30 metalsickle to castrate his father Ouranos. I have already referred to this as the remains of an old separation myth. With line 211 after the first part of the succession myth, Hesiod returns to a more descriptive part: Νὺξ δ’ ἔτεκε στυγερόν τε Μόρον καὶ Κῆρα μέλαιναν καὶ Θάνατον, τέκε δ’ Ὕπνον, ἔτικτε δὲ φῦλον Ὀνείρων. δεύτερον αὖ Μῶμον καὶ Ὀιζὺν ἀλγινόεσσαν οὔ τινι κοιμηθεῖσα θεῶν τέκε Νὺξ ἐρεβεννή, Ἑσπερίδας θ’, αἷς μῆλα πέρην κλυτοῦ Ὠκεανοῖο χρύσεα καλὰ μέλουσι φέροντά τε δένδρεα καρπόν· καὶ Μοίρας καὶ Κῆρας ἐγείνατο νηλεοποίνους, ... ... αἵ τ’ ἀνδρῶν τε θεῶν τε παραιβασίας ἐφέπουσιν, οὐδέ ποτε λήγουσι θεαὶ δεινοῖο χόλοιο, πρίν γ’ ἀπὸ τῷ δώωσι κακὴν ὄπιν, ὅστις ἁμάρτῃ.
30 161: αἶψα δὲ ποιήσασα γένος πολιοῦ ἀδάμαντος, / τεῦξε μέγα δρέπανον καὶ ἐπέφραδε παισὶ φίλοισι. ‘At once she created an offspring, of gray adamant, and she fashioned a big sickle and showed it to her dear sons.’ (Most 2006). Metals, according to an old concept, are born in the earth or grow on it like plants, cf. West 1966: 215 on line 161.
Categorization in Hesiod’s Theogony
31
τίκτε δὲ καὶ Νέμεσιν πῆμα θνητοῖσι βροτοῖσι Νὺξ ὀλοή· μετὰ τὴν δ’ Ἀπάτην τέκε καὶ Φιλότητα Γῆράς τ’ οὐλόμενον, καὶ Ἔριν τέκε καρτερόθυμον. Night bore loathsome Doom and black Fate and Death, and she bore Sleep, and she gave birth to the tribe of Dreams. Second, then, gloomy Night bore Blame and painful Distress, although she had slept with none of the gods, the dark Night, and (she bore) the Hesperides, who care for the golden, beautiful apples beyond glorious Ocean and the trees bearing this fruit. And she bore (a) Destinies and (b) pitiless punishing Fates, … … who (b) prosecute the transgressions of men and gods; and the goddesses never cease from their terrible wrath until they give evil punishment to whoever commits a crime. Deadly Night gave birth to Nemesis (Indignation) too, a woe for mortal human beings; and after her she bore Deceit and Fondness and baneful Old Age, and she bore hard-hearted Strife.31 Hesiod here switches back to the first generation after Chaos, which he had described in line 123. Night itself was a somewhat sinister offspring of the dark and gloomy Chaos, and this negative character is inherited by nearly all her children. Loathsome Doom, black Fate and Death are all elaborate aspects of Death which in epic diction are nearly synonyms. Moros comes from the verb μείρομαι ‘to allot, assign to’ and expresses the assigned time or fate of death.32 Ker is the feminine goddess of doom in a personal sense, Thanatos is her male counterpart and also the most abstract way of expressing death. It is not entirely clear if Hesiod had a concept of synonyms, but more probably he did not. Therefore he lists all the words that circumscribe the aspects of death as three different children of Night. And although to us death mostly is an abstract phenomenon (although it has very concrete notions), Hesiod’s way of describing and understanding it again has very strong personal notes. After Death in all its aspects, Night produces Sleep and the tribe of Dreams. Night and Death are a famous pair of brothers in early epic33 and Hesiod returns to these two later in his Theogony (758–766), where their character becomes more evident. He
31 This translation is taken with only one change (220: ‘hold fast to the transgressions’ Most) from Most ([2006], except for the brackets around lines 218 and 219). 32 In later Greek thought, this is assigned to the Moirai, who are named in 217. 33 E.g. Iliad 14. 231 (Hera flies to Hypnos): ἔνθ᾿ Ὕπνῳ ξύμβλητο, κασιγνήτῳ Θανάτοιο, ‘... ‘there she met Hypnos, the brother of Death, ...’.
32
Jochen Althoff
says that they both live under the surface of the earth near dark Tartaros and that the sun never looks upon them. Sleep comes up and moves around the earth and the sea at night, calmly and sweet for men. But the heart of his brother Death is made of iron, hard and never moved by pleas. Whenever Death gets hold of someone, he never lets him go again. All these features demonstrate his strong and inescapable, never faltering grip. This makes him an enemy even of the immortal gods, whose minds can be changed by prayers and offers. This is a marvelous piece of poetry. We learn that Sleep is a child of night, because men and gods sleep at night. The relation is one of synchronicity. But sleep’s character is not like his parent. Sleep is calm and sweet to humans, he is welcome as a time of rest and recreation. This, once again, is the principle of contrast which we already found at the beginning when day and light came to be out of night. Death, on the other hand, is of course not restricted to nighttime. He shares the sinister, horrible character of his parent and even enhances it, because he is without any exception horrifying and unwelcome to everyone; he never has a calm and positive side to him. Like a hunter he hunts his prey. Once he captures it, he never lets go again. Even the gods, who do not have to fear death, because they are immortal, hate death because of his pitiless character. Like Moros, Ker and Death as three different aspects of Death, Death and Sleep represent two different aspects of Night: one positive, the other one negative in terms of emotional value. Whereas the relation between night and sleep is mostly one of synchronicity, the same holds for dreams. We dream when we sleep and therefore the Dreams (understood as single animals of a tribe) are also children of Night. It is a little surprising that Hesiod at this point mentions not only one dream or a god of dreams (which does not exist in Greek mythology), but a whole tribe of similar entities.34 In terms of classification, this means that a whole class of things is described as related to the parent Night. In this respect, this example comes closest to a hierarchical order of certain groups of beings. We have already encountered such a way of speaking with regard to the mountains in line 129 and will find the same shortly in the case of rivers and Nymphs. It presupposes groups of things for which a common name exists, and where it does not make sense to focus on each and every single item of the group. In the case of dreams, this is caused by the sheer number of different dreams all individuals dream. It did not seem feasible to Hesiod to go into further detail, although he could have distinguished e.g. nightmares and sweet dreams or other sub-groups. All these children of Night are born without a father, just like the rest of her offspring. A second group, consisting of Blame (Momos) and painful Distress (Oizyn), gives us a hint why this is so. Like death, these abstract forces can be understood as negative, destructive phenomena of human society. Distress is a consequence of hateful Blame among humans. The anti-social effects can best be explained as off-
34 Here I follow a remark by West 1966: 198 f. on line 129.
Categorization in Hesiod’s Theogony
33
spring not of a sexual encounter involving love, but of a single parent without any partner. Dire adjectives (‘dark’, ‘painful’, ‘loathsome’) stress the negative character of these children. We remember that in the case of Earth’s first children another reason obtained. They seemed to be fatherless because they were structures and aspects of the earth. The attribute of lacking a father has different meanings in different families. Like the positively connoted sleep, the Hesperides do not share the negative sides of their siblings. They fit in here because they guard the famous golden apples far in the west (hespera means ‘evening’) in a land even beyond the Okeanos, which is the circular big river forming the outermost border of the flat disk of the earth (born 133 as a child of Earth and Sky). Here we learn in passing that Okeanos even has an outer bank where famous figures of Greek mythology are located. It is a kind of paradise, which is normally closed to mortals. The Hesperides, then, are children of Night, because night begins when the sun sets in the west. The relation between children and parent is one of local identity or neighborhood. Further dark and fearful children follow (217). Their names sound like Doom and Fate in line 211 (Moros and Ker), only that here we have very similar plural forms: Moirai and Keres.35 They, nevertheless, seem to be different forces: Doom and Fate represent individual death, whereas the plural groups have a different function. They are a kind of moral police for humans and gods. Without any pity they punish every transgression and crime, and fiercely prosecute their revenge until it is fulfilled. In classical Greek mythology, this is the job of the Erinyes,36 but here it is done by Fate and Doom (the plural means that each offender or even each offence has its own doom and fate). Hesiod’s idea seems to be that every crime inadvertently by law of fate leads to punishment. The stubbornness with which Moirai and Keres prosecute wrongdoers reminds us of the stubbornness of death at the beginning of this passage, and this seems to be considered a negative trait of character by Hesiod. It is a negative feature inherited directly from Night. Nemesis has been translated by ‘indignation’, but it is much more ‘just retribution’. The word stems from νέμω ‘to assign, allot to’ (like μόρος above from μείρομαι). It is basically the same as the Moirai and Keres just mentioned. Like Night, which here is given the epithet oloe ‘destructive, deadly’, Nemesis is an evil for mankind. The character of parent and child are the same. The same holds for Deceit (Apate), accursed or wretched old age and hard-hearted Strife (Eris). Whereas old age is more a private
35 It is doubtful if lines 218 and 219 really belong here, because they also appear in 905 f. in a more appropriate context. They contain the correct names of the classical Moirai, but it seems that Hesiod here presents a different concept of these Moirai. Cf. West 1966: 228 f. on line 216 and on 217 and 218 f. 36 The Erinyes are mentioned 185 as offspring from the blood-drops that fell to earth after the castration of Ouranos by Kronos. They are practically the same as the Keres here.
34
Jochen Althoff
hardship, deceit and strife, like blame and distress above, destroy human society and are therefore considered negative forces. Quite surprisingly, Love (Philotes) is mentioned as a child of Night, too. This abstract force does not belong to the series, because it is mainly a positive force. In Hesiod’s pragmatic view of life, deceit and love belong together (cf. 205f.), because lovers tend to deceive each other and themselves. So Deceit and Love form a pair like Sleep and Death. Thus, love has a slight tendency towards a negative force, but this is not what links it to its parent Night. Love and sex happen at night, and therefore, like sleep above, Love is similar to its parent in terms of synchronicity. If we look at this passage in general, the most important feature is that Hesiod no longer divides major geological or cosmic structures into groups. Instead, he designs a single big group of abstract factors and forces which exert an influence on human beings as individuals and members of a society. This is most remarkable, because men have not yet come into existence at this point of cosmogony. But Hesiod very often applies the perspective of his own age to the beginnings of the world and thus tends towards somewhat anachronistic statements. If one had asked him about this, he would certainly have answered that there is no contradiction in assuming the abstract forces to develop before mankind. We find certain repetitions within this group, or at least unclear divisions, e.g. Moros, Ker and Death, which all seem to be aspects or mere synonyms of death. The same with Moiras and Keres and Nemesis, who all punish crimes and to us at least seem to be synonyms. Hesiod indicates that these entities are very similar, because they are siblings. Most entities of this group share a negative character from the perspective of humans and even of the gods, which is also a consequence of their genealogical proximity. Only sleep and love have a positive connotation; their relation to their parent is not one of emotional value, but of synchronicity. Dreams and the Hesperides do not have a value in themselves or are of ambivalent nature (dreams can be nightmares and the opposite); they are related to night by time and location. The anti-social force of most members of this group seems to be stressed by the parthenogenetic way of birth. Another interesting passage has the following lines 226–232, which present the offspring of Strife (Eris), mentioned as the last child of Night in 225: Αὐτὰρ Ἔρις στυγερὴ τέκε μὲν Πόνον ἀλγινόεντα Λήθην τε Λιμόν τε καὶ Ἄλγεα δακρυόεντα Ὑσμίνας τε Μάχας τε Φόνους τ’ Ἀνδροκτασίας τε Νείκεά τε Ψεύδεά τε Λόγους τ’ Ἀμφιλλογίας τε Δυσνομίην τ’ Ἄτην τε, συνήθεας ἀλλήλῃσιν, Ὅρκον θ’, ὃς δὴ πλεῖστον ἐπιχθονίους ἀνθρώπους πημαίνει, ὅτε κέν τις ἑκὼν ἐπίορκον ὀμόσσῃ. And loathsome Strife bore painful Toil
Categorization in Hesiod’s Theogony
35
and Forgetfulness and Hunger and tearful Pains, and Combats and Battles and Murders and Slaughters, and Strifes and Lies and Tales and Disputes, and Lawlessness and Recklessness, much like one another, and Oath, who indeed brings most woe upon human beings on the earth, whenever someone willfully swears a false oath.37 Because Strife has been defined as an anti-social and negative force, all her offspring share the same character. In the case of Lawlessness and Recklessness, this is explicitly stated: they are related or akin to each other.38 Painful Toil (Ponos), Forgetfulness (Lethe), Hunger (Limos), Pains (Algea), Fighting of all sorts (Hysminas te Machas te39), Murder (Phonos), Lawlessness (Dysnomie) and false Oaths (Horkos) are among others named as children of Strife (once again often in the plural, which makes them a subclass like the dreams above). These are mostly the effects and consequences of strife. One forgets to take care of one’s daily work and to cultivate the fields. This in turn leads to poverty, hunger and pain. Fighting and war with the killing of people is quite often the consequence. And once again these consequences are relevant mostly for human beings, certainly not for gods, who do not have to work the fields to earn their living. In the last section (211–225), we found a mixture of mythological persons (Momos, Hesperides, Moirai, Keres etc.) and more abstract nouns. Thanatos and Hypnos can be disputed, because they are also mythological figures, but Love and Old Age perhaps tend more to the abstract side.40 But in 226–232, we almost exclusively find abstract nouns with little personal potential. This distinction, however, is modern and becomes blurred in Hesiod. It is another consequence of his genealogical model that he treats abstract nouns and personal gods on the same level. In line 233 Hesiod returns to Pontos, the Sea, last mentioned in line 132 as a son without father of the Earth. The eldest son of Pontos is Nereus, a very gentle and reliable god of the sea who in classical Greek mythology functions also as a fortune-teller. Whereas Nereus’s brothers and sisters are horrible sea monsters, Nereus represents the helpful and welcome aspects of the Mediterranean Sea. Nereus himself has no less than 50 daughters with Doris, a daughter of Okeanos, who are all born within the
37 Translation by Most 2006. 38 This is at least what the adjective synethes signifies later. Here it may be understood ‘probably in the sense of ‘neighbours’, having common haunts (ἤθεα)’ (West 1966: 232 on 230). 39 To modern ears these are mere synonyms, just like ‘Murders and slaughters’ in the same line. It remains unclear whether this holds for Hesiod, too. 40 But West in his commentary rightly reminds us of vase paintings that represent Deceit and Old Age as persons; the same holds true e.g. for Hunger of line 227: 1966: 230 f. on lines 224, 225, 227.
36
Jochen Althoff
sea and live there. In a marvelous catalog, Hesiod names them from 243 until 262, 50 names in perfect hexameters, a masterpiece of catalog-poetry.41 Let me just pick a few examples to show the common features of this group: Eudore (‘the giver of fine presents’, 244), Galene (‘the calmness of the sea’, 244), Glauke (‘the blue colour of the water’, 244), Speio (‘the cave’ at the seaside, 245), Pherousa (‘the carrier’ of the ships, 248), Nesaia (‘belonging to the island’, 249), Aktaie (‘belonging to the cliff’, 249),42 Kymodoke (‘she who takes in the waves’, 252), Kymatolege (‘she who soothes the waves’, 253), etc. These daughters all represent aspects of the sea, be it the water or the beach and cliffs where the sea borders on land, or the islands in the sea, the sea as carrier of the ships, which were the fastest means of travel, and so on. And whereas the ancient Greeks mostly did not like the sea because of its dangerous instability, they used it all the time for traveling and fishing.43 The daughters of the mild and gentle Nereus all share their father’s character: they are mild and gentle, too, and represent the friendly face that the sea can show to men depending on it. What modern scientists would describe as positive attributes of the sea with regard to men, Hesiod puts into one genealogical group, namely the daughters of Nereus. Here again we find the tendency to personalize abstract features of a major geological structure and present them as a group of siblings. But it is one step further in terms of abstraction. The Sea as a geological reality may quite naturally be understood as a personified deity. But it is something different to summarize the positive aspects of the Sea as a number of daughters of the parent Pontos. The forming of groups in Hesiod’s poem is obviously not restricted to geological realities, but also to clusters of attributes and abstract forces that determine human society. It has to be admitted, though, that not all the names listed as the daughters of Nereus are connected to the sea.44 In cases such as Eukrante (‘good sovereign’, 243), Proto (probably ‘fortune-teller’, cf. Proteus, another old man of the sea), and Poulynoe (‘with much reason, wise’, 258), the names are much too general. They retain general character traits of their father Nereus. Here the systematic coherence of the
41 Catalog-poetry forms a recurring part of narrative epic, e.g. the catalog of ships, Iliad 2. 484–760, where all the Greek troops before Troy are listed with their ships and generals. These catalogs aim at impressing the recipient by the sheer number of items listed and demonstrate the wealth, strength etc. of a party. 42 West comments 1966: 239 on 249: ‘the Nereids live in the sea, but they are worshipped on shores and islands.’ 43 Cf. Lesky 1947: Kap. 1: ‘Ursprüngliche Seefremdheit der Griechen’, 1–7, about Hesiod: 31–37. Hesiod himself offers some of the most important testimonies for this ambivalent attitude, e.g. Erga 236 f.: In the just city, people do not have to sail the sea to earn their income; Erga 618: ‘But if desire for stormtossed seafaring (ναυτιλίης δυσπεμφέλου, lit.: ‘hard, rough, distressing’) seize you …’ West in his commentary (as note 21), 313 f. cites additional passages from other authors for this distaste. 44 In some cases it is not even clear whether what we read is a name or an adjective, e.g. 245: Σπειώ τε θοὴ Θαλίη τε (‘Speio and swift Thalie’) or, with different word-division: Σπειώ τε Θοή θ’ Ἁλίη (‘Speio and Thoe and Halie’). In other cases, the reading of the name is uncertain.
Categorization in Hesiod’s Theogony
37
group is a little loose, although for ancient Greek thought all these features belong to the sea. Whereas in the passage 226–232 the children of Eris are just abstract nouns used as personal names, Hesiod here has invented most of the ca. 50 names, which are no abstract nouns, but composed like Greek personal names.45 Nevertheless, they are all telling names whose connection to the sea can most of the time easily be detected. This, then, is a third strategy of naming, besides the traditional mythological names (or names of gods) and the personalization of abstract nouns. One last example may be given. After having spoken about the birth of a whole bunch of monsters (265–336: e.g. the Graiai, Gorgones, the snake Echidna, Chimaira, mixed out of three different animals, the Sphinx), from line 337 onwards Hesiod lists the sons of Tethys and Okeanos. Tethys and Okeanos are children of Ouranos and Gaia (133 and 136), and accordingly their children belong to the great geological structures of the world. The sons of Okeanos and Tethys are all the rivers on earth, and 25 of them are named in the passage 337–370, once again in the form of a catalog of names (like the Nereids). It is difficult to say why Hesiod chose just the rivers he names. From lines 367–70 it becomes clear that he is only giving examples, not a complete list, because there he says: τόσσοι δ’ αὖθ’ ἕτεροι ποταμοὶ καναχηδὰ ῥέοντες, υἱέες Ὠκεάνου, τοὺς γείνατο πότνια Τηθύς. τῶν ὄνομ’ ἀργαλέον πάντων βροτὸν ἄνδρα ἐνισπεῖν, οἳ δὲ ἕκαστα ἴσασιν οἳ ἂν περιναιετάωσιν. There are just as many other rivers (sc. as there are Okeanides) flowing with a roar, sons of Okeanos who bore Lady Tethys. It is hard for a mortal man to name them all, but those who live around each of these (rivers) know it well. Moreover, we cannot be certain which rivers he knew from his own experience and which he had only heard of. The river Phasis, for example (340), is from very early times onwards connected with the story of the Argonauts who sail in search of the Golden Fleece, located near the river Phasis.46 Later it was identified with the modern
45 Cf. West 1966: 236: ‘Of the individuals named by Hesiod, Thetis and Amphitrite at least, and perhaps others, had an independent cult, and cannot have been his own invention. But the majority appear to be invented, whether by him or by a predecessor (…). It is sometimes said that they express different aspects of the sea; better to say that most of them are suggested by the sea.’ 238 he refers to Bechtel 1902: 73–75, who shows that Nereid names were used by mortal women. 46 Our only complete literary source is the Hellenistic epos Argonautika by Apollonios Rhodos, but the Argo (as the first ship invented just for this purpose) is already mentioned in the Odyssey 12. 70.
38
Jochen Althoff
river Rion in Georgia. Most probably Hesiod knew that river only as a literary reminiscence, but did not know its exact whereabouts. This points to a further problem. Quite often the exact location of rivers of early epos or mythology is not clear. The common localizations found in the lexica and handbooks mostly go back only to Hellenistic times, when scholars tried to identify them with known real rivers.47 But after all, rivers share a common outlook (and sweet water as opposed to the salty sea-water) and it makes good sense to describe them as sons of the big river Okeanos. The sons of Okeanos and Tethys are complemented by an equally great number of daughters (346–366). Again a catalog of 41 names is given, but even these are only examples, for in line 364 Hesiod says that 3000 daughters exist. Their identity is not absolutely clear, but West has made a strong point for them as Nymphs of springs and groves.48 The catalog is made up like that of the Nereids (and sometimes the same names recur49), but the names are not clearly connected to water or springs. Often they represent good wishes for the children which are reared by the Nymphs.50 That all sweet water springs stem (by subterranean connections) from Okeanos is the underlying assumption, though not expressed by Hesiod. He defines the group by stating its responsibility for the nurturing and upbringing of children on earth (346f.), which is a traditional task of the Nymphs. Once again, their importance for human beings is stressed before the existence of human beings. It may sound puzzling to a modern reader that the obvious connection between rivers and springs is not mentioned (except for the fact that Nymphs and rivers are siblings); after all, every river starts with one or more springs. But for Hesiod it must have been more prominent that the rivers all start from Okeanos and the springs are separate entities, though they share the same quality of water. Further examples could be given, but I will stop here for a conclusion.
47 An example is the well-known Eridanos (Ἠριδανός), the river where Phaeton died after having burnt almost the whole world with the sun-car stolen from his father Helios. Since Pherekydes of Athens (who wrote after 508/07 BC, fr. 3 F 74 Jacoby) it has been identified with the Po in northern Italy. But the Rhone in France has also been proposed (Aeschylus fr. 73 a Radt). Cf. West 1966: 261. 48 This would be made clearer by reading τίκτε δὲ Κουράων ἱερὸν γένος in 346 (‘they bore the holy race of Nymphs’) instead of τίκτε δὲ θυγατέρων ἱερὸν γένος (‘they bore the holy race of daughters’) as all the manuscripts do (a conjecture by West 1966: 263). 49 E.g. Doris (350) is the name of the mother of the Nereids (241) and a Nereid as well (250); Eudore (360) is also found in 244; Polydore (354) is similar to Eudore (360 and 244). 50 E.g. the just mentioned Doris is the ‘giver of goods’, Idyia (352) is ‘the one looked at’, like Klymene (351) ‘the famous’, Plouto (355) and Tyche (360) refer to ‘wealth’, Zeuxo (352) to marriage (zeugnymi ‘to couple’), and Metis (358) to wisdom. Amphirhoe (‘flowing in a circle’, 360) and Okyrhoe (‘flowing fast’, 360) pick up qualities of their father Okeanos.
Categorization in Hesiod’s Theogony
39
Conclusion In terms of modern natural science, Hesiod’s Theogony may not be very convincing. This impression is caused by the very wide range of entities which he puts into genealogical interrelations. But actually he never says that he is going to write a history of nature and he most probably does not even have a concept of ‘nature’.51 What he is interested in is reality (although he does not have that word either), everything which, to his eyes and his mind, is real in the world he knows. This includes everything that we today would call nature, but also the many gods of Greek mythology and religion. He even takes abstract forces into account, which dominate the realm of men (and partly of the gods, too): love, hate, death, strife, fighting, etc. His perspective in doing so is human throughout. Even when looking at the attributes of the sea or the rivers and springs, he is mostly concerned with their relevance for humans. His interest is not so much historical, although he writes about how the world came to be from the very beginnings until his own time. His interest is structural: he wants to understand the inner relations of things, attributes and forces in order to explain their importance for men. The model he uses to describe structure is the genealogical model. He takes this model not only from the observation of human reality, but also from the literary track in which he moves. Epic poetry is a means of documenting a glorious past of superhuman heroes and by that reassuring the value of aristocratic individuals as the main audience. Genealogy plays a central role within epic: the acting heroes feel the obligation to their forefathers’ heroic deeds which they have to live up to.52 In the same way, the aristocratic audience defines its own role as descendants of these same heroes which for them was a historical truth. Hesiod transfers this genealogical model and adapts it to his specific purpose. By explaining the world in terms of begetting and births, he establishes relations between the entities he focusses on. These relations are primarily taken over from the human sphere of parents and children, namely, similarity as the normal relation and opposition as a deviant mode of this. But even relations such as synchronicity and local neighborhood are quite naturally taken over from the human sphere. The
51 Not even the word: φύσις, the standard term for ‘nature’ becomes frequent only in the 5th cent. BC. Homer’s Odyssey 10. 303 is the only example before the 5th cent. (φύσις φαρμακοῦ: Hermes gives the magic Moly-plant to Odysseus to save him from the spell of Kirke; the meaning is ‘the grown form’ of the plant [the substantive stems from the verb φύεσθαι ‘to grow’]). Cf. Heinemann 2001. The classic study is Heinimann 1945: 92. 52 Just one famous example may suffice. When Glaukos meets Diomedes on the battlefield before Troy, he identifies himself by recalling his famous ancestors (Iliad 6. 150–211), until he reaches his father Hippolochos, who sent him to Troy, advising him thus (6. 208): ‘Always be the first and outstanding above the others, and never bring to shame the line of ancestors (γένος πατέρων), who were by far the best men in Ephyra and in broad Lykia. Of this ancestry (ταύτης τοι γενεῆς) and blood I am proud to be.’ Cf. Grethlein 2006: 63–84.
40
Jochen Althoff
lifetime of children and parents necessarily overlaps, and the birth and growing up of children normally take place at the same location. Even when Hesiod constructs relations such as that between starting point and consequences (Eris and her children), this follows naturally from the relations in the human sphere, because children are necessarily the consequences of their parents’ acts. And as children and parents in human life form a family with manifold individual characters, but a single identity, so the sea and its different attributes can just as well be taken as a family with a local identity and many different, but similar characters. Taking over a genealogical model to describe reality is much easier for Hesiod than for us, because he understands most of the things around him as persons or anthropomorphic gods. Even the abstract forces or natural states, like darkness, light, love, hate, death, sleep, etc., are understood as persons or gods. Once again Hesiod’s systema naturae (if I may say so) includes many more entities than anything we would subsume under the term nature, but they all nevertheless form a unity by personification and relations. We still miss a systematic, logical approach and some more details. We hear of the major structures of the world: earth, sea, sky, mountains, rivers, etc., but we lack explanations for e.g. living beings such as plants and animals. This is due to the macroscopic view typical of a cosmogonic approach. Hesiod is content with having explained the macrostructure of the world, including all the gods (per se mighty players) and forces that are relevant for him. Although he always takes the perspective of a human, he never even explains when or how human beings entered the scene. But the basic idea of reducing the colourful multitude of phenomena to only one single beginning, Chaos, has, since Hesiod, been an attractive model for centuries. The same holds true for his idea of structuring and explaining the world in terms of a genealogical model. I finally would like to look very roughly at some points of the further career of this method of categorization. Just a few lines above, I used the expression systema naturae to describe Hesiod’s overview of the world. This is the title of a famous book by the natural scientist Carl von Linné (or in Latin: Linnaeus, 1707–1778); the complete title is: Systema naturae sive regna tria naturae systematice proposita per classes, ordines, genera, et species, Leiden 1735. The Latin means ‘System of nature or the three kingdoms53 of nature, systematically put forth by classes, orders, genera, and species’. This very influential work in its 10th edition became the basic instrument for modern botanical and zoological classification.54 Linné started with assigning two-part names to plants (earlier, now called binary names) and animals (later, now called binomes),
53 By this he means the animals, plants and minerals. 54 C. Linnaeus 1758 f., the Latin title translated: ‘System of nature through the three kingdoms of nature, according to classes, orders, genera and species, with characters, differences, synonyms, places’.
Categorization in Hesiod’s Theogony
41
of which the first part states the genus, the second one the species (e.g. homo sapiens for human beings: homo as the genus comprises also distinct species such as homo Neanderthalensis; sapiens refers to the modern species). For our context, it is more important, however, to know what kinds of categories Linné used. Of the five listed in the title (regnum, classes, ordines, genera and species), Linné believed the first three to be artificial, whereas he thought of genera and species as naturally given. Genera is the plural of the Latin word genus, which is etymologically the same as the Greek word genos (τὸ γένος, ‘race, stock, kin’),55 whose stem is also present in the word ‘genealogical’, by which the explanation of natural entities by Hesiod has been described above.56 In early epic, where the word genos is quite common, it mostly denotes ‘race, stock’ of a hero, also ‘offspring’,57 and is cognate to the verb gignesthai (γίγνεσθαι) ‘to become, come into being’. Starting from some more general usages in epic, such as ἡμιθέων γένος ἀνδρῶν (‘the race of half-divine men’, Iliad 12. 23), genos from the 5th century BC onwards was used as ‘class, sort, kind’.58 The use goes one step further, mostly in combination with eidos (εἶδος, ‘species’, since Aristotle).59 Genos there is the supreme class, under which a number of different eide (‘species’) is grouped. Aristotle then transferred this classification to the animal kingdom, within which he distinguishes seven ‘greatest classes’ (μέγιστα γένη): birds, fishes, cetacea, shellfish, soft-shelled animals, soft animals and insects.60 Besides these groups, he positions men as a separate group with only one species, and also the four-footed viviparous and oviparous animals. Of the viviparous part of the latter group, he says: ‘There
55 Cf. Beekes 2010: 273 (s.v. γίγνομαι); de Vaan 2008: 260 f. (s.v. gigno, -ere). 56 The verb genealogeo (γενεαλογέω) is already used by Herodotus (2. 146), meaning ‘to trace a pedigree’. The noun genealogia (ἡ γενεαλογία) is found in Plato, Cratylus 396 C 4, where Socrates etymologically explains the names of Ouranos. Then he says: ‘Could I only remember the genealogia of Hesiodos, which other ancestors of these (sc. mentioned gods) he names, I would not stop to show how adequate their names are.’ This, then, is the first use of the word to describe Hesiod’s Theogony that we know of. 57 Cf. note 52 above, where genos and its synonym genea are used. Note that the phrase ‘genos of fathers’ already shows a slight tendency towards the more abstract use ‘class’. An example of ‘offspring’ is given in Iliad 19. 124. 58 E.g. Plato, State VI. 501 E 2 f.: ‘the philosophical class’; Timaeus 17 C 6–8: ‘the class of farmers opposed to the class of warriors’; Xenophon, Cynegetica 3. 1: ‘races of dogs’. 59 Cf. Aristotle, Metaphysics K 1. 1059 b 36f.: τὰ γένη δ’ εἰς εἴδη πλείω καὶ διαφέροντα διαιρεῖται. ‘The gene are divided into more and different species.’ 60 Historia animalium I 6. 490 b 8–14. The Greek names are: ὄρνιθες, ἰχθύες, κῆτοι, ὀστρακόδερμα, μαλακόστρακα, μαλάκια, ἔντομα. These groups are partly similar to the six established by Linné: 1. Mammalia (missing in Aristotle, but partly overlapping with the tetrapodes), 2. Aves (birds), 3. Amphibia, 4. Pisces (fishes), 5. Insectes (insects), 6. Vermes. Some problems are connected with these groups, cf. Zierlein 2013: 67f. and the commentary 236–240. Cf. Kullmann 1998: 167–171. Inaccurate in this respect is the otherwise useful article ‘Systematik’, in: Toepfer 2011, Bd. 3: 443–468, esp. 444 f.
42
Jochen Althoff
are many kinds (eide) included in the group (genos) of viviparous quadrupeds …’61, and the same holds true for the other groups, although normally a threefold classification can be observed: megiston genos (group), genos (sub-group), eidos (species). Whereas the term eidos (‘species’) has been taken over from Plato’s theory of ideas, the term genos certainly has its roots in the genealogical way of thinking as represented in Homer’s epic and, most prominently, in Hesiod’s Theogony. Linné’s system is inspired by Aristotle,62 and Aristotle’s use of the term genos finally reaches back – by way of Plato – to Hesiod’s first attempt to classify the entities of the world. It is an irony of the history of ideas that Charles Darwin (1809–1882) returned to the genealogical model in its basic form. For him the manifold species we find in nature are the result of a real genealogical evolution from earlier ancestors. In his view ‘all true classification [is] genealogical’.63
References Bechtel, Friedrich. 1902. Die attischen Frauennamen nach ihrem Systeme dargestellt. Göttingen: Vandenhoeck & Ruprecht. Beekes, Robert. 2010. Etymological Dictionary of Greek. Vol. I. Leiden/Boston: Brill. Blümer, Wilhelm. 2001. Interpretation archaischer Dichtung. Die mythologischen Partien der Erga Hesiods. Bd. 1: Die Voraussetzungen: Autoren, Texte und homerische Frage. Münster: Aschendorff. Boys-Stones, George R. 2010. “Hesiod and Plato’s history of philosophy.” In Plato and Hesiod, edited by George, Boys-Stones/J. H. Haubold. Oxford: Oxford University Press. Bremer, Dieter. 2013. In: Flashar, Bremer, Rechenauer (2013) Darwin, Charles. 1859. On the Origin of Species. London: John Murray. de Vaan, Michiel. 2008. Etymological Dictionary of Latin and the other Italic Languages. Leiden/ Boston: Brill. Diller, Hans. 1946. “Hesiod und die Anfänge der griechischen Philosophie.” Antike und Abendland 2, 140–151. (repr. in: Heitsch [1966], 688–707) Flashar, Hellmut, Dieter Bremer and Georg Rechenauer (eds.). 2013. Die Philosophie der Antike Bd. 1: Frühgriechische Philosophie. (Grundriss der Geschichte der Philosophie) Basel: Schwabe. Grethlein, Jonas. 2006. Das Geschichtsbild der Ilias. Eine Untersuchung aus phänomenologischer und narratologischer Perspektive. Göttingen: Vandenhoeck & Ruprecht. Heinemann, Gottfried. 2001. Studien zum griechischen Naturbegriff, Teil I: Philosophische Grundlegung: Der Naturbegriff und die “Natur”. (AKAN Einzelschrift Bd. 2) Trier: WVT. Heinimann, Felix. 1945. Nomos und Physis. Herkunft und Bedeutung einer Antithese im griechischen Denken des 5. Jahrhunderts. Basel (Nachdruck Darmstadt 1987): Wissenschaftliche Buchgesellschaft.
61 Historia animalium I 6. 490 b 31f.: τοῦ δὲ γένους τοῦ τῶν τετραπόδων ζῴων καὶ ζῳοτόκων εἴδη μέν ἐστι πολλά ..., translation by Peck 1979: 33. 62 The details are much more complicated, cf. Toepfer 2011. 63 Darwin 1859: 499. Cf. also Toepfer 2011: 450 f.
Categorization in Hesiod’s Theogony
43
Heitsch, Ernst (ed.). 1966. Hesiod (Wege der Forschung Bd. 44) Darmstadt: Wissenschaftliche Buchgesellschaft. Koning, Hugo H. 2010. Hesiod: The Other Poet. Ancient Reception of a Cultural Icon. Leiden/Boston: Brill. Kullmann, Wolfgang. 1998. Aristoteles und die moderne Wissenschaft. Stuttgart: Steiner. Lesky, Albin. 1974. Thalatta. Der Weg der Griechen zum Meer. Wien: Rohrer. Linnaeus, Carl. 1758f. Systema naturæ per regna tria naturæ, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis, 2 vols. Stockholm. Lloyd, Geoffrey Ernest Richard. 1992. Polarity and Analogy. Two Types of Argumentation in Early Greek Thought. Bristol: Bristol Classical Press. Montanari, Franco, Antonios Rengakos and Christos Tsagalis (eds.). 2009. Brill’s Companion to Hesiod. Leiden/Boston: Brill. Most, Glenn W. 2006. Hesiod, Theogony, Works and Days, Testimonia, edited and translated by Glenn W. Most. Cambridge, Mass./London: Harvard University Press. Paulsen, Thomas. 2004. Geschichte der Griechischen Literatur. Stuttgart: Reclam. Peck, Arthur. L. 1979. Aristotle in 23 vols. Vol. IX: Historia animalium, I (books I-III), with an Engl. transl. by Arthur L. Peck. Cambridge, Mass./London (first imprint 1969). Philippson, Paula. 1966. “Genealogie als mythische Form. Studien zur Theogonie des Hesiod.” In Heitsch (1966), 651–687. Originally published as Fasc. Supplet. VII of the Symbolae Osloenses, 1936. Reichel, Michael. 2011. „Epische Dichtung – Homer.“ In Handbuch der griechischen Literatur der Antike, Erster Band: Die Literatur der archaischen und klassischen Zeit, edited by Bernhard Zimmermann. München: Beck. Schulz, Raimund. 2005. Die Antike und das Meer. Darmstadt: Wissenschaftliche Buchgesellschaft. Snell, Bruno. 61986. “Die Welt der Götter bei Hesiod.” In Die Entdeckung des Geistes, edited by Bruno Snell, 45–55. (repr. in: Heitsch [1966], 708–725). Göttingen: Vandenhoeck & Ruprecht. Toepfer, Georg. 2011. Historisches Wörterbuch der Biologie, Bd. 3. Stuttgart/Weimar: Wissenschaftliche Buchgesellschaft. West, Martin L. 1966. Hesiod, Theogony, edited with Prolegomena and Commentary by M. L. W. Oxford: Clarendon Press. West, Martin L. 1978. Hesiod, Work and Days, edited with Prolegomena and Commentary by M. L. W. Oxford: Clarendon Press. (repr. 1990) Zierlein, Stephan. 2013. Aristoteles, Historia animalium Buch I und II, übersetzt, hg. und kommentiert v. S. Zierlein. (Aristoteles, Werke in dt. Übers. Bd. 16) Berlin: Akademie-Verlag.
Orly Goldwasser
What is a horse? Lexical Acculturation and Classification in Egyptian, Sumerian, and Nahuatl Summary: The Egyptian, Mesopotamian, and Amerindian cultures all had to respond to a similar challenge: a new, ‘game-changing’ animal appeared in their worlds at some point in their respective histories. How was the new ‘item‘ analyzed and incorporated into the languages and scripts of these cultures? The Ancient Near Eastern material will be put in this article into the general framework of Lexical Acculturation theory, as it has lately been comprehensively defined by Cecil Brown in his book: Lexical Acculturation in Native American Languages.
1 The arrival of the horse in Egypt Archaeology tells us that horses were unknown in Egypt until the end of 17th century BC. As far as we know, the animal was introduced into Egypt during the early Hyksos period, as a few horse bones were discovered in Avaris, Tell el-Dab’a.1 Somewhat later, in first half of the 16th century, a mare, probably the beloved animal of one of the kings, was found in an almost intact burial within the Hyksos palace in Avaris (fig. 1). The careful burial in the palace itself behind the throne room reveals the importance of this specific animal for someone in the royal circles. The context of the first horses is to be expected — the Eastern Delta had been heavily settled since the end of the 12th dynasty by foreigners of various Levantine origins.2
1 A single bone was found in the 13th dynasty context. Boessneck and von den Driesch 1992: 24 f. The ‘Buhen horse burial’ is dated frequently in publications to the Middle Kingdom. This is, however, impossible as the horse was buried on top of the Middle Kingdom rampart pavement. This could have happened only after the fortress was abandoned for a very long period so that sand could have accumulated in such a way as to cover even the top of the wall. This horse dates most likely to the time of the re-conquest of Nubia at the beginning of the 18th Dynasty, see Bietak 2010a: 170. Vernus suggests that the horse may date to the Hyksos period, see Vernus & Yoyotte 2005: 535. On horses and chariots as new technology in Egypt, see Shaw 2012: 99–102 with bibliography. For possible circumstantial evidence for chariots already in the 13th dynasty, see Helck 1978. 2 On horses in Mesopotamia already in the 3rd Millennium, see recently Makowski 2014, with extensive bibliography. DOI 10.1515/9783110538779-003
46
Orly Goldwasser
Fig. 1: Horse burial in the Hyksos palace in Tell el-Dab’a
What is a horse? Lexical Acculturation and Classification
47
1.1 The arrival of the horse into the Egyptian lexicon and script The horse makes its appearance in the Egyptian lexicon in two forms. The earliest attestations (very end of the 17th Dynasty, the beginning of the 18th dynasty) reflect two different paths of admission into the recipient culture lexicon. On the first track, a new differentiated meaning is added to an old local lexeme. The second track shows the adoption of a foreign word, which will be activated as an integrated loanword in the Egyptian lexical system. Both tracks enjoy great success, and the words appear in hundreds of attestations in various Egyptian texts of various genres and registers from the 18th Dynasty on. In the script, two new conspicuous hieroglyphs appear: and . In a recent encompassing study, Pascal Vernus has offered a thorough analysis of the lexical history of the horse in Ancient Egypt. This work includes a detailed listing of occurrences, spelling variations of ‘horse words’, a discussion of their possible origins, genre analysis, etc.3
1.1.1 Arrival of the horse into the Egyptian lexicon through the lexeme ḥtr – referential extension4 Through the strategy of ‘insertion’, a new meaning referring to a new signified is embedded into an existing lexeme.5 How an embedding lexeme is selected is still a matter of debate, and different cultures show different paths.6 However, the lexeme that was chosen by the Ancient Egyptians to introduce the horse is ḥtr — ‘team of yoked quadrupeds’.7 A few additional Egyptian lexemes may refer to horses by extension, but they are rare cases.
3 See Vernus 2009. Special consideration will be given here to the classifier system. The Egyptian script contains an elaborate system of unpronounced graphemic classifiers that are shared by all script variations from hieroglyphs to hieratic (Goldwasser 2002; 2006; Goldwasser and Grinevald 2012; Lincke and Kammerzell 2012). Certain tendencies of classifier use can be detected as typical of a certain script variety, such as a tendency to choose iconically more simple signs in hieratic or a tendency to multiply classifiers in cursive varieties; however, there are not a few exceptions to these generalizations. 4 See Brown 1999: 28, and below. 5 For the cultural integration of the horse on all levels of Egyptian society as reflected by its texts, see Vernus 2009, also Vernus and Yoyotte 2005: 202. 6 See here Eco’s discussion of the perceptual cognitive and semiotic considerations that may be involved, in Eco 1996, and Brown 1999: 53 with analysis of many cases in different languages. 7 The French word Attelage and the German word Gespann may provide a closer meaning, as it includes the team and the yoke in a single word. Earlier interesting treatments of the word may be found in Schulman 1963; Schulman 1980; Gardiner AEO: Vol. 1, *113 f.
48
Orly Goldwasser
1.1.2 The lexeme ḥtr in the lexicon before the extension into the meaning horse The most important recipient lexeme for horse(s) in Egyptian is ḥtr, ‘the yoked ones’ which describes two draft quadrupeds yoked to a plough. The word is known from texts since the Old Kingdom, and it frequently appears in proximity with the verb skꜢ ‘to plough’. The root ḥtr ‘yoke’ is known rather early (already in the Old Kingdom) to receive an extended metaphorical meaning: ‘tax’. The tax seems to be conceptualized as ‘the yoke on the back of the people.’8 The yoked pair of draft animals were usually cattle, but rarely also donkeys (see below).
1.1.3 Classifiers of the lexeme ḥtr before the horse: ‘team of yoked quadruped’ In most Old and Middle Kingdom attestations, the word ḥtr — referring to a team of hard-working draft animals — takes a bovine classifier. This manifests sometimes as a single hieroglyph + plural strokes, and other times as a pair of bovines. A rare logogram of the animals attached to their plough is known from a 1st Intermediate Period stela (DZA 27.540.720, and fig. 2). A few examples of two consecutive classifiers – [yoke] and [cattle] (DZA 27.540.810) or (DZA 27.540.750) — always in this order — are also known. However, the most common classification until the New Kingdom is by far by the animals themselves. The fact that the classifiers never show the plough alone shows that the semantic load of the lexeme indeed rests on the animals and their ‘yoked’ state.
Fig. 2: The Stela of Uha, First Intermediate Period, Dynasties 7-10 (2250–2134 BC)
In the Middle Kingdom, a rare example shows donkeys (DZA 27.540.830) instead of cattle as a classifier, thus exemplifying the ‘open door’ within the semantic core of the word for ‘a pair of (any) yoked draft quadrupeds’. The donkey may reappear in disguise during the New Kingdom when the hieratic manuscript dating to the 19th dynasty (Prophecy of Neferti) shows an example with the imaginary Seth animal as classifier followed by the [hide & tail] (DZA 27.540.890).9
8 The classifier for this root in this meaning when used in verbal and nominal forms is (Faulkner 1962: 181). However, during the 19th dynasty the word takes also the usually classifier , which puts the word into the category of [(Administrative) power], see DZA 27.542.980, DZA 27.542.970. This classification will become more prominent in later periods. 9 Here I disagree with the reading of Goedicke 1977: 73. The Canaanites are described as disturbing the process of harvesting and stealing the draft animals on their ploughs.
What is a horse? Lexical Acculturation and Classification
49
It is indeed possible that donkeys were used occasionally for ploughing, but it seems that the prototypical quadrupeds were bovines. Ploughing scenes in Egyptian tombs always show ploughing by cattle. Nevertheless, another rare and challenging classifier appears in the tomb of Senmose in Aswan from the New Kingdom. Here (DZA 27.540.880) appears in a broken context which reads ‘…skꜢ.i m ḥtr m tp n rmṯ’ — ‘(I did not?) plough with a human team’ (lit. ‘head of man’). The unusual referent for the yoked team is marked by the [human/male] classifier for the word 10 ḥtr, followed by a lexical explanation. By the New Kingdom, following the general diachronic trend in classifier development of all lexemes referring to cattle, the word ḥtr starts to take the [hide & tail] classifier (see below), instead of bovines, or in addition to bovines.11
1.1.4 The new referential extension — the embedded lexeme ḥtr — ‘team of horses’ Ḥtr is the first lexeme that appears in the Egyptian texts by the end of the Hyksos period with a clear new reference to horses & chariot. The choice of this extension is both utilitarian and morphological. It anchors the horse in its utilitarian function as (part of a) ‘team of draught animals’, but also reflects a perceptual-morphological basis for the equation. Yoked animals are naturally domesticated and thus the horse in this version is inserted directly into the realm which we would call ‘domesticated quadruped’. or more specifically, ‘draught quadrupeds’. The two earliest known references for ḥtr referring to horses emerge from the domain of royal texts — the Kamose stela and the parallel hieratic version of the same text on the Carnarvon tablet, circa 1550 BC.12 The word probably referred to a team of chariot horses, which seems to be one of the earliest forms in which horses were encountered by the Egyptians, at least in Egypt in direct contact (see below). Literally, the horses are another example of ‘The Yoked Team’. Nevertheless, the horses are yoked not to a plough, but to another new artefact (and new hieroglyph ) imported into the Egyptian universe: a chariot.13
10 See Edel 2008, Tafel LIX, QH 351, Text 1. For ‘referent classification’ in the context of ‘lexeme classification’ see Lincke and Kammerzell 2012: 91–101 and compare Goldwasser and Grinevald 2012, passim. 11 All ‘bovine-words’ are late arrivals into the [hide & tail] category, see the explanation and discussion in Goldwasser 2002: 82–85. 12 See below 1.1.6. During this early stage of the New Kingdom, the repertoire of what would be the official linguistic register of Egyptian royal texts is not yet organized and canonized. The Kamose stela shows the variation tꜢ-nt-ḥtr which may be translated as ‘chariotry’, see discussion below. 13 Schneider and others favor a loanword, Zeidler (2000) sees in it a new Egyptian creation, see Schneider 2004: 20 f.; 2008: 184.
50
Orly Goldwasser
For this impressive new vehicle, a new lexeme was created (or adopted) in Egyptian: wrryt.14 At a later stage, an additional Semitic loanword, mrkbt, is introduced in hieratic administrative texts, probably as a result of strong language contact and a partially bilingual situation in the military sphere.15 Nevertheless, mrkbt does not replace wrryt and they continue to be used side by side, mostly in texts of different genres. However, already in the 18th Dynasty, we find the word ḥtr with a more direct reference to the animals themselves. This is the imminent meaning in an example from the tomb of Paheri (fig. 2).
Fig. 3: ḥtr in the tomb of Paheri
The inscription above the team of horses and chariot reads: pꜢ-ḥtr iḳr n ḥꜢty-Ꜥ mry nb.f ꜤbꜤ.n. (?)...ḥꜢty-Ꜥ im.f n bw nb — ‘The excellent team of the prince, beloved of his master, of which the prince boasts to everyone’.16 It seems that the signifieds in this case are the animals, as the patient of love in the common phrase mry nb.f ‘beloved of his master’ is almost always animated.
14 The word wrryt belongs to the royal domain and tends to appear in hieroglyphic texts where it mostly carries the as a unique classifier. The hieroglyph may also serve a logogram. In not a few examples in hieratic, but in hieroglyphic texts as well, the word takes the [wood] classifier, showing alternative classification into a typical meronymic classifier that describes the relation of ‘made of’ between the host and its classifier, Goldwasser 2002: 39–55. The choice of this classifier in hieratic manuscripts may be due to the complexity of the new chariot hieroglyph. 15 The loanword mrkbt always shows the [wood] classifier, Hoch 1996: 145–147. For a recent discussion of the linguistic variety of the Egyptian army lexicon in the New Kingdom and its extensive use of foreign words, see Schneider 2008. 16 DZA 27.540.940, see the interesting discussion on the Zettel.
What is a horse? Lexical Acculturation and Classification
51
There are some examples of insertion of the horse into other Egyptian lexemes, even though these are very marginally attested in comparison to the high frequency of ḥtr.17 Such a case is the root rnp ‘young’, in a nominal use as ‘young one’, referring to a young horse. Originally it referred to bovines. This noun may be extended into the meaning ‘young horse’, as both signifieds are large domesticated quadrupeds that share somewhat similar functions and perceptual qualities.
1.1.5 Nomenclatural manifestation of referential extension — overt marking18 — tꜢ-nt-ḥtr The referential extension of ḥtr into horse may have resulted, in the spoken language at least, in a polysemy. This linguistic situation might have promoted an overt lexical marking through a compound lexeme. The compound tꜢ-nt-ḥtr ‘that(?) of the Gespann’ comes to denote the team of horses and chariot very early19. When this compound is used, the speaker/hearer is not hindered by polysemy. However, the compound form variation is in use together with ḥtr alone throughout texts (AEO I: A237, Schulman 1963; 1980: 112 f.).
1.1.6 The classifiers of ḥtr when referring to horse As we saw above, the moment the horse was embedded into the lexeme ḥtr, the word became polysemic, with two very different meanings: ‘a team of draught animals (with a plough)’ and ‘a team of horses (with a chariot)’. The difference is manifested not only in the daily use, but also in the social context — mundane versus prestige.20 In Egyptian we can follow the tendency to clarify ambiguity, at least in the written language, by the classifier system. In not a few cases, the word, when referring to horses, takes the classifier . Although horses were imbedded into the lexeme ḥtr, which included the team and its chariot, the lexeme is never classified by the chariot, or chariot and horses.
17 Faulkner 1962: 150. For the rare use of nfrw and nhp in reference to horses, see Vernus 2009: 29. 18 Brown 1999: 28. 19 See above, note 12. 20 On the horses and chariot as prestige symbols, see Vernus 2009: 38–45.
52
Orly Goldwasser
Fig. 4: Horse sign on top of the Kadesh ‘poem’, pSallier 3, 6. See also lines 6, 9 as classifiers of ḥtr (the horse classifiers are in the wrong direction!)
In most cases, the lexeme is classified by the horse(s) alone, or by the [hide & tail] classifier, which is used since the New Kingdom for the classification of ḥtr referring to the cattle team and their plough, as well as to the team of horses and chariot. When classified by the horse hieroglyph the reader is pulled immediately into the extension of ‘team of horses’. In this case the classifier, even if not pronounced, has an active linguistic role in modelling the correct semantic signified of the polysemic word. Nevertheless, what may be one of the first known occurrences in the texts of the lexeme ḥtr (overt marking tꜢ-nt-ḥtr) in the Kamose stela, shows the superordinate [hide & tail] classifier according to Habachi. However, the picture in the publication shows a break in the inscription, and the remains allow easily the reconstruction of a hieroglyph. The ears and the front leg of the animal may still be visible (fig. 5). The word ḥtr in the Carnarvon tablet indeed shows the superordinate which is the preferred classifier in hieratic texts (See Habachi 1972: 36, fig. 22 and Gardiner 1916: 106, line 16).
Fig. 5: ḥtr in the second stela of Kamose
What is a horse? Lexical Acculturation and Classification
53
1.2 Loanwords for HORSE 1.2.1 The Loanword ssm(t) However, the conceptual tie between horse and chariot did not stand in the way of the most impressive actual animal. One or two decades after the appearance of the word ḥtr in the texts, a loanword ssm first appears in Egyptian sources, possibly addressing the ‘need’21 to refer more directly to this sensational new animal. And indeed the lexeme ssmt or ssm mostly refers throughout its long history in the Egyptian texts to the animal itself. The ‘need’ might have been influenced by direct interaction and care for the prestigious animal, as may be surmised, for example, from the careful single mare burial mentioned above. The Canaanite king would probably refer to his choice pet animal by its proper ‘word’ in his source culture. Vernus (citing Vycichl) suggests that the loanword ssmt was ‘imported’ into Egyptian before the New Kingdom, probably through Semitic intermediaries. The mimation of the foreign word known in as sisi in Sumerian, sisium in Akkadian and ssmt in Egyptian may indicate an early loan in Egypt, already in the beginning of the Middle Kingdom, maybe even before the arrival of the actual animal.22 This suggestion offers the probability that the loanword settled in Egypt through a foreign language rather early and might have entered the Egyptian lexicon via a colloquial register. Such reconstruction would also explain its relatively late appearance in the written repertoire. When eventually written down, the well-integrated ssm(t) does not show a version in ‘syllabic writing’, i.e. it is probably not felt to be a loanword anymore. As both the Egyptian recipient culture and the Mesopotamian donor cultures of the Northern Levant were ‘writing cultures’, it is feasible that the word entered through an unknown textual venue.23 However, this textual material has not yet been discovered. Another possibility is to surmise a scenario such as the one that Casson terms ‘pidgin trade language’, where special lexical items would move among culture agents such as travellers, merchants, soldiers, etc. (Casson 1994: 108).24 The first of attestation of ssmt in an Egyptian text comes from the hieroglyphic inscription of Ahmose, son of Ibana, that describes the Egyptian struggle against the Hyksos in Canaan. Ahmose chose to refer to the three elements of the war chariot
21 Here ‘need’ as a building block in cultural interference theory, see Even-Zohar 2005. 22 The correspondence between the Canaanite and the Egyptian s is unusual. For an in-depth discussion of the origins of this loanword with extensive bibliography, see Vernus 2009. 23 On similar issues in Burmese, see Lehman 1994: 111. 24 Practically a Wanderwort.
54
Orly Goldwasser
separately — the horse, the chariot, and its charioteer, all brought by him to Egypt (fig. 4). This specific lexical dismantling of the complex Gespann may enhance and detail his achievement.25
Fig. 6: Detail of horse hieroglyph from the tomb of Ahmose son of Ibana
However, it is the indigenous Egyptian lexeme ḥtr which survived in Coptic as the signifier for horse, while the loanword ssmt disappeared after the 3rd Intermediate Period (see Vernus 2009: 3; 36).
1.2.2 Classifiers of the loanword ssmt In the script, the written loanword ssmt is analyzed correctly by the Egyptians in terms of its classification; it is classified either by the newly introduced hieroglyph of the horse (in this case activated in the role of a pictorial ‘repeater’), or by the superordinate [hide & tail]. In the Egyptian script system, the [hide & tail] classifies the category of animals that ‘have hide and tail’ (see below). The new hieroglyph became an active participant in the hieroglyphic system. It plays the role of logogram for a few ‘horse words’, or of classifier also for some additional words denoting different types of horses.26 In some rare cases it may be extended to classify words strongly connected to horses, such as iḥ ‘stable’ (Vernus 2009: 28 f.).
25 Probably spelled ssmt, see Davies 2009: 156. 26 The iconic signifier may refer to different horse-word signifieds, on this phenomenon see Vernus 2009: 26–29. However, this semiotic phenomenon in not rare in the Egyptian script system and occurs with the [bovine] iconic signifier which may refer to several different signifieds, i.e. different names for different types of cattle, see Goldwasser 2002: 82–85.
What is a horse? Lexical Acculturation and Classification
1.2.3 The Canaanite loanword i-b-r
/
55
‘stallion’27
The other prominent loanword for horse, the Canaanite i-b-r surfaces in the lexicon about a hundred years after ssm, in the days of Thutmosis III in the list of booty from Megiddo in Canaan. Its use in the Egyptian sources is, however, much more limited. Hoch suggests that it refers specifically to stallions (Hoch 1996: 18). If this is indeed the case, it may be the necessity for gender marking that pushed this loanword into the written aggregate. Unlike ssmt, the word is written in the syllabic writing, i.e. marked as ‘foreign’. The lexeme i-b-r also takes the [horse] or the [hide & tail] as classifiers.
2 Comparable cases The introduction of the horse into two other cultures is briefly discussed below. The first comparison comes from the Sumerian and Akkadian languages and the cuneiform script. The second parallel comes from the description of the introduction of the horse into Nahuatl, an Aztec language.
2.1 The introduction of horse into the Mesopotamian lexicon and cuneiform writing28 The horse was also a new arrival in Mesopotamia. There is no consensus as to when exactly horses were first introduced into the Mesopotamian world and cuneiform writing, but it was clearly known from the very beginning of the second millennium, i.e. earlier than the recorded finds in Egypt.29 The same challenge of incorporating the new animal into language and script faced the writers of cuneiform texts. In a stratified process, related to the different languages operating in the cuneiform script, we find the horse in the Sumerian-Akkadian lexicon in two forms: as anše.kur.ra = ‘donkey of the mountain’ and sisi an apparent loanword.
27 Examples after Hoch 1996: 18, from Urk. IV 663,18 and pAnastasi IV 17,9. 28 I am grateful to Gebhard Selz for all the data on the introduction of the horse into the cuneiform script. See also Selz, Grinevald and Goldwasser (in press) for additional discussions. 29 See above note 3 and bibliography there.
56
Orly Goldwasser
Gebhard Selz describes the introduction of the horse into the script system in the following words: ... the process of extension of a class by absorption of new items that have entered the culture is illustrated by the case of the classifier anše [donkey] (Equus asinus asinus) – ... extended beyond its prototypical value of the specific species of domestic animal to head a class that includes wild donkeys (probably the Asiatic onager (Equus hemionus), also (Equus hemionus hemippus) and perhaps even deer. It finally comes to head a class that has absorbed various foreign or wild animals, such as the horse (called the ‘donkey from the mountain’ anše.kur.ra = read / sisi/), and later occasionally even the camel as anše.a.ab.ba = ‘donkey of the sea (lands)’ (Selz, Grinevald and Goldwasser (in press)).
The cuneiform anše.kur.ra is a clear case of overt marking of a referential extension. It is similar, for example, to the case of sheep which is called ‘cotton “deer”’ = stunimal chig in Tzeltal (Brown 1999: 28). Like the above case of sheep in the Amerindian language, anše.kur.ra is the name for the horse embedded in an animal of similar form and function – the donkey, while sisi is the loanword which is introduced into Sumerian and Akkadian, known in Egyptian as ssm or ssmt (see above).30 The Sumerian cuneiform system, like the Egyptian script, includes a system of classifiers. However, the Sumerian classifier system differs from the Egyptian system in a few important aspects. Firstly, the Sumerian system is strictly a ‘noun classifier’ system — it does not classify verbs as the Egyptian, Anatolian and Chinese writing systems do, and most classifiers are in pre-position.31 Secondly, the number of classifiers is considerably smaller. Moreover, the Sumerian system probably emerged from a system of pronounced compounds. Two examples from the cuneiform texts show a distilled combination of the process of embedding and word-loaning in a single written word. In the Old Babylonian Akkadian writing [anše] sí-s[í-i] (18th century BC) and probably also in the much earlier Sumerian writing anše-zi-zi (in UR III period, 21st century BC),32 [anše] is activated as an unpronounced classifier for the loanword written as si-si or zi-zi.
30 Compare also the Biblical Hebrew loanword 31 See Selz, Grinevald and Goldwasser (forthcoming), Chen 2016, and Payne (in press). 32 E.g. ELA/Ur III/Drehem 4(diš) anšesi2-si2 BIN 03, 085 1; 2(diš) anšesi2-si2 BIN 03, 454 1 (from the electronic Pennsylvanian Sumerian Dictionary [ePSD]). In the case of Sumerian it is difficult to know whether the classifier was pronounced or unpronounced. All cuneiform examples have been provided by Gebhard Selz.
What is a horse? Lexical Acculturation and Classification
57
2.2 The Aztecs and the horse In his book Kant and the Platypus, Umberto Eco approaches many aspects of the theory of categorization and classification. He dedicates a special, colourful chapter to ‘Montezuma and the horses’ (Eco 2000: 127–130). Basing himself on indigenous Spanish chronicles recording the event, he reconstructs the reaction of the locals to the arrival of the first horses to America with the Spanish invaders. The first messengers were sent to Montezuma to tell him of the landing and of the ‘… terrifying marvels they were witnessing ... one scribe … explained that the invaders were riding deer as high as the roofs of houses’ (Eco 2000: 128).33 Nahuatl, the language spoken by Montezuma, is a special case in Latin America. It remains widely spoken in Mexico today. The language, however, is best known as the speech of the Aztecs who ruled large parts of Middle America during the Spanish invasion in 1519 AD. The Spanish promoted Nahuatl as a vehicle of administration. As such, textual materials written in Nahuatl using Spanish orthography were produced abundantly immediately after the conquest. As a result, unlike other cases, the post-contact history of Nahuatl is relatively well documented. We know that ‘horse’ was not unique in being embedded in an extended lexeme; other new-comers to the continent underwent the same process, such as totolin ‘turkey’ for chicken. Nahuatl is also rich in what may be called ‘coinage’, when a descriptive compound becomes the ‘name’ of the new item. Such an example is tentzone, ‘bearded one’, for goat. Being under colonial rule, the local speakers probably became at least partially bilingual. The extended local maçatl ‘deer’ that hosted horse, is replaced by cahuallo, the Nahuatl version of the Spanish caballo. Yet, not all extended local lexemes disappeared. The extended Nahuatl mizton ‘little cougar’ = cat, for example, has not been replaced by a Spanish loanword (Brown 1999: 96).
3 Discussion — the emergence of horse in the three cultures From their cognitive-anthropological and socio-linguistic points of view, Brown (and his commentators, Brown 1994; 1999; Oswalt 1994; Lehman 1994) try to trace the linguistic and social settings that are present in a repetitive linguistic situation: a new item appears on the horizon of a culture. How would this item be treated linguistically? What will it be ‘called’? The paradigm of Linguistic acculturation enumerates five types of possible venues for the acculturation of a new item or concept:
33 Eco’s interest lies in the analysis of the cognitive and perceptual reasons that caused the beholders to call the new animal ‘deer’.
58
Orly Goldwasser
1. Extension of native term 2. Compound extension – native term + modifier.34 3. Coinage – a newly invented ‘name’ in the recipient language. 4. Loan words 5. Loan translation Brown, following Casagrande, suggests a clear division between the options. ‘Primary accommodation’ is represented by types 1–3. They are usually the earliest linguistic reaction to a new item that appears in a certain culture’s universe. Types 1–3 do not entail any type of bilingualism. ‘Secondary accommodation’ is represented by types 4 and 5. In many cases, types 4–5 require at least some level of language contact or a minimal level of bilingualism, sometimes even only in the case of a single ‘key person’, (e.g. king) or other influential culture agent(s) (Brown 1990: 160).35 The Nahuatl example shows a clear phase of ‘primary accommodation’, followed by the introduction of the Spanish loanword, which eventually replaces the original maçatl in a ‘secondary accommodation’ process. The Ancient Near Eastern material paints a somewhat more complicated picture. In Egypt, we find firstly an extension of a native term, ḥtr, as expected by the general theory. Yet, the horse indeed enters a local lexeme, but not a lexeme that refers to an animal alone, but a lexeme that refers to a team of draught animals in a specific role — ‘a team in ploughing’. In this case, the extension of the meaning is in the field of function. The new draft animals are ‘a team leading a chariot’. Nevertheless, the perceptual information is also active, as it keeps the general structure of two quadrupeds drawing a heavy item. Interestingly, the embedding process of ḥtr ignores completely the different social setting and the role in society of the two signifieds of the word ḥtr. Unlike the slaving cattle and the field workers, the horses and chariotry were elite and prestige markers, and kings and scribes were keen to describe themselves as chariot riders and horse owners. When classifiers are added to the word ḥtr, they guide the reader to the ‘animal itself’, either through the classification by the iconic horse hieroglyph, or by the classifier of [hide & tail] that is reserved mainly for classification of those ‘having hide and tail’, which is roughly parallel to our category of animals. In a very early period of use, we also find an example of extension + modifier. We encounter the variation tꜢ-nt-ḥtr ‘belonging to the ḥtr’. The compound contains an overt marking confined to the horses and chariot signified alone, which solves any polysemy that might be created by the simple extension of the old ḥtr into the new meaning.
34 See here also ‘reversal marking’, Brown 1999: 159. 35 For the notion of ‘key’ persons, see Lehman 1994: 110.
What is a horse? Lexical Acculturation and Classification
59
No ‘coinage’ is known for the horses themselves. Yet the chariot, called wrryt, if indeed this is a new linguistic creation in Egyptian as some scholars believe, would fall into the type 3 – coinage.36 After a few decades, the Egyptian texts show, for the first time, the loanword ssm(t). It is not known how this word was transmitted to Egypt, when, from where, or by whom. It appears in the texts as a ‘second accommodation’, fitting the generally prescribed expectations of linguistic acculturation theory. However, due to the spelling considerations mentioned above, it may be a wrong impression created by the written material. The loanword might have entered the lexicon in the spoken language much earlier. Interestingly, a replacement37 process never took place in Egyptian, and the ‘horse lexemes’, the extended forms, and the loanwords all lived on, side by side, for hundreds of years. In Mesopotamia, we find an interlacing process. On the one hand, there is a clear lexical extension where the horse is embedded in a local, prototypical pack animal — the donkey. This kind of lexical extension into a ‘similar’ animal is identical in essence to the examples known from the languages studied by Brown.38 The animal was used for long-distance transfer and riding, and even pulled war chariots. The donkey is a closer, more accurate prototype for the horse than cattle. Like the donkey, the horse is not used regularly for meat consumption or for leather-goods production. However, the earliest occurrences of the horse, which are known already from the 21st century BC in cuneiform, show a sort of ‘blending’ — an embedding in anše donkey with the addition of the loanword sisi or zi-zi. These examples should testify to early encounters and linguistic contacts with some carriers of the loanword sisi. Spellings such as [anše]-zi-zi solve all problems of possible polysemy between donkey and horse. Interestingly, the compound extension anše.kur.ra ‘donkey of the mountain’, appears in the cuneiform texts only a few hundred years after the loanword sisi. It contains an explanation and specification of the special ‘donkey’. Looking at the results of the cuneiform and Egyptian texts, in the light of the theoretical framework suggested above, it seems that in both cultures the results favour an early linguistic contact and maybe even limited bilingualism with some carriers of the foreign ‘horse word’. This linguistic contact might have pushed the loanword forward. In the cuneiform system, the loanword sisi is the earliest to emerge as referring to
36 A unique early example for a coinage (1850 BC) may be tbt n ḫntš ‘box for walking around’, see Schneider 2004: 20. 37 For this term and its definition see Brown 1999: 92–104. 38 In Brown’s large collection of Amerindian languages, Brown shows examples of the embedding of ‘horse’ in the designations of different local animals. A few languages show an extension of ‘dog’ into horse, see Brown 1999: 53.
60
Orly Goldwasser
horse. However, the loanword does not always appear by itself. The new loanword is combined with a lexical extension to create a compound anše-zi-zi or anše-sí-sí. Prima facie, the Egyptian evidence favours the general framework suggested by Brown – Extension (ḥtr), extension + modifier (tꜢ-nt-ḥtr) and later the loanword (ssm). However, here too one must be vigilant, as the spelling of the word points to a possible earlier loan of ssm not attested in the existing textual material. In both systems, cuneiform and Egyptian scripts, the classifier system plays an important role. Even if anše is an unpronounced classifier in the combination [anše] sí-s[í-i], the extension into the donkey is firmly expressed and cognitively takes place in the mind of the reader. When the unpronounced classifies the lexeme ḥtr, it resolves immediately all possible polysemy (cattle or horses), at least in the written lexicon.
4 Words of conclusion — animal classifiers in Egypt and Mesopotamia. A study of animal classifiers in the Egyptian scripts and in Sumerian cuneiform carried out on another occasion should be mentioned here.39 If we look at Sumerian knowledge organization as evidenced by the classifier system of the cuneiform script, there is, first and foremost, a noticeable absence of any classifier that represents a generic category of the sort of ‘animal’. The animal world is divided in the script by the classifiers into five classes, which seem to correspond to a utilitarian perception of the animals beyond some physical similarities.40 This clear utilitarian classification distinguishes between classes of mušen [bird], ku [fish], gud [bovine], udu [sheep], and anše [donkey]. The role of the classes of [bird] and [fish] is clear, as they form the basis of the nutrition system and have a clear economic value, as shown by numerous economic documents from the earliest phases of writing. The same could be said about the livestock classes of [bovine], [sheep] and [donkey] which have an immeasurable industrial role, for food consumption, leather products and wool production, as well as transport.
39 Selz, Grinevald and Goldwasser (in press). 40 See Denny 1976 for an early discussion in classifier studies of the three principles of categorization: by essence, by form or by function. For utilitarian considerations in classification, see Hunn 1982.
What is a horse? Lexical Acculturation and Classification
61
Large wild carnivore quadrupeds such as lions and panthers and smaller domesticated quadrupeds such as dogs, cats, monkeys or even mice, remain unclassified in cuneiform. Reptiles such as snakes or invertebrates such as worms remain unclassified, too. Egypt, on the other hand, presents us with a ‘silent show’ in the script, in which we can watch the slow cognitive formation of a pre-verbal concept somewhat parallel to our Western animal. Nevertheless, this concept never matured into a lexeme and was never born into the lexicon of the Egyptian language. The unifying quality that caught the attention of the Egyptians was not the ‘breathing’ ability of the animals, but a more morphological aspect: ‘having hide and tail’. In its cognitive formation process, this pre-verbal assessment, it is closer to Western morphological concepts such as mammal (having mammae) or quadruped (‘having four legs’). From the Old Kingdom onwards, the classifier , which carried the iconic meaning [hide] or [hide & tail], was used as a classifier for different types of hide. Somewhat later, it was extended to classify artefacts ‘made of hide’. All members in this class (stage 1) were always inanimate, and almost all members that were not a ‘kind of hide/ leather’ could be defined as manufactured (‘made of’). In the meantime, during the Old Kingdom, animals were classified by their own icon (e.g. ‘dog’ = phonograms – ṯsm + [dog] classifier) (see Müller 2002: Appendix II, 25). In these cases, the classifying icon repeats via iconic depiction the same information given by the previous hieroglyphs, which are themselves used as phonograms. These cases correspond to the well-known type of classifier in classifier languages known as ‘repeater’.41 This kind of classifier does not represent a concept ranking higher in the taxonomic hierarchy and seems to ‘repeat’ the lexical information. By the end of the Old Kingdom, the classifier was suddenly extended to the classification of living animals that ‘have [hide & tail]’. One of the first animals known to receive the classifier was the lion (Goldwasser 2002, 64 f.), a big carnivore with no obvious utilitarian qualities. Another early example appears on clay tablets recently published, and found in the provincial town of Balat, dating to the end of the Old Kingdom. The [hide & tail] is found as a classifier in a place-name mw-mꜢ ‘the water of the antelope’ (Pantalacci and LesurGebremariam 2009: 247). The antelope species may have played here the lexical role of prototype for all desert animals that come to the pools of the oasis. Moreover, the appearance of [hide & tail] in this early example cannot be due to a reluctance to present the full animal hieroglyph. The tablets from Balat, although written in hieratic on clay, never refrain from presenting the names of the various cervine animals with rather iconic repeaters or logograms showing in detail the differences between various quadrupeds, marked by their different horn forms.
41 See Appendix by Grinevald in Goldwasser and Grinevald 2012, with examples in Yakaltek. See also Goldwasser 2006.
62
Orly Goldwasser
By the Middle Kingdom the extension into the living animals had become a clear tendency in the script, with dog, cat, mouse, as well hippo and pig, getting the , all recorded in lapidary hieroglyphic inscriptions, as well as in cursive script and hieratic. To sum up, the main differences between the classification of animals in the Egyptian script and cuneiform scripts in the 2nd millennium are: 1. Unlike the sensitivity of the cuneiform system to the differentiation of domesticated vs. non-domesticated, the Egyptian [hide & tail] classifier is not sensitive to this criterion. The Egyptian classifier system also is insensitive to differences between carnivores and herbivores. Lions, antelopes, mice, donkeys, dogs and cats belong to the same category of . 2. Unlike the cuneiform classification, the Egyptian classification of animals in the script is basically non-utilitarian. It is a sortal classification based on observable characteristics of the animals, disregarding their function or specific relations to man. Members of the category are those ‘having hide and tail’. 3. By extension processes of the category, almost all quadrupeds may be classified in the script by by the New Kingdom.42 Newcomers such as the horse, the main topic of this study, are immediately analyzed as and put into the script with the correct classifier. Even some clear fringe-members, such as the turtle in the Middle Kingdom and the flea and scorpion in the New Kingdom, are occasionally classified in this class.43 The only animal that has a kind of ‘hide & tail’ but keeps being classified almost always with a repeater is the crocodile.44 4. [bird] 45 and [fish] have no ‘hide & tail’ and show separate categories. As we have seen above, [fish] and [bird] also appear as separate categories in cuneiform, which, in both Egyptian and Sumerian systems, have dozens of members. By the end of the New Kingdom it seems that the generic category [hide & tail] starts to be further extended, to occasionally include also birds - thus moving towards our generic concept [animal]. 5. Unlike the Sumerian system, snakes and worms are classified by the Egyptian system. The classifier is hosted by words describing all kinds and sizes of snakes and worms — termed [Swarms] in my earlier work (Goldwasser 2002: 57, 68). By the end of the New Kingdom, the category [hide & tail] begins to occasionally embrace [Swarm] members, thus confirming the extension of the
42 The Egyptian script system always keeps the parallel option of classifying an animal by repeater or unique, or both by repeater and the generic [hide & tail]. See the tables by Müller 2002, Appendix II. 43 See discussions in Goldwasser 2002, passim. Fringe members are defined as such if they only rarely show the [hide & tail] classifier. 44 This may be due to the special place of the crocodile within the Egyptian culture. It may be parallel to the case of the dog in Yakaltek, see Selz, Grinevald and Goldwasser (in press). 45 The word ḥtm translated as ‘provide’ may take the [bird] classifier, see Faulkner 1962: 180 and Goldwasser 1999.
What is a horse? Lexical Acculturation and Classification
63
[hide & tail] category into a higher superordinate concept that now also encompasses reptiles and similar. To sum up, the Sumerian and Egyptian classifier systems reflect rather a dissimilar mental organization of the animal world. The reasons for these differences should be reserved for another study.
References AEO= Gardiner, Alan H. 1947. Ancient Egyptian onomastica, 3 vols. Oxford: Oxford University Press. Bietak, Manfred. 2010a. “From where came the Hyksos and where did they go.” In The Second Intermediate Period (Thirteenth – Seventeenth Dynasties): Current research, Future prospects, edited by M. Marée, 139–181. Leuven: Peeters. Bietak, Manfred. 2010b. “Le Hyksos Khayan, son palais, et une lettere cuneiform.” CRAIBL: 976–978. Boessneck, Joachim, and Angela von den Driesch. 1992. Tell el-Dab’a VII, Tiere und historische Umwelt im Nordost-Delta im 2. Jahrtausend anhand der Tierknochen der Ausgrabungen 1975–1986, UZK X. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Brown, Cecil H. 1994. “Lexical acculturation in Native American languages.” Current Anthropology 35: 95–117. Brown, Cecil H. 1999. Lexical acculturation in Native American languages. New York: Oxford University Press. Casson, Ronald W. 1994. “Comment.” In: Brown 1994: 108 f. Chen, Jongsheng. 2016. “The prototypical determinatives in Egyptian and Chinese writing.” Scripta 8: 101–126. Davies, W. Vivian. 2009. “The tomb of Ahmose son-of-Ibana at Elkab. Documenting the family and other observations.” In Elkab and beyond. Studies in honour of Luc Limme, edited by Wouter Claes, Herman De Meulenaere, and Stan Hendrickx, 139–176. Leuven: Peeters. Denny, John P. 1976. “What are noun classifiers good for?” Papers from the Regional Meeting. Chicago Linguistics Society 12: 122–132. DZA= Digitized Slip Archive, http://aaew.bbaw.de/tla/servlet/DzaBrowser. [Accessed 1/2017] Eco, Umberto. 2000. Kant and the platypus: Essays on language and cognition. Orlando: Harvest Books. (Translated by A. McEwen). Edel, Elmar. 2008. Die Felsgräbernekropole der Qubbet el-Hawa bei Assuan. I. Abteilung. (Band 1–3, Pläne und Tafeln). Architektur, Darstellungen, Texte, archäologischer Befund und Funde der Gräber QH 24 – QH 209. Paderborn: Ferdinand Schöningh. ePSD= From the electronic Pennsylvanian Sumerian Dictionary, http://psd.museum.upenn.edu/ epsd/nepsd-frame.html. [Accessed 2/2017] Even-Zohar, Itamar. 2005. “Laws of cultural interference.” In Papers in culture research. Tel Aviv: The Culture Research Laboratory. Faulkner, Raymond. O. 1962. A concise dictionary of Middle Egyptian. Oxford: Griffith Institute. Gardiner, Alan H. 1916. “The defeat of the Hyksos by Kamōse: The Carnarvon tablet, No. I.” Journal of Egyptian Archaeology 3: 95–110. Goedicke, Hans. 1977. The protocol of Neferyt: (the prophecy of Neferti). Baltimore: Johns Hopkins University Press. Goldwasser, Orly. 1999. “The determinative system as a mirror of world organization.” Göttinger Miszellen 170: 73–93.
64
Orly Goldwasser
Goldwasser, Orly.2002. Prophets, lovers and giraffes. Wor(l)d classification in Ancient Egypt. (Classification and Categorization in Ancient Egypt 3; Göttinger Orientforschungen, 4. Reihe: Ägypten 38,3) Wiesbaden: Harrassowitz. Goldwasser, Orly. 2006. “A Comparison between classifier language and classifier script: The case of Ancient Egyptian.” In A Festschrift for Hans Jakob Polotsky, edited by G. Goldenberg, 16–39. Jerusalem: Magnes Press. Goldwasser, Orly, and Colette Grinevald. 2012. “What are determinatives good for?” In Lexical semantics in Ancient Egyptian, edited by Eitan Grossman, Stéphane Polis, and Jean Winand, 17–53. (Lingua Aegyptia Studia Monographica, 9) Hamburg: Widmaier. Habachi, Labib. 1972. The second stela of Kamose and his struggle against the Hyksos ruler and his capital, edited by Eitan Grossman, Stéphane Polis, and Jean Winand. (Abhandlungen des Deutschen Archäologischen Instituts Kairo, Ägyptologische Reihe, Bd. 8) Glueckstadt: J.J. Augustin. Helck, Wolfgang. 1978. “Ein Indirekter Beleg Für Die Benutzung Des Leichten Streitwagens in Ägypten Zu Ende Der 13. Dynastie.” Journal of Near Eastern Studies 37/4: 337–340. Hoch, James E. 1994. Semitic words in Egyptian texts of the New Kingdom and Third Intermediate Period. Princeton: Princeton University Press. Hunn, Eugene. 1982. “The utilitarian factor in folk biological classification.” American Anthropologist 84: 830–847. Lehman, Frederic K. 1994. “Untitled comment.” Current Anthropology 35: 110 f. Lepsius, Karl R. 1849. Denkmaeler aus Aegypten und Aethiopien nach den Zeichnungen der von Seiner Majestät dem Koenige von Preussen, Friedrich Wilhelm IV., nach diesen Ländern gesendeten, und in den Jahren 1842–1845 ausgeführten wissenschaftlichen Expedition auf Befehl Seiner Majestät, Volume 3. Berlin: Nicolaische Buchhandlung. Lincke, Eliese-Sophia, and Frank Kammerzell. 2012. “Egyptian classifiers at the interface of lexical semantics and pragmatics.” In Lexical Semantics in Ancient Egyptian, edited by Eitan Grossman, Stéphane Polis and Jean Winand, 55–112. (Lingua Aegyptia Studia Monographica 9) Hamburg: Widmaier. Makowski, Maciej. 2014. “Terracotta equid figurines from Tell Arbid. New evidence on equids, their equipment and exploitation in North Mesopotamia during Third and First Half of Second Millennium BC.” Études et Travaux (du Centre d‘ArchéologieMéditerranéenne de l‘Académie Polonaise des Sciences) 27: 257–278. Müller, Matthias. 2002. “Appendix.” In Prophets, lovers and giraffes. Wor(l)d classification in Ancient Egypt, edited by O. Goldwasser, *iii–*46. (Classification and Categorization in Ancient Egypt 3; Göttinger Orientforschungen, 4. Reihe: Ägypten 38,3) Wiesbaden: Harrassowitz. Oswalt, Robert L. 1994. “Comment.” In: Brown 1994: 111 f. pAnastasi IV = Alan H. Gardiner. 1937. Late-Egyptian miscellanies. Bruxelles: Édition de la Fondation Égyptologique. Pantalacci, Laure, and Joséphine Lesur-Gebremariam. 2009. “Wild animals downtown: Evidence from Balat, Dakhla Oasis (End of the 3rd millennium BC).” In Desert animals in the Eastern Sahara: Status, economic significance, and cultural reflection in antiquity. Proceedings of an interdisciplinary ACACIA workshop held at the University of Cologne December 14–15, 2007, edited by Heiko Riemer et al., 245–259. Köln: Heinrich-Barth-Institut. Payne, Anick. In press. “Compound determinatives.” In Proceedings of ‘Die Sprache des Bewusstseins und das Bewusstsein von Sprache im Alten Orient’, Berlin 17/11/2014, edited by J. Cale Johnson, Eva Cancik-Kirschbaum and Jörg Klinger. Schulman, Alan R. 1963. “The Egyptian chariotry: A reexamination.” Journal of the American Research Center in Egypt 2: 75–98. Schulman, Alan R. 1980. “Chariots, chariotry and the Hyksos.” Journal of the Society for the Study of Egyptian
What is a horse? Lexical Acculturation and Classification
65
Antiquities 10/2: 105–153. Schneider, Thomas. 2004. “Nichtsemitische Lehnwörter im Ägyptischen. Umriss eines Forschungsgebietes.” In Das Ägyptische und die Sprachen Vorderasiens, Nordafrikas und der Ägäis. Akten des Basler Kolloquiums zum ägyptisch-nichtsemitischen Sprachkontakt, Basel 9.-11. Juli 2003, edited by Thomas Schneider, Francis Breyer, Oskar Kaelin, and Carsten Knigge, 11–31. Münster: Ugarit-Verlag. Schneider, Thomas.2008. “Fremdwörter in der ägyptischen Militärsprache des Neuen Reiches und ein Bravourstück des Elitesoldaten (Papyrus Anastasi I 23, 2–7).” Journal of the Society for the Study of Egyptian Antiquities 35: 181–205. Shaw, Ian. 2012. Ancient Egyptian technology and innovation. London/New York: Bloomsbury. Selz, Gebhard J., Colette Grinevald and Orly Goldwasser. In press. “The question of Sumerian determinatives: Inventory, classifier analysis, and comparison to Egyptian classifiers.” In Proceedings of the Conference ‘Crossroads: whence and whither?’ 17–20 February, 2016, Berlin. Taylor, Joseph J., and Francis L. Griffith. 1895. Wall drawings and monuments of El kab: the tomb of Paheri. London: Egypt Exploration Fund. Teeter, Emily. 2003. Treasures from the collection of the Oriental Institute. Chicago: The Oriental Institute of the University of Chicago. Urk. IV = Kurt Sethe. 1909. Urkunden der 18. Dynastie. IV, Band III, Heft 9-12: Historisch-biographische Urkunden. Leipzig: J. C. Hinrichs. Vernus, Pascal. 2009. “Réception linguistique et idéologiqued’une nouvelle technologie: le cheval dans la civilisation pharaonique.” In The knowledge economy and technological capabilities. Egypt, the Near East and the Mediterranean 2nd millennium BC – 1st millennium AD, proceedings of a conference held at the Maison de la Chimie Paris, France, 9–10 December 2005, editey by Myriam Wissa, 1–46. Sabadell: AUSA. Vernus, Pascal, and Jean Yoyotte. 2005. Le bestiaire des pharaons. Paris: Agnès Viénot; Perrin. Zeidler, Jürgen. 2000. “Zur Etymologie von wrry.t ‚Wagen‘. Mit einigen Bemerkungen zur ‘syllabischen Schreibung’.” Göttinger Miszellen 178: 97–111.
Figures Fig. 1: After Bietak 2010b: 978, fig.5 Fig. 2: After Oriental Institute Museum, 16956, Teeter 2003: 33 f. Fig. 3: After Taylor and Griffith 1895: pl. 3. Fig. 4: Photo courtesy British Museum, 10181.4. Fig. 5: After Habachi 1972: pl. VI. Fig. 6: After Davies 2009: fig. 19.
Sonja Gerke
All Creatures Great and Small – The Ancient Egyptian View of the Animal World Summary: Animals formed in many ways a basis of daily life in ancient Egypt. This is evident from their presence in numerous fields like mythology and religion, in the hieroglyphic script or in magic and medicine, to cite only a few examples. This outstanding interest in the animal world is accompanied by a very special attention to details which makes it possible (in most cases) to determine exactly what animals are depicted in reliefs, paintings and three-dimensional images or even in written descriptions, according to modern zoological- and taxonomic criteria. Such a detailed view of the animal world implies a differentiated, explicit interest which leads beyond simple attention to the given natural facts. Thus, the question arises of whether this particular interest leads to some kind of classification or categorization of the animal world and, if so, on which concepts, models or types this classification is based. Furthermore, this leads to the question of whether a system of hierarchy and canonization can be determined and these are the subjects the present paper deals with.
1 The Egyptian categorization of animals – primary thoughts The main challenge in tracing some kind of a system of animal-categorization lies – as is often the case – in the involvement of a modern point of view or the transfer of what are modern approaches, systems or models to ancient sources. Especially in the case of animal-categorization, it seems very hard to abandon the clear, regular modern taxonomic system that was created primarily by Carl Linnaeus, and the nomenclature and perception of the animal world that are still taught nowadays, even at primary school level. It has to be kept in mind, – especially when dealing with concepts of the animal world, – that every view of the ancient sources is in some way biased by modern thoughts and interpretations, even though we are aware of these problems and try to eliminate them.
DOI 10.1515/9783110538779-004
68
Sonja Gerke
From the start, it has to be considered that the ancient Egyptians had no educational texts1 dealing with the animal world such as those known, for example, from Classical Antiquity, written by Aristotle or Pliny the Elder.2 This means that it is hardly possible to find some kind of ancient Egyptian self-definition or -statement concerning this subject, and therefore the investigation of the ancient Egyptian view of the animal world can only be a modern construct, derived from several ancient sources. The distinct selection of the sources consulted is, – as will be shown in the conclusions of this paper, – very influential on the results of the investigation. The sources that the present paper deals with are taken from three different levels of evidence that can give information about ancient Egyptian concepts: 1. Textual sources; 2. iconographical sources; and 3. the combination of both. Considering the textual sources – as said before – there are no zoological texts as such; however, some ancient Egyptian texts correspond to this in very broad sense. These are texts that simply list the specific names of animals (onomastica), on the one hand, and, on the other, such texts that describe them in a more detailed way (lexica3), though their context or usage is quite insecure. Apart from these few distinct texts, which will be discussed below, when considering the question of the ancient Egyptian categorization of animals, it is important to look for sources that do reveal information on the matter in a more covert sense. This means, for example, figurative scenes on tomb- or temple-walls, whose primary purpose was not to communicate distinct knowledge about animals, but which do it nonetheless by the way in which animals are depicted, e.g. subspecies (to use the modern vocabulary) are differentiated by iconographical details. In some cases it is also possible to match language and image, meaning nomenclature and depiction, for example when – again in the scenes of temple- and tomb-decoration – a painting or relief of an animal carries its name in hieroglyphs as an addendum nearby. The broad subject Egyptian animal-categorization cannot be covered in its entirety within the scope of this paper. The ambition is rather to give an overview of the topic in general and to present some ideas and concepts of the ancient Egyptian view of the animal world. The examples treated in the following text are thus not to be seen
1 This excludes texts such as the so-called teachings / instructions which can be seen as educational texts (wisdom or advice considering good behaviour and references for life in general, but which give no reference to animals at all). For these texts and further reading see e.g. Burkard and Thissen 2003: 72–109 and id. 2008: 99–140. 2 It should be noted of course that in the case of the texts by Aristotle or Pliny, too, it is still questionable as to whether they can actually be seen as educational or scientific texts. But here it is important to point out the difference between the sources from Egypt and those from Greece and also the differences in recording or maybe even the transfer of knowledge within the society. 3 In the present context, the term lexicon is used after Fischer-Elfert 2008: 115 f., note 5, in contrast to onomasticon, which gives only single words or names see Gardiner 1968: 1 f.
All Creatures Great and Small
69
as a complete compilation of sources. However, the cited examples can be seen as more or less representative to outline some general chains of thought.4
1.1 Textual sources – onomastica and lexica Only very few texts exist which include information about animals in particular. First of all, there are the so-called onomastica, simple lists consisting of words that contain specific names/nomenclatures of persons, tribes, occupations, towns, specific products or plants, etc., without any further information (Gardiner 1968: esp. 1–5; Osing 1982: 572, and Pommerening 2015: 125–166). Until now, strictly speaking, only two texts are known that list the names of animals: the so-called Ramesseum Onomasticon5 from the Middle Kingdom (18th–17th cent. BC6; Gardiner 1968: 6–23), and a papyrus from Tebtynis7 from the Roman Period (Tebtynis I; 2nd cent. AD; Osing 1998: 121–140, Taf. 9–11). Both texts list names of animals that are obviously grouped together according to specific attributes or common criteria. The (identifiable) animals in the Ramesseum Onomasticon (Gardiner 1968: 9 and pls. I f.) are part of the following groups: birds,8 fish,9 again birds10, and quadrupeds11 (see fig. 1), whereas Tebtynis I (Osing 1998: 121–140, Taf. 9–11) lists cattle,12 small animals, livestock, and the like,13 predators,14 snakes and other vermin, insects,15 birds,16 fish, and finally turtles. The naming of those groups with general terms such as ‘birds’ or ‘fish’ might be somehow misleading, as none of these groups is separated from the other clusters or carries a specific heading. The only exceptional case is the fish-group in Tebtynis I, where the
4 Note that this paper does not deal with animal categorization in the hieroglyphic script by determinative/classifier; see instead the paper by Orly Goldwasser in this book. 5 pBerlin 10495. 6 The chronological dates of this paper are based upon Hornung, Krauss and Warburton 2006: 490– 493. 7 pCarlsberg 180+pBerlin 10465 u. 14475+PSI I 76; in the following called Tebtynis I after Osing 1998. 8 E.g. (122) rꜤ ‘bean goose’, (123) trp ‘greater white-fronted goose’, (127) mnt ‘swallow/dove (?)’, (132) ḳḳ ‘cuckoo (?)’, (133) wšꜢ.t ‘fattened fowl’. 9 E.g. (137) nꜤ[r] ‘catfish‘, (138) jms[kꜢ?] ‘Nile perch’, (146) Ꜥḏ ‘mullet‘. 10 E.g. (154) ḏꜢ.t ‘crane‘, (155) dšr ‘flamingo’, (160) bjk ‘falcon’. 11 E.g. (163) gḥs ‘gazelle’, (166) šsꜢ.w ‘hartebeest’, (167) jbꜢ.w ‘barbary sheep’, (169) mmj ‘giraffe’, (170) b[f…] ‘monkey (?)’. 12 Probably sequenced by age/size as supposed by Osing 1998: 121 see also his note 546. 13 E.g. (N 9,23) jꜤꜢ ‘donkey’ šꜢj ‘pig’, (N 9,24) sẖꜤ.t ‘hare’. 14 E.g. (N 11,2) [m]Ꜣj.t ‘lioness’, (N 11,3) ḥtm.t ‘bear’, (N 11,7) ḫfw ‘crocodile’. 15 E.g. (O 2,1) fnd ‘worm’, (O 2,4) kꜢr ‘chameleon (?)’, (O 2,5) pꜢw … n ṯsm.t ‘dog flea’, (O 2,11) ḥḏr ‘jerboa (?)’ ḳꜢdy ‘cockroach (?)’, (O 2,15) hfj ‘snake’. 16 E.g. (O 2,16) rꜤ ‘bean goose’, (O 2,17) [tr]p ‘white-fronted goose’, (P 3) ḥnw.t ‘pelican’, (Z 1) sḥsḥ ‘heron (?)’, (Z 3) mrrw ‘stork (?)’.
70
Sonja Gerke
beginning has a headline in red ink (S2.18) rm.w ‘fish’ and at the end of the section a short summary: (U 14) [rm] nb(t) jmj.w mw rn=w ‘all [fish], that are in the water, their name […]’ (Osing 1998: 121, and 136, Taf. 11 (U 14)). Also noteworthy on this Roman onomasticon is the bird-section. The birds’ names are not only listed, but the birds are described occasionally, with short explanations or addenda concerning the colour of their plumage or their natural habitat, which gives this section the form of a lexicon (see below).17
Fig. 1: Animal section of the Ramesseum Onomasticon
In the case of the Ramesseum Onomasticon, it is also interesting to note that the whole layout of the papyrus refers to the grouping of animals, as some space is made between the actual name of the animal and the last sign, the so-called determinative or classifier (bird , fish , sitting Seth-animal , recumbent hartebeest , hide and tail and monkey ),18 which gives the text a clear structure (see fig. 1). This would have probably allowed the reader to scan the text and search for special words or groups of words more easily.
17 On the fragments O and P 3 f., only the birds’ names are simply listed as in the other animal-groups of the papyrus, whereas on fragments R, Q S 1 and perhaps P 1 the explanatory additions are made. For the term lexicon see note 3. The text is in most parts very fragmentary, as an example here some better preserved lines: (R 2,14) jsm (?) sdm […] rn=f wnn=f m jnm ḥḏ km dšr jj[.n=f] (R2,15) ḥndj(?) Ꜥ Ꜣ(?) Ꜥnḥbw(?) rn=f wnn=f m jnm km ḥḏ ẖ.t=f mj […] ‘(R 2,14) jsm (?) sdm: ‘[…-goose] is his name. He is of white, black and red colour. [He has] come. (R 2,15) Great kingfisher (?): Pied kingfisher (?) is his name. He is of black and white colour. His belly is like […]’. 18 In the Egyptian script this is – in most cases – a single sign that appears at the end of a word and gives specific information on the meaning of the related word. The categorization of animals in the Egyptian hieroglyphic script is omitted in this paper, for the work with and the analysis of classifiers and their significance for classification, see the paper by Orly Goldwasser in this book.
All Creatures Great and Small
71
Besides these two texts, two further lists exist, but, their animal-sections are slightly different from those in the previously cited documents. The first text was inscribed on a wooden writing board found in Giza and dating back to the Old Kingdom (2400–2100 CE; Brovarski 1987: 27–48, plate I; Reisner 1910).19 It lists, in hieroglyphic columns, the names of kings, gods, and places, each repeated several times; at the end – again in vertical columns – twelve small depictions of six birds and six fishes can be seen.20 It is doubtful whether the name ‘onomasticon’ is appropriate for this writing board, due to the repetition of words and especially the images of animals, instead of listing their names, which is why the board could be seen as a ‘product of the schoolroom’ or some kind of ‘copy book for the scribe, with lists of difficult signs for the learner to copy’ (Brovarski 1987: 51).21 E. Brovarski however states that the board shows the clear ‘tendency to fuller record’ and an ‘interest in classification’ (Brovarski 1987: 51) as can also be seen in the almost contemporary ‘chamber of the seasons’ in the sun temple of Neuserre (2400–2370 BC; Edel 1961; id. 1963; Edel and Wenig 1974; see also below), or other true onomastica found from the Middle Kingdom onwards (see Gardiner 1968: 1–5). Another text to be cited in this context is the so-called Onomasticon of Amenemope (Gardiner 1968: 24–63)22 from the New Kingdom (1200–1100 BC). It is special, as it lists not animal-names, but parts of animals, such as jwf ‘meat’, ḏꜢḏꜢ ‘head’ or nḥb.t ‘neck’
19 Cairo, Egyptian Museum, JE 37734. 20 Description after Brovarski 1987: 47 f.: ‘The drawings of six different birds in division 4 are too schematic to assure identification in any one instance. The outlines are confidently drawn but taxonomic features such as colour or character of feathers which distinguish the species from one another are necessarily abbreviated on so small a scale. In spite of the fact that it has neither crest nor wattle, the overall feathering of the first bird allows the possibility that a Sennar Guinea-fowl is depicted. The absence of a wing outline also differentiates this bird from the succeeding (…). The relative size of the other five birds suggests three geese followed by two ducks. On the basis of tail-profile alone, the fourth bird might be a greylag (…) or white-fronted (…) goose. Considering the slightly longer central rectrices of the penultimate fowl, a pintail duck might be represented. (…) The compartments in division 5 contain three different fish (repeated 2 times). Even though he has done better by the fishes than the birds, the scribe has erred in several particulars. While the posterior dorsal fin is suppressed, the silhouette of the body, as well as the pinous anterior and rounded causal fins of the first fish, argue that a Nile perch is depicted (…). The body of the second fish is stream-lined like a mullet rather than deep-bodied like Barbus bynni, but the placement of the dorsal fin near the middle of the back with the anal fin about under it on the lower side of the body, the pectoral fin behind the head placed rather low, and the forked caudal fin suggests this species of carp is intended (…). The third fish, Tilapia nilotica with its dorsal fin of strong spines and soft rays and its rounded caudal fin, is difficult to mistake (…).’ 21 Onomastica in the sense of Gardiner 1968: 5, and Osing 1982: 572, are attested since the Middle Kingdom, so the writing-board from Giza is the first known onomasticon-like list (considering animals) from the Old Kingdom. 22 The text is preserved in several manuscripts, see Gardiner 1968: 26, and also Liszka 2010: 316, with note 8.
72
Sonja Gerke
(Gardiner 1968: 237*–238*, N°s. 579–581), to cite only the first three entries, which are written as a rubrum in red ink to mark the beginning of a new section.23 A very similar list is also found on the already mentioned Ramesseum Onomasticon (Gardiner 1968: 15–19, and plates III–IV, N°s. 271–311).24 It seems bizarre, when looking at these four texts, that each originates from one of the four great eras of Egyptian History – the Old Kingdom (writing board), the Middle Kingdom (Ramesseum Onomasticon), the New Kingdom (Onomasticon of Amenemope), and the Roman period (Tebtynis I). These texts, then, could show a useful average throughout Egyptian history, if it were not for their uniqueness, which narrows, or rather disables their scientific representativeness. The use and purpose of these lists are largely uncertain, but all are supposed to be some kind of handbook or compendium for priests or scribal students.25 The onomastica show a clear intention to group specific animals together to bring them into groupings by special attributes or characteristics they have in common. However, it must be noted that obviously no effort has been made (or there was no need?) to give these groupings some kind of headline or superscription, e.g. with a generic term that subsumes the species listed. In addition to the onomastica, there also exist the lexica of animals, as already mentioned above. These lexica differ from the former by the explanatory additions presented after each animal name, but again very few examples of these texts have survived down to the present day. The most popular one is the so-called Snake Papyrus from the Late Period (650–330 BC; Sauneron 1989; Leitz 1997; von Lieven 2004: 156–160).26 This papyrus lists and describes distinct species of snakes? in 38 (only 25 are preserved) paragraphs according to their physical appearance, how dangerous they are,
23 Those terms are sometimes also used for parts of the human body, but as Gardiner 1968: 237* states: ‘That parts of an ox are here intended, not parts of the human body, seems clear, not only from the three kinds of meat with which On.Am. comes to an end (Nos. 593.597.604) are not found anywhere in the many lists of human limbs and organs that we possess, whereas they do occur in connection with slaughtered oxen.’ 24 The spatially separated classifier is in all cases the sign for meat. 25 See e.g. Brovarski 1987: 51: ‘Indeed William Stevenson Smith saw the writing board (…) as a copy book for the scribe, with lists of difficult signs for the learner to copy. He was, however, undecided whether it represented a sample original to serve as a guide or a student’s practice tablet [reference to Smith, History of Egyptian Sculpture and painting, 358].’ Tebtynis I is called ‘Onomasiologisches Wörterbuch und Kompendium priesterlichen Wissens (Onomastikon)’ by Osing 1998: 25, and further id., 17: ‘Es handelt sich um Handbücher priesterlichen Wissens, wie es in römischer Zeit in Ägypten gültig war, sonst aber nur sehr selten auf Papyrus, etwas häufiger dagegen in die Dekoration von Tempelwänden umgesetzt überliefert ist.’ (For the dispatch of such handbooks on temple walls see von Recklinghausen 2014, and Pries 2014). 26 Brooklyn Museum inv. 47.218.48+85.
All Creatures Great and Small
73
and especially according to the gods which are connected to, or manifestations of, the specific snakes, for example (Sauneron 1989: 9, 26–27): (I.16) jr ḥf ꜤꜢ n ꜤꜢpp27 wnn=f dšr.t(j) r-ḏr=f ẖt=f ḥḏ jw (j)bḥ.w 4.t m rꜤ=f psḥ=f sj mt=f ḥr-Ꜥ
(§15)28 What concerns the great snake of Apophis:29 She30 is entirely red, her belly is white. There are four teeth/thorns in her mouth. When she bites a man, he will die at once.
(II.6) jr fjj šrj wnn jnm=f mj pꜤr nn wp.t ḥr tp=f wnn Ꜥ.t nb n ẖr.t=f ktkt jw=k r nḥm=f ꜤḥꜤ=f n Ḥr
(§29) What concerns the small viper:31 Her colour is like (the colour of) a quail, without a horn on her head. Every limb of whom has been bitten by her (lit. ‘who is under him’) is shaking. You will rescue him. He stands for (i.e. is a manifestation of) Horus.
This part of the papyrus has been interpreted as some kind of handbook of a physician, to give quick information about the toxicity of several snakes and possible prognosis for the bitten person.32 In this context, it is probably very important for the physician in charge and the patient to know the god or gods which are associated with the specific snakes, to exactly address a prayer or offering to the right deity.33 The link to religion and mythology is also visible in the other preserved texts of this genre, for example, the dog-section of pJumilhac (Greek-Roman Period).34 This papyrus represents a compilation of mythological details of one specific Egyptian nome, typical for the Graeco-Roman period – a so-called Gaumonographie – listing holy places, plants, objects or animals of a distinct region. Within this text, a section appears where altogether ten dog species are described in a very similar way to the Snake-Papyrus, giving information about which deity is manifested in which dog,
27 The first sentence of each paragraph is written in red ink (rubrum) to mark the beginning of a new section. 28 The numeration in paragraphs is taken from Sauneron 1989. 29 This snake can be identified as Naja haje (Egyptian cobra/uraeus-snake), see Leitz 1997: 52–58. 30 The Egyptian word for snake ḥf is male, so the suffix =f is used, whereas in English the female pronoun has to be used. 31 This snake can be identified as Cerastes vipera after Leitz 1997: 72–78. 32 The cited description of 38 snakes are followed by medical instructions for the treatment of the wounds. 33 But von Lieven 2004: 160, indicates rightly that snakes whose bite is fatal have an assigned deity, so according to her ‘der Gedanke mediko-magischer Nutzanwendung ist also durchaus nicht von der Hand zu weisen, greift aber zu kurz’. 34 Paris, Musée du Louvre E17110: von Lieven 2004: 160–162; Vandier 1960: 81–96; 127f.
74
Sonja Gerke
where the animal is worshipped, where and how it shall be buried and in some cases also a mythological explanation for the colouring or behaviour of a specific dog.35 However, one has to admit that these two very specific texts – pBrooklyn and pJumilhac – are the only two known texts that describe animals in such a detailed way.36 A few more texts have been recognized, in some cases quite fragmentary lexica such as pHal. Kurth37 (Ptolemaic Period; Fischer-Elfert 2008: 123–127), the Geographic Papyrus38 from Tanis (Roman Period; Griffith and Petrie 1889: 21–25), or Tebtynis II39 (Roman Period; Osing 1998: 257–258; von Lieven 2004: 162 f.) that also list animals. Still, in these cases, the animals are simply allocated to specific deities, as in pHal. Kurth (Fischer-Elfert 2008: 126): II, y+14 II, y+15 II, y+16 II, y+17
hꜢb Ḏḥw.ty pw bnw RꜤ pw ... (?) [...] nrj.t jr.t RꜤ pw ... (?) [...] Ꜣḫ.t jr.t RꜤ pw ...(?) [...]
Ibis: It is Thoth. Heron: It is Ra ...(?) [...] Female vulture: It is the eye of Ra ...(?) [...] Female northern bald ibis: It is the eye of Ra ...(?) [...]
It thus becomes clear that all lexica that list animals also position them in a context that is somehow religious or mythological, the only exception being the specific part of the above-mentioned ‘bird-section’ of Tebtynis I (Osing 1998: 124–136), where no gods occur.40 However, in this case, the religious component is given by the general context of the papyrus, which is to be seen in connection with a priestly or temple
35 For example: (XVI.2) (...) jr ṯsm (gloss: km) ḥḏ.t nḥb.t jw/r ẖ.t=f ttb/tbtb(?) (XVI.3) =f tp s n (??) sd=f wbḫ ꜤḥꜤ=f n Ḥr-sꜢ-Ꜣs.t jr pꜢ wbḫ nḥb.t=f ḥnꜤ ẖ.t=f sḏw(?) n ḥbs pw ḏr n (XVI.4) nsb=f ḥꜤ.w jt=f Wsjr m wꜤb.t jw=f (ḥr) nmms=f jm=f jw=f (ḥr) mr=f r ḫḫ=f (…) ‘Concerning the (gloss: black) ṯsm-dog with white neck (down) to (?) his belly (?). His enemy (?) head for (??) and his tail is bright. He stands for Harsiese (i.e. Horus, son of Isis). Concerning the brightness of his neck and his belly, he is (??) in this garment because of his licking of the limbs of his father Osiris in the hall of embalming. He wrapped himself in it (i.e. the mummy-linen). He bound it as far as to his throat (…).’ For this section see also Quack (2008: 15). 36 Another text that could be cited in this context is another passage from the already mentioned Ramesseum Onomasticon, where a list of 20 sorts of cattle (differentiated by colour) is presented (see Gardiner 1968: 22 f.; III, pl. V). The animals are listed one after the other, each allocated to a special sign or combination of signs at the beginning of the line. The meaning of these signs is unclear, perhaps they ‘served as abbreviations in inventories’, as Gardiner 1968: 22, supposes. For another, hitherto unpublished text, see von Lieven 2004: 167, note 39. 37 University of Halle-Wittenberg, Robertinum (formerly private collection Julius Kurth), inv. 33 A–C. 38 pLondon 10673. 39 pCarlsberg 182+PSI I 77. The fact that this text has parallels in Tebtynis III, pTanis and an unpublished hieroglyphic text shows, according to Osing 1998: 220: ‘(…) dass dieser Text standardisiert war und sowohl in Tanis wie in Tebtunis, wahrscheinlich sogar in ganz Ägypten Geltung hatte. Auf die Dekoration einer Tempelwand übertragen, findet sich die materia sacra der ägypt. Gaue, wie sie in Abschnitt 2 der Papyri aufgeführt ist, ja auch in Edfu.’ 40 See note 17. But this could also be due to the very fragmentary state of preservation of the papyrus.
All Creatures Great and Small
75
environment, to which actually all cited texts can be linked. This can be explained on the one hand by the fact that temple and priesthood in ancient Egypt represent the centres of knowledge and education; hence, it is not surprising to find such documents in this context or to link them with it. On the other hand, the connection between animals and gods or mythology lies in the ‘divinity of nature’, which was confirmed and worked out by von Lieven. She states:41 Alle relevanten Quellen gehören, soweit überprüfbar, in das Umfeld der Tempel. Dieser Glaube entspringt also nicht etwa zurückgebliebenen Unterschichten, sondern findet sich gerade auch in höchsten Kreisen der wissenschaftlichen und priesterlichen Elite (…). Der Tierkult in seiner bekannten (…) Form ist nur eine logische Weiterentwicklung der Vorstellung von einer allgemeinen und grundsätzlichen Göttlichkeit der Natur. Die Tierköpfe der ansonsten anthropomorph dargestellten Götter, die so charakteristisch sind für die religiöse Ikonographie des Alten Ägypten, vervollständigen das Dreieck zutiefst miteinander verflochtener Konzepte (von Lieven 2004: 167).
In summary, it can be stated that the textual sources do reveal some kind of categorization of the animal world – at least this can be said for the onomastica that group distinct animals together: fish, birds and quadrupeds. However, these groupings of animals do not seem to need any further explanation or titles/headings, but are seen to be self-explanatory. Because of the very few and unique examples of such texts to have survived, it is very difficult (or speculative) to make any generalized propositions on this basis, considering the whole of ancient Egyptian culture and history.
1.2 Iconographical sources – scenes of tomb decoration Ancient Egyptian culture is famous for its richness of images and pictures as is hardly any other ancient civilization. Whether in two-dimensional decoration of tomb- or temple-walls, or three-dimensional statues, statuettes or other sculptures: animals are present in any genre of figurative representations, be it as part of dramatic scenes, as sacrificial offerings, or in connection with a metaphorical meaning (see e.g. Guichard 2014). For these themes, especially the scenes of tomb decoration of the Old Kingdom (c. 2500–2100 BC), reveal a richness of detail, which primarily becomes obvious in the very detailed and natural depiction of animals that remains unsurpassed in the following periods of Egyptian history (Arnold 1995: 61).
41 For a general overview about animals in connection with the ancient Egyptian religion see Vernus and Yoyotte 2005: 20–49.
76
Sonja Gerke
As mentioned before, animals can appear in many different contexts in the scenes of tomb decoration. However, the following remarks will focus on the dramatic scenes in which animals act with or beside other animals or humans.42 For this case it is possible to differentiate between two types of animal representations. On the one hand, there are depictions of animals which stand by themselves, and on the other hand, depictions which have written addenda nearby, revealing the Egyptian names of those animals (for the latter see below). Why particular depictions, and which, in the scenes of tomb decoration are accompanied by inscriptions and which are not, is still a matter for debate; no prove or evident explanation has yet been found. At this point, it can only be stated that it is a fact that some animals never have addenda, whereas others do.43
Fig. 2: Fishing in the marshes, Tomb of Mereruka (Saqqara)
42 For all scene-types of tomb decoration of the Old Kingdom cited below, the reference to the online Database Oxford Expedition to Egypt: Scene-details Database; Linacre College, Oxford, 2006 will be given (in the following abbreviated: Scene-details Database plus the particular subject headings), [http://archaeologydataservice.ac.uk/archives/view/oee_ahrc_2006/ (DOI: 10.5284/1000009; accessed: 2/2015)]. 43 But even for those animals which do have their names as addenda (see below), there is no consistency recognisable in consequent use, e.g. with all depictions of a specific animal species.
All Creatures Great and Small
77
Animals which never have an addendum are, for example, the different fishes44 that are depicted in fishing-scenes with several fishermen45 or scenes that show the tombowner fishing in the marshes (see fig. 2).46 Although they cannot be identified by their Egyptian names, it is nevertheless possible to determine the exact species, owing to the very detailed and natural realization of the depictions (Sahrhage 1997: 60–62, figs. 19 f.; Gamer-Wallert 1970: pls. I f., and VII).47 In some cases, it is also obvious that the people who painted the scenes or who were responsible for the design were aware of the specific behaviour and habitus of distinct animals. For example, in several scenes a fish is depicted that is obviously upside-down in the water (Kanawati 2011: pls. 67 f.).48 What appears at first glance quite bizarre or somehow a mistake or carelessness on the part of the designer, turns out to be a very detailed observation of nature when examined more closely. This fish can be clearly identified in all examples as the Nile catfish (Synodontis batensoda), which possesses the very peculiar characteristic of swimming on its back, especially underneath the surface of the water, to scavenge for food (Houlihan 1996: 43 f.; Gamer-Wallert 1970: 12, 52; Sahrhage 1997: 57 f., 70). This behaviour can only be observed in free-swimming specimens, since animals that feel endangered or threatened do not show this. To depict such a peculiar behaviour of fish in the scenes of tomb decoration (Gamer-Wallert 1970: 52) assumes very exact observation of the fish, beyond what would be expected in fishing. It obviously appears to be a clear intention to differentiate between the particular fish-species and not to portray just a fish or a prototype49 of a fish, although both would be possible by highlighting some ‘typical’ characteristics, e.g. fish-tail, fins and scales. The illustrations are carried out in such a detailed way that an exact identification of the species is easily possible, even without the written tagging of the specific fish-name.
44 Nevertheless, the names of fishes are known from other sources, see Gamer-Wallert 1970: 16–46. 45 See Scene-details Database (note 42), 1.15–1.21: dragnet or seine-netting scene; fishing with a large net, set from boats; fishing with a funnel trap; fishing with a rounded basket trap, or carrying fish in a basket trap; fishing with a hand-held net; boatmen spear-fishing in a marsh (excluding the major figure); angling from a boat. See also Klebs 1982, 74–76. 46 See Scene-details Database (note 42), 1.1: Spear-fishing and/or fowling scene. See also Klebs 1982: 37. 47 This is also true for three-dimensional representations of fish, see e.g. Guichard 2014: 60 f, cat. 45a–g). 48 Saqqara, 6th dyn. (ca. 2300 BC). For further examples, see Sahrhage 1997: 54, fig. 16; 60 f., fig. 19; pl. 14. 49 On the use and significance of the term prototype see Lakoff 1987: 5–11, and Goldwasser 2002: 19–24.
78
Sonja Gerke
Similar predications can be stated for dramatic scenes where birds appear in connection with human interaction, that is, fowling scenes,50 or when birds are shown as inhabitants of the papyrus marshes51 while the tomb owner is shown e.g. fishing or carrying out other related activities. The most popular scene in this context is probably the hippopotamus-hunting scene in the tomb of Ti at Saqqara (5th dynasty; Wild 1953/1939/1966: pl. CXV–CXIX; Houlihan 1996).52 Here, the oversized tomb owner is shown standing in a papyrus boat poled by two other men. Next to him, two smaller boats can be seen. While the left boat is manned with just one person fishing with a string and a club, the four men in the right boat are hunting a hippopotamus with long spears. Below the boats, a small waterway is shown with some of its inhabitants: ten fishes (representing nine distinct fish species)53 and four hippopotami (one of the latter has a crocodile between its jaws). The much bigger part of the scene – all in all, four and a half meters high (Houlihan 1996: 20) – is occupied by the actual background illustration, the thicket of high papyrus stems and the inhabitants of this: about seventy different birds, two genets and several butterflies. The birds in the lower parts are sitting in their nests, in some cases protecting their eggs with their wings, or some nestlings crying for food are visible, while the two genets are scavenging for prey. In the upper parts of the scene, the birds are depicted flying or sitting on the papyrus flowers. Here again, the multiplicity of animals is represented in such a detailed and differentiated way that it is possible (in most cases) to identify the animals, as has been done by Houlihan (1996: 24–48). Furthermore, none of the animals – no matter whether flying or sitting within the papyrus thicket or swimming in the water – carries an addendum or an inscription that would reveal its ancient Egyptian name. It is also noteworthy that, even in scenes where animals are shown with inscriptions (see below), other animals may also appear which are not labelled and never have been. This phenomenon shall be outlined briefly in the following, and clarified by a specific example: the hedgehog. As von Droste zu Hülshoff pointed out, the hedgehog is an animal that is present in many sectors of Egyptian life and plays a not unimportant role, as the quantity of small amulets, figurines and vessels in the shape of hedgehogs, which can be found throughout Egyptian history (von Droste zu Hülshoff 1980; Vernus and Yoyotte 2005: 145 f.; Aufrère and Erroux-Morfin 2001; Leonard Jr. 2000). Additionally, the hedgehog is also present in many scenes of tomb decora-
50 See Scene-details Database (note 42), 1.1: Spear-fishing and/or fowling scene; 1.13. clap-net or bird-trapping scene. See also Klebs 1982: 35; 38; 70–74. 51 See Scene-details Database (note 42), 1.2. Pleasure, cruise in a marsh. 52 Similar scenes, but much smaller in size, with fewer animals depicted, can be found e.g. in the tomb of Mereruka at Saqqara (6th dynasty; Kanawati 2011: pls. 67–70). 53 Clarias anguillaris (or Clarias lazera), Synodontis batensoda, Labeo niloticus, Citharinus sp., Malapterurus electricus, Mugil sp., Anguilla vulgaris, Gnathonemus cyprinoides, Tilapia sp. after Houlihan 1996: 43–47.
All Creatures Great and Small
79
tion, e.g. in desert hunting scenes,54 where it may be depicted in its natural habitat as a desert dweller in or beside its burrow55 or while eating an insect (e.g. a locust).56 Those scenes seem to be some kind of frame depiction for hunting in the desert to show the natural environment (Gerke 2014: 48 f.). However, other scenes clearly show that the hedgehog can also be a beast of prey itself (Droste zu Hülshoff 1980: 22 f., 81–92), when it is depicted with other small animals in baskets being transported to the city57 or presented as an offering to the tomb owner.58 The appearance of hedgehogs in offering scenes leads to the assumption that these animals were also used as a daily food resource (Droste zu Hülshoff 1980: 23, with note 4). It seems astonishing that the name of this little animal – although it is depicted quite often in two and three-dimensional representations – is only known from two medical recipes from the New Kingdom.59 With this small example, it should have become clear that the tagging of animals with their names in hieroglyphs in the scenes of tomb decoration cannot be explained only by the economic relevance60 or importance of the particular animal within Egyptian society, but has to be seen in a broader context, and also probably individually for each example (see also below).61 To sum up, it can be said – as has been stated above and also for the textual sources – that a clear will or need for categorization and classification is traceable in the scenes of tomb decoration, meaning that it is important to differentiate between distinct fish or bird species, and not only to depict fish or birds. But it has to be kept in mind that the focus of these scenes lies not in the depiction of animals, but in the particular topic of the scene, including the human interaction (in most cases involving the tomb owner or his relatives and staff), meaning, for example, fishing and fowling in the marshes or hunting in the desert etc., so that the human actors form the actual core
54 Droste zu Hülshoff 1980: 17–21, 53–92. 55 E.g. in the tomb of RꜤ-m-kꜢ (Saqqara, 5th dynasty), see Droste zu Hülshoff 1980: 61, Nr. 7, Abb. 8; Decker and Herb 1994: 302, J23 (here with detailed bibliography). 56 E.g. in the tomb of Ptḥ-ḥtp (Saqqara, 5th dynasty), see Droste zu Hülshoff 1980: 63, Nr. 9, Abb. 10; Decker and Herb 1994: 304, J30 (here with detailed bibliography). 57 E.g. in the tomb of Ptḥ-ḥtp (Saqqara, 5th dynasty), see Droste zu Hülshoff 1980: 82, Nr. 31, Abb. 27; Decker and Herb 1994: 307, J31 (here with detailed bibliography). For the action-chain hunting – transportation – slaughtering see also Herb and Förster 2009. 58 In most cases, the hedgehogs are also presented in baskets, e.g. in the tomb of Ꜣḫt-ḥtp (Saqqara, 6th dynasty), see Droste zu Hülshoff 1980: 87, Nr. 37, Abb. 32. 59 However, the correlation hedgehog = ḥntꜢ is very insecure, as this term could also be the name for porcupine, see Droste zu Hülshoff (1980: 13–16), and also Vernus and Yoyotte 2005: 146. 60 For animals and their economical relevance, see esp. Herb and Förster 2009: 25–33, with fig. 14. 61 Note the addendum for a hare (sẖꜤ.t) in the desert hunting scene in the tomb of Baqet III, unique to my knowledge, see below (Newberry 1893: 47, pl. IV; Decker and Herb (1994: 319 f.; J66; Taf. CXLVI. J66).
80
Sonja Gerke
of the scenes, whereas the animals are more or less to be seen as an attachment – a necessary and important attachment, but nonetheless second-rank.
1.3 Combination of image and text As mentioned above, scenes of temple and tomb decoration can be found which show animals with addenda, giving the names of the specific species in hieroglyphs. The most famous example for this is the so-called chamber of the seasons (Edel 1961; id. 1963; Edel and Wenig 1974) in the sun temple of Neuserre, that has been mentioned above already in the context of the organization of knowledge (with regard to onomastica and lexica) and ‘interest in classification’ (Brovarski 1987: 51). The iconographic program of the chamber of the seasons consists of natural scenes which show specific events happening cyclically during the different seasons,62 e.g. the mating and birth of animals, the migration of fish and birds, as well as agricultural activities carried out by men, such as sowing and harvesting. In the case of the scenes which show animals, it is conspicuous that nearly all depictions have an addendum with the specific animal names, for example, the scenes where different animals give birth parallel to each other in space (which means at the same time) during the šmw-season (i.e. summer) (see fig. 3) (Edel 1961: 245; Edel and Wenig 1974: Taf. 14).63 The animals are described as smꜢ.t ‘wild cow’, nwḏ.w ‘antelope (addax)’, gḥs ‘gazelle’, bꜢ ‘leopard’, and nṯr.t ‘cheetah’. Especially in the cases of the two felines, the inscriptions are very helpful, as they are hard to distinguish from each other only by means of iconography (see Kleinsgütl 1997).
Fig. 3: Giving birth of different animals in the chamber of the seasons (Abu Gurab)
62 šmw = summer; Ꜣḫ.t = inundation. The third ancient Egyptian season (pr.t = winter) is missing in the reliefs. 63 Relief Z 250, Berlin 20036.
All Creatures Great and Small
81
Other examples of tagged animals derive again from the huge number of scenes of tomb decoration. Here, hunting scenes especially that show the tomb owner and his sons hunting in the deserts offer, in many cases, tagged animals, e.g. the desert hunt of Baqet III of Beni Hassan (Newberry 1893: 47, pl. IV; Decker and Herb 1994: 319 f., J66). The different desert animals are depicted in most cases in pairs, one after the other, and nearly all of them have short addenda that give the names of the single animals or describe in a few words the portrayed action.64 However, not all animals show an addendum, for example, the gazelle on the left or the animals at the outer right end of the scene lack descriptions. In contrast, even the three – to modern thinking non-existent – mythical creatures Seth-animal, griffin and serpopard that appear quite naturally among the real desert animals, have their names as a supplement.65 In this scene, the name for the hare (sẖꜤ.t) can also be found. Except for this example in the tomb of Baqet III, only two other sources for sẖꜤ.t are known: the first one is from a further contemporary tomb in Beni Hassan (Newberry 1893: pl. XIII),66 the second from the already mentioned onomasticon Tebtynis I (Osing 1998: 121 (N 9, 24)).67 The whole hunting-scene makes the impression of some kind of list, in which the animals in most cases are shown by pairs with their names written nearby, but primarily it represents a dramatic scene depicting human and animal interaction. Another significant and somehow comparable example of animal classification can be found in two scenes in the lower half of the same tomb wall of Baqet III:68 In the lowermost register of this wall, some aquatic activities are depicted, showing men on boats fishing with nets, tearing out papyrus-stems (‘Papyrusraufen’), or poking each other with poles (‘Fischerstechen’).69 In the waterway beneath the boats, different fishes (as well as one hippopotamus) can be found again. Above the fishermen hover altogether 29 different birds and three bats70 (two in top-view, the one in the middle in side-view). All animals are standing on four small half-registers, which
64 The inscriptions from right to left are: jb ‘goat’, wnš ‘jackal’, sꜢb ‘jackal’, bꜢ-šmꜤ ‘Upper Egyptian leopard’, bꜢ-mḥ ‘Lower Egyptian leopard’, šsꜢ ‘bubalis antelope’; nḏr ḥbn ‘seizing the antelope’, sẖꜤ.t ’hare’, hnn ‘deer’, jꜢm.t ‘female ibex’, nrꜢ.w ‘ibex’, bꜢ ‘leopard’, nḏr […] ‘seizing […]’, šꜢ ‘Seth-animal’, sfr ‘griffin’, s(w)ḏꜢ ‘serpopard’; Ꜣbw ‘elephant’; nḏr.t nrꜢ.w mꜢ-ḥs ‘seizing the ibex (by) the lion’; sṯt ‘pouring out the semen’. 65 For the name of the griffin and its meaning as animal in ancient Egypt see Gerke 2014. 66 See also Erman and Grapow 1955: IV, 268.11. 67 The few sources for the name of the hare is comparable with the missing nomenclature of the hedgehog, mentioned above. The hare also played a small, but not unimportant role in ancient Egypt, and also found its way into the hieroglyphic sign-system (E34); interestingly, this never stands for the hare itself but is used only as phonetic sign wn. See also Brunner-Traut 1977: 1023 f. 68 Newberry 1893: 47, pl. IV. 69 For these scene-types, see Herb 2001. 70 Davies 1949: 14, calls the depicted animals ‘bats’, but the more precise zoological term would probably be ‘fruit bats’ or ‘megabats’.
82
Sonja Gerke
are not in a scenic context and have no point of contact to the surrounding scenes, and are therefore even more obviously to be seen as some kind of listing (Hoffmann 2005: 197). All animals have their names written nearby; most of them could be taxonomically identified by N. M. Davies (Davies 1949: 15–20). Very interesting is the compilation of bats – flying mammals – with birds in one context, leading to the conclusion that these different species were seen by the ancient Egyptians as connected or even related to each other, as they are put together in one context.71 Sources that include a combination of image and text reveal similar conclusions concerning the categorization of animals as the investigation of the textual and iconographical sources. The examples cited above show that distinct animals in distinct situations or contexts can be accompanied by their names as hieroglyphic inscriptions, which shows a clear will and need for differentiation between animals or distinct animal species. But there is no system or standard visible that could reveal a general use or ‘non-use’ of these addenda. Furthermore, examples like the ‘chamber of the seasons’ or the bird list of Baqet III are unique examples that have, as yet, no equivalent counterpart. Even the hunting scene of Baqet III (see fig. 5) represents a peculiarity with a lot of inscriptions and the list-like representation of animals by pairs.
2 Models of categorization Following the investigation of the different levels occasionally showing classifications of animals in ancient Egypt, the question arises of whether the ancient Egyptians classified or categorized the animal world according to distinct models or systems.
2.1 Categorization by habitat E. Hornung states in his paper on the meaning of the animal in ancient Egypt: Es entspricht dem Sinn des Ägypters für klare, großzügige Gliederung, dass sich seine Systematik des Tierreiches in erster Linie an den Lebensbereichen der Tiere orientiert, also an den großen Linien der Schöpfungswelt, nicht an einzelnen Merkmalen der Lebewesen selbst (Hornung 1967: 69 f.).
Actually, the categorization of animals by habitat seems to be the most obvious and can be traced throughout all Egyptian epochs (Vernus and Yoyotte 2005: 15). J. Assmann, for example, identifies a clearly visible categorization of the animals men-
71 The depictions of bats in ancient Egypt are very rare, see Störk 1977: 263 f.
All Creatures Great and Small
83
tioned in the Egyptian hymns in creatures of the water, creatures of the sky, and creatures of the earth, while creatures of the earth are again sub-divided into creatures that live upon the earth (cattle, small livestock, etc.) and creatures that live in the earth (mice, vermin, etc.; Assmann 1983: 206, note n). The three great environments earth, sky, and water are also represented in hieroglyphic signs ( , , ) and this system can also be retraced in the nomenclature of animals itself. This can occur, for example, in paraphrasing terms:
jmj.w-mw ‘those who are in the water’ (Erman and Grapow 1955: I, 74.7; TLA Lemma-Nr. 25480). The two classifying signs fish and crocodile show that the expression includes both types of animals, which share the common habitat water.72
jr.w-p.t ‘those who belong to the sky’ (Erman and Grapow 1955: I, 104.4; TLA Lemma-Nr. 28640). It must be noted here that the classi fier ‘bird’, can also be used for other flying animals like insects, which could therefore also be meant by this expression (see e.g. Meeks 2010: 273 f.).
jmj-tꜢ ‘those who are in the earth’ (Erman and Grapow 1955: I, 75.17; TLA Lemma-Nr. 25860). Although this one has no specific animal-classifier, the context clarifies an interpretation of this expression as one to be used for snakes.
Additionally, further terms which name specific animals do exist, as seen in the following two examples: sꜢ-tꜢ ‘son of the earth’,73 a name for a snake. gr-n-p.t, which means literally ‘the fowl of the sky’, but is said to develop later to Demotic grmp (Erichsen 1954: 585) and Coptic (Crum 1939: 828), which means dove and is probably already used with this connotation in the New Kingdom (Erman and Grapow 1955: V, 181.1 f.; Vernus and Yoyotte, 2005: 81)
Finally, two examples shall be cited here to exemplify that specific animals are not thought to be limited to only one habitat:
72 But this term may also appear with only one classifier (fish), as within a hymn on a stela of Ramses IV (Cairo JE 48831; 20th dynasty): Kitchen 1983: 404.8, l. 4; Assmann 1999: Nr. 220, l. 9, see below. 73 Later Coptic snake/basilisk (Crum 1939: 359), see Erman and Grapow 1955: III, 410.16 f.; TLA Lemma-Nr. 126130; Vernus and Yoyotte 2005: 81.
84
Sonja Gerke
Ꜣpd.w-n-mw ‘the birds of the water’ (TLA Lemma-Nr. 113) shows that the birds are not only limited to the sky, but that there is some kind of waterfowl. jmj-tꜢ-mw ‘those who are in the earth and in the water’ (TLA Lemma-Nr. 500045) makes clear that snakes can live both on the earth and in the water.
2.2 Categorization by appearance Despite the statement of E. Hornung that the ancient Egyptians classified their animals only ‘an den großen Linien der Schöpfungswelt, nicht an einzelnen Merkmalen der Lebewesen selbst’ (Hornung 1967: 70), classification by the outward appearance of several animals can also be proved to have existed in ancient Egypt. This becomes particularly apparent, for example, in the administrative title jmj-rꜢ Ꜥb wḥm(.t) šw(.t) nšm(.t) ‘Overseer of horned, hoofed, feathered and scaled’, which is attested for officials of the Middle and New Kingdoms (Ward 1982: Nr. 64; Ayedi 2006: Nr. 68; Loret 1916 f.; Quirke 1996: 666–669).74 For this title, a significant part of four specific animals (horn, leg with hoof, feather and scale) is used and stands pars pro toto for the whole creature, or rather the whole species, like cattle, bird or fish. However, it is not entirely certain which actual duties the official who bore this title had to perform,75 nor what specific creatures are meant exactly by the abbreviations. The connections between ‘horn and cattle’, ‘feather and bird’, or ‘scale and fish’ in general seem to be quite obvious, although the distinct kinds of animals – which kind of cattle, bird or fish is supposed exactly – must remain uncertain. In the case of the hoof (or the hoofed leg, respectively), the identification of the linked animal is much more problematic. In his hieroglyphical sign-list, A.H. Gardiner claims the sign (F25) to be the ‘leg and hoof of an ox’, but states that it can also be transferred to donkeys (Gardiner 1957: 464). Given the fact – or better the assumption – that the horn stands for cattle, it would not be reasonable to cite another part for cattle. Clearly a much wider range of animals that could be meant by exists, especially when looking at iconographical sources, where the legs of cattle, ibex, gazelles, sheep, donkeys or even horses are depicted in a way very similar to each
74 See also Jones 2000: Nr. 338: jmj-rꜢ Ꜥb nb ‘Overseer of all horned’; Ward 1982: Nr. 62: jmj-rꜢ Ꜥb.w ‘Overseer of horned’, and Nr. 63: jmj-rꜢ Ꜥb wḥm(.t) ‘Overseer of horned and hoofed’ (= Ayedi 2006, Nr. 66); Ayedi 2006: Nr. 67: jmj-rꜢ Ꜥb jmj-rꜢ wḥm.t šw.t ‘Overseer of horned, overseer of hoofed and feathered’, and Nr. 69 jmj-rꜢ ob jmj-rꜢ šw(.t) nšm.w ‘Overseer of horned, feathered and scaled’. 75 Quirke 1996: 669, suggests that the person thus designated was ‘an official charged with transport of livestock and/or meat’.
All Creatures Great and Small
85
other and resembling the sign .76 It is also possible that the ‘exact’ nomenclature of the species was not important at all, and so the terms wḥm(.t) šw(.t) nšm(.t) can be seen as some kind of ‘generic terms’.
2.3 Categorization by locomotion A third possibility of animal-classification is the classification by locomotion. So the paraphrase ntj nb ḥr tꜢ šm.w ḥr rd.wj ntj m Ꜥḫ ḥr pꜢjj m ḏnḥ.w=sn ‘all which is on earth and walks on its feet and which is ‘on high’ and flies with its wings’ (Assmann 1999: Nr. 92 [l. 81 f.]; Sandman 1938: 93–96 [l. 8])77 shows clearly the differentiation between ‘walking’, i.e. movement on the ground with legs and feet, and ‘flying’ i.e. movement off the ground with wings. This again is reflected in the nomenclature of creatures, for example: pꜢj as a verb means ‘to fly’ (Erman and Grapow 1955: I, 494.1–12; TLA Lemma-Nr. 58780) but it is also used as a general term for birds pꜢjj.t (Erman and Grapow 1955: I, 494.13) resp. pꜢjj.w (Erman and Grapow 1955: I, 494.15–17; TLA Lemma-Nr. 59070) ‘those who fly (up)’ often with the addition ḫnn.t ‘those who come down’ (Erman and Grapow 1955, III, 288.4–6, see also Erman and Grapow 1955: 494.14; TLA Lemma Nr. 118310). ḥfꜢ ‘to snake/twist’ (Erman and Grapow 1955: III, 73.6–8; TLA Lemma-Nr. 104340) resp. ḥf ‘to revere/praise‘ (Erman and Grapow 1955: III, 73.9–12; TLA Lemma-Nr. 104350)78 becoming with ḥfꜢ.t (Erman and Grapow 1955: III, 72–73.5) a general term for vermin, snakes and worms. Those three briefly presented models of animal classification represent a very selective choice; many more types of models could be added here, for example, classi-
76 In modern taxonomy cattle, ibex, gazelle and sheep belong to the order of ‘even-toed ungulates’, whereas donkeys and horses belong to the ‘odd-toed ungulates’. Though the hoof-forms of both orders are easily and visibly distinguishable from each other, in Egyptian reliefs and paintings they are obviously hardly differentiated. This may be because the animals are always shown in side-view and so whether the foot has one or two toes is not visible. 77 Great Hymn to Aten, New Kingdom, 18th dynasty (Amarna). 78 The connection between the two meanings of the word may derive from the crooked or submission posture while praising, which can be compared with creeping on the ground.
86
Sonja Gerke
fication by gender, by purpose, or by function.79 Nevertheless, it becomes obvious that in classification by habitat, appearance and locomotion – although these are seen separately and clearly distinguished from each other – each also determines the other. This means, strictly speaking, whatever lives on the ground needs feet to walk, whatever lives in the air needs wings and feathers to fly, and whatever lives in the water needs the ability to swim and usually has scales. In simple words: there is no bird with scales, no fish with legs, and no quadruped with wings.80
3 Hierarchy and canonization – the hymns of the New Kingdom The last part of this paper deals with the question of whether some kind of Egyptian hierarchical system can be traced which would arrange the animal world in a dependant order (also considering the connection to other creatures, such as mankind and gods) and whether this system, or rather the whole ancient Egyptian view of the animal world, can be looked at as a standardized or canonized concept. To turn to the creation of hierarchy regarding the animal world in general, some textual sources appear to be quite appropriate owing to their general structure, meaning the listing of animals within a continuous text: the Geschöpfkataloge (meaning catalogues of creatures), so called by J. Assmann, considers specific sections especially within the hymns of the New Kingdom (Assmann 1983: 206, note n; id. 1999, 254).81 Hymns and prayers to gods and goddesses can be found throughout a period of more than 2500 years in ancient Egypt, written on tomb and temple walls, on papyri, statues, stelae, ostraca and coffins. As variable as the text media are the texts themselves. Generally, they praise the god or the goddess they are addressed to, for his or her effectiveness as the sun, life or creation deity and therefore form a medium of communication between gods and humans (Knigge Salis and Luiselli 2013: 145 f.). The texts are grouped together by the appearance of special keywords which can be found at the beginning of each text and have something to do with adoration, worship or veneration.82 Some of these texts, especially from the New Kingdom, include small chapters in which the creation initiated by the addressed deity – in most cases the sun
79 See in particular Vernus and Yoyotte 2005: Bestiaire, 15–19; 62–93. 80 Excluded are mythical creatures such as the griffin, which has four legs (feline body) and the wings and head of a bird (in most cases that of a falcon); for this, see Gerke 2014 on the treatment of fantastic animals, in general, see also Vernus and Yoyotte 2005: 632–697. 81 Here called ‘Listen der Geschöpfe’. 82 In most cases, the term dwꜢ is used.
All Creatures Great and Small
87
god – is praised. This praising is expressed by an enumeration of the specific deeds and creatures the deity has made, such as gods, men, animals, plants, towns, etc. Lists, however, are predestined to function as markers of hierarchies, or at least are supposed to be. When listing several objects, topics or creatures, the human mind searches automatically for a deeper sense, an organization or a meaningful order in the structure – be it chronological, alphabetical, hierarchical, or whatever. The ‘catalogues of creatures’ can vary extremely in their length, as well as in the number and order of the creatures and topics listed. For hymns including catalogues of animals, J. Assmann cites only five examples (Assmann 1999: 254)83 but it is possible to add more texts, which, however, cannot be discussed entirely within the scope of the present paper.84 At this point, only a few examples will be given, which make no claim to completeness. Nevertheless, it is still possible to discuss some general chains of
83 Tomb of Tjanefer, pBoulaq 17, Hibis 2, stela of Baki and pChester Beatty IV. For references, see the following note 86. 84 Hymns of the New Kingdom that include lists of animals: 1. pBoulaq 17 (for further sources of this text, see Luiselli 2004, XX–XXI; 17th/18th dynasty): Luiselli 2004: I.18–20, I.24, III.30–39, III.66–68; Assmann 1999: Nr. 87A (l. 15–25), 87E (l. 107–120), 87F (l. 145–148). 2. Tura-Hymn (18th dynasty): Bakir 1943, 83–91 (l. 16; l. 22–23); Assmann 1999: Nr. 88 (l.40 f., l. 55–58). 3. Hymn in the Tomb of Merire, (further sources: Ahmes, Haia; 18th dynasty): Sandman 1938: 7 (l. 1 f.); Assmann 1999: Nr. 95 (l. 8–10). 4. Tomb of Apy, Small Hymn to Aten (further sources: Merire, Ani, Mahu, Tutu; 18th dynasty): Sandman 1938: 11–12 (l. 3); Assmann 1999: Nr. 91 (l. 16–20). 5. Tomb of Aya, Great Hymn to Aten (18th dynasty): Sandman 1938: 94 (l. 8); Assmann 1999: Nr. 92 (l. 79–82). 6. Stela of Baki (Louvre 2660; 18th dynasty): Moret 1909: pl. XIX (l. 3, l.5 f., l. 12); Assmann 1999: Nr. 170 (l. 10–14, l. 30 f.). 7. pLeiden I 344 (18th/19th dynasty): Zandee 1992: Taf. 2 (l. II.1–3); Assmann 1999: Anhang Nr. 1, 4–13. 8. Tomb of Tjay / TT 23 (19th dynasty): Assmann 1983: Nr. 16 (l. 3 f.); Assmann 1999: Nr. 104 (l. 7–11). 9. Book of the Dead of Hunefer / pBM 9901 (19th–20th dynasty): Budge 1899: 8 (l. 7–10); Assmann 1999: Nr. 42A (l. 7–12). 10. pChester Beatty IV (19th–20th dynasty): Gardiner 1935: pl. 15 (rto l. 7.5–7); Assmann 1999: Nr. 195 (l. 115–123). 11. Tomb of Amenmessu / TT 373 (19th–20th dynasty): Assmann 1983: Nr. 253 (l. 18–25); Assmann 1999: Nr. 97 (l. 20–27). 12. Tomb of Tjanefer / TT 158 (20th dynasty): Assmann 1983: Nr. 156 (l. 20–26); Assmann 1999: Nr. 108 (l. 20–26). 13. Stela of Ramses IV. (Kairo JE 48831; 20th dynasty): Kitchen 1983: 404.8 (l. 4–5); Assmann 1999: Nr. 220 (l. 9–11).
88
Sonja Gerke
thought even by comparing only a few texts85 which are interesting for the topic of classification and creation of hierarchy of the animal world. To show the general structure of these texts, two examples are cited here:86 You made men, small cattle and cattle, all that came into (16) jr=k rmṯ.w Ꜥw.t being and all that exists. (…) mnmn.t ḫpr.t wnn.t nb.t (...) (22) jr=k n=sn Ꜥw.t You made for them small cattle [and cattle and eve[mnmn.t] pꜢjj [ḫnn nb] (...) rything], which flies up [and lowers down(?)] (…) [Tura-Hymn (18th dynasty)] (VII.5) (...) {m} ḏd (VII.6) (...) He who spoke with his mouth and those who exist came into being: Men and gods, cattle and all small cattle rꜤ=f ḫpr m wn.w m rmṯ.w nṯr.w mnmn.t Ꜥw.t nb together with all that flies up and lowers down. mj-ḳd=st pꜢjj.t ḫnn{=j}(.t) [pChester Beatty IV (19th–20th dynasty)] (VII.7) r-Ꜣw J. Assmann suggests that the creatures listed in this context are to be seen as the ‘children of the deity’ (Assmann 1999: 254) and he outlines the structure of the catalogues with a more or less hierarchically organized scheme including (I) men, (II) gods, and (III) animals, the latter being further subdivided into animals of the water (fish), animals of the air (birds), and animals that live in the earth (vermin, mice, and fleas) or on the earth (cattle and small livestock), respectively (Assmann 1983: 206, note n; see also above). Furthermore, he emphasizes that ‘Auffallend ist der Vorrang der Menschen vor den Göttern’ (Assmann op.cit.)87, leading him to the question: ‘Sind die Menschen vor den Göttern entstanden?’ (Assmann op.cit.).88 Thus, according to J. Assmann, these catalogues obviously carry hierarchical information about the
14. pBerlin 3048 (22nd dynasty): Möller 1970: Taf. 42, l. VIII.2; Assmann 1999: Nr. 143, l. 156–158. The examples cited are not to be seen as a complete catalogue of sources for Geschöpfkataloge. As said before, those lists of creation-acts and animals can also be found in other contexts. A famous example is a section in the Teachings for Merikare (cf. Assmann 1999: 29 with note 27), where a hymn-like speech is embedded that also includes a small catalogue of creatures, naming, among others, cattle, birds, and fish. The reason for the limited focus solely on the sections of the hymns of the New Kingdom lies with the limited space of the present paper, but as the hymns of the New Kingdom represent a reasonable and comprehensible, more or less self-contained selection, they are suitable for a separate and, within this scope, manageable investigation. They were chosen after the compilation by Assmann 1999. 85 See note 84. 86 For particular bibliography, see note 84. 87 However, a few exceptions are cited here. 88 Here with reference to Sauneron and Yoyotte 1959: 75.
All Creatures Great and Small
89
creatures listed, at least as far as their chronological order of creation is concerned.89 But can those text-sections really fulfil these expectations? When comparing the above-mentioned fourteen text-sections90 to come to probable conclusions concerning the order or hierarchy of the listed elements, it is necessary to take a closer look at the particular creatures, which shall be done in the following, taking J. Assmann’s scheme (as mentioned above) as a basis (the numbers in brackets show the number of appearances in the fourteen investigated texts, see note 86):91 I. Men: 92 / 93 / 94 / 95 / 96 rmṯ.w ‘men’. This term is the most common term used to describe mankind and appears in nearly all text passages. In most cases, it is used in parallel – or rather: contrast – to the gods, e.g ḳd rmṯ.w jr nṯr.w ‘he who built men, he who made the gods’97 (the precedence can vary, see below), but it can also be mentioned alone, especially in the texts of the Amarna period. (12/14) 98 / 99 / 100 ḥnmm.t, often translated as ‘heaven’s folk‘ or ‘sun-folk’. The origins of this designation are not altogether clear, but it is also used as a collective term for mankind, maybe in a more
89 See also Assmann (1983: 354, note n): ‘Konzeptionen einer Reihenfolge der Schöpfung, wie sie z.B. das 6-Tage-Schema der Genesis darstellt sind in ägyptischen Texten äußerst schwer faßbar. (...) Die wenigen Texte, die einen narrativen Bericht und daher eine zeitlich entfaltete Konzeption der Schöpfung geben, wie z.B. der Turahymnus ÄHG Nr. 88 und der Hymnus Hibis 32 (=ÄHG Nr. 129) lassen zuerst die Götter als die Glieder, Worte, Gedanken des Schöpfergottes entstehen (vgl. zu Text 149), sodann den Raum mit Himmel und Erde und schließlich die Lebewesen, wobei die Menschen den Anfang machen.’ 90 See note 86. 91 The particular references of the different dictions are given in the footnotes; the line numbers refer to the editions cited in note 86. Only the creatures (men, gods and animals) are cited in the following. Other elements, such as plants, trees, cities or abstract terms (e.g. ‘all that exists’), are omitted. 92 pBoulaq, I,18; III, 39; Tura-Hymn, 16. The hieroglyphs as shown here are adopted from Bakir 1943. The specific signs are not recognisable on the photographs, so the exact spelling must remain uncertain; Stela of Baki, 3, 5; Book of the Dead of Hunefer, 10; pChester Beatty IV, 7.6; Stela of Ramses IV., 4; pBerlin 3048, VIII.2. 93 Hymn in the Tomb of Merire, 1. 94 Tomb of Apy ‘Small Hymn to Aten’, 3; Tomb of Aya ‘Great Hymn to Aten’, 8. 95 pLeiden I 344, II.1. 96 Tomb of Tjanefer, 3. 97 pLeiden I 344, II.1 f. 98 pBoulaq III.31; pLeiden I 344, II.3. 99 Tomb of Tjay, 3. 100 Tomb of Amenmessu, 7.
90
Sonja Gerke
mythological context (Serrano 1999; Nibbi 1991; Assmann 1983: 354, note n; a new lexicographical study is Jacoby 2015, unpublished master’s thesis ‘Zur Identität des sogenannten Sonnenvolkes (ḥnmm.t)’). In most cases of the appearance within the catalogues, the nourishment of the ḥnmm.t is mentioned.101 (4/14) 102 rḫjj.t, lit. ‘peewit (people)’, is traditionally to be seen in the sense of subject, concerning the people who subjects of the king, meaning again mankind (see Kaplony 1980: 417–422). (1/14) II. Gods: 103 / 104 / 105 nṯr.w ‘gods’. This term is always used in plural and in most cases in parallel – or rather: contrast – to mankind (see rmṯ.w). (6/14) III. Animals: of the water: 106 /
107 /
108 rm.w ‘fish’ (2/14)
109 jmj.w mw ‘those who are in the water’ (1/14) of the air: 110 /
111 /
112 Ꜣpd.w ‘birds’ (2/14)
113 ḫnws ‘gnat(?)’114 (1/14)
101 But in the Tomb of Amenmessu, 7, the term by context seems to be used as a synonym for rmṯ.w: (7) ḫpr.n ḥnmm.t (8) ḥr-sꜢ ḳmꜢ.n=k ḫpr.n Ꜥw.t nb.t ḫft-wḏ.n=k ‘The sun-folk came into being after what you created, all small livestock came into being in accordance to what you commanded’. 102 Book of the Dead of Hunefer, 7. 103 Stela of Baki, 3; Tomb of Tjanefer, 3. 104 pLeiden I 344, II.2; pChester Beatty IV, 7.6; pBerlin 3048, VIII.2. 105 Stela of Ramses IV., 4. 106 pBoulaq III.32, here with the addition: jtr.w ‘(of) the stream’, see also note 110 (stela of Baki). 107 Stela of Baki, 6. 108 Stela of Baki, 12, here with the addition: ḥr jtr.w ‘in the stream’, see also note 108 (pBoulaq). 109 Stela of Ramses IV, 4. 110 pBoulaq III.33, here with the addition: gnẖ p.t ‘which fly (in) the sky’. 111 Stela of Baki, 6. 112 Tomb of Tjanefer, 3. 113 pBoulaq III.36. 114 See Erman and Grapow 1955: III, 290.2 f. and TLA Lemma-Nr. 117830.
All Creatures Great and Small
91
115 pwjj ‘flea’ (1/14) 116 ḳjj ‘bug(?)’117 (1/14) 118 /
119 /
‘what flies up’ (6/14) 124 / 125 /
120 / 126 /
121 /
122 /
123pꜢjj
127 ḫnn ‘what comes down’ (4/14)
128 ntj m Ꜥḫ Hr pꜢjj m ḏnḥ.w=sn ‘what is ‘on high’ and flies with his (lit. their) wings’ (1/14) in the earth: 129 sꜢ Ꜥpnn.t ‘son of the Ꜥpnn.t-snake(?)’ (1/14) 130 pn.w ‘mouse’ (1/14) 131 /
132 ḏdf.t ‘vermin/snakes’ (2/14)
133 ḥfꜢ.t ‘vermin/snakes’ (1/14)
115 pBoulaq III.37. Assmann 1983: 206, note n, counts the flea among the animals that live in the earth (together with mice). But the determinative supposes that the ancient Egyptians saw fleas as more related to the birds, i.e. the flying animals, and they are therefore here counted as animals of the sky. 116 pBoulaq III.39, here with the addition m ḫt nb.t ‘in every wood/tree’. 117 Luiselli (2004: 30) translates ‘birds’ according to Erman and Grapow 1955: V, 17.2; here the translation of Assmann 1999: 87E l. 120 ‘bug’ is preferred by context. 118 Tura-Hymn, 22. The hieroglyphs as shown here are adopted from Bakir 1943. The specific signs are not recognisable on the photographs, so the exact spelling must remain uncertain. 119 Hymn in the Tomb of Merire, 1. 120 pLeiden I 344, II.3. 121 Tomb of Tjay, 3. 122 pChester Beatty IV, 7.6. 123 Stela of Rames IV., 4. 124 Hymn in the Tomb of Merire, 1. 125 pLeiden I 344, II.3. 126 Tomb of Tjay, 3. 127 pChester Beatty IV, 7.6. 128 Tomb of Aya, Great Hymn to Aten, 8. 129 pBoulaq III.35. 130 pBoulaq III.38, here with the addition m bꜢbꜢ=sn ‘in their holes’. 131 pBoulaq III.37. 132 Hymn in the Tomb of Merire, 1. 133 Tomb of Tjanefer, 3.
92
Sonja Gerke
on the earth: 134 / 141 /
135 /
136 /
142 /
137 /
143 /
144 /
138 / 145 /
139 /
140 /
146 Ꜥw.t ‘small livestock’
(10/14) 147 jwꜤ ‘small cattle(?)’148 (1/14) 149 ḥw.t ‘small cattle(?)’150 (1/14) 151 / 156 /
152 /
153 /
154 /
155 /
157 mnmn.t ‘cattle’ (8/14) 158 ntj nb ḥr tꜢ šm.w ḥr rd.wj ‘all which is on earth and
walks on its feet’ (1/14)
134 pBoulaq I.18. 135 pBoulaq III.66. 136 pBoulaq III.68 here with the addition nt ḫꜢs.t ‘of the desert’. 137 Tura-Hymn, 16. The hieroglyphs as shown here are adopted from Bakir 1943. The specific signs are not recognisable on the photographs, so the exact spelling must remain uncertain. 138 Tura-Hymn, 22. The hieroglyphs as shown here are adopted from Bakir 1943. The specific signs are not recognisable on the photographs, so the exact spelling must remain uncertain. 139 Hymn in the Tomb of Merire, 1 (determinative uncertain). 140 Tomb of Apy, Small Hymn to Aten, 3. 141 Tomb of Aya, Great Hymn to Aten, 8. 142 Stela of Baki, 5. 143 pLeiden I 344, II.2. 144 pChester Beatty IV, 7.6; Gardiner 1935: pl. 15 gives the hieroglyphic transcription with, but notes that ‘(...) is represented by a sign completely identical with (...)’. 145 Tomb of Amenmessu, 8; after this entry appears a gap, where originally also some kind of cattle must have been quoted: [Ꜥw.t] nb ḥr-tp ḫꜢs.wt=sn ‘every [type? of small livestock?] within their deserts’. 146 Tomb of Tjanefer, 3. 147 Stela of Ramses IV, 4. 148 Probably alternative writing for Ꜥw.t, see TLA, Lemma-Nr. 22370; Erman and Grapow 1955: I, 50.6 f. quote one text source from the 20th dynasty and three sources from the Late Period. 149 pBerlin 3048, VIII.2. 150 Probably alternative writing for Ꜥw.t, see TLA 102380; Erman and Grapow 1955: III, 45.7 quote only this text. 151 pBoulaq III.30. On this exceptional and unique writing, see also Goldwasser 2002: 74. 152 pBoulaq I.20; Tura-Hymn, 16. The hieroglyphs as shown here are adopted from Bakir 1943. The specific signs are not recognisable on the photographs, so the exact spelling must remain uncertain. 153 Tomb of Apy, Small Hymn to Aten, 3. 154 Tomb of Aya, Great Hymn to Aten, 8; Book of the Dead of Hunefer, 10; Tomb of Tjanefer, 3. 155 Stela of Baki, 5. 156 pLeiden I 344, II.2. 157 pChester Beatty IV, 7.6. 158 Tomb of Aya, Great Hymn to Aten, 8.
All Creatures Great and Small
93
The compilation of the several elements of the catalogues of creatures shows the enormous variation of creatures mentioned and their nomenclatures. pBoulaq 17 can certainly be seen as the most detailed text, giving detailed and partly unique examples of creation, from men to cattle and mice to bugs and fleas. However, as e.g. pBerlin 3048 shows, the act of creation (of living creatures) can also be described with only three elements: ḳd nṯr.w rmṯ.w Ꜥw.t nb ‘He who built gods, men and all small livestock’. But not only the number of creatures mentioned and the general length of the catalogues may vary; even when the same elements appear in different texts, they can change positions. Elements that often appear together in pairs are e.g. mnmn.t ‘cattle’ and Ꜥw.t ‘small livestock’159 as well as nṯr.w ‘gods’ and rmṯ.w ‘men’.160 In the case of the former pair, mnmn.t appears in five of seven texts in the first place, in the case of the latter pair nṯr.w stands first in three of six cases, which shows that there seems to be no purpose of order or no purpose of hierarchy in the listing of elements, as J. Assmann (among others) has supposed, concerning the chronological order of creation of men and gods (Assmann 1983: 206, note n; Sauneron and Yoyotte 1959: 75). Another inconsequence lies in the different nomenclature for several creatures, i.e. the use of paraphrases for e.g. rm.w ‘fish’ (jmj.w mw ‘those who are in the water’) and Ꜣpd.w ‘birds’ (pꜢjj ‘what flies up’; ḫnn ‘what comes down’; ntj m Ꜥḫ ḥr pꜢjj m ḏnḥ.w=sn ‘what is on high and flies with his (lit. their) wings’). But also the subjects of creation differ from text to text (and also sometimes within one and the same text). The first issue is the creation of beings for their own sake, as in the Tura-Hymn (16) jr=k rmṯ.w Ꜥw.t mnmn.t ‘You made men, small livestock, and cattle’. The second one is the creation of the beings for someone, as again in the Tura-Hymn (22) jr=k n=sn Ꜥw.t ‘You made for them small livestock’.161 The third subject is the creation of nutrition or supply for the beings, as e.g. in pLeiden I 344, II.3 jr Ꜥnḫ m ḥnmm.t pꜢjj ḫnn nb ‘he who makes the supply for the sun-folk (as well as for) those who fly up and lower down’.162 Finally, there are different techniques visible, exemplifying how the listed creatures came into being. This is expressed by different Egyptian verbs.163 There is the
159 In Tura-Hymn, 16; Tomb of Apy, Small Hymn to Aten, 3; Tomb of Aya, Great Hymn to Aten, 8; stela of Baki, 5; pLeiden I 344, II.2; pChester Beatty IV, 7.6; Tomb of Tjanefer, 3. 160 In stela of Baki, 3; pLeiden I 344, II.1 f.; pChester Beatty IV, 7.6; tomb of Tjanefer, 3; stela of Rames IV, 5; pBerlin 3048, VIII.2. 161 The context gives no clear information about who is meant in this context with =sn ‘them’. 162 Nearly the same line is to be found in the tomb of Tjaj, 3. 163 The variation of verbs is also to be seen in the context of textual semantics and rhetorics, as there is no concept recognisable of which verbs are used in which context.
94
Sonja Gerke
active creation with verbs such as jrj ‘to make’,164 ḳd ‘to build’165 or ḳmꜢ ‘to create’,166 whereas in some cases it seems as if the creatures literally ‘come into being’ on their own without any further assistance – expressed by the verb ḫpr167 – or they do it with assistance expressed by the causative form of the same verb sḫpr ‘to bring into being’,168 and additionally, life may simply exist as it is expressed with the Egyptian verb wn.169 As has been shown with this brief examination and comparison of these texts and their elements, it becomes clear that every catalogue differs from every other to a greater or lesser extent and in many features. Nevertheless, they can obviously and easily be compared to each other by the same subject and contents, i.e. the creation of the world and its inhabitants by a deity. For this purpose, it does not seem to matter which specific creatures are named and in which order they are put. What was important was the emphasizing of the act of creation in general, so a long, detailed list of created beings and things or a short expression like ‘he who made all that exists’ can, after all, express the same message, which can be understood in both ways quite clearly.
4 Conclusions – the Ancient Egyptian view of the animal world? At the beginning of this paper, the questions were asked of whether some kind of classification or categorization of the animal world can be traced in ancient Egypt, on which concepts, models or types this classification could be based, and whether a system of hierarchy and canonization can be found within the sources. As has been stated already at the beginning, to answer these questions the selection of sources is basically decisive. As the particular sections of this paper have shown, examples of animal classification in image and script and also classification models by which animals were categorized can actually be found. However, these examples were found because they were specifically searched for. The cited examples show classifications and models that are basically needed when wanting to depict or talk about animals, but they have no intention to sequence, put
164 Tura-Hymn, 16, 22; Tomb of Merire, 1; Tomb of Apy, Small Hymn to Aten, 3; pLeiden I 344, II.2, 3; Tomb of Tjay, 3. 165 pLeiden I 344, II.1; pBerlin 3048, VIII.2. 166 Tomb of Apy, Small Hymn to Aten, 3; Tomb of Aya, Great Hymn to Aten, 8; Book of the Dead of Hunefer, 7; Tomb of Amenmessu, 8. 167 Stela of Baki, 3; pChester Beatty IV, 7.6; Tomb of Amenmessu, 7, 8. 168 Stela of Baki, 3; Book of the Dead of Hunefer, 9; Tomb of Tjanefer, 3. 169 Stela of Ramses IV, 4.
All Creatures Great and Small
95
in order, or examine the animal world, and are all taken from different contexts that have primarily nothing to do with knowledge about animals. The physical description of animals in some textual sources (lexica) is not primarily to be seen as being for the animal’s own sake, but must be looked at in the broader, superordinate context of theology and mythology which are tightly linked to it. Furthermore, it should be noted that such examples as the list of birds in the tomb of Baqet III have no parallels at all; the same is true for many text examples or quotations given in this paper. I have tried to emphasize in this paper that it is very important to differentiate between single examples and generalized statements or assumptions. It is indisputable that all examples cited above show a will to differentiate between several animals and subspecies. But they also show their diversity, that there was no general or the ancient Egyptian view of the animal world, and there was no standardized or centralized canonization of nomenclature or hierarchy of animals, as this was obviously not needed. Egyptological papers dealing with the view of the animal world often communicate the impression that the ancient Egyptians had a very clear and structured system of the animal world (see e.g. Hornung 1967; Vernus and Yoyotte 2005: 15–19, 62–93; Hoffmann 2005), an impression which can easily be made by choosing and citing adequate sources, as was also shown in the present paper. But what look like general models or types are, in fact, special, and in most cases, single phenomena that depend on and are determined by their specific context and purpose, and which form both the basis and the deciding factor for the need of a categorization or a classification model.
References Arnold, Dorothea. 1995. An Egyptian Bestiary. (The Metropolitan Museum of Art Bulletin, 52, 4) New York: The Metropolitan Museum of Art. Assmann, Jan. 1983. Sonnenhymnen in thebanischen Gräbern. (Theben, 1) Mainz: Philipp von Zabern. Assmann, Jan. 21999. Ägyptische Hymnen und Gebete. (Orbis Biblicus et Orientalis) Freiburg/ Göttingen: Universitätsverlag; Vandenhoeck & Ruprecht. Aufrère, Sydney H., and Marguerite Erroux-Morfin. 2001. “Au sujet du hérisson: aryballes et préparations magiques à base d’extraits tirés de cet animal.” In Encyclopédie religieuse de l’univers végétal: croyances phytoreligieuses de l’Égypte ancienne 2, edited by Sydney H. Aufrère, 521–533. Montpellier: Université Paul Valéry-Montpellier III. Al-Ayedi, Abdul Rahman. 2006. Index of Egyptian administrative, religious and military titles of the New Kingdom. Ismailia, Egypt: Obelisk Publications. Bakir, Abd el-Moḥsen. 1943. “A hymn to Amon-Rē’ at Ṭura.” Annales du Service des Antiquités de l’Égypte 42: 83–91. Brovarski, Edward. 1987. “Two Old Kingdom writing boards from Giza.” Annales du Service des Antiquités de l’Égypte 71: 27–52. Brunner-Traut, Emma. 1977. “Hase.” In Lexikon der Ägyptologie Band II, edited by Wolfgang Helck, 1023 f. Wiesbaden: Harrassowitz.
96
Sonja Gerke
Budge, Ernest A. Wallis. 1899. The Book of the Dead, Facsimiles of the Papyri of Hunefer, Anhai, Kerâsher and Netchemet with Supplementary Text from the Papyrus of Nu. London: British Museum. Burkard, Günter, and Heinz-Josef Thissen. 2003. Einführung in die altägyptische Literaturgeschichte. I: Altes und Mittleres Reich. (Einführungen und Quellentexte zur Ägyptologie, 1) Münster: Lit. Burkard, Günter, and Heinz-Josef Thissen. 2008. Einführung in die altägyptische Literaturgeschichte. I: Neues Reich. (Einführungen und Quellentexte zur Ägyptologie, 1.) Münster: Lit. Crum, Walter E. 1939. A Coptic Dictionary. Oxford: Clarendon Press. Davies, Nina M. 1949. “Birds and Bats at Beni Hassan.” Journal of Egyptian Archaeology 35: 13–20. Decker, Wolfgang, and Michael Herb. 1994. Bildatlas zum Sport im Alten Ägypten. Corpus der bildlichen Quellen zu Leibesübungen, Spiel, Jagd, Tanz und verwandten Themen. Teil 1: Text und Teil 2: Abbildungen. Leiden: Brill. Droste zu Hülshoff, Vera von. 1980. Der Igel im Alten Ägypten. (Hildesheimer Ägyptologische Beiträge, 11) Hildesheim: Gerstenberg. Edel, Elmar. 1961. “Zu den Inschriften auf den Jahreszeitenreliefs der ‘Weltenkammer’ aus dem Sonnenheiligtum des Niuserre, I. Teil.” Nachrichten der Akademie der Wissenschaften zu Göttingen, Philologisch-Historische Klasse 8: 209–255. Edel, Elmar. 1963. “Zu den Inschriften auf den Jahreszeitenreliefs der ‘Weltenkammer’ aus dem Sonnenheiligtum des Niuserre, II. Teil.” Nachrichten der Akademie der Wissenschaften zu Göttingen, Philologisch-Historische Klasse 4/5: 89–217. Edel, Elmar, and Steffen Wenig. 1974. Die Jahreszeitenreliefs aus dem Sonnenheiligtum des Königs Ne-User-Re. (Mitteilungen aus der Ägyptischen Sammlung, 7) Berlin: Akademie-Verlag. Erichsen, Wolja. 1954. Demotisches Glossar. Kopenhagen: Munksgaard. Erman, Adolf, and Hermann Grapow. 1955. Wörterbuch der ägyptischen Sprache Bände I–V. Berlin: Akademie-Verlag. Fischer-Elfert, Hans-Werner. 2008. “Weitere Details zur Göttlichkeit der Natur – Fragmente eines späthieratischen Lexikons (Pap. Hal. Kurth Inv. 33 A-C (Halle/Saale)).” Zeitschrift für ägyptische Sprache und Altertumskunde 135: 115–130. Gamer-Wallert, Ingrid. 1970. Fische und Fischkulte im alten Ägypten. (Ägyptologische Abhandlungen, 21) Wiesbaden: Harrassowitz. Gardiner, Alan Henderson. 1935. Chester Beatty Gift, Vol. 2. (Hieratic Papyri in the British Museum Ser. 3) London: British Museum Press. Gardiner, Alan Henderson. 31957. Egyptian Grammar. Being an introduction to the study of hieroglyphs. London: Oxford University Press. Gardiner, Alan Henderson. 1968. Ancient Egyptian Onomastica, 3 vols. London: Oxford University Press. Gerke, Sonja. 2014. Der altägyptische Greif. Von der Geschichte eines “Fabeltiers”. (Studien zur Altägyptischen Kultur, Beihefte 15) Hamburg: Buske. Goldwasser, Orly. 2002. Prophets, lovers and giraffes: wor(l)d classification in ancient Egypt. (Classification and Categorization in Ancient Egypt, 3. Göttinger Orientforschungen, 4. Reihe: Ägypten, 38) Wiesbaden: Harrassowitz. Griffith, Francis Ll., and William M. Flinders Petrie. 1889. Two hieroglyphic papyri from Tanis: I. The sign papyrus (a syllabary) / by F. Ll. Griffith. II. The geographical papyrus (an almanack) / by W. M. F. Petrie. With remarks by Heinrich Brugsch. (Memoir of the Egypt Exploration Fund) London: Trübner. Guichard, Hélène (ed). 2014. Des animaux et des Pharaons. Le règne animal dans l’Égypte ancienne [cet ouvrage accompagne l’exposition Lens, Musée du Louvre-Lens, 5 décembre 2014–9 mars 2015; Madrid, CaixaForum, 31 mars 2015–23 août 2015; Barcelone, CaixaForum, 22 septembre 2015–10 janvier 2016]. Paris: Somogy Éditions d’Art.
All Creatures Great and Small
97
Herb, Michael 2001. Der Wettkampf in den Marschen: Quellenkritische, naturkundliche und sporthistorische Untersuchungen zu einem altägyptischen Szenentyp. (Nikephoros Beihefte 5) Hildesheim: Weidmann. Herb, Michael, and Frank Förster. 2009. “From desert to town: The economic role of desert game in the Pyramid Ages of ancient Egypt as inferred from historical sources (c. 2600–1800 BC). An outline of the workshop’s inspiration and objectives.” In Desert animals in the eastern Sahara: status, economic significance, and cultural reflection in antiquity. Proceedings of an interdisciplinary ACACIA workshop held at the University of Cologne, December 14–15, 2007. Colloquium Africanum 4, edited by Heiko Riemer, Frank Förster, Michael Herb and Nadja Pöllath, 17–44. Köln: Heinrich-Barth-Institut. Hoffmann, Friedhelm. 2005. “Das Göttliche in der Natur – Biologie im alten Ägypten.” MatthiasGrünewald-Gymnasium Würzburg, Jahresbericht 2004/2005, Wissenschaftliche Beilage 4: 196–205. Hornung, Erik. 1967. “Die Bedeutung des Tieres im alten Ägypten.” Studium Generale 20: 69–84. Hornung, Erik, Rolf Krauss and David A. Warburton (eds). 2006. Ancient Egyptian Chronology. (Handbuch der Orientalistik, erste Abteilung: Der Nahe und Mittlere Osten 83) Leiden: Brill. Houlihan, Patrick F. 1996. “A guide to the wildlife represented in the great swampland scene in the offering-chapel of Ti (No. 60) at Saqqara.” Göttinger Miszellen 155: 19–53. Jones, Dilwyn. 2000. An index of ancient Egyptian titles, 2 vols. (BAR International Series, 866 (1 f.)) Oxford: Archaeopress. Kanawati, Naguib. 2011. Mereruka and his family. III.1 The tomb of Mereruka. (Australian Centre for Egyptology: Reports 29) London: Aris & Philipps. Kaplony, Peter. 1980. “Kiebitz(e).” In Lexikon der Ägyptologie Band III, edited by Wolfgang Helck, 417–422. Wiesbaden: Harrassowitz. Kitchen, Kenneth A. 1983. Ramesside inscriptions, vol. VI. Oxford: Blackwell. Klebs, Luise. 1982. Die Reliefs des alten Reiches (2980–2475 v. Chr.). (Abhandlungen der Heidelberger Akademie der Wissenschaften, Philosophisch-Historische Klasse, 3) Hildesheim (et al.): Olms. Kleinsgütl, Dagmar. 1997. Feliden in Altägypten. (Veröffentlichungen der Institute für Afrikanistik und Ägyptologie der Universität Wien, 80; Beiträge zur Ägyptologie, 14) Wien: Afro-Pub. Knigge Salis, Carsten, and Maria Michela Luiselli. 2013. “Ägyptische Hymnen und Gebete.” In Hymnen, Klagelieder und Gebete, edited by Daniel Arpagaus, 145 f. (Texte aus der Umwelt des Alten Testaments, NF 7) Gütersloh: Gütersloher Verlagshaus. Lakoff, George. 1987. Women, fire, and dangerous things. What categories reveal about the mind. Chicago (Ill.): University of Chicago Press. Leitz, Christian. 1997. Die Schlangennamen in den ägyptischen und griechischen Giftbüchern. (Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse/Akademie der Wissenschaften und der Literatur in Mainz, 6) Stuttgart: Steiner. Leonard Jr., Albert. 2000. “Why a hedgehog?” In The archaeology of Jordan and beyond essays in honour of James A. Sauera, edited by Lawrence E. Stager, Joseph A. Greene and Michael D. Coogan, 310–316. Winona Lake, Indiana: Eisenbrauns. von Lieven, Alexandra. 2004. “Das Göttliche in der Natur erkennen: Tiere, Pflanzen und Phänomene der unbelebten Natur als Manifestation des Göttlichen; mit einer Edition der Baumliste P. Berlin 29027.” Zeitschrift für ägyptische Sprache und Altertumskunde 131: 156–172. Liszka, Kate. 2010. “‘Medjay’ (no. 188) in the onomasticon of Amenemope.” In Millions of Jubilees: Studies in Honor of David P. Silverman, vol. 1, edited by Zahi Hawass and Jennifer Houser Wegner, 315–331. Cairo: Conseil Suprême des Antiquités. Loret, Victor. 1916–1917. “Le titre .” Recueil de travaux relatifs à la philologie et à l’archéologie égyptiennes et assyriennes 38: 61–68.
98
Sonja Gerke
Luiselli, Maria Michela. 2004. Der Amun-Re Hymnus des P. Boulaq 17 (P. Kairo CG 58038). (Kleine ägyptische Texte, 14) Wiesbaden: Harrassowitz. Meeks, Dimitri. 2010. “De quelques ‘insectes’ égyptiens: entre lexique et paléographie.” In Perspectives on ancient Egypt: studies in honor of Edward Brovarski, edited by Zahi A. Hawass, Peter Der Manuelian and Ramadan B. Hussein, 273–304. Cairo: Conseil Suprême des Antiquités de l’Égypte. Möller, Georg. 1970. Hieratische Papyrus aus den Königlichen Museen zu Berlin. II: Hymnen an verschiedene Götter. Zusatzkapitel zum Totenbuch, edited by Adolf Erman (unveränd. fotomech. Nachdr. d. Ausg. Leipzig 1905). Leipzig: Zentralantiquariat der DDR. Moret, Alexandre. 1909. Catalogue du Musée Guimet. Galérie égyptienne: Stèles, Bas-reliefs, Monuments divers. (Annales du Musée Guimet, 32) Paris: Leroux. Newberry, Percy E. 1893. Beni Hasan II (Archaeological Survey of Egypt: Memoir 2) London: Egypt Exploration Society. Nibbi, Alessandra. 1991. “A note on the Hnmmt.” Discussions in Egyptology 21: 49–57. Osing, Jürgen. 1982. “Onomastika.” In Lexikon der Ägyptologie Band IV, edited by Wolfgang Helck, 572. Wiesbaden: Harrassowitz. Osing, Jürgen. 1998. Hieratische Papyri aus Tebtunis I, The Carlsberg Papyri 2. (CNI Publications, 17) Copenhagen: Museum Tusculanum Press. Pommerening, Tanja. 2015. “Bäume, Sträucher und Früchte in altägyptischen Listen – eine Betrachtung zur Kategorisierung und Ordnung.” In Die Liste. Ordnungen von Dingen und Menschen in Ägypten, edited by Susanne Deicher, 125–166. Berlin: Kulturverlag Kadmos. Pries, Andreas Henning. 2014. “Onomastika und Soubassements: erkennbare Wege einer intertextuellen Bezugnahme.” In Altägyptische Enzyklopädien. Die Soubassements in den Tempeln der griechisch-römischen Zeit: Soubassementstudien I.2, edited by Alexa Rickert and Bettina Ventker, 1001–1015. Wiesbaden: Harrassowitz. Quack, Joachim Friedrich. 2008. “Lokalressourcen oder Zentraltheologie? Zur Relevanz und Situierung geographisch strukturierter Mythologie im Alten Ägypten.” Archiv für Religionsgeschichte 10: 5–29. Quirke, Stephen. 1996. “Horn, feather and scale, and ships: on titles in the Middle Kingdom.” In Studies in honour of William Kelly Simpson 2, edited by Peter Der Manuelian, 665–677. Boston: Dept. of Ancient Egyptian, Nubian and Near Eastern Art, Museum of Fine Arts. von Recklinghausen, Daniel. 2014. “Monographien in den Soubassements.” In Altägyptische Enzyklopädien. Die Soubassements in den Tempeln der griechisch-römischen Zeit: Soubassementstudien I.1, edited by Alexa Rickert and Bettina Ventker, 29–50. Wiesbaden: Harrassowitz. Reisner, George A. 1910. “A scribe’s tablet found by the Hearst Expedition at Giza.” Zeitschrift für ägyptische Sprache und Altertumskunde 48: 113 f. Sahrhage, Dietrich. 1998. Fischfang und Fischkult im alten Ägypten. (Kulturgeschichte der Antiken Welt 70) Mainz: Philipp von Zabern. Sandman, Maj. 1938. Texts from the time of Akhenaten. (Bibliotheca Aegyptiaca, 8) Bruxelles: Édition de la Fondation Égyptologique. Sauneron, Serge, and Jean Yoyotte. 1959. “La naissance du monde selon l’Égypte ancienne.” In La naissance du monde: Égypte ancienne – Sumer – Akkad – Hourrites et Hittites – Canaan – Israel – Islam – Turcs et Mongols – Iran préislamique – Indie – Siam – Laos – Tibet – Chine, edited by [Anonymous], 17–91. Paris: Éditions du Seuil. Sauneron, Serge. 1989. Un traité égyptien d’ophiologie: Papyrus du Brooklyn Museum No. 47.218.48 et .85. (Bibliothèque Générale, 11) Le Caire: Institut Français d’Archéologie Orientale. Serrano, Jose M. 1999. “Origin and basic meaning of the word Hnmmt (the so-called ‘sun-folk’).” Studien zur Altägyptischen Kultur 27: 353–368. Störk, Lothar. 1977. “Fledermaus.” In Lexikon der Ägyptologie Band II, edited by Wolfgang Helck, 263 f. Wiesbaden: Harrassowitz.
All Creatures Great and Small
99
Vandier, Jacques. 1960. Le Papyrus Jumilhac. Paris: Le Centre national de la recherche scientifique. Vernus, Pascal, and Jean Yoyotte. 2005. Bestiaire des Pharaons. Paris: Agnès Viénot Perrin. Ward, William A. 1982. Index of Egyptian Administrative and Religious Titles of the Middle Kingdom. With a Glossary of Words and Phrases Used. Beirut: American University of Beirut. Wild, Henri. 1953/1939/1966. Le Tombeau du Ti. (Mémoires publiés par les membres de l’Institut Français d’Archéologie Orientale du Caire, 65) Le Caire: Imprimerie de l’Inst. Français d’Archéologie Orientale. Zandee, Jan. 1992. Der Amunhymnus des Papyrus Leiden I 344, Verso. Bände I–IV. (Collections of the National Museum of Antiquities at Leiden, 7) Leiden: Rijksmuseum van Oudheden.
Figures Fig. 1: After Gardiner 1969, pls. I f. Fig. 2: After Sahrhage 1998, 61, fig. 19. Fig. 3: After Edel and Wenig 1974, pl. 14.
Iolanda Ventura
Classification Systems and Pharmacological Theory in Medieval Collections of Materia Medica: A Short History from the Antiquity to the End of the 12th Century Summary: Starting from the two major authorities (auctoritates) of Ancient and Medieval pharmacology, Dioscorides and Galen, my paper provides an overview of the systems of classification of medicamina simplicia derived from plants, animals, metals, and precious stones as recorded in the Latin pharmacological literature from Late Antiquity until the middle of the 12th century, including its intellectual and philosophical background as it determines the rational criteria that regulate the acquisition of knowledge and the systematic ordering and structuring of such medicamina according to their nature, their effect, and their therapeutical properties. In three chronologically structured paragraphs, the paper first examines the two main pharmacological texts written during the Antiquity, viz. Dioscorides‘ De materia medica and Galen‘s De simplicium medicamentorum temperamentis et facultatibus. From there, it moves on to the main types of pharmacological collections produced in the Late Antiquity, the so-called Antibalomena, Dynamidia, and Quid pro quo lists. The period of transition from Late Antiquity to the first centuries of the Middle Ages was marked by the redaction of some well known herbals that will dominate over the Latin pharmacological literature until at least the end of the 12th century, viz. the spurious Alphabetum Galeni, the Herbarium attributed to the Pseudo-Apuleius, and the medical poem De viribus herbarum written by Odo of Meung, but better known under the title of ‘Macer floridus’. The main contribution to the perception and classification of natural elements provided by those works lied in the criteria of structuring and ordering nature according to its relevance and use in medicine. In contrast, the Arabic-Latin pharmacological literature reaching the Western world thanks to Constantine the African‘s translations of Al-Majusi‘s Pantegni and Ibn al-Jazzar‘s Liber de gradibus provided medieval Latin medicine and pharmacology with a deeper and stronger theoretical background that gave contemporary physicians and medical authors belonging, among others, to the Medical School of Salerno, the chance to reason about the rational criteria and elements of recognition and classification of the nature of medicamina, their qualities, and their effects. The ‘theoretical turn’ initiated by Constantine‘s translations and further developed by the authors belonging to, or connected with, the Medical School of Salerno (Bartholomew of Salerno, Platearius, the Magister Salernus, John of Saint-Paul) played therefore a decisive role in the history of rational pharmacology, and will be the object of a long discussion in the third paragraph. My overview ends in the same section with what can be considered the most impressive and influential account of rational pharmacology proDOI 10.1515/9783110538779-005
102
Iolanda Ventura
duced and read during the Middle Ages, viz. the first treatise of the second book of Avicenna‘s Liber canonis, which represented, with the discussion of its long sections on the acquisition of pharmacological knowledge per experimentum and per ratiocinationem, the most complete, the deepest, and the most problematic and debated pharmacological manual of the Late Middle Ages, whose reception and meaning in Medieval universities was exemplified, among others, by John of Saint-Amand and his pharmacological works.
1 Introduction In his 2002 essay on Standardwerke mittelalterlicher Drogenkunde, G. Keil offered a valuable overview of the philosophical and scientific criteria that permeated the classification of plants, animals, and minerals during Antiquity and the Middle Ages (Keil 2002). Starting from the Empedoclean, the Aristotelian, and the Galenic system, he explains how theories such as the basically warm nature of animals and the equally basic cold one of plants, the presence of one or more parts and facultates animae, or of one or more spiritus/pneumata (vegetativa/-vus, naturalis, and animalis), or the combination of primary qualities (cold, warm, dry, humid) helped to distinguish between the three elements of the natural world, and characterize or classify each natural object. He also recalls that, on a more technical (read: medical) level, the long tradition of medical compendia of materia medica is characterized by a preeminence of three different criteria of distinction, classification, and structure of medical substances: the naturalistic (secundum species), the Galenic (secundum qualitates), and the alphabetical (secundum nomina). In the following paragraph, I will attempt to recall these criteria, to introduce the main difficulties and problems they provided, and to refer some of the ways in which they evolved during the Middle Ages. Widely represented during Antiquity, and particularly exemplified in Dioscorides’ De materia medica, is the naturalistic principal of organization and classification of medical substances according to their placement within the three realms of nature and, within each realm, in a particular species (he calls this placement Familienzugehörigkeit). This principle, which is not always completely clear and rationally understandable, especially in the case of the herbs listed in books III and IV of De materia medica, was accompanied, in Dioscorides’ text, by the attention and the space devoted to the description of the external appearance of the natural element (e.g., the plant). The description of the image of the natural element did not become, however, the main interest of compilers of pharmaceutical collections. Such information is often neglected in medieval herbals or pharmaceutical collections, and not just because of the lack of interest of the compilers, their ignorance of the appearance of the natural element (or, maybe, their incapacity to recognize and identify a plant), or – more rarely – the presence of an illustrative corpus summarizing and delivering
Classification Systems and Pharmacological Theory
103
a clear image of a plant. The main reason for the absence of information related to the look of an element of nature can be found in what Keil describes, while talking about the ‘Macer floridus’ (viz., Odo of Meung’s De viribus herbarum), as the distinction between Arzneipflanzen and Arzneistoffen (Keil 2002: 352). What is hidden behind this opposition can be described as follows. In pharmacological literature dominated by the Galenic system (which I will deal with in the next paragraph), the source and the abode of therapeutic qualities, powers, and effects are not the natural element (a plant, an animal) as a whole, but its parts (leaves or roots for plants, blood or organs for animals) and the stuff that can be derived from them (juices, wax, fat etc.). In such a perception of the natural world and its therapeutic power, the external appearance of the natural element may become a secondary detail, or be completely disregarded. A second criterion originated from the Galenic system, and was exemplified in Galen’s De simplicium medicamentorum temperamentis et facultatibus by the content of each entry, viz. the distinction of each substance according to the primary qualities (cold, warm, dry, humid). Although Galen himself did not really use this principle in his De simplicium medicamentorum temperamentis et facultatibus, both the mention of the qualities (e.g., ‘calidum et siccum in primo gradu’) within the description of each substance and the organization of the whole compendium according to the primary qualities will characterize the collections De materia medica from the 11th century onward, as we will see later. Nevertheless, it is necessary to remind the reader that the simple mention of the primary qualities does not entirely explain the historical and theoretical development of the ‘Galenic’ principle. Primary qualities are nothing but one of the principles of identification and classification of substances according to qualities. During the last centuries of the Middle Ages, primary qualities were accompanied and defined more precisely by secondary and tertiary ones. To provide the reader with a clear definition of what a secondary and a tertiary quality is, we may suggest that whereas a secondary quality refers to a general definition of a therapeutical property or of an effect operated on every part of the body (e.g., purification, hydration, strengthening), the tertiary quality defines an action (and therefore an effect) involving a specific organ, section of the body, or secretion (e.g., strengthening the heart, ejecting urine, purifying blood)1. The co-existence of primary, secon-
1 For a definition of secondary and tertiary qualities, cf. the prolog of John of Saint-Amand’s Areolae, where the author gives his definition of the proprietates communes and the propriae operationes, and shows the different orders (alphabetical, a capite ad calcem) he uses to list and describe them. Iohannes de Sancto Amando, Areolae, Prologus, ed. Pagel 1893: 1 f.: ‘Voco autem operationem communem: adustionem, congelationem et putrefactionem et ceterae quae eis similes sunt, apertio et abstersio etc. Primo ergo ponemus istas operationes communes secundum ordinem alphabeti. Quidquid autem imprimis dicemus de operationibus istis, sunt verba Avicennae in II°, et si aliquid postea adijciamus actor proprio nomine describetur. Proprias operationes voco juvamenta et nocumenta quae faciunt simplices medicinae membris propriis ut pectori, capiti etc. et sic de aliis et circa hoc huic ordini insistemus primo ponendo medicinas proprias capiti et membris ejus et sic
104
Iolanda Ventura
dary, and tertiary qualities was not, however, without difficulties, and the delimitation between secondary and tertiary qualities not always clearly defined. The third main classification and organization principle is, in actual fact, the most abstract one: the alphabetical or semi-alphabetical order. According to this principle, substances that could be considered different according their qualities, placement, or both within the three realms of nature can be ordered and described in a sequence depending on the initial letter (semi-alphabetical) or the initial letters (aa, ab, ac etc.; alphabetical) of their names. This principle, constantly applied from the 11th c. onward, could not be used by authors and dealt with by readers without difficulty. From the point of view of the authors, the main problem was represented by the difficulty of application of alphabetical and semi-alphabetical order, and of creating and showing connections between similar species with different names (e.g., domestic garlic or allium and wild garlic or scordeon). From the perspective of uses and readers, one may quote the difficulty of converting information included in this collection into practice (e.g., to find a therapy), and, if the work was translated from one language to another, the problems in keeping the alphabetical order. An example of such struggles can be found in Gerard of Cremona’s translation of book II, treatise 2 of Avicenna’s Liber canonis, for the translator did not always manage to identify and translate the substance mentioned by the Arabic physician, and, when translating, could not keep either the Arabic alphabetic order, or apply a Latin one. The alphabetical system was, however, often used by Late Medieval authors, such as the compiler of the Salernitan collection Circa instans or, in the 14th c., by Mattheus Sylvaticus for his mammoth collection, the Liber pandectarum medicinae2. So far for a general overview of the criteria of classification of natural substances and organization of pharmacological collections, viz. in systematic works attempting to offer a clear, coherent structure of the information and a linear description of the properties. Such an overview excludes de facto the rich textual and scientific tradition represented by the numerous collections of recipes to be found in Ancient and Medieval manuscripts, in Latin as well as in vernacular languages. By looking at the overview, a reader may gather that the three principles outlined here could co-exist or substitute for each other in the course of the historical development of Ancient and Medieval pharmacological culture. Actually, this conclusion should be challenged or, at least, reconsidered by examining, at the same time, the historical development of the classification criteria of natural elements exemplified by the most representative collections of materia medica and medicamina simplicia, the theoretical debate deve-
per singula descendamus; deinceps ad medicinas decorationis et ulcerum veniemus et post hoc de humorum digestione in quocumque loco contineatur et de evacuativis quae medicina, cui humori conveniat et repressivis earum determinabimus’. On John of Saint-Amand, cf. below, 140–144. 2 On the Salernitan collection Circa instans, see below, 129–131. On Matthaeus Sylvaticus, see Bottiglieri 2007 and García González 2008.
Classification Systems and Pharmacological Theory
105
loped around the principles leading the process of ‘structuring nature’ in pharmacological collections, and the formal application of standard rules with the aim of identifying each element in a complete and correct fashion, and of classifying the whole set of natural elements (e.g., all plants, animals, and minerals). This last aspect needs some clarification. In actual fact, although I acknowledge the criteria (e.g., name and synonyms, description, qualities, employment and effects) identified, applied, and used in Antiquity and the Middle Ages to identify and describe each natural element and to classify a whole set of elements, such as the plants listed in a herbal, I believe that we should not confuse the role they played when they were used as principles of description of single items or of classifications of several objects and the organization of entire texts. My paper cannot deal with both of these aspects, but will exclusively focus on the first one. In this sense, I will attempt to provide a historical overview of the texts dealing both with the description of properties of medicamina simplicia and the criteria of setting them into categories (in other words, classifying them)3. My survey will result in a necessary incomplete history of collections of medicamina simplicia, for I will not consider all texts, but only the ones offering both ‘theoretical’ reasoning on the principles of classification and a ‘practical’ section describing some simple remedies (= drugs). I will divide my historical account into two chronologically ordered parts, the first of which addresses the main collections of Antiquity, whereas the second describes the ‘theoretical turn’ of the 12th century. I will leave out of this essay the ‘new approaches’ emerging during the 13th century which would deserve, in my opinion, a specific analysis. I will therefore exclude the introduction, in Medieval therapeutics, of the ‘theory of purgation’, and the consequent reorganization of pharmaceutical literature around the simple medicines providing the necessary purification of the body from excessive or unnatural humors. This new concept in pharmacology represents the theoretical basis of the pharmacological and pharmaceutical corpus of writings attributed to the ‘Pseudo-Mesue’, and particularly of his work De consolatione simplicium medicinarum, a collection of 54 ‘purgative’ medicines divided into two classes, the mild and the violent – or potentially poisonous – ones4. The work, as well as the pharmacological theory represented by it, became extremely successful during the Late Middle Ages and the Early Modern period, and constituted the basis for the development of a complex Purgantienlehre witnessed by pharmacological and pharmaceutical literature up to the 18th century that largely surpasses in size and complexity the limits of a single article.
3 On the history of pharmacology and pharmacy, see Schmitz 1998 and 2005. A short and sometimes superficial overview of the history of pharmacology in the Antiquity is provided in Schulze 2003. See also Scarborough 2010 and Hardy and Totelin 2016. 4 On the ‘Pseudo-Mesue’, see Klimaschewski-Bock 1987; Heuken 1990; Lieberknecht 1995, and De Vos 2013.
106
Iolanda Ventura
2 Classification and description systems in Antiquity and the Middle Ages: a short history If we wish to examine the various systems of classification selected and applied during Late Antiquity and the Middle Ages, we have to take into account several elements. First of all, there is the relevant increase of the number of known and used substances that took place from the 11th century onward. This phenomenon has its roots both in the intensification of the commercial connections between East and West across the Mediterranean, and in the transmission, in Western Europe, of Arabic medical and pharmacological writings via the translations into Latin from the Arabic. An immediate and relevant consequence of this phenomenon was the application of systems of classification that could ‘manage’ the information overload. Secondly, we have to consider that not only the material increased, but also the number and the types of texts. The last centuries of the Middle Ages experienced in fact a real ‘explosion’ in the production of pharmacological collections featuring different content, form, and structure; they not only competed, but were integrated reciprocally in manuscripts. It is therefore possible to find, in miscellaneous codices, several herbals or collections – especially short ones – one of which is alphabetically organized, another as a quid pro quo, the third according to qualities or effects. Therefore, we cannot really speak of substitution of classification principles and criteria, but of reciprocal convergence, overlapping, integration, and co-existence. Thirdly, if we wish to detect a trend in the evolution and transformation of the classification systems, we may emphasize that, from the Early down to the Late Middle Ages, the general tendency was toward the increasing application of ‘abstract’ principles instead of ‘pragmatic ones’. In other words, from the 12th century onwards, we notice that pharmacological collections increasingly tend to show organizational criteria which are not connected to the type or the species of the natural substance(s), but to less perceivable properties, such as the name (a principle providing the fundament of alphabetical collections), the primary and the secondary qualities (the last two systems represent both a way to organize ‘Galenic’ collections and to introduce them with more theoretical chapters). An overview of the history of classification systems and the works that represented them may help to clarify the principles I have just formulated. This historical overview will necessarily appear incomplete, for I will be obliged to leave aside some texts and types of texts, despite their importance. My overview will leave out, for example, Isaac Israeli’s De diaetis universalibus et particularibus, despite the significiance of its organizational criteria and of the system resulting from them (the first section, De diaetis universalibus, enunciates the theoretical principles of dietetics, the second one, De diaetis particularibus, describes the
Classification Systems and Pharmacological Theory
107
single kinds of food – vegetables, meat, drinks etc. – by grouping them together)5. My exclusion is simply motivated by my intention to focus exclusively upon pharmacological collections. My survey will also exclude short texts such as the Kräutersammelkalender (viz., the calendars that indicate when plants should be collected in order to keep their full therapeutic power), a type of work classifying plants according to the time of their collection that was copied, for instance, in the well-known Lorscher Arzneibuch6. In this case, the reason for my exclusion lies in the difficulty of tracing the Fortleben (continued survival) of such short collections in medieval medical manuscripts, and therefore of putting together a historical account of the literary genre. Having explained my perspective, it is now time to turn to the overview I have just announced. a. Dioscorides, De materia medica, and Galen, De simplicium medicamentorum temperamentis et facultatibus The two main pharmacological collections written during Antiquity, and the most influential during the Middle Ages, are without doubt Dioscorides Pedanius’ De materia medica (hereafter, DMM) and Galen’s De simplicium medicamentorum temperamentis et facultatibus (hereafter, DSF). They virtually eclipsed other works, such as the herbal written by Diokles of Carystos, and Theophrastus’ botanical oeuvre (two texts that I will also leave aside here). The two works share a remarkable authority, but nothing more. First of all, a big difference can be found in their nature and place within a corpus of writings, and within the idea of materia medica and pharmacology defended by the author. In Dioscorides’ case, DMM is with all probability the only pharmacological work produced by this author (the treatises Theriaca and Alexifarmaka being now considered spurious), and therefore reflects a coherent vision of pharmacy7; in Galen’s case, DSF is only one, albeit the most important, among the physician’s writings on simple remedies and, more generally, on pharmacology8. Other works on drugs include a De compositione medicamentorum secundum locos divided into 10 books9, a De compositione medicamentorum secundum genera libri VII consisting of seven books (both works were written around 191–192 AD, and became part of the same treatise De compositione medicinarum)10, a De purgantium medicamentorum
5 Edition: Opera Ysaac 1515. On Isaac, see Veit 2003; 2015a, and 2015b. 6 Edition: Stoll 1989. See also the digitized version of the manuscript at http://bibliothecalaureshamensis-digital.de/view/sbbam_mscmed1?ui_lang=eng. 7 Edition (Latin): Lugduni 1598; (Greek): Berlin 1907–1914. 8 Edition: Kühn 1821–1833, here vol. XI, 379–892 and vol. XII, 1–377. On Galen’s pharmacology, see Debru 1997 and Vogt 2008. On the manuscript tradition of the text, see Petit 2013. 9 Edition: Kühn 1821–1833, here vol. XII, 378–1007, and vol. XIII, 1827, 1–361. 10 Ibid., vol. XIII, 362–1058.
108
Iolanda Ventura
facultate11, two works De theriaca (one ad Pisonem, the other ad Pamphilianum)12, a short treatise Quos, quibus catharticis medicamentis et quando purgare debeant13, and a De antidotis libri II14. Of all these works, only DSF was translated early enough to be accessed by medieval readers. The others were either not rendered into Latin at all, or were made accessible relatively late (this is the case with De compositione medicamentorum secundum locos, translated by Niccolò of Reggio), and did not circulate widely. Other works on pharmacology circulating under the name of Galen were a treatise De succedaneis, and the so-called De medicinis simplicibus ad Paternianum, a short alphabetical collection circulating in Latin until the end of the 12th century, to which I will return later. Therefore, it is improper to speak of just ‘one’ Galenic pharmacology in the same way as we would speak of ‘a single’ Dioscoridean one. However, if seen from the perspective of transmission and reception, both Dioscoridean and Galenic pharmacological thinking are closely related to the destiny of their sole (or most representative) works, DMM and DSF. If Dioscorides had left nothing but one work (other works circulating in Greek and Latin such as the Theriaca or the Ex herbis femininis being spurious), the reductio ad unum experienced by Galen’s corpus should be stressed, because it practically narrowed his pharmacological lore. The two works represented, from the point of view of both classification and description of medical substances, two opposite poles. According to A. Touwaide, the former text champions the ‘médicocentrisme’, the latter endorses the ‘pharmacocentrisme’ (Touwaide 1997). More concretely speaking, Dioscorides’ collection focuses on the specific characteristics of the natural element, and connects them with a large variety of effects and properties (dynameis); particular attention is paid to the intensity of the effect produced by each simple remedy. Each drug is therefore described by emphasizing first its natural shape, and then its own characteristics; concretely speaking, each entry is divided into several sections, dealing respectively with the name of the simple remedy, the description of the natural element from which the substance is derived, the kind of remedy (leaves, powders, etc.) and its preparation, its therapeutical properties and effects, the composition of the medicaments including the substance, the strategies to determine its authenticity (or, on the contrary, if it has been faked), other uses apart from human therapeutics (e.g., cosmetics), level of toxicity, and possible confusions with other remedies. All characteristics and properties listed by Dioscorides are not, however, structured according to a coherent system (viz., the Hippocratic humoral theory), or distinguished as general effect or specific actions, but are considered as a specific feature of the substance, and explicitly related to
11 Ibid., vol. XI, 323–342; see also the new edition in Ehlert 1959. 12 Edition: Kühn 1821–1833, vol. XIV, 1827, 210–294 (De theriaca ad Pisonem), and 295–310 (De theriaca ad Pamphilianum). 13 Ibid., vol. XI, 343–356; alternative edition in Raeder 1928, 221–227. 14 Edition: Kühn 1821–1833, vol. XIV, 1–209.
Classification Systems and Pharmacological Theory
109
its nature. In this sense, the properties cannot work as a strategy to group substances, nor as a tool to connect them with medical theories, but more to characterize each substance in isolated fashion. The only system the compiler uses to link (more than to structure) various natural elements and simple remedies derived from them is the similarity of their aspect and of the type and the intensity of the effect (the so-called ‘theory of affinity’). The system adopted by Dioscorides, and the function the properties, is perfectly explained in the Prologue to DMM, and, more precisely, in the words used to criticize his predecessors and fellow compilers of collections on materia medica: Quoniam, amicissime Aree, conscripsere multi, non modo veteres, sed et recentiores, de medicamentorum cum praeparatione, tum etiam facultate atque examine demonstrare tibi conabor, me neque frustra neque citra rationem ad hanc commentationem impulsum fuisse, quod ex iis ipsis alii nihil absolutum reliquerint, alii pleraque ex narratione descripserint. […] Cratevas autem herbarius et Andreas medicus (qui quidem accuratius quam reliqui hac in parte versati existimantur) radices multas apprime utiles, herbasque nonnullas non satis designatas reliquerunt. Verumtamen priscis hoc dandum elogii, etsi pauca tradiderint, accurate tamen in iisipsis elaborasse. At recentioribus (quo in numero sunt, Iulius Bassus, Niceratus, Petronius, Niger et Diodotus, Asclepiadae omnes) assentiendum non est, ut qui vulgatam et unicuique cognitam materiam qudantenus accurata descriptione dignati sint, medicamentorum vero facultates et examina cursim tradiderint, ac minime vim eorum ad experientiae normam examinantes, sed inania in assignandis causis verba funditantes, singula in controversiarum acervum extulerint, praeterquam quod alia pro aliis descripserunt (DMM, ed. 1598: 1 f.).
With these words, Dioscorides affirms his intention of adopting a more reliable method of compiling by connecting the rational and learned description of natural elements and the experience of their therapeutical powers and effects as simple remedies, and to consider description and experience as the main criteria, both as tools to structure each entry and to associate the objects of multiple entries. Galen’s system is, on the other hand, much more difficult to describe. His DSM is a long and complex work, written over several years, which employes different strategies to perceive and describe the nature of drugs in general, and to classify them specifically. First of all, we should recall that, compared to Dioscorides’ search for strategies to describe each simple remedy and to create links and associations between some of them, Galen’s pharmacological manual focuses on the search and the identification of general rules and principles that can be valid for discovering and describing the qualities and powers of any drug, can explain their effects from a rational perspective, and can structure them into a coherent system. As a consequence of this, if Dioscorides attempted to emphasize the specificity of each drug by using a large variety of definitions of qualities and properties before linking small groups of substances to each other, Galen aimed to describe all simple remedies by reducing their properties to a small number of valid and all-encompassing definitions, and to gather all substances into a fixed structure.
110
Iolanda Ventura
This goal is reached by Galen in different ways. On a first level, this is accomplished by attempting to structure nature and effects of simple remedies according to the Hippocratic system of the four ‘humors’; by doing this, he reduced the ‘primary’ qualities to four related to elements, and the variety of effects to a compact range of ‘secondary’ qualities by taking into account the humor that was affected by the drug and the sort of action the same drug performed on the human body. Secondly, this is accomplished by reducing the sections of the entries and the properties, powers, and effects to be described to a small number (e.g. by eliminating the description of the natural element from which the drug was derived). Thirdly, by substituting the link created by Dioscorides between classes or groups of simple remedies sharing the same nature and effect, albeit on different levels of intensity, with other, more abstract, but also openly reader-friendly systems of classification, the typological (viz., the one based on the realm of nature: vegetal, animal, and mineral world) and the alphabetical one within each type. In this sense, Galen operated a massive revolution of both the criteria of classification of simple remedies and of the description, interpretation, and assessment of their qualities, powers, and effects. A further step made by Galen toward a rationalization and a rational solidification of the criteria of description and classification of simple remedies can be identified in the elaboration of a consistent theory of pharmacology. Without doubt, Galen bases his pharmacological theory on the Hippocratic theory of elements and humors; however, he does not rely on Hippocrates’ empirical way of establishing the link between elements, humors, and natural substances, but insists on the rational process the ideal physician should undertake to determine the nature and the properties of each substance, as well as its place in the ‘great chain of being’. Furthermore, the foundation of his rationally based pharmacology is laid on the idea that the power (dynamis) of a drug is closely related to its innate elementary qualities and its complexion. By doing that, he strengthens the link between the primary and the secondary qualities, which, on the other hand, he carefully distinguishes. Finally, he introduces the idea according to which qualities and properties of a simple remedy cannot be determined just in themselves or with respect to its complexion, but also (or better, principally) with reference to the body assimilating it and its (healthy or pathological) condition. By doing that, Galen opens the path to a long discussion of potential and actual qualities, and of ‘substantial’ and ‘accidental’ properties and effects. The same connection to the body (and to physical perception) is underlined with reference to the intensity or degree of the same qualities. As far as classification systems are concerned, I have already pointed out that Galen uses two different orders, the typological and the alphabetical. These two systems helped him to order and to structure the drugs. Now, a more important question would concern the ways in which Galen structures and classifies properties and effects. As we have seen before, Dioscorides offers a large variety of properties reflecting an equally large range of powers and effects more or less limited to a specific part of the body, a humor, a condition. Galen does not, and could not, follow Dioscorides
Classification Systems and Pharmacological Theory
111
on this path. His aim is not to assess a large variety of single properties empirically, but to determine and structure a restricted range of main properties rationally. The main consequence of this kind of reasoning is that Galen firstly restructures the four main classes of effects that had already been established by the tradition (and, by the way, used by Dioscorides) – that is, heating, cooling, loosening and astringent, and then connects them with four different degrees of intensity. All other properties, powers, and effects, are structured around the four main ones, and all degrees of intensity are indicated with respect to a ‘medium effect’ (see Jacques 1997: 122–124). Those properties are not, however, described one by one, for Galen is not interested in showing the identity of single properties, but the rational structure that encompasses them. They are not even ‘extracted’ from nature and the use of simple remedies by deriving them from their use, in an inductive fashion, but should be established and known before using them, in a deductive way. Therefore, he did not really gather together a set of effects, and probably did intend to do so; rather, he focused on the identification of the right way and moment to determine them. By doing that, he introduced the theoretical discussion on the rational/experimental way to determine them, whose consequences we will find in the introductory chapters of al-Majusi’s/ Constantine the African’s and Avicenna’s works. But what are the practical consequences of Dioscorides’ and Galen’s approach? What does it imply in term of principles of both classification of sets of natural objects and description of single ones? On a superficial level of investigation, it may simply mean that, while Dioscorides focuses on the description of the substance and of its therapeutic properties (in DMM we find a large array of definitions such as softening, astringent, diuretic, sleep-inducing, emetic), and attempts to assembly a schematic, linear structure of his material by grouping its items according to their location in nature and the intensity of their effects, Galen goes beyond the external appearance of the natural substance and/or its place in nature, and tries to extract from it its qualities, grades, and power(s), and to connect them with the Hippocratic system of the four ‘humors’. Furthermore, he pushes the level of abstraction of the order by listing the substances according to the alphabetical order, and by treating vegetal, animal, and mineral substances separately. On a deeper level of analysis, however, the presence of the two approaches to the natural world and its medical use not only led to the creation of two competing systems of classification (secundum species, secundum qualitates), but also to the emergence of different forms and strategies of description taking into account the naturalistic elements and the pharmacological ones. According to the Prologue, Dioscorides’ goal is to provide his readers with complete information concerning external aspects, qualities, ways of preparation and use, and therapeutic properties of the natural element. In this sense, he fills the entries with all-encompassing data, including external description, strategies of preservation, qualities and properties, therapeutic properties and uses. A different, perhaps more complex, way is pursued by Galen: In his DSM, the Greek physician selects four main elements of identification and description of natural substances: the nomencla-
112
Iolanda Ventura
ture, the qualities, the therapeutic powers and effects (facultates), and the uses. Of course, neither Dioscorides nor Galen manage to build all entries by giving all sorts of information. More often, they only succeed in providing partial information, leaving several entries in what we could define as an incomplete state. Whatever the situation of the single entries, it is clear that the ‘naturalistic’ and the ‘pharmacological’ approach had consequences on the principles of organization of the collection, as well as on the criteria of description of the isolated items. The reception and the Fortleben of the two collections were not, however, without difficulties. Dioscorides’ DMM, on the one hand, turned out to be sometimes too complicated and difficult to grasp; its structure was consequently reworked and modified, and its content summarized. That led on both the Greek and Latin side to the redaction of new versions, e.g. of alphabetical recensions and alphabetical herbals interpolated by other sources, and to the compilation of abridged versions and collections of excerpts15. If we want to limit ourselves to the Latin cultural world, we know of at least three different versions circulating between the 6th and the 11th centuries, viz. (1) a complete Latin translation of the ‘original’ five-book version, the so-called Dioscorides langobardus, (2) another complete Latin translation of the five-book version, now lost, but vouched for by two small compendia, the Ex herbis femininis and the Curae herbarum preserved in the codex Lucca, Biblioteca Statale, MS 296, and (3) the alphabetical collection based on a selection of entries taken from the original DMM (or, perhaps, from the Dioscorides langobardus), and completed by some additions, the Dioscorides alphabeticus16. This last version, whose translation/redaction has been attributed to Constantine the African, entirely eclipsed the previous ones, and reached the peak of its diffusion during the 12th century. Galen’s DSF, on the other hand, although undoubtedly the ultimate source of the ‘Galenic’ system of classification and description, was not known in its entirety before the 14th century, and, even then, only rarely transmitted. In actual fact, his Latin tradition is based on the partial translation from the Arabic by Gerard of Cremona, of the first 5 books of the work (to which an anonymous translation of book VI was added), to which Niccolò da Reggio could add a new, Greek-Latin, version of the books VI–XI. Therefore, it is easy to see that, from the point of view of the transmission and affirmation of the classification and description principles and their influence with respect to that matter, neither Dioscorides nor Galen could act in a direct way, but, as for Dioscorides, via the sequence of Latin collections and rearranged (read: alphabetical) versions we have mentioned above, and, as for Galen, via a group of pharmacological
15 On Dioscorides in Byzanz, see Touwaide 2006, with further bibliography; Cronier 2006; 2010a. 16 Editions of the Dioscorides langobardus: Hoffmann et al.: 1883–1903 (book I also in Mihaescu 1938). Edition of the Dioscorides alphabeticus: Colle 1478. On the Latin tradition of Dioscorides’ DMM, cf. Cronier 2010b; 2013; Ventura 2006; 2013a; 2016.
Classification Systems and Pharmacological Theory
113
works written – mostly in Arabic – during the Middle Ages, and translated between the 11th and the 13th centuries, such as Ibn al-Abbas al-Majusi’s Kitab kamil as-Sina’a at-Tibbiyya, translated as Pantegni Theorica sive practica by Constantine the African, Ibn al-Jazzar’s Kitab I’timad al-adwiya al-mufrada, rendered into Latin by the same translator as Liber de gradibus, or Avicenna’s Liber canonis in the translation produced during the 12th century by Gerard of Cremona in Toledo. However, before focusing on the return of the Galenic system of classification via the Arabic medical treatises, let us explore the pharmacological lore of the first centuries of the Middle Ages. b. The Early Middle Ages: Herbals, Dynamidia, and Quid pro quo According to general opinion, the first centuries of the Middle Ages (viz., the ones comprised between the 6th and the 11th century, that is, before the beginning of the wave of Arabic-Latin and Greek-Latin translations) were characterized by the redaction and the diffusion of small pharmacological collections which can either be described as compendia of material and information ultimately derived from Dioscorides or Galen, or as little herbals built on Late Antique sources but, contrary to them, showing a peculiar lack of structure and organization. The number and the variety of texts is considerable, and it is impossible to treat them all in a single article17. However, if we wish to examine the history of pharmacological collections during the first centuries of the Middle Ages not by simply putting together a list of texts, but by taking into account the creation, the presence, and the affirmation of criteria of classification and description, we can emphasize the role played by three different types and traditions of texts: (1) the Late Antique and Early Medieval herbals, (2) the so-called Dynamidia, and (3) the Antibalomena sive Quid pro quo. Alternative systems of classification and organization of herbals that we can consider more ‘empirical’ or ‘para-scientific’ (at least, if we compare them with the genuinely medical and pharmacological literature) are featured in the herbals connecting plants and their therapeutic powers with the influence of the planets. The traditions of medico-astrological or medico-magical herbals, which, according to L. Voigts, show the place of plants in Medieval cosmos (see Voigts 2008), apply a specific logic of relation between plants and elements of the superlunar universe that deserve specific enquiries, and therefore will be left aside in this article. Nevertheless, I consider it necessary to mention their existence. To the first category we can refer, among other texts, two main collections, the Herbarius attributed to Apuleius Platonicus (hereafter, PAH)18, and Odo of Meung’s De
17 On pharmacological literature in Late Antiquity, cf. Ferraces Rodriguez 1999; 2007; 2009; Santamaría Hernández 2012; Maire 2014; with further bibliography. See also Fischer 2011; 2015. 18 Edition: Howald-Sigerist 1927.
114
Iolanda Ventura
viribus herbarum (also known as ‘Macer floridus’; hereafter, ‘MF’)19. I acknowledge that my decision to place the two works into the same group may sound far-fetched, or inexplicable. Their chronology, cultural background, structure, and form do separate them. However, if we do not look at their form (one is prose, one poetry), sources, or content, but simply at the principles and criteria of structuring the material and describing the items, we may find some points in common. The first, main point in common is their apparent anarchy and the lack of classification and structuring of the material and, more generally, of the world of nature. Both works are small, and present a short selection of natural substances. PAH, written between the end of the 4th and the beginning of the 5th century and preserved in two redactions, includes 131 plants; ‘MF’, whose original core may date back to the 10th century, only 77; the criteria that determined the selection and the creation of the sequence are, however, impossible to grasp. This feature, viz. the lack of order or the apparent anarchy, deserves to be stressed in a study about classification criteria, because it has without doubt some historical reasons. Both works were written during centuries that were permeated by a general decadence of culture, and especially of scientific and medical culture. Small, reader-friendly manuals were needed, rather than large, expensive, and sophisticated collections. However, I do not believe that this can be the sole explanation of these small manuals. I add that these writings are based on the principle of selecting items that were easily available, rather than describing a larger number of objects the reader could not always find or use. So, I believe that, in a certain sense, they share the same selective, reader- and user-oriented principles of construction and organization. Closely connected to the first feature and to its raison d‘être is the second characteristic the two collections have in common, viz. the fact that they result from a kind of copy-paste procedure, from filtering and collecting of material derived from different sources, and thus are compendia gathering together and summarizing knowledge from other sources. The third point they share is the presence, in their corpus or manuscripts, of criteria of description inherited from Dioscorides and Galen which, in their works, either show their survival during the first centuries of the Middle Ages, or their re-appearance for the first time after some centuries. At the same time, the presence of such criteria show that, if the principles of classification seem to have been neglected by the compiler and to have disappeared from the intellectual horizons of both authors and readers, the same cannot be said for the principles of description and definition of natural elements when considered as simple remedies. If we examine one entry of PAH and one of ‘MF’, we easily realize that their compilers did understand the importance of structuring information in a clear, and, whenever possible, constant way, and did everything that was in their power to respect this rule. The author of the PAH, for instance, organizes each entry in following way: he
19 Edition: Choulant 1832. On the tradition of the Macer floridus, see Black 2012, and with some reservations, Jansen 2013 (critical review in Black 2015).
Classification Systems and Pharmacological Theory
115
first lists the therapeutical properties of the plant, and then, at the end of the entry, offers information about the place where the plant is to be found, the way in which it should be collected and prepared for use (sometimes, rituals and charms connected to those occasions are recalled). Data concerning the appearance of the plants, which were included by Dioscorides in the entry, are supplied in PAH by a rich illustrative corpus (illustrated copies of ‘MF’ are, on the other hand, quite rare, and generally date back to the last centuries of the Middle Ages; a good example of an illustrative corpus attached to ‘MF’ can be found in the so-called ‘Herbarius des Vitus Auslasser’, written and illustrated during the 15th century20). On the other hand, the text does not show any attempt to include any theoretical information, or to gather the different therapeutical properties into a more general idea of ‘(Galenic) quality’. No mention is therefore made of any sort of power or effect. ‘MF’, on the other hand, starts each entry, whenever necessary, with the explanation of the name of the plant (that is, the etymology), and, on a more common basis, with the indication of the qualities of the natural substance. He then continues by indicating the area of the body and/or the kind of pathology the natural element most aptly heals or more successfully fights. Then the therapeutical properties and uses are listed. In both cases no apparent order can be detected in the lists of pathologies. We can thus speak, even in those ‘small collections’ apparently deprived of any logical structure and order, of the presence of principles of description. As for the surviving or re-appearing criteria we have just mentioned above, it is important to recall that the two collections show the incorporation of some additional data that have become criteria of description and definition, one as if responding to a ‘logic of relation’, the other to a ‘logic of abstraction’. The first type (which I call ‘logic of relation’) is exemplified in PAH by the presence, in the manuscripts belonging to the second redaction, of a network of synonyms related to the Scholia Dioscoridis. The presence of synonyms and their later inclusion shows that one of the main criteria used by Dioscorides and, more generally, during Antiquity not only to define the natural object, but also to allow its reliable identification in various languages – and therefore, not just to describe its appearance, but also to refer and connect its name to other places and contexts –, is now treated by the author of the herbal as a sort of additional information to be joined to the text in order to accompany it. The second type (which I call ‘logic of abstraction’) is illustrated by ‘MF’, where the mention of the Galenic qualities and grades opens the entries; here, we are dealing with a sort of information that had virtually disappeared from the intellectual horizon of pharmacology, and, with Odo’s poem, re-enters it, and that, not as additional information, but as essential, first-rate information. The example of the surviving or reappearing criteria offered by PAH and ‘MF’, two works I have placed together in order to emphasize their common features, shows that, whereas strong
20 Edition: Mayer-Goehl 2001; digital reproduction of the manuscript on www.muenchener-digitalisierungszentrum.de.
116
Iolanda Ventura
and rigid classification principles might have disappeared during the first centuries of the Middle Ages, description criteria prove to be, historically speaking, more volatile and subject to temporary oblivion and comeback. At any rate, we can conclude that, despite such addition, the nature and the horizon of PAH and ‘MF’ herbals are essentially practical, and do not really show any attempt at converting therapeutical properties into a more abstract conception of therapeutic qualities. The second type and tradition of pharmacological collection featured in Early Medieval manuscripts can be identified in the so-called Dynamidia. According to the seminal study published in 1936 by L. Mackinney, the use of the term dynamidia is attested by several Early Medieval, Pre-Salernitan medical collections (MacKinney 1936). Those texts can be divided into two different classes, according to the meaning of the term dynamis given by Hippocrates and by Galen (and by Galen-attributed texts). In the first case, dynamis refers to the vis and the potentia of different elements of nature and life to be used to preserve and restore health, such as food, drinks, exercise, etc., not just drugs. The latter type of text, on the other hand, does not offer such a broad idea of dynamis, but seems to restrict it to simple remedies. Thus, Hippocratic-oriented Dynamidia collections, such as the Pseudo-Hippocratic Dynamidia, Gargilius Martialis’ Medicinae ex oleribus et pomis, the Diaeta Theodori, or the Vatican Liber dynamidiorum published by A. Mai in 1835, do not limit their scope to pharmacology and pharmacological strategies, but aim to find and describe therapeutic effects and power, for example, in food, drinks, and medicines (especially of vegetal origin), as well as in exercise, sleep, and geographic location21. On the other hand, Galenicoriented Dynamidia collections were distinctly and strictly pharmacological, their horizon being clearly limited to natural elements belonging to the materia medica, and at the same time clearly excluding dietetics and other aspects of human life. To this category of Dynamidia we can attribute, still according to MacKinney, two subcategories of writings, the first of which is in the form of treatises, whereas the second shows a more schematic nature of – usually alphabetical – lists of simple remedies. The first sub-category is represented by the Pseudo-Galenic De simplicibus ad Paternianum (hereafter: SAP), the second by various collections generally sharing the main characteristics of being alphabetically structured and accompanying the names of the drugs with a more or less short and schematic account of their therapeutic virtues. According to its recent editor, N. Everett, SAP appears to be an important step in the development of the theory of dynamis, because it is situated between the Dioscoridean and the Galenic perspective22. The position of the collection in the history of the idea of dynamis can be summarized as follows. Although the work displays a vocabu-
21 On those texts, cf. Touwaide 2012; Mazzini 1992; Maire 2007; 2009. 22 Edition: Everett 2012. See also the critical reviews of this edition in Fischer 2013; 2015; 2017 (forthcoming).
Classification Systems and Pharmacological Theory
117
lary of properties that is essentially similar, albeit reduced, to the one featured in De materia medica, it also shows that these properties are not, as in Dioscorides’ collection, only used to define isolated drugs without creating groups of similar substances or providing a general system), but are structured into a stable and logically based paradigm pairing four major categories of heating/loosening (excalefaciens/relaxans) and cooling/astringent (refrigerans/stringens), and opposing one pair to the other (see Everett 2012: chapter 2, ‘Pharmacology’, esp. 45–52). This system is, on the other hand, nothing but an attempt to translate into a more fixed theoretical structure the principle of ‘allopathy’ set by Dioscorides at the basis of his pharmacology (e.g., a drug that can ultimately be defined as ‘cooling’ fights several diseases or pathological states that can be defined as ‘inflammation’ or ‘unnatural heat’). In this sense, the Pseudo-Galenic Dynamidia witnessed a process of theorization of pharmacology that took place during Late Antiquity and the High Middle Ages. However, the fact that the title Dynamidia could be given to different kinds of texts, independently of the presence of a theoretical definition of what a dynamis was or of the simple mention of dynameis, did not contribute toward a coherent definition of the literary genre and of its principles of description and classification. Consequently, medieval library catalogues would define as ‘Dynamidiorum liber’ the various kinds of alphabetical collections of simple remedies, no matter if they contained an attempt to list, describe, or classify them. Besides, the fact that the Hippocratic-oriented idea of Dynamidia collections had already gone out of fashion in the first centuries of the Middle Ages, while the Galenic one continued to be used, became the ‘main’ conception of this literary genre, and led to the multiplication of small pharmacological texts listing only items of materia medica, albeit in various fashions. What had such an unstable tradition of Dynamidia to offer for the history of principles of classification and description? I think the question can be answered in two different ways. The first, and most obvious one, may be related to the identification of a sub-category of Dynamidia with texts structured as lists. This apparently harmless and perhaps obvious statement has, in fact, important consequences, because, if considered as a principle and a system of structuring material, the ‘list’ answers various questions, and fulfills various needs. It crystallizes a selection, it summarizes, it structures in a clear, reliable way what could not be organized in other ways, it helps with understanding and structuring the content of other texts, and can (or must) therefore accompany them. Even SAP is accompanied, in some manuscripts, by a short Liber de dynamidiis listing 21 classes of drugs, opened by a short vademecum recalling the origin of substances (animal, vegetal, mineral), the determination of qualitates and potentiae by the four senses, the uses of contraria and combinations of substances, the incurability of some diseases23. This short text, which was not written by the same author of SAP, and shows some important differences from it, might have functioned both as a summary
23 On the Liber de dynamidiis, cf. Everett 2012: 18–21, here esp. 19.
118
Iolanda Ventura
of the current categories of dynameis, and as an integration of the material included in the Pseudo-Galenic collection. Therefore, the introduction and the affirmation of the principle and the structure of the ‘list’ – which, I need to stress, is not a consequence of the redaction and diffusion of the Dynamidia, but a phenomenon related to them – should not be seen as a simple reduction of the size, the background, and the scope of a pharmacological text, or as a type of text created in a time of decadence, but as a literary and hermeneutic strategy found and applied in order to obtain clarity by summarizing, reducing to the essential, and clearly structuring the material, and to integrate and comment upon the information assembled in pharmacological collections preserved in manuscripts. The second, and most important, answer, is that the Dynamidia collections, and more particularly the Galenic-oriented ones, despite using other ways (read: the alphabetical order) to structure the material externally, emphasize the role played by the virtus (dynamis) to characterize the natural element. In other terms: no matter if the collection is organized for consultation according to the alphabetical order, the clue the reader has to retain from its content, and the ‘real way’ to interpret and use the items displayed in it is/are the ‘power(s)’ possessed by the natural element. In my opinion, we can use (and I will use) this apparently obvious statement concerning the new and raising meaning of dynamis/virtus in two ways. In a first, more general way, it will be used to stress the affirmation of this principle as a strategy to describe and classify substances. The virtus becomes, in fact, not just a way to identify, describe, and characterize a single natural substance or simple remedy, but also, and especially, a strategy to gather together several substances sharing the same power, and therefore creating other, and probably more practice-oriented, regroupments and systems of classification. This is a principle that will permeate, for instance, the lists or tabulae produced during the Late Middle Ages (e.g., the Tabulae Salerni which I will deal with later). The other, perhaps more limited, consequence, involves the vocabulary of the dynameis/virtutes. In other words, if the dynamis/virtus becomes the core of the information given by each entry, and the main way to describe and characterize the single substance and to regroup and classify several substances sharing the same properties and effects, then it will be interesting to see whether this system stays stable and fixed throughout time, or experiences development and transformation. In this sense, SAP could be considered, as well as Dioscorides and Galen’s works, as a point of departure for historical research on the development of the definitions of medical and therapeutical properties. If we consider, for instance, the categories listed in the Liber de dynamidiis, we easily realize that there is a significant discrepancy between the ones listed in this short work and the ones presented in SAP (Everett 2012: 19 f.). Only 10 out of 21 categories listed in the former can, indeed, be found in the latter. This discrepancy, to which I will return in the next section, is very important not only to determine the authorship of the two texts and their connections and differences, but also to sketch the historical and typological development of the categories of dynameis/virtutes, to identify which categories appeared, persisted, transformed themselves, and ceased to
Classification Systems and Pharmacological Theory
119
exist. Whatever the case for the single properties, I hope I can make the reader aware of the meaning of the idea of dynamis for the history of principles and criteria of classification and description in Ancient and Medieval pharmacology. The third literary genre that helped to create a network of connections between natural elements considered simple remedies, more than a real system of classification, is represented by the writings known as Antibalomena or Quid pro quo. Thanks to the recent survey by A. Touwaide (Touwaide 2012), we may now emphasize both the practical value and the subtlety and complexity of the structure of Quid pro quo texts, a still widely unknown literary genre consisting of lists of simple remedies followed by a possible substitution. Contrary to what is commonly thought, the Quid pro quo lists cannot simply be considered lists displaying a one-to-one correspondence between the substituted substance and the substituting one, but can assume more complex forms, allowing cross-connections and mutual references between substances (e.g., if a substance ‘x’ is substituted by a substance ‘y’, the substitution process will work in the opposite direction, too, and the bi-directional process will be made evident by the presence of two distinct entries), and even the construction of ‘chains of two or more complex substitutions’ (see Touwaide 2012: 22 f.). The reason for the creation of such lists (especially of the ‘chains of substitutions’) cannot simply be reduced to the intention to replace a more rare or expensive item of materia medica with a better and more available one, but involves more complex mechanisms of perception of similar therapeutic properties in different elements of nature, of reasoning in creating a system of correspondences of substances because of their effects, and achievement of pragmatic goals (for example, to substitute a potentially toxic substance with a harmless one, or to reach a more flexible and adaptable therapeutic strategy by taking into account a wider range of ingredients). For our purpose, Touwaide’s analysis is important to acknowledge that the appearance and the diffusion of Quid pro quo lists did not simply fulfill a practical goal (viz., helping the apothecary in finding an available ingredient to be used instead of a rarer, more expensive, dangerous or simply unavailable one), but testify to the existence of an alternative system of classification of substances. This system, which we can define as a sort of sophisticated ‘network’ of substances connected because of their effects and therapeutic properties, should not be considered a competitor to the other ones, but, more realistically, a tool created and copied to help readers and users to find equivalences and similarities between items of materia medica that could not be easily perceivable by reading a pharmacological collection describing each substance in full. The value of the system is emphasized, in my opinion, by the presence of Quid pro quo lists in manuscripts preserving antidotaria (e.g., the Antidotarium Nicolai), a type of text handing over sometimes complex recipes without an explanation for the presence of a specific ingredient. In such manuscript contexts, the Quid pro quo not only helped to find an immediate substitute for an item that was not – for many reasons – imme-
120
Iolanda Ventura
diately available, but also to grasp which substances could be related to other ones, and therefore placed into a network. c. The ‘Theoretical Turn’ of the 12th Century: Constantine the African, Liber de gradibus and Practica Pantegni (book II); The Reception of ‘Constantinian Pharmacology’ in Salerno; Avicenna, Liber canonis (book II) The systems of classification and description I have just mentioned and described are attested by pharmacological collections written during Antiquity and the first centuries of the Middle Ages; more precisely, by works dated between the 1st and the 11th centuries AD. The medical literature produced during these centuries is relatively easy to survey, if not in all its attestations, at least in its general trends. Much more complicated is the work of the historian attempting a definition of the development of pharmacology between the 12th and the 15th centuries. The reasons for such difficulties can be found in the large number of writings, in the dimension and the complexity of some of them, and in the presence of large and intricate manuscript traditions leading to the production of several versions of the same text. If the ‘quantitative’ analysis of the attestation of Late Medieval pharmacy is difficult, the ‘qualitative’ one has its challenges, too, for we are confronted with a wider range of textual structures and organizational systems, and with works characterized by a deeper theoretical background. By using F. Wallis’ words, we can speak of a ‘theoretical turn’ of 12th century pharmacology that was fostered by the translation of some important Arabic works into Latin (e.g., Avicenna’s Liber canonis)24, and resulted in the elaboration of more scholarly manuals of pharmacology (as well as of pharmacy, by the way; see Wallis 2012, 107). Here, I cannot deal with the entirety of Late Medieval pharmacological literature, but attempt to exemplify, with the help of some selected writings, the general trends of the development of systems and principles of classification and the description of simple remedies. My survey on the development of a ‘theoretical pharmacology’ and of the consequences it had on the development of new and/or more complex systems of classification and identification/description of simple remedies begins with the translations produced during the second half of the 11th century by Constantine the African. The significance of Constantine’s translations for the progress of medical culture during the Late Middle Ages has been often stressed by scholars, who have emphasized his contribution to the introduction of basic manuals of medicine such as the Pantegni in Western Europe, the enlargement of the medical vocabulary, and, last but not least, the development of both a medical theory and practice25. The same scholars have not
24 On Arabic pharmacology, cf. Pormann 2001, and Pormann and Savage-Smith 2007. 25 Edition: Opera Ysaac, Lugduni 1515. On Constantine the African and the Pantegni, see Jacquart 1990, Jacquart and Micheau 1996; Brunett and Jacquart 1994 (especially the articles written by Green,
Classification Systems and Pharmacological Theory
121
failed to underline the numerous problems Constantine’s translations still pose to their Medieval and modern readers. Medieval readers – in the first instance, the later translator of al-Majusi’s manual, Stephen of Antiochia – strongly criticized Constantine’s behavior as a translator, accusing him of a lack of precision in rendering the Arabic original, and of plagiarism while attributing the translated texts to himself. On the other hand, modern readers and scholars have struggled with the textual situation of Constantine’s translation and with the philological issues it raises. The complicated situation can be summarized as follows: Constantine did not manage to translate the entirety of al-Majusi’s Kitab, but only to produce a Latin version of the entire Theorica Pantegni and of three small sections of the Practica, viz. the first book, the first section (De probanda medicina) of the second, and the ninth on surgery. This small torso, commonly defined as ‘Ur-Practica’, was later integrated, by the same Constantine, by his collaborators, and perhaps even after by later authors, with further sections that created an ‘enlarged Practica’, including ten books dealing with the same subjects as the Arabic original. However, those integrated sections cannot always be considered translations of the Kitab, but are mostly identified with interpolations taken, for example, from Constantine’s Viaticum, Iohannes Afflacius’ Liber aureus, or Isaac Israeli’s De febribus. The extant Latin manuscripts of the Pantegni attests the various stages of creation of the ‘enlarged Practica’, as well as the circulation of portions of it in an independent fashion26. Constantine’s translations also played a significant role in the development of Medieval pharmacology, particularly via the translation of Ibn al-Jazzar’s Kitab I’timad al-adwiya al-mufrada, which he rendered as Liber de gradibus (hereafter, LDG)27, and of the second book of the second section (known in the Latin version as Pantegni Practica; hereafter, PP) of al-Majusi’s Kitab kamil as-Sina’a at-Tibbiyya. The first text is a collection of simple remedies derived from plants, animals, and metals, the second is a more complex structure showing, in its most complete form, a theoretical section concerning the different categories of simple remedies (the De probanda medicina; hereafter, DPM), and two collections of materia medica, the De simplici medicina (hereafter, DSM), and LDG. Both translations have not come down to us without problems. LDG, which was structured in the Arabic original according to qualities and degrees, is transmitted by Latin manuscripts in four different versions, all attested by 12th century manuscripts, and therefore close to the date of the production of the Latin translation28. The first version, which appeared in print in an edition
Jacquart, Newton, Wack); Wallis 2011; Burnett 2013; Veit 2003a; 2003b; 2006; Veit-Ferre 2009. On the cultural milieu in Montecassino, see Newton 1999. 26 Monica Green is currently preparing a book-length study on Constantine the African and the circumstances of the translation and the manuscript tradition of the Pantegni. I thank her for sharing the first results of her research with me. 27 Edition: Constantinus Africanus, Opera, 1536. 28 Wallis 2012: 109, records 2 versions.
122
Iolanda Ventura
published in Basel in 1536, is arranged secundum gradus according to the main qualities (warm and cold); both the second and the third are, on the other hand, arranged in ‘rough alphabetical order’, but, if the first of them, attested in print by the edition Lyon 1515, re-arranges each section secundum gradus according to the alphabetical order, the other one, only attested in manuscripts, re-structures the whole work into an alphabetical herbal (e.g., the codex München, BSB, Clm 19429). Finally, the fourth version, attested by the sole manuscripts Bruxelles, KBR, 2419-31 and Paris, BnF, lat. 7029, transmits a fragment of the translation, and can be considered the ‘Ur-Version’ or a first draft of the translation in fieri29. Even more challenging is the situation of PP II. The Lyon printing shows, in fact, the final stage of development of the book and a patchwork of writings, for it includes a first section, DPM, that corresponds to the one translated from the Arabic text of al-Majusi by Constantine, a section, DSM, which does not derive, as commonly stated, from an unknown source, but reflects the original Arabic text that both Constantine and, later, Stephen of Antiochia translated, and a third one which is nothing but LDG. The manuscripts I have been able to see so far show different stages and combinations of the three sections, which I will not deal with here. However, I am able to affirm, as a preliminary result, that if DPM had been directly received in manuscripts since the release of the ‘Ur-Practica’, LDG was interpolated into the body of the text at a later stage; however, its connection with PP is already attested in 12th century manuscripts (for instance, by the ‘Codex Gigas’ or by the manuscript Roma, Biblioteca Casanatense, 369). The final step in the construction of the ‘complete PP II’ was the insertion of DSM. Manuscripts preserving the complete Pantegni are commonly attested from the 13th century onward30. The scope of this article obliges us to deal exclusively with the systems of classifications and the principles of description exemplified by Constantine’s translations, and not with their textual history. However, it is necessary to recall the issues related to the history of the redaction and the transmission of the two texts, because these data are necessary to frame and define the real impact both LDG and the various sections of PP II had on medieval pharmacological knowledge, and to avoid superficial and risky statements, the first of which will be to speak about the influence of ‘Constantine’s pharmacology’ as if it was a single thing represented by a single text. Instead, I believe that we should speak about different aspects of ‘Constantine’s pharmacology’, about the texts that represented it in different ways and in different historical phases, and about the various systems and principles they convey.
29 A later version, which never succeeded in breaking the monopoly of the Liber de gradibus, was produced during the 13th century by Stephen of Saragossa. On this version, see Ventura 2011. I am currently preparing an overview on Stephen of Saragossa, in collaboration with Sandro Passavanti (University of Pisa/EPHE, Paris). 30 I am currently preparing an article on the manuscript tradition and the development of the text of PP II.
Classification Systems and Pharmacological Theory
123
Let us start with what we can consider the most successful output of ‘Constantine’s pharmacology’, LDG. The collection consists of 205 entries, each of which describes an item of materia medica. Each item strictly focuses on the role played by the natural element in therapy, neglecting its appearance in nature. Therefore, the entry does not provide any external description of the plant or of the mineral, but only includes the following information related to its ‘therapeutic identity’: qualities and their degrees, indication of the humor and/or the part of the body that the simple remedy is to help bring back into balance (by purifying, heating, strengthening, etc.), and therapeutic properties (see, for example, the entries Rosa or Viola in LDG, ed. 1536, 344). Significantly, the effects of the remedy are not expressed in form of ‘Galenic’ primary qualities, nor by using the definition of the dynameis given by Dioscorides (a strange omission, for a text supposed to be transmitting ‘Galenic pharmacology’, and for an author considered to be at the origin of the redaction of the Dioscorides alphabeticus!). The lack of such data, and, particularly, of the special technical vocabulary used to express it (e.g., ‘vim/virtutem calefactivam/calefaciendi’ or ‘vim/virtutem purgativam/purgandi’) is particularly worth noticing, because it will shortly after become more diffused (during the 12th century, as attested by Salernitan texts such as the Circa instans). The work begins with a Prologue which, as correctly stated by F. Wallis, was not translated from the Arabic original, but more probably composed by Constantine himself. This Prologue has an important function, for it explains the theoretical framework of the collection, and, more specifically, the theory of the ‘four degrees’. It explains how ‘ancient physicians’ (viz., Galen’s followers) constructed a theory according to which the complexio of each simple remedy was characterized by the presence of a dominant quality whose intensity could be measured on a scale of four degrees; those degrees could be, in turn, divided into three sections, the beginning, the middle, and the end. Natural elements can show nine different complexiones, one equalis, and eight inequales. Among the latter, four are simplices (with one quality prevailing), and four compositae (with two qualities prevailing). Those complexiones are present both in natural substances and in the human body; they cannot be detected by the senses, but only by the intellect. Physicians attempt to keep or restore the natural balance of such qualities by using natural elements both as food and as medicines, and, more precisely, by selecting opposite substances compared to the complexio they want to change or improve, and applying them. The intensity of the quality is established in comparison with the complexio of the human body, in a simple or composite way. In a composite, when compared with the equal complexion in general; in a simple, when something is compared to the equal complexion of no matter which quality (‘uni qualitati cuiuslibet eque complexionis’). In this way, the intensity (degree) of a quality of a natural element is not just compared to the complexion in general, but to the single qualities of the complexion. In this way, Constantine offers a ‘Galenic’ explanation of the qualities that can justify the choice of the natural element with respect to the quality (or the qualities) that has (or have) to be changed
124
Iolanda Ventura
in the human body in order to heal it31. In practice, this conception of the qualities is given practical application by arranging the simple remedies according to the degree of their main qualities (here, warmth and coldness, whereas dryness and humidity only play a secondary role). The second book of PP is, on the other hand, a more complex patchwork that has not been analyzed in full yet. Its complexity depends upon its structure, as well as upon its historical development. As I have pointed out before, this section of PP, as we read it in the Lyon printing, was the result of a combination of three different units that took place between the end of the 11th and the 13th century. Therefore, we cannot really speak of this section as a textual unity delivering one system of classification and coherent principles of description, nor as a collection of texts intentionally providing several alternative systems. On the contrary, we can define it as a combination of texts gradually fitting together over time, and experiencing different receptions. In order to make the reader better acquainted with the state of the text and its content, I will describe each section separately. The first section, DPM, consists of 34 chapters devoted to the theoretical definition and interpretation of the virtutes medicinarum. The first two chapters perform an introductory function. The first recalls the role played by pharmacology within therapy and toward the restitutio sanitatis and cura egritudinum, and lists the main issues dealt with by this branch of medicine, viz. the general definition and discussion of the principles of pharmacology (‘simplicis medicine disputatio’), and the punctual description of the properties of each simple remedy (‘uniuscuiusque medicine ostensio’), while the second clarifies the reasons why the physician ought to know the therapeutic properties of medical remedies. The other 32 chapters can be structured into two main sections. The first one includes chapters III–VIII, devoted to the experimental ways of perceiving and defining the main properties and effects32.
31 A second Prologue, which has been so far overlooked by scholarship, can be found at the beginning of the section de quarto gradu (Constantinus, Liber de gradibus: 380 f.). This Prologue can be considered, at least in part, a repetition of the first, and deals again with the issue of the determination of the degree in comparison with the complexion of the human body. 32 It is interesting to notice that, in chapter III of DPM, De probatione medicine per experimentum, Al-Majusi/Constantine list the same criteria to be used to perceive and the determine the properties and the effects of simple remedies later mentioned by Avicenna in LC (see below, 146–148), with the sole exception of the last one: ‘Certissima et veracissima probatio est in corporibus probanda, sed tamen cum magna cautela, sicut antiqui dixerunt. Secundum hoc placitum confirmavit Galienus in simplicium libro medicinarum. Medicina ergo octo modis probatur. Unus vero est, ut sit omnibus accidentibus munda, ut actio sua nature patefiat. Alius, ut sit morbus simplex, cui medicina sit probanda. Tertius vero modus est, ut medicina probetur multis diversisque modis. Quartus, ut nec debilior nec fortior fit morbo. Quintus, ut actio eius investigetur utrum citissima aut tardissima sit. Si enim tarde calefacit corpus, prius tamen illud infrigidans, manifestum est actionem illius accidentaliter esse. Similiter est intelligi in ceteris qualitatibus. Sextus, ut perquiras utrum actio eius sit in omnibus corporibus equalis, aut non. Si enim est equalis, constat eam naturalem esse; nam si
Classification Systems and Pharmacological Theory
125
The central thesis emerging from this section, which we can define as one of the first, at least, indirect ways to introduce Galenic pharmacology into Western medicine, is that simple remedies can be classified the first instance according to their general complexion, and particularly according to their capacity to transform themselves more rapidly or slowly, and therefore of generating two main effects, viz. heating and cooling (e.g., via their speed in dissolving within the human body and heating it or, contrariwise, in coagulating and cooling it), and, secondly, via the message they send to the senses through their colour, smell, taste etc. The first of these definitions, which recalls the distinction made by Galen into drugs leptomera and pachymera33, will resurface again, at least in the form of allusion, in the chapter devoted by Avicenna in the Liber canonis to the perception and definition of drug effects per ratiocinationem34. On the other hand, the second section comprises chapters IX– XXXIV, and describes the virtutes secundae (secondary properties and effects), here considered not as a sequence of isolated effects, but as properties derived from the two main effects of heating and cooling and from their main qualities (cold, warm, dry, and humid), each of which does not share the same intensity with another one in the same natural object, that is, in the parts of the human body and in medical remedies (‘Quas virtutes possumus cognoscere intellectis earum complexionibus, que sunt calida, frigida, humida et sicca, que in omnibus corporibus in qualitate et quantitate non sunt equaliter, sed necessario una res non est eiusdem virtutis cuius alia, ideoque quedam medicina est aperitiva, quedam mollitiva, et quedam alterius virtutis’). Each simple medicine or category of simple medicine has the power, because of its complexion and action, of provoking an alteration in the complexion of the body. These alterations are schematized as secondary properties. According to this scheme, each property or effect, no matter if it may in potentia affect several parts of the body or a specific humor, organ, or element of the body, can be reduced to the result of two main qualities, a cooling or heating one. These functions can be possessed by the simple medicine in itself, or with respect to the complexion of the human body, or both. In a certain sense, this way of reasoning mirrors what is said in the Prologue of LDG, albeit from the perspective of the effects rather than that of the main qualities.
non equalis, accidentalis est. septimus, ut probetur in corporibus quibus compatitur. […] Octavus vero modus, ut ipsa medicina differat a cibo. Nam medicina calefacit et refrigerat qualitate, cibus vero cum tota sua substantia’ (Constantinus Africanus, Practica Pantegni, II, 2, in Opera Ysaac: f. 65vab). The presence of the same criteria in both al-Majusi’s and Avicenna’s manuals seems to have escaped the attention of modern scholars, who attribute these criteria only to the latter, turning them into one of the main elements of Avicenna’s rational pharmacology, ignoring the first formulation provided by the older manual of medicine. 33 On this distinction, see the previous paragraph, 109–111. On that matter, cf. Debru 1997b. 34 Edition: Avicenna, Liber canonis, 1507. On rational perception and definition of drugs, see LC, II,1,3; on this chapter, see, 147–149.
126
Iolanda Ventura
On the other hand, the section known as DSM consists, in the Lyon printing, of 38 chapters (in this printing, identified as chapters 37 to 75); they can be divided into two groups. The criteria of division are described in chapter 35, which distinguishes items of materia medica according to their vegetal, animal, and mineral origin. This way of structuring is not as simple as it may appear, but rather very sophisticated. First of all, the author distinguishes the mineral world from the vegetal and the animal one, because the first consists of simple remedies existentes per se (that is, the metal or the precious stone is the medical remedy), whereas the other two are to be divided further into different items, such as roots, leaves, juices (for the vegetal), and substances coming from the membra, the superfluitates, the humiditates, the capilli, of both animals and human beings (for the animal). Therefore, we cannot affirm that the author of DSM simply structures the materia medica according to the three realms of nature, but we have to acknowledge that he introduced further ways of classifying them within the three realms. According to this double system of classification (realm of nature and origin of the item), the following chapters providing descriptions of simple remedies are firstly organized secundum genera, and then once again secundum virtutes. This double division means that the items of materia medica are firstly grouped and described with respect to their nature and origin (that is, the vegetal world is divided into herbs, seeds, grains, leaves, flowers, fruits, oils, juices, gums, woods, barks, and roots, the mineral into earths, stones, salts, and metals, the animal into blood, milk, eggs, sweat, urine, fat, etc. of both animals and human beings), and then grouped and described again with regard to their effects. In the manuscripts I have seen so far, no clues are given to allow the readers to trace the same items within the two lists. As far as the criteria of description are concerned, we may observe that, while the descriptions of the first section are short, and generally limited to the nomenclature, the effects, and the indication of few selected properties, the ones included in the second sections are longer, and include qualities and degrees, effects, and some indications of therapeutic employments. In other words, if the descriptions secundum virtutes recall, in their structure if not in their content, the ones provided by LDG, the ones secundum genera comply with different standards, and concern the natural element more than the medical remedy. By offering two systems of description and classification, the author of DSM respects the commitment to which he had given expression at the beginning of the section. Besides, he manages to complete the circle he had started to draw in DPM, where he had focused on the theoretical explanation of properties, by listing the substances according to their properties (they are not the same in the two sections, though). Therefore, we may say that, from the point of the view of the structure, the organization, and the content of the work, the two sections belong together. And, indeed, far from being the translation of an ‘unknown source’, as stated by M. Wack, DSM is more probably the translation of the second part of the second book of AlMajusi’s Kitab, or, at least, of the same version of the Kitab Stephen of Antiochia
Classification Systems and Pharmacological Theory
127
translated in his Regalis dispositio35. If taken together, the two sections do not necessarily need the presence, in the same collection, of LDG. Then, some questions have to be put: What is the connection between DPM, DSM, and LDG? Why did they end up in the same ‘enlarged PP’? I fear there is no answer to that. From a purely historical point of view, the answer could appear to be easy: LDG was added by necessity, to give a more complete shape to the book, as long as DSM was not translated or widely available in translation. As M. Wack and F. Wallis have correctly stated, LDG was added to PP in order to fill the void created by the absence of the second section of the second book of the Arabic original in the version translated by Constantine, and by the necessity of attaching a pharmacological collection on simple medicines (Wack 1994; Wallis 2012). The connection created between the two works could be alluded to by the beginning of the prologue of LDG, which not only makes reference to an intention to continue a discussion started elsewhere, but uses the same words we can read in the Prologue of DPM, ‘simplicis medicine disputatio’ (‘Quoniam [alternative reading: Cum] simplicis medicine disputationem prout ratio postulavit, explevimus, restat ut ordo sequatur de unaquaque specie singulariter’). A further piece of evidence in favor of a connection between DPM and LDG can be found in the combination of the two texts in the most ancient manuscripts of the ‘enlarged PP II’, where DPM is followed by the prologue of LDG and by its (often fragmentary) text. However, if the situation is clear from an historical perspective, it is not clear from an epistemological one. In actual fact, we must notice that, from the point of view of the classification systems used in DPM and LDG, the two sections do not really fit with each other, despite the attempt by the author of the prologue of the latter text to connect them (if that really was his goal!). The situation can be explained as follows. The main principle used by the author of DPM to classify medical remedies is not the complexion, but the effect. On the other hand, LDG does not mention the properties described by the DPM as a consequence of the main effects of heating and cooling. Therefore, we may assume that, if the connection between classification and organization system displayed by DPM and LDG can be justified historically, for the latter work was inserted to fill a gap made evident by the former (a work that, we should remember, pointed to the necessity of describing nature and effects – ‘virtutis et nature’ – of simple remedies), it could not be easily justified logically, for they address different ways of perceiving, describing, organizing materia medica. As a con-
35 On th structure of the second maqala of al-Majusi’s Kitab, see Ullmann 1970: 267 f. Wack 1994: 180s fails to connect this section to the Arabic original and to the text translated by Stephen of Antiochia. This section also circulated independently in a slightly different form compared to the Lyon print. This version is witnessed by three 12th century manuscripts, the codices Città del Vaticano, BAV, Pal. Lat. 1080; Křivoklát (Czech Republic), Biblioteca Castri Křivoklát (Křivolátska Knihovna), MS 144 (I d 61), and Edinburgh, National Library of Scotland, Advocates, 18.6.11. I thank Monica Green for drawing my attention to these codices and to the independent circulation of DSM. I am currently preparing an edition of the independent DMS according to these three manuscripts.
128
Iolanda Ventura
sequence of that, Constantine’s contribution to the development of a classificational and organizational system, and of principles of interpretation and description of properties and effects of simple remedies, cannot be identified with the offer of a single system or of a list of principles, but with the emergence, in parallel with the textual development of his pharmacological oeuvre, of different systems and principles that were connected together in different fashions throughout the 12th and the 13th centuries. The presence of these alternative systems (a ‘Galenic’ one centered on the nature of the substance, its qualities and degrees exemplified in LDG, a perhaps equally ‘Galenic’ one focusing on properties and effects represented by DPM, and a mixed one combining a classification secundum genera and secundum virtutes in DSM) in chronological sequence, and not at the same time (at least, not before the 13th century) also had some consequences for the reception and influence of the different ‘Constantinian models’. The situation can be summarized as follows. To my knowledge, there is no evidence of any use or reception of DSM, neither of its structure, nor of its content. No collection of materia medica, as far as I can see, shows the double distinction secundum genera and secundum proprietates, nor has any later compiler excerpted any of the entries. With all probability, the text was made available and circulated too late to become a serious competitor for the more complete, reliable, and authoritative collections of materia medica. Therefore, although we acknowledge its existence, we cannot think of this work as an influential one. Generally speaking, the organizational system secundum species is not widely featured in Medieval herbals, nor is the description of the natural object (read: the plant) often considered and used as a criterion to identify the item of materia medica and to gather together similar items into a group. I have only seen a few Late Medieval herbals describing the appearance of the plant before focusing on its qualities and properties. This is the case, for instance, for the anonymous herbals preserved in the manuscripts Città del Vaticano, Biblioteca Apostolica Vaticana, Reg. Lat. 1788, and München, Bayerische Staatsbibliothek, Clm 775, both dated to the 15th century. The compilers of both herbals place the description of the plant at the beginning of the entry. By doing this, they seem to anticipate the renewed interest for botany that characterized European learned science during the Renaissance36. The situation is different for LDG, which became, both as an independent text and as part of the Practica Pantegni, a great success. On a more general level, we can easily acknowledge that it contributed to conveying the Galenic idea of qualities and degrees, although its interpretation of the theory of degrees presented in its Prologue was often deemed to be confused and difficult to understand. However, we can go further than that. In my opinion, the influence of LDG can be seen in the affirmation
36 On Renaissance botany, see Ogilvie 2006; Egmond et al., 2006; Egmond 2010.
Classification Systems and Pharmacological Theory
129
of various aspects of its organizational system and of some of its principles of description of simple remedies. Let us focus, for example, on the version structured in alphabetical order. A major evidence of its success can be found in the wide use made of this collection in Salernitan pharmacology. The Salernitan collection known as Liber de simplici medicina (or, from its incipit, as Circa instans), written between 1150 and 1170 and attributed to a certain ‘Platearius’, provides a good example of the influence of ‘Constantinian pharmacology’. The Salernitan pharmacological collection shows a broad use of material (whole entries, as well as isolated data) derived from LDG. In actual fact, if we compare the list of entries comprised in each letter, we see that the two collections deal with the same items, with few exceptions. It therefore seems possible that the compiler of the Circa instans had an alphabetical version of LDG at his disposal which he re-structured into his collection. Therefore, although it is considered the champion of the affirmation of alphabetical order as a pattern for the organization of pharmacological texts during the Middle Ages, it must be admitted that the Circa instans was the most famous and perhaps the most user-friendly example of the use of alphabetical order in pharmacology. It must also be admitted, however, that it might have taken inspiration from the alphabetical version of Constantine’s LDG. But there is more, as we can see by focusing on the principles of description of each simple remedy. The internal organization of the entries, as well as the selection and the placement of the principles of description, is essentially the same. Both compilers leave aside the description of the natural object as a whole and concentrate first on the identification of the Galenic qualities, then on the humors and the organs of the human body it affects, and finally on the therapeutic properties and uses. Hence, we will be tempted to consider the Circa instans as a sort of imitator of the LDG, but that is not correct, for the Salernitan collection introduces an element lacking in LDG. This element consists of the secondary qualities, the general definition of properties and effects we can define as vis/virtutes maturativa/maturandi, attractiva/attrahendi, mundificativa/mundificandi etc. The existence of such qualities, which could be subsumed by the definitions of therapeutical properties (or, better, operations) included in the entries of LDG, were not formulated in that fashion in Constantine’s text. Therefore, we can argue that, if LDG could act as a model for the description of the Galenic primary qualities and the therapeutical properties of single substances, it did not (and could not) act as a reference text for the definition and the use of the criteria to be used to describe secondary qualities and, in consequence, to gather together, structure, and organize simple remedies because of those qualities. Those elements of description had to be added from other sources, or derived from the practical experience of the compiler, and be formulated with the specialized form characterized by the suffix ‘-ivus/iva/ivum’ or ‘-andi/endi/indi’. It may seem that LDG could easily be used as a pattern to follow for practical description and definition of the characteristics of simple remedies, but not as a theoretical basis for a more general pharmacological discourse aiming to provide a structure into which the remedies could be placed. This lack of ‘theoretical influence’ can be confirmed by another interesting feature of
130
Iolanda Ventura
the influence of Constantine’s pharmacology, viz. the lack of interest for the ‘theory of degrees’ and their perception and interpretation enunciated in the Prologue. To my knowledge, no author of pharmacological collection ever mentioned this theory, nor made implicit use of it in his work. Finally, it would be interesting to analyze the influence of DPM. Undoubtedly translated by Constantine and already available in manuscripts of PP since its first redaction (‘Ur-Practica’), it would be easy to argue that DPM was widely read and used. If it was, however, no reader left a trace of any use of the collection. In fact, it seems that it did not enjoy a wide reception, or, more probably, that its nature of theoretical text did not bring many authors to quote openly from it. In order to see whether its classification system did meet with some success, I have tried to find its traces in contemporary (read: Salernitan) medical literature, and, more specifically, in therapeutical and pharmacological literature of the 12th century. I divided my search into three strands, the first of which focused on the ‘experimental ways’ to recognize the main properties and effects of simple remedies listed in chapters III–VIII of PP II (viz.: velocitas/tarditas alterationis/coagulationis, the smell, the colour, the taste), while the second examined the list of secondary virtues derived from heating and cooling described in chapters IX–XXVII (viz., the virtus mollitiva, indurativa, oppilativa, aperitiva, mundificativa, rarificativa, solidativa, largativa, constrictiva, uritiva, putrefactiva, liquefaciens carnes, cicatritiva, carnis curativa, attractiva, laxativa, expulsiva, custodiens corpus a mortifera potione, mitigans dolorem), the third the tertiary virtues the author dealt with in chapters XXVII–XXXIV. (viz., the virtus lapidem frangens, urinam provocans, menstrua provocans, lac generans, sperma generans, lac auferens et desiccans, pulmonem mundificans). For each group, I tried to identify points in common and differences in representative texts of Salernitan medicine; the results of my search are by no means exhaustive. As for the the first strand, I noticed no trace of the experimental ways listed in the DPM in the Prologue of the Circa instans, but found those criteria listed among the modi probationis medicine in egrotantibus discussed by Bartholomew of Salerno – by the way, an author supposed to have been well acquainted with Constantine’s work – in his Practica37. At the beginning of the work, we read a chapter De diffinitione medicine, where he lists eight ways (modi) to experiment on the efficacy of simple remedies on ailing people, among which he includes the ‘velox mutationis seu tarda’, the ‘tarda seu cita coagulationis’, viz. the first two categories we had already met with in DPM (Bartholomaeus Salernitanus, Practica: 322 f.). He then devotes two little chapters to tastes and smells, but none to colour (Ibid.: 323–325). We may therefore argue that the list of the modi was at least in part influenced by the Constantinian lore. However, some characteristics of Bartholomew’s account reveal more. First of all, the two modi apparently derived from DPM are not presented by
37 Edition: Bartholomaeus Salernitanus, Practica, in De Renzi 1856. On Bartholomew, cf. Kristeller 1986; Wallis 2007; Fortuna and Urso 2010.
Classification Systems and Pharmacological Theory
131
Bartholomew tout court as experimental ways to identify the effects of simple remedies, but, more specifically, as a part of a strategy especially devoted to distinguishing food from medical remedies. Secondly, the ‘experimental’ ways mentioned by Bartholomew, that is, the purity of the medicine (‘puritas medicine’), the simplicity of the illness (‘simplicitas egritudinis’), the opposition of the medicine to the disease (‘contrarietas morbi ad ipsam medicinam’); the others being the proportionality of illness and remedy – ‘proportionalitas medicaminum ad ipsam egritudinem’ – and the capacity of the medicine to heal the same disease in different parts of the body – ‘identitas medicaminum’ – and the limitation of the action to specific parts of the same body – ‘operatio medicaminum ad corpora’), correspond to the ones mentioned in chapter III, De probatione medicine per experimentum, of DPM38. Finally, if we move from the criteria used to define medical remedies experimentally to the lists of properties and effects, we notice that Bartholomew distinguishes 18 types of virtutes simplicium medicinarum, the first of which – probably the most relevant ones, although considered to be ‘non adeo perfecte sicut complexiones’ - are the virtues maturativa, mollificativa, indurativa, opilativa, rarificativa, inspissativa, and indurativa (Bartholomaeus Salernitanus, Practica: 325 f.); to these virtues, he attaches the mundificativa and the remedia ad fluxus humorum. A second category (alie virtutes) includes the liquefactiva, putrefactiva, cauterizativa, carnis generativa, attractiva, salvativa, doloris mitigativa (Ibid.: 326–328); to them, the laxativa and the consumativa venenum are attached. If we compare the two lists provided by Constantine and Bartholomew, we can see some sporadic points in common, but no consistent influence of the former upon the latter. The result of this comparison can be therefore summarized as follows: although the list of virtues may be different, the theoretical background used to determine them is the same, viz. the one inherited by Bartholomew from Constantine. The careful preservstion of the same theoretical structure, despite some formal differences, can be proved by the fact that Bartholomew preserves the threefold distinction of medical virtues secundum complexionem, secundum actionem ad proprietatem precedentem, and secundum operationem (viz., the primary, the secondary, and the tertiary virtues) provided by the DPM, albeit in other words, even though he does not really explain the difference between the properties secundum actionem and the ones secun-
38 This passage bears some resemblances to the Prologue of the Salernitan collection Circa instans: ‘Questio enim non occiosa proponitur, cur medicine fuerint invente composite, cum omnis virtus, que compositis inest, in simplicibus reperiatur. Medicina propter morbi causam inventa fertur extitisse. Omnis autem morbi causa aut est ex humoris superhabundantia aut ex inanitione aut ex fluxu aut ex debilitate virtutis aut ex alteratione qualitatis aut solucione continuitatis. Invenitur autem medicina simplex repletionis solutiva, inanicionis restaurativa, fluxus constrictiva, debilitationis confortativa, alterationis immutativa, solucionis consolidativa. Solucio: Compositionis medicinarum multiplex extitit causa, scilicet morbi violencia, morborum contrarietas, membrorum contraria dispositio, nobilitas membri, violentia medicine’ (Platearius, Circa Instans, in Wölfel 1939). On the Circa instans, see Ventura 2007; 2013b; 2015.
132
Iolanda Ventura
dum operationem, and does not acknowledge the role played by heating and cooling as main properties that had been stressed in Constantine’s translation. Therefore, we can argue that, although the formal classification system provided by DPM was not re-used by Bartholomew, the theoretical structure at its base was a Constantinian one. The results provided by Bartholomew can be compared with, or rather integrated by, the analysis of two other writings more or less directly connected with the Salernitan milieu, the Tabulae Salerni (hereafter, TS) and John of Saint Paul’s De medicinis simplicibus (hereafter, DMS), two texts I selected because of their common nature of lists of simple remedies structured according to primary and secondary qualities. By analyzing them, I hoped to find some – at least indirect – traces of the division of properties and effects provided by DPM. In actual fact, this is true only for the second text. But, before starting their analysis, let us first summarize the information we have about the works. The TS are attributed to a ‘Magister Salernus’, and are considered to have been written around 1160; at the end of the 12th century, they were ‘commented on’ – or, better, added to by information concerning simple remedies – by Bernhard of Provence39. They consist of 12 thematic ‘tables’, each of which is divided into ‘columns’ dealing with an aspect of the subject of the ‘table’. According to the index following the Prologue, the first seven ‘tables’ deal with secondary qualities or effects, the tables 8–11 with primary ones, and the last table (the 12th, which is not always copied in manuscripts) with poisons. The first seven tables concern the following qualities and effects: purgative (1st), diuretic (2nd), emetic (3rd), excretive (4th), astringent (stipticus; 5th), sweat-provoking (6th), nourishing (7th). The second group of tables (8-11) concern primary qualities, warmth (8th), coldness (9th), humidity (10th), and dryness (11th). As far as the content is concerned, each column provides nothing but a list of simple remedies of vegetal, animal, and mineral origin sharing the same qualities and effects, without describing the substances and/or the natural element they originated from; such description was later provided by the commentary written by Bernhard of Provence. On the other hand, John of Saint Paul’s DMS, a work written during the last decades of the 12th century, is composed of 70 chapters divided into two main sections, the first of which (chapters I–XVII) focuses on primary qualities, the second (chapters XVIII–LXX) on secondary and tertiary ones.40 The first section falls into
39 Editions: Salernus (magister), Tabulae, and Bernardus Provincialis, Commentum super Tabulas Salerni, in De Renzi 1859. On the ‘Magister Salernus’, cf. Ausécache 2008. On Bernard of Provence, cf. the entry ‘Bernardus Provincialis’, in Clavis Auctorum Latinorum Medii Aevi (C.A.L.M.A.), accessed via the platform Mirabileweb (www.sip.mirabileweb.it. [Accessed 4/2016]), with further bibliography. 40 Editions: Constantini de virtutibus simplicium medicinarum, in Opera Ysaac: f. 186va–189rb; Kroemer 1919. On Iohannes de Sancto Paulo († 1214), cf. Kroemer 1919; Paravicini Bagliani 1987: 72 f.; Nicoud 2007, vol. I: 36–45; Martín Ferreira – García González 2010; Martín Ferreira 2013. A preliminary list of manuscripts preserving the DMS was included in Frippiat 2006–2007 (I thank Baudouin Van den Abeele for attracting my attention to this unpublished thesis).
Classification Systems and Pharmacological Theory
133
five groups according to the pre-eminence of a single quality. Thus, the first chapter coincides with the first group, because it concerns the equal medicines (the perfecte temperata), which do not display any prevailing quality. The second group (chapters II–V) is devoted to warmth; its four chapters list warm simple remedies in the first (chapter II), the second (chapter III), the third (chapter IV, and the fourth degree (chapter V). The same structuring criteria are applied to cold substances (third group, chapters VI–IX), dry (fourth group, chapters X–XIII), and humid (fifth group, chapters XIV–XVII). Then, the second section (chapters XVIII–LXX) list the secondary and tertiary qualities. The range of qualities taken into consideration by John is far larger than the one examined by the ‘Magister Salernus’, or in the DPM. Here, I cannot list and describe all properties and effects included by John in his collection, but only offer a glimpse of them. The collection comprises both secondary qualities regarding the totality of the body and tertiary focusing on a part of it. For instance, it includes attractive, laxative, astringent (constrictiva), hardening (indurativa), emollient (mollificativa), abscess-opening (maturativa), obstructive (opilativa), opening (aperitiva), cleansing (mundificativa), replacing flesh in wounds, corrosive, ulcerative, burning, aiding formation of scars, sticking (conglutinativa), softening (attenuantia), incisive, rarefying, sudorific (diaforetica), repercussive, soporific, stinging (mordicantia), strengthening (confortantia), but also poison-fighting, pain-killers repressing pain derived from warmth and coldness, blood vessel expanding, soothing, narcotic, eyehealing, worm-killing, lung-cleaning (mundificantia pectus), milk-generating, and aphrodisiacs (confortantia coitum). If we now move from the structure to the content, we notice that John’s work shares the same internal organization of the chapters of the TS. In fact, the chapters of John’s DMS are nothing but lists of items of materia medica. Each chapter, however, carefully separates substances of different origin by creating small regroupments of herbs, flowers, seed, fruits, gums, gums, minerals, and stuff derived from animals and parts of their bodies. If we now analyze the structure of the two works, we may single out some of their features. The structure of TS presents two interesting characteristics. First of all, it shows a ‘subverted Galenic system’; secondly, it offers, both in the choice of the secondary qualities and in the internal organization of the tables in ‘columns’, an interesting way to perceive and define those qualities that we can define as ‘pragmatic’ or ‘action oriented’. The first characteristic is easy to define: whereas the ‘Galenic’ legacy of Constantine’s LDG or of the Circa instans would consider primary qualities as the first principle of definition and classification of simple remedies, recall them at the beginning of each entry and, at least in LDG, use them as a way to structure the collection, the ‘Magister Salernus’ does not seem to share the same opinion, for he deals with them after he has illustrated a selection of what he considers the main secondary qualities. From his choice we can gather that, in his conception of pharmacology, the primary qualities are not seen as the first criterion and principle of definition and classification of substances, but as a subsidiary one; the most important – perhaps from a practical or therapeutical point of view, are the secondary ones,
134
Iolanda Ventura
viz. the ones related to perceivable effects. The second characteristic requires to be explained by discussing the internal organization of the tables, starting with the first one. The first table, for instance, deals with purgative effects, and is divided into four ‘columns’, according to the humors purged by the groups of substances. Hence, the first ‘columns’ list the haemagoga (that is, the medical remedies purging blood), the second the phlemagoga (the ones purging phlegm), the third the cholagoga (the ones purging the yellow bile), the fourth the melagoga (the ones purging the black bile), and the fourth the hydragoga (the ones draining out excess water). A little appendix is then added to list the remedies purging various kinds of ‘bad, non-natural humors’, such as the phlegma vitrum or salsum. The first four ‘columns’ of the second table, devoted to diuretics, show that these simple remedies can be subdivided into categories according to their nature (warm, cold, humid, and dry), whereas the fifth isolated the ‘diuretica forinseca’, that is, the ones that let liquids be expelled outside of the body. Other tables and columns are structured according to other criteria, e.g. to the intensity of the effects, or to the kind of natural element they originate from (herbs, fruits, and so on). If we now turn our attention to John of Saint Paul’s DMS, we notice that, on a macro-level, the organization of the collection is the opposite one. Here, simple medicines are first gathered together and classified according to their primary qualities, and then according to their secondary and tertiary qualities, therapeutical properties and effects. In this sense, he seems to have been following the message sent out by Constantine the African in his LDG, a work which could be considered his main source. However, I think that the structure of DMS reveals more than a simple imitation of Constantine’s ‘Galenic pharmacology’ as delivered in LDG, but shows that the compiler followed the principles displayed in DPM, too. If we search for evidence of my argument, we should focus on the Prologue of the collection. Here, John reveals to us the main reason why he had decided to put together the collection. His main goal was, he tells the readers, to offer a list of the properties and effects of simple medicines, and, more specifically, to react against the chaos permeating the tradition of the Galenic Dynamidia (which books he had exactly in mind, we cannot say) by rewriting that type of book again: ‘Legens itaque in dianamidiis (sic!) Galeni pene ita hunc inveni scriptorum errore et vitio depravatum, ut potius auctorem quam correctorem exigeret’ (DMS, Prologue, f. 186va). He then recalls what the virtus of a simple medicine is, viz. its capacity to change the state of the human body. Such a capacity is due either to its nature and complexion, or to its specific effect on the human body, or on the action: ‘Hec triplex est: prima nomen complexionis, secunda virtutis vocabulum specialiter, tertia vero operationis suscipit vicem’ (Ibid.). This threefold distinction, which we can translate into a definition of primary, secondary, and tertiary quality, reproduces the idea of the virtus medicine placed at the beginning of DPM, viz. in chapter II, ‘Quod medicus scire debet vires simplicium medicinarum’, where we read:
Classification Systems and Pharmacological Theory
135
Virtus itaque simplicis medicine tribus modis est: una est complexio et natura. Secunda est actio a prima procedens, id est maturativa, mollitiva, indurativa, dissolutiva, oris venarum aperitiva, carnium minutiva, attenuativa, dolorum mitigativa. Tertia est lapides frangens, urinam et menstrua provocans, pectus mundificans, lac generans (PPII, DPM, II, in Opera Ysaac: f. 65va)
.
Constantine’s influence can be detected in what follows, viz. the definition of the unique complexio equalis and of the eight complexiones inequales that we have already encountered in the Prologue of LDG41, and in the separation openly established by John between virtutes and operationes, while DPM only implicitly postulates it when opening chapter XXVII De virtute tertia medicine lapidem frangente with the words ‘Quemadmodum actionem secunde virtutis a prima diximus procedere, ita tertiam a secunda oriri dicamus que quidem lapidem frangunt, urinam et menstrua provocant, pectus et pulmonem a screandum adiuvant, lac et sperma augmentant [...]’ (PPII, DPM, XXVII, in Opera Ysaac: f. 67rb). The results of my analysis of TS and John of Saint Paul’s DMS have shown, I hope, that Constantine’s influence on Salernitan pharmacology is not limited to the Galenic perception and description of simple remedies delivered in LDG, but also extends to DPM. On the other hand, it has shown the size and the limits of the reception of Constantine’s DPM. The TS do not seem to have re-used any element from DPM, neither the system of classification of simple remedies, nor its theoretical background. Besides, their author seems to have applied very pragmatic strategies of regroupment and classification of items of materia medica, for he organizes every table in columns according to different principles (nature, effects, humors purged), without trying to distinguish them or to put them into a paradigm or a hierarchy. In the case of TS, we should at least argue that their cultural background and their library of sources may be found in works dealing with clinical medicine (practicae, or compendia), more than in writings that offered theoretical fundaments of pharmacology. The situation is different for John’s DMS. In this work, we find clear traces of the Constantinian heritage, and the use of the same categories. That heritage, that has led some scholars to consider Constantine to be the real author of the work, is evident in the definition, the distinction, and the classification systems of the virtutes medicinarum, of the dynameis which, in the author’s opinion, had become the object of confusion and misunderstanding. In this sense, we may venture that Constantine’s influence was larger than the one of a simple source, for it became the fil rouge the author used to restructure pharmacology in a rational and reliable way. Constantine the African was certainly not the only author who endorsed the ‘theoretical turn’ of pharmacology during the 12th century. Another text had, according to the general opinion, a considerably deeper influence on Late Medieval theoretical pharmacology: Avicenna’s Liber canonis (hereafter, LC), the last text I will consider
41 See above, 124.
136
Iolanda Ventura
in this survey. Written during the 9th century by one of the most illustrious Medieval Arabic physicians, and translated in Toledo in the middle of the 12th century by Gerard of Cremona, LC became, from the middle of the 13th century, one of the most widespread manuals (if not the standard one) of medicine, and a pillar of the Academic medical curriculum. After the extensive studies carried out during the last decades, we now know more about the organization, the content, and the impact of Avicenna’s manual on Eastern and Western medical culture42. The scholars I have just mentioned have stressed Avicenna’s contribution to the emergence and the development of the learned, academic, theory-oriented, ‘scholastic’ medicine in the Latin Middle Ages. Avicenna’s LC provided physicians with abundant material and reliable (Aristotelian) theoretical background on subjects belonging to the domain of medical theory, e.g., on the nature of medicine, the structure of the human body, the meaning, the role, and the function of its main organs (liver, heart, and brain), or the activity of the senses, the human soul, and the intellect. Less studied, but not less important, is Avicenna’s contribution to pharmacology and pharmacy through the equally abundant information provided by him, respectively, in book II and V of LC. In this essay, I will deal exclusively with book II, and, more specifically, with the theoretical discussion of the principles of pharmacology included in its first treatise, and principles of description and the criteria of classification of simple remedies employed in the second. However, before starting my discussion, it is necessary to recall, however briefly, the structure and the content of book II of LC. Pharmacology is the object of the second book of LC. The section is particularly long, for it occupies 83 folios of the Latin folio-edition published in Venice in 1507 which I will use in my study. It consists of 2 treatises, the first of which, divided into 6 large chapters, deals with the main principles of theoretical pharmacy and with the basic theoretical knowledge the physician should possess about simple remedies. More specifically, the chapters deal with the nature and complexio of simple medicines (chapter I), how to identify and recognize properties and effects of simple remedies per experimentum (chapter II), how to identify and recognize the same properties rationally (per ratiocinationem; chapter III), the therapeutical actions (operationes) of simple remedies (chapter IV), the external circumstances that influence and transform the effects of medicines (chapter V), and the correct ways to collect and preserve the medicamina (chapter VI). On the other hand, the second treatise is a collection of materia medica listing and describing around 760 substances. The collection was structured, in the Arabic original, in alphabetical order according to the initial letter of the Arabic name of the substance. The Latin translator, Gerard of Cremona, strugg-
42 On LC and its Latin translation and reception, cf. Ullmann 1970; Jacquart 1992; 1997; 1998; 2002; Jacquart and Micheau 1996; Pormann and Savage-Smith 2007; Gutas 2014; Jacquart 1998; Burnett 2001; 2005; 2009; Chandelier 2007; Paavilainen 2009; Veit 2011; Ayari-Lassueur 2012; Adamson 2013; Van’t Land 2014; Chandelier 2017.
Classification Systems and Pharmacological Theory
137
led to keep this order, and his choices as translator lead to inconsistencies in the text and to difficulties in reading and using it. The use of the alphabetical order is not, however, the most important characteristic of Avicenna’s collection of simple remedies. Its most intriguing feature is the internal organization of the entries and the disposition of each kind of information and category of information, as well as the theoretical background displayed in the first treatise, to which we will return later. In the second treatise, each lemma is structured with the aim of separating different elements of the description very clearly, and to structure the content in a precise sequence. The sequence can usually be divided into three parts. A first, descriptive one aims to offer useful data for identifying and describing the item. Each section is opened by a word summarizing the type of information. In other terms, each section has its own header, and each header deals with a specific aspect of the nature and the therapeutic value of the simple remedies. The headers are the same, and are constantly applied in the same way. That provides all entries with a fixed structure, and with the same selection and organization of the content. The entry usually opens with the mention of the name (Nomen) and the physical appearance (Complexio), and continues with the criteria of selection of the best sort of natural remedy (e.g., of the best plant; Electio); this section usually opens with the general expression ‘quid est’. The second section includes the description of its primary qualities, grades and properties (Qualitates et natura), of its general effects (Proprietates operationum), and ends with its specific actions and effects (‘Operationes’); the information delivered in this section is usually accompanied by the expressions natura and operationes. Then follows a third, ‘pathological’ one, describing various kinds of actions and effects against specific categories of pathologies: the first on skin diseases, introduced by the expression Decoratio, the second on tumors and abscesses, opened by the expression Apostemata et bothor, the third the wounds, ulcerations, and fractures, starting with the words Ulcera et vulnera et fracturae, the fourth the pathologies of nerves and joints, starting with the words Egritudines iuncturarum et nervorum, the fifth the pathologies of the head, introduced by the expression Egritudines membrorum capitis, the sixth the ones of the eye, marked by the words Egritudines membrorum oculi, the seventh the ones of the digestion, introduced by the words Egritudines membrorum digestionis, the eighth the ones of the intestines and the bowels, starting with the words Egritudines membrorum expulsionis. The series is closed by the pathologies and the accidents involving the whole body, such as the fevers (section 9; opening words: Febres) and the poisons (section 10; opening words: Venena). The entry usually ends with the mention of substitute drugs to be used if the described one is not available (Permutatio; see Paavilainen 2009, 51 sqq.). Of course, it is extremely rare to find an entry including all the sections; while the ones defining the nature and the general and special effects and actions are always available, not all pathologies are represented. The main reason for this lack is evident. No simple remedy can be applied to any part of the body, or cure any disease! Another, less evident, reason may be found in the relative novelty of the described drugs, whose
138
Iolanda Ventura
properties and effects were not extensively known. However, some entries, such as the one devoted to the rose (Rosa) are particularly long, because of the great use and the excellent reputation of the plant and the remedies derived from it (LC, II,2,575: f. 146ra–va). The merits of this structure are easy to grasp: First of all, the entry is much easier to browse, and the information the reader is looking for can be found without effort. Hence, the first advantage can be found in its user-friendly character. But there is more. The structure employed by Avicenna makes an implicit structure explicit with the help of the opening words. We have already seen how other collections of simple remedies organized the information on each simple remedy by following a fixed and established path throughout the entries. The second treatise of the second book of LC is the first collection, at least in the Western Middle Ages, that makes this organization visible and explicit through the constant use of some fixed introductory expressions which are visually emphasized in manuscripts, e.g. by copying them in a larger size. The structure selected by Avicenna has another interesting characteristic that we can define as the application of an all-encompassing, but strictly rational system. Both the macro-sequence of the three sections, examining name and nature, qualities, and single operationes against various pathologies, and the micro-structure of each section, going from name to external aspect and strategies to collection (the first), from primary to secondary and qualities (the second), and then to more specific applications and uses from skin to the organs a capite ad calcem, to finish with the diseases totius substantiae, (the third) follows a clear path going from origin to employment of the simple remedy on one hand, and from rational knowledge to practical experience on the other (I will come back to this connection between rationality and experience later). And, last but not least, the constantly applied, rationally motivated structure selected by Avicenna allows two different ways to consult, read, understand and use the collection, that is, by extracting the single features of each item with reference to each section, sub-section, and category, and by cross-reading the entries, searching for similar sorts of information (e.g., all the items including information on Apostemata et bothor) and regrouping all data related to that category. In this way, the reader can create his own collection of materia medica structured on categories of information (e.g., pathologies or properties and effects), rather than on the alphabetical order of the nomenclature, or aggregate different substances because of the similarity of their natural characteristics and/or properties, effects, and actions. This is, for instance, what the anonymous compiler of the collection preserved in the manuscript Bologna, Biblioteca Universitaria, 586 (1082) does. The Bolognese manuscript, a large paper codex consisting of 227 folios written by a single, probably Italian, hand during the 15th century and formerly belonging to Ulisse Aldovrandi’s private library, is identified by L. Frati, compiler of the catalogue of the Latin manuscripts of the University Library, as ‘Simplicium liber manuscriptus’ (Frati 1908, 306, nr. 586). Frati, however, being unaware of the principles of description listed and used by Avicenna, failed to recognize the real identity of the collection,
Classification Systems and Pharmacological Theory
139
which is nothing but a re-organization of the second treatise of the second book of LC combining the alphabetical order of the medicamina simplicia and the categories that structure Avicenna’s entries. The collection is then organized in 15 sections dealing with Quid et quod species (the first: f. 1r–38v), Electio (the second: f. 41r–49v), Natura (the third: f. 51r–66v), Operationes et proprietates (the fourth: f. 67r–89r; this section is accompanied, on f. 89v–92v, by some tables displaying Galenic qualities and grades of simple remedies), Decoratio (the fifth: f. 95r–102v), Apostemata et bothor (the sixth: f. 103r–109v), Ulcera et vulnera (the seventh, f. 111r–119v), Instrumenta iuncturarum (the eighth: f. 120r–127r), Egritudines membrorum capitis (the ninth: f. 128r–140r), Egritudines oculi (the tenth, f. 144r–151r), Egritudines membrorum hanelitus et pectoris (the eleventh: f. 152r–159v), Egritudines membrorum nutrimenti (the twelfth: f. 162r–174v), Egritudines membrorum expulsionis (the thirteenth: f. 176r–198r), Febres (the fourteenth: f. 203r–205v; some folios seem to be lost in the quire), and Venena (the fifteenth: f. 207r–214r). Each section is, in turn, structured in form of tables, and arranged as an alphabetical list of the simple remedies that, in LC, offer information on the subject; each entry reproduces, in shortened form, Avicenna’s account, sometimes integrated by other sources. A complex system of numbering accompanies each entry; it was probably planned to allow the reader to navigate across the collection and find other mentions of the same item of materia medica. The collection was, with all probability, put together by the scribe and first owner of the manuscript for his personal use, and shows a long, and unfortunately unachieved, work of excerpting and reorganizing Avicenna’s pharmacological account in table-form by using all criteria of classification and description of simple remedies provided by the Arabic physician. The name of the compiler will remain unknown, for the first folios are now lost. If a more complete Prologue or a simple preface including the name of the author or the explicit mention of his goal existed, we will likely never know. However, we can claim that this work could not have been possible if Avicenna had not structured his pharmacological account in such a rational, all-encompassing, stable fashion. The aim of providing a rational and stable description of the materia medica is pursued by Avicenna by gathering all definitions belonging to some of the headers in a small collection of 12 Areolae placed at the beginning of the second treatise. Avicenna’s choice to collect these definitions deserves to be described and assessed in full, in order to understand its function and value for ‘rational pharmacology’ delivered by the Arabic physician. First of all, it is necessary to recall what the ‘areolae’ are. Perfectly represented by John of Saint-Amand’s work bearing the same title, the Areolae are collections of short definitions (the word areola means ‘a small area’ used to gather and collect) organized in order to foster quicker learning and memorizing of basic ideas of medicine43. John of Saint-Amand’s work is consequently divided
43 On John of Saint-Amand, cf. Schalick 1997; Jacquart 1998; McVaugh 2011.
140
Iolanda Ventura
into three sections, the first of which is devoted to materia medica, the second to the parts and the organs of the human body, the third to the principles of pharmacy and therapy. Without doubt, we may assume that John, a physician who was in all probability related to the Parisian milieu of the last decades of the 13th century, took the principle and the text type of areola from Avicenna’s LC. Now, to go back to LC, we notice that Treatise 2 of book 2 of Avicenna’s manual opens with 12 collections of short definitions called areolae. Those 12 short collections, each of which has the length of a paragraph, are nothing but lists of all kinds of properties extracted from the sections of each entry opened by the same header. The 12 collections correspond to the sections of the entry from operationes to venena. Here, I would like to offer an example of the content of one of the aureolae: the first deals with one of the main aspects of classification of materia medica, the secondary qualities. This is its content: Subtilis, spissa, viscosa, exsiccativa, subtiliativa, inspissativa, inviscativa et resolutiva, abstersiva, applanativa, aperitiva orificiorum venarum, laxum faciens, incisiva, disrumpens ventositates, attractiva, mordicativa, repercussiva, mundificativa, sedativa dolorum. Et est prohibens effusionem materiarum, rubificativa, pruritum faciens, ulcerativa, corrosiva, adustiva, rectificativa putredinis, putrefactiva, cauterizativa, roborativa, digestiva, crudum faciens, stupefactiva, opilativa laxi et rari et inflativa, lavativa, lubrificativa, expressiva, contractiva, extinctiva, clarificativa, sanguis, sudoris provocativa, retentiva, sanguinis retentiva, sudoris laudabilis, chimi vituperabilis, chimi expulsiva, nocumenti cibi vel membrorum, multi nutrimenti, pauci nutrimenti, confortans membra, confortans ulcera mali humoris, convertitur ad malum vel omnem humorem, confert egritudinibus melancolie, generat melancoliam, generat coleram citrinam, generat flegma, expellit nocumentum flegmatis, confert senibus. Operationes exstrinsece: operatio eius in aere et facit penetrare solutiva (LC, II,2, areola I: f. 87rb–va).
The list I have just reproduced does not make any hierarchy of powers and effects evident. Only small groups or small clusters of therapeutic effects can be recognized; among others, the little sequence related to food and chimus, or the last one concerning humors (melancolia, colera citrina, flegma). Whether these connection are accidental or not, we cannot say. There is another characteristic of this list of operationes that is impossible to explain: the absence of any mention of a category of medical remedies sharing the same property of being ‘cordialia’, viz. of being able to comfort the heart and the soul, and to contribute to the spiritual well-being and strength of a patient who we would define nowadays as ‘depressive’. This absence is much more surprising when we consider not only that Avicenna mentions this particular therapeutical property, but also that he wrote a treatise on this specific category of simple remedies and their effects, the De medicinis cordialibus (also known with the title of De viribus cordis). Perhaps an explanation of this absence can be found in the fact that the De medicinis cordialibus was written when LC had already been completed. Nevertheless, it is interesting to notice that Avicenna mentions the effect ‘confert cordi’ with reference to some substances such as ambergris (II,2,63), gold (II,2,79), or musk (II,2,464) that he later mentioned in his treatise on medicamina cordialia (Ibid.: respectively f. 94vb, 96vb, and 133ra). Thus, we could venture that, while
Classification Systems and Pharmacological Theory
141
writing LC, he had already recognized from experience, or gathered from his sources, the link characterizing these substances and their actions, but had not converted it into a specific, theoretically well-founded category to be isolated and dealt with in a specific account. The category of the medicamina cordialia was, after the introduction of Avicenna’s treatise into Western medicine via an anonymous translation sometimes attributed to Arnald of Villanova, received as a fait accompli, and the substances belonging to them used and prescribed – at least, to the patients who could afford them! – to heal all sort of melancholic and depressive conditions. Whatever the situation of the list, and no matter which operationes are lacking, one conclusion can be derived from reading it. Avicenna offers his readers two different ways to structure the entries and their content, viz. by employing, on a more practical level, the same parts of the description (and headers) throughout the collection, and, on a more theoretical one, by collecting and listing the definitions we can find within these headers. In this sense, the reader becomes not only aware of the elements of the system of classification used by Avicenna in each section of the entry, but becomes acquainted with their definition and their constant and coherent meaning. He can therefore be sure that the same definition has one and the same meaning throughout the text, and that he can relate the simple medicines sharing the same characteristics. Just to give an example: the secondary quality defined as repercussiva is not only employed with the same meaning, but is included in the list of the Operationes et proprietates gathered in the first areola. The reader can then obtain the following kinds of information and help: 1) the operatio/proprietas repercussiva is a secondary quality because it is included in the list; 2) no matter how many times it is featured in the collection, it indicates one and the same quality; 3) the simple remedies defined as repercussive share the same powers and effects, and can, at least with respect to that, be connected. As we may easily imagine, not all headers and sections of description were treated in the areolae; only the sections concerning the therapeutical qualities and the single operationes, that is, the sections 3–10 out of 14, have their main definitions reduced into keywords. Avicenna’s choice can be easily explained by the fact that these sections were the most important, and include most of the definitions the reader needed to bear in mind and to memorize. Besides, they were the sections that allowed the reader to connect the second book to the wide, rich pathological (book III and IV) and pharmaceutical (book V) accounts of LC, and to grasp the reason why a specific simple remedy was used to heal a disease (e.g., if a substantia repercussiva was required, one may know which simple remedy has this property)44, or was included
44 E.g.: Among the general therapies (Canones universales curationis) against head diseases, Avicenna recommends the use of ‘odorifera aperitiva sternutationem facientis et caputpurgia, scilicet pulveres vel liquores qui per nares mittuntur et embrocas que materias de capite deponant’ (Ibid.: III,I,i,29, f. 171ra). Obviously, the odorifera and the aperitiva could be searched and found among the
142
Iolanda Ventura
among the ingredients of a compound because of its operationes45. In this sense, we can venture that the ‘categories’ and principles of description and classification of the simple remedy were not only meaningful for the pharmacological account, but for the whole manual, and work as ‘intra-textual’ categories. We do not know whether Avicenna’s system of classification of the secondary qualities became a constant or, at least, a widespread paradigm to distinguish and categorize simple remedies according to their therapeutical properties and effects. Nevertheless, we can argue that his influence was great in making Late Medieval medicine recognize the meaning of the secondary qualities as a system of classification tout court. In that sense, we should not forget that the already quoted Areolae written by John of Saint-Amand are organized according to the secondary and the tertiary qualities (the secondary are the first to be dealt with, though, and should therefore be considered to be necessary introductory information), rather than according to the more accessible alphabetical order, which was possibly more widespread during the 13th and the 14th centuries. At the same time, we should remember that the same John opens his work with a Prologue recalling Avicenna’s LC as the ultimate source and inspiration for the system he had adopted. The same influence can be detected in another master of medicine active in the Parisian milieu, Pierre de Saint-Flour, whose Concordantiae, an alphabetic lexicon of medical concepts written between the end of the 13th and the beginning of the 14th century, and planned to be a complement to the homonymous work of John of Saint-Amand, includes 20 entries bearing the title Medicina and devoted to various kinds of secondary qualities. In these entries, several of which only contain a cross-reference to medical authorities such as Galen, Ali ibn-Ridwan, the Pseudo-Mesue, or Avicenna, we notice that the Arabic physician is widely quoted, and represents, together with Avicenna, the main source of the entry Medicina temperata, viz. on the simple remedy of equal, perfectly balanced complexio46. If the constant employment of the same structure and organizational system is the first strategy applied by Avicenna to construct a rationally based pharmacology, and the extraction and collection of all definitions used in the various sections of the entries concerning simple remedies is the second, the third, and perhaps the most
simple remedies listed in book II. 45 It must be recalled that the fifth book opens with a chapter De necessitate compositarum medicinarum, where Avicenna explains the reasons of the necessity of using compound medicines because illnesses are a mixture of different unbalances, or because the power and the effect of one simple remedy has to be integrated, increased, or accompanied by the ones of other medicines (Ibid.: V, Tractatus scientialis: cap. I, f. 507ra–va). On pharmacological theories concerning compound medicines, cf. Tzivi Langermann 2003. 46 Petrus de Sancto Floro, Concordantiae, in: Pagel 1896. On Pierre de Saint-Flour, cf. Jacquart 1998: 188–194. On the Late Medieval development of a ‘rational pharmacology’, cf. the ‘Introduction’ to McVaugh 1975 (= Arnaldus de Villanova, Aphorismi de gradibus).
Classification Systems and Pharmacological Theory
143
relevant, one is represented by the canones universales, viz. by his discussion of the basic rational criteria that constitute the basis of theoretical pharmacology. These rational criteria are described and discussed in LC in the first treatise of book II, which consists of six long chapters. The sequence of the subjects approached by Avicenna is, at first glance, in some ways similar to the one shown by the first chapters of Constantine’s DPM; the same subjects, such as the role played by smell, flavours, and colours in the classification of simple remedies, are approached both by Avicenna and Al-Majusi (the latter was, it must be noticed, one of the sources of the former!). The content and the intellectual horizon are, however, significantly different. This is, in short, the content of the six chapters: the first (De complexionibus medicinarium singularium) deals with the composition of simple remedies, with the circumstances and the factors that can change them, and with their relationship to the composition of the human body. The second chapter, possibly one of the most relevant parts of LC, deals with the ways to identify and understand effects and powers derived from the composition through practical experience (De cognitione virtutum complexionum medicinarum singularium per experimentum). This chapter has often been mentioned and described by scholars, who generally consider it the demonstration of the specificity of Avicenna’s pharmacological theory, that is, his double approach, combining rationality and practical experience. The third chapter marks a further step, for it deals with how to identify and understand the effects and powers of simple remedies with the help of a rational process (De cognitione virtutum medicinarum singularium per ratiocinationem). After dealing with the complexio, the Arabic physician now turns his attention to the operationes, which become the object of the fourth chapter (De operationibus medicinarum singularium). The fifth chapter is, on the other hand, devoted to the right assessment of the external circumstances or ‘accidents’ that can modify the qualities, effects, and action of simple medicines (De iudiciis quarundam medicinarum extrinsecis). Finally, the sixth chapter achieves the ‘theoretical instruction’ of the rational physician and ‘theorizing pharmacologist’ by offering some instructions on the right way to collect and keep natural substances (De collectione medicinarum et ipsarum conservatione). The second book of LC became one of the fundaments of Late Medieval academic pharmacology. Although not widely commented upon, it attracted the attention of two of the main interpreters of Avicenna, Dino del Garbo and Gentile da Foligno, two magistri medicinae who actively contributed to the diffusion and the discussion of Avicenna’s medicine within Italian universities, and to the creation of a ‘medical scholasticism’47. Whereas Gentile commented upon the whole LC, Dino’s broad commentary, the Expositio super canones generales de virtutibus medicinarum simplicium secundi Canonum Avicenna, is especially devoted to the first treatise of book II, and therefore deserves special studies, for it is a perfect
47 On Avicenna in Italian universities, cf. Jacquart and Micheau 1996: 167–203; Siraisi 1981; 1987; 2001; Chandelier 2007 and 2017. On Gentile da Foligno, cf. French 2001.
144
Iolanda Ventura
example of the impact LC had on the development of a debate concerning the principles of rational pharmacology in the academic milieu48. In this study, I will only discuss the chapters I–IV which, by taking as a subject the principles of composition and the effects of simple remedies and the criteria of their identification, supply some useful data to assess Avicenna’s theoretical framework of pharmacology, from nature to the effect of medical remedies. The first chapter of treatise I, De complexionibus medicinarum singularium, deals with the correct definition and perception of the composition, taking into account that of the human body and of the elements of nature interacting with it. Thanks to J. Chandelier’s and A. Robert’s recent study on the emergence and the development of a medico-philosophical discussion on the idea of complexio in 14th century Italy, a discussion that was largely based on the commentaries of Taddeo Alderotti, Gentile da Foligno, and Ugo Benzi upon LC, we know more about the attempt made by medical scholasticism to separate the notion of ‘composition’ from the one of ‘soul’, and to make the former a ‘substantial quality’ identifying the human being as such, and therefore overcoming all specificities of each body and person (ChandelierRobert 2013). However, as the two scholars rightly emphasized, the debate concerning composition is not limited to the definition of the substance of the human body, and therefore to human anatomy and physiology, but extends to other branches of medicine, viz. therapeutics. More specifically – and, for our discussion, more interestingly – the discussion touches on the elements of the natural world, especially when used as medical remedies. The content of the first chapter of the first treatise of LC, book II, takes up that subject. However, as Avicenna states right at the beginning, the issue is not new in LC; it had been dealt with before, viz. in the well-known section of Book I De his que proveniunt ex his que comeduntur et bibuntur concerning the distinction between food and medicine, where he had defined how a substance ingested will operate in the human body (LC, I,2,2,1,15, f. 33va–34va). In this widely commented section, Avicenna had distinguished, following the path traced by Galen at the same time as applying the connection of matter and form postulated by Aristotle, three ways of action of a medical remedy, according to a quality (1), to the matter (the mixture of its components, producing a composition (2), and the ‘whole substance’ and the forma specifica (viz. the specific virtue and quality acquired by a compound, not just as a consequence of the mixture of its ingredients and their qualities, but because of the perfection it acquires while attaining to its own identity; these qualities and virtues are, however, not always detectable by the senses (349). Avicenna summarizes what he had said before by declaring:
48 Editions: Avicenna 1520; Dinus del Garbo 1514 (digital version in: www.muenchener-digitalisierungszentrum.de). 49 Cf. Weill-Parot 2013: here 41–61. On Avicenna and the forma specifica, cf. also Jacquart 2008.
Classification Systems and Pharmacological Theory
145
Iam declaravimus in libro primo intentionem sermonis nostri, scilicet hec medicina est calida, et hec medicina est frigida, et hec medicina est humida, et hec medicina est sicca. Et ostendimus illic quod illud est per complexionem ad corpora nostra, et premisimus quod omnium compositorum mineralium et vegetabilium et viventium elementa sunt materiae IIII, et quod ipse commiscent ad invicem et agunt in se donec consistant secundum equalitatem aut secundum dominium in eo quod est inter eas. Et cum consistunt secundum illud, est complexio vera, et quod complexio cum est contenta in composito preparat ipsum receptioni virtutum qualitatum quarum proprium est ut sint in ipso. Et ostendimus universaliter secundum quod divisiones est complexio et quid queritur per complexionem equalem in omnibus et quid per complexionem equalem in medicinis. Et ostendimus quod per hoc non intelligitur nisi quod corpus hominis cum obviat ei et agit in ipsam calore innato non prevalet, ita ut in corpore hominis imprimat infrigidationem aut calefactionem aut humectationem aut desiccationem ultra ea que est in homine. […] Et scias quod complexio est secundum duos modos, scilicet complexio prima et complexio secunda. Complexio itaque prima est complexio proveniens ab elementis. Et complexio secunda est complexio que pervenit a rebus que in seipsis habent complexionem, sicut verbi gratia complexio medicinarum compositarum, ut complexio tyriace; unaqueque enim medicinarum singularium tyriace complexionem habet sibi propriam, deinde cum miscentur et componuntur ita ut uniantur et adveniat eis commixtio advenit complexio secunda. Et hec quidem complexio secunda non est, ita quod non sit tota nisi ab arte, immo est quandoque etiam a natura. Lac enim commixtum in veritate est ex aqueitate et cascitate et burrositate. Et unumquodque horum trium est non simplex in natura, immo est commixtum et habet complexionem sibi propriam. Et hec quidem complexio secunda est ex operatione nature, non ex operatione artis (LC, II,1,1, f. 81rb–va).
In this way, he clearly shows that what will be said in the first treatise of Book II had its roots and principles already presented in the first book of LC, and therefore that pharmacology is a direct consequence of a more general discourse on the definitions of qualities, properties, and modes of action put together in the section where the fundaments of theoretical medicine are gathered together. What he had said before is, at the same time, the basis upon which he builds up his account of the complexio medicinarum singularium that constitutes the first introductory chapter of Book II. The rest of the chapter is therefore not devoted to the definition of the first and the second composition – that is, to the composition of the simple medicines as such, and of the one resulting from the mixture of several simple medicines, each of which has its own composition – but to the different kinds of either strong and weak connections and balances that may result from the mixture of simple substances having different, or, better said, contrary qualities, to the ways in which each single complexion of each simple remedy can be discovered, identified, and interpreted within the mixture, the ways in which it performs, and the ways in which different properties resulting from each complexion act. The outcome of such an approach is simple: Avicenna does not aim to define single properties of simple remedies and to classify them, but to understand the structure of a compound medicine and the role played by the composition of each simple remedy in each compound. In other words, he does not provide his readers with information useful to characterize each building block of a construction, but with principles necessary to grasp and identify its function and role in keeping
146
Iolanda Ventura
the whole edifice more or less stable and to perform its duty, and to crack the unity of the same edifice to recognize the same role and function. The same aim of offering general principles can be found in the second chapter, probably the most famous and quoted of the whole treatise, De cognitione virtutum complexionum medicinarum singularium per experimentum (LC: II,1,2, f. 82ra–va). The reason for its fame resides in its reputation as the methodological fundament of Medieval rational pharmacology on the one hand, and in the particular meaning given by its author to the ‘experimental way’ to determine the properties of a simple remedy and to obtain reliable knowledge of this, on the other. In other words, we could say that Avicenna provided contemporary medicine with a rationally based ‘experimental protocol’ to be used to gain certain evidence of the power of a simple remedy. According to Avicenna, the physician who wishes to ‘prove’ the value and the properties of simple remedies should conduct six different tests under specific circumstances. First of all, in order to be correctly evaluated, the medical remedy should be deprived of any accidental quality, such as induced warmth or coldness, or of any quality obtained through an alteration of its essence: Una est ut fit medicina vacua a qualitate acquisita est caliditate accidentali aut frigiditate accidentali, aut qualitate que accedat ei cum alteratione in sua substantia aut vicinitate alterius (Ibid.: II,1,2, f. 82ra).
Secondly, the medical remedy should be proved against an elementary, not ‘composed’ or mixed disease requiring the application of different, if not contrary, therapeutical strategies and medicaments, to avoid confusion in determining the property: Et secunda est ut illud, in quo experitur, egritudinem habeat singularem; nam si egritudo est composita et in ea sunt due res significantes duas curationes contrarias et experiatur in ambabus medicina et conferat, nescimus causam in illo vere (Ibid.: II,1,2, f. 82rab).
Thirdly, the physician should take care to apply the medicament in the cases of two opposed diseases (or, more probably, two opposed types of ill people), in order to ascertain that the substance heals them both, is not opposed in its composition to only one of them, and does not heal in one case because of its own qualities, and in the other one by chance: Et tertia est ut medicina sit experta in duobus contrariis, ita ut cum utrique eorum simul conferat, et non iudicetur quod sit contraria complexioni alicuius eorum, et fortasse confert uni eorum per se et alteri per accidens (Ibid.: II,1,2, f. 82rb).
Fourthly, care should be taken to apply a medical remedy that is opposite in its effects to the illness, and should be at least as powerful as the illness it is supposed to heal; then, in order to determine the degree of its efficacy, it should be tested at different levels of seriousness, until the physician can determine how strong the substance is:
Classification Systems and Pharmacological Theory
147
Et quarta est ut sit virtus que est in medicina opposita ei quod equalis ei de potentia egritudinis. Scientia enim medicine quedame quarum caliditas minoratur a frigiditate egritudinis alicuius, quare potentius in eam non agit, et quandoque cum administratur in infrigidatione aliqua levior est illa fac** calefactione. Oportet ergo ut experiatur prius in debiliori, et secundum gradus fiat processio paulatim donec sciatur virtus medicine et non dubitetur (Ibid.: II,1,2, f. 82rb).
Fifthly, the physician should take the time into account, in order to establish whether the therapeutical effect is the same throughout time, or varies: Et quinta est, ut observetur tempus in quo apparet eius impressio et operatio. Nam si in primis, cum administratur, apparte et confidere facit quod efficit illud per se, et si fuerit operatio que in primis apparet ex ea contraria ei que apparet in fine, aut in principio rei non apparet ex ea operatio, deinde in fine rei apparet ex ea operatio, tunc erit locus ambiguitatis et dubietatis (Ibid.: II,1,2, f. 82rb).
Sixthly, time is considered again, but in terms of continuity, since the physician should consider whether the simple remedies keeps the same effect throughout time, or that effect vanishes, for in the latter case, the effect could be of accidental nature: ‘Et sexta est, ut observetur processus operationis eius secundum assiguationem, aut secundum plurimum. Nam si non fuerit ita, tunc processio operationis ab ea est per accidens, quoniam res naturales procedunt principiis suis aut semper aut secundum plurimum’ (Ibid.: II,1,2, f. 82rb–va). Finally, the last rule to follow is to test the simple remedy on the human body, because the balance between the composition of the remedy and that of the human body can be different from the balance between the composition of the remedy compared to an animal one; besides, the test on two bodies may let properties and effects emerge that a single one cannot: Et septima quidem est, ut experimentum sit in corpore hominis. Nam si experiatur in corpore non hominis possibile est, ut fallat duobus modis: unus eorum est quoniam possibile est, ut fit medicina comparata ad corpus hominis, calida et comparata ad corpus leonis et equi frigida cum hec medicina est calidior homine et frigidior equo vel leone […]. Et secundus quoniam quandoque comparata ad unum duorum corporum habet proprietatem quam non habet comparata as corpus secundum (Ibid.: II,1,2, f. 82va).
As the skilled reader has noticed, the criteria are more or less the same that we find in Constantine’s PP (DPM, chapter III), with the notable exception that, in Avicenna’s view, those criteria are not just a ‘list of useful tests’, but each of them is part of a logical process. Whatever the theoretical degree of Avicenna’s reasoning, we are allowed to argue that, in this case, the author of LC is not only describing his experimental criteria, but is embarking on a dialogue with Al-Majusi and the Arabic medical tradition predating him. And, at the same time, he is giving the ‘practical experience’ a particular form by fitting into a logical process, and by proposing a logical regulation of it. After dealing with ‘general’ principles, such as complexio and virtus that can help with recognizing, interpreting, testing, and defining the nature and the therapeutic
148
Iolanda Ventura
effects of all simple remedies – or, better, after making his reader aware of what the composition of any simple remedy is and how it affects the therapeutical outcome of a compound medicine, and of the way in which any substance should be tested in order to have some reliable knowledge of its powers –, Avicenna focuses, in the third chapter, on the criteria that can help to distinguish and classify each simple remedy in isolated fashion. At first glance, the chapter does not seem to be very innovative, since traditional topics like the distinction of tastes and colours are widely discussed. However, if we follow the logical development of Avicenna’s argumentation, the originality of its ‘rational’ system becomes more evident. Interestingly enough, the chapter starts with an allusion directed toward some unspecified alii who use criteria such as the speed and slowness displayed by drugs in transforming themselves into ‘heat’, viz. into energy and into active factors of healing, to classify simple remedies, or take into account features such as smell, taste, and colour to derive from these external features an opinion about their therapeutical properties: Cognitionis autem virtutum medicinarum singularium per ratiocinationem canones alii sumunt ex velocitate conversionis earum ad ignem et calefactionem et ex tarditate conversionis ipsarum, et ex velocitate congelationis earum tarditate congelationis ipsarum. Et alii sumunt ex odoribus, et alii ex saporibus, et quandoque sumuntur ex coloribus, et quandoque sumuntur ex operationibus et virtutibus notis, ex quibus acquiruntur significationes sue manifeste super virtutes ignoratas (LC: II,1,3, f. 82va).
We do not know who may be hiding behind the allusion to the ‘alii’; nevertheless, I would like to recall that the capacity to heat and cool, to dissolve or to coagulate were the main criteria used by DPM (that is, the first section of book II of Constantine’s PP) to define the main properties of a drug in an experimental sense (and not a rational one, for rationality does not seem to be involved in Al-Majusi’s/Constantine’s treatise), as well as colours, smells, and tastes (PPII, DPM, IV–VIII, in Opera Ysaac: f. 65vb–66va). Maybe we can venture that Avicenna, by opening his discussion about the distinction and classification of drugs per ratiocinationem, aims to go beyond those simple, or, better, simplistic criteria, and to propose a new, rational, theoretically based method explaining, in the first instance, what speed and slowness of ‘transformation’ mean and imply, and how you can distinguish and structure smells, tastes, and colours in a logical way in order to make them something more than simple external ‘accidents’. In other words, it seems to me that Avicenna’s ‘rational’ method aims to propose a ‘causality’ of the main effects of cooling/heating and dissolving/ coagulating, and a clear, solid, logical distinction of the characteristics perceivable to the senses that are, in the last instance, a clear emanation of the primary qualities (heat, coldness, humidity, dryness). It is in this sense that we can read the long passage that represents the first subsection of the chapter, where Avicenna explains how the constitution and the consistence of a substance connects with its qualities and effects. By probably using a
Classification Systems and Pharmacological Theory
149
link that had already been established by Galen between leptomera and pachymera qualities and complexions (the Greek physician, as I have said before, claimed that a light, rare substance is also warm in quality and effect, whereas a heavy, thick one is cold; cf. on this matter Debru 1997b), Avicenna states that this characteristic also connects with the capacity of a substance to become warmer or colder because of external influence, and therefore to change its status, or can influence the status of another substances: Via autem prima est quoniam rerum similium in constitutione essentie, scilicet in raritate et spissitudine, quecumque recipit calefactionem velocius est calidior, et quecumque recipit infrigidationem velocius est frigidior. Et una earum in hoc est, quoniam res quandoque calefit velocius alia, agente existente uno, quoniam in seipsa est calidior altera, et neque infrigidat ipsam nisi frigus accidens. Cumque advenit caliditas exterior et consentit ipsi virtus calida innata que est ipsa equatur alii in causa exteriore et superfluit illi in virtute que est in ipsa et fit calidior, et secundum hoc cognosces dispositionem eius que infrigidatur velocius (LC: II,1,3, f. 82va).
This apparently obvious statement is, in fact, of crucial meaning, and not only because it compares various substances and to place them on a scale going from cold to warm, but, because its states the permeability they have to the influence of another element or substance, to consider and plan the combination of simple remedies within a compound. A similar reasoning follows some lines after, where Avicenna illustrates the second rational principle allowing to discerning the qualities of drugs rationally, the rapidity/slowness in freezing or warming: Res autem quarum proprietas est ut congelentur et ille quarum proprietas est, ut inflammate igniantur, possibile est ut ad invicem comparentur; ergo quecumque est velocioris congelationis et eius essentia est sicut essentia alterius est frigidior, et quecumque est velocioris inflammationis et eius differentia est sicut essentia alterius est calidior propter illud quod diximus prius (Ibid.).
According to Avicenna, discriminant factors like nature (cold and warm) and complexion as well as the capacity to change status and to influence the change of another substance build a system that allows to order simple remedies into a structure and to create a system of relation between them, for a drug is not just warm/cold, light/ thick, fast/slow in coagulating or loosening as such, but has the same characteristics in relation to other drugs, or when compared with them. That connection is not established, according to Avicenna, by medicine, but by natural philosophy, and therefore depends on the laws of nature more than on pharmacological principles. And it is according to natural philosophy that possible exception should be considered (e.g., drugs that are warm in their qualities but thick in their substance). It is only after dealing with the rational principles determining the substance and the main qualities and properties of drugs – in other words, with their essence – that Avicenna can go back to the criteria that were commonly used by pharmacology to dis-
150
Iolanda Ventura
tinguish between substances and to determine their nature, quality, and powers, that is, the ones connected with the sensible perception: tastes, smells, and colours. The rest of the chapter is devoted to the distinction of the eight tastes, the various colours stretching from white (a sign of dryness and warmth) and black (a sign of humidity and coldness), and smells. With reference to tastes, smells, and colours, Avicenna’s account does not seem to be different from the one we can find in other medical manuals. Therefore, it is not necessary to dwell on that in detail. Smells, tastes, and colours are described separately with their qualities, properties, and effects; they are connected with qualities (warm or cold, for instance, in case of smells; ibid.: II,1,3, f. 84rb), and their degrees of intensity (e.g., sapor acutus is the warmest one, followed by amarus and salsus; ibid.: II,1,3, f. 83va). In this way, Avicenna provides a description of perceivable qualities that, if not original, forms a compact, rationally structured, reliable system. Nevertheless, it is interesting to notice some aspects of his discussion that are useful to our discussion of rational criteria of identification and classification of substances. First of all, it is interesting to notice that taste, smell and colour are not considered by Avicenna to be three equal characteristics, but are organized into a hierarchy of qualities that can be perceived by the senses. According to him, taste comes before smell and colour, for its perception is clearer and more reliable, even in cases of mixture, because a dominant characteristic and quality would prevail over the others: Harum quidem qualitatum sensatarum quedam sunt digniores ut efficiat in eas impressionem manifestam illud quod eis permiscetur ex contrario, et dum perseverant qualitates, earum vere sensate et non sentiuntur in eis contrarie earum sunt vincentium virtutum, et hoc quidem est in saporibus non secundum quod sit necessarium, sed secundum quod est plurimum, et post sapores est in odoribus, et post utroque est in coloribus, et in coloribus quidem est sicut sine fiducia cum eo. Et ex causis quidem in quibus sapores vincunt odores, in hoc calido est adventus eorum ad sensum cum obviatione. Sunt ergo digniores ut perducant ex omnibus partibus medicine virtutem, sed odores et colores imprimuntur sine obviatione partium suorum corporum, quare possibile est ut perveniat ad sensum ex partibus habentibus odorem vapor ex subtili partium eius et inobediens fiat vapor spissus partium eius, et non vaporet, et possibile est ut perveniat ad ipsum color manifesti, et est vincens sine vicio occulto (Ibid.: II,1,3, f. 83rb)50.
A second passage where we notice how tastes, smells, and colours are not just recalled, but structured into a rationally organized system, and therefore become part of the acquisition of ‘rational pharmacological knowledge’, is represented by the way in which Avicenna relates perceivable qualities and effects. While dealing with taste, Avicenna creates a network of associations between tastes and properties/effects. This association is exemplified by following passage:
50 On that subject, cf. Bénézet 1999: 456 f.
Classification Systems and Pharmacological Theory
151
Operationes autem dulcis sunt digestio et lectificatio (sic!) et multiplicatio nutrimenti, et natura diligit ipsum et virtus attractiva attrahit ipsum. Et operationes amaritudinis sunt abstersio et exasperatio, et operatio ponticitatis est constrictio si est debilis, et exasperatio si sit fortis. Et operationes stipticitatis sunt contractio et inspissatio et induratio et retentio. Et operationes unctuositatis sunt lenificatio et lubrificatio et coctio parva. Et operationes acuitatis sunt resolutio et incisio et putrefactio. Et operationes salsedinis sunt abstersio, ablutio et exsiccatio, et prohibet putrefactionem. Et operationes acetositatis sunt infrantio et incisio (Ibid.: II,1,3, f. 84ra).
This passage is important for our discourse, from different points of view. First of all, because of the clear and fix system it describes. The structure created by Avicenna with the aim of relating tastes and properties/effects can be translated visually into a diagram in which one or more tastes are connected with one or more effects. This diagram could be translated, in a certain sense, into a system of axes (horizontal, vertical), where each taste has its ‘zone’ filled with its effects. Secondly, this passage is interesting because it shows that the effects of tastes are not limited to their effect on senses; their power of diluting, purifying, irritating, drying, or burning, are referred to the material (read: humors) of which the body shows an excess, as well as to the effect the substance has on it. Therefore, taste is not a simple ‘characteristic’ or external quality of a substance, but is an essential element, maybe the main reason, of its therapeutical property, effects, and powers. In this sense, tastes (and, in a more limited way, smells and colours) are not simple ‘accidental features’, but perceivable signs of the real, intrinsic power of a simple remedy. Acknowledging, learning, and using the connection between tastes and effects on the body is a rational act that can be applied to every substance; therefore, taste, smell, and colour are elements that belonged to the domain of the ‘rational knowledge of simple remedies’. Finally, a more complex way to include taste, smell, and colours into a rational systematization of pharmacology can be found in a passage in which Avicenna discusses the nature of colours. Here, Avicenna observes that colour should not only be considered an element of distinction, characterization, identification, and classification of single substances, but that colour is also the result of a mixture (commixtio) of several substances in a compound, and therefore should be evaluated within this process, too: Et similiter odores boni acuti non sunt, nisi in substantia calida, et colores albi in corporibus congelatis in quibus est humiditas. Et colores nigri non sunt, nisi in substantia frigida. Et colores albi, in quibus est siccitas et pulverizatio, non sunt, nisi in substantia calida et nigri in utrisque rebus sunt econtrario. Frigus enim albificat humidam substantiam et denigrat siccam, et calor denigrat humidam et albificat siccam. Et hoc quidem verum est necessarium. Verum his est causa alia propter quam significationes iste quandoque diversificantur et proprie in odore et colore et proprie in colore suo. Et illud est, quod nos iam ostendimus quod corpora medicinalia commiscentur ex elementis contrariis quando commixtione primaria et quandoque commixtione non primaria, immo magis digna ut vocetur complexio secunda, et possibile est ergo in hac complexione secunda ut uni duorum elementorum iam advenierit complexio qua mereatur colorem aut odorem aut saporem. Et advenerit illud quod iam meretur, et ut alii elemento iam advenerit complexio diversa illi complexioni, et sit possibile, ut in ea mereatur colorem contra-
152
Iolanda Ventura
rium illi colori, aut odorem aut saporem contrarios primis. Et sit possibile, ut ea non mereatur contrarium primo. Quod si meretur colorem contrarium illi, deinde sint equalis quantitatis, pervenit in commixto secundo color compositus ex duobus coloribus, et si sunt diversa continetur in secundo commixtio (!) color magis declivis ad unum duorum colorum. Si autem secundum non meretur colorem omnino, et similiter odorem aut saporem et sint equalia, qui in eis reperitur color est primus et odor est primus. Et si sint iam fracta propter commixtionem partium privatarum minutarum vel inimica aliis contrariis et colori secundo non sit impressio penitus, quoniam hoc etiam frangitur fractione parvii permixti colorato, et quasi illud corpus videtur verbi gratia album (Ibid.: II,1,3, f. 83ra).
The consequences of this observation are, again, not small, for Avicenna is the medical author and thinker who initiated the long debate around the theory of the socalled forma specifica, that is, the specific nature and property of a compound medicine resulting from the perfection achieved by a compound that cannot be reduced to the combination of all the qualities contributed by each ingredient and their ‘fermentation’ and ‘dissolution’, leading to the creation of the compound. In this sense, his discussion of the commixtio of colours and its result in the compound (e.g., the white of a simple remedy is not the same white of a compound resulting from the mixture of several drugs, each of which had its colour and related quality!) should be read not as a simple question of criteria of classification, but, once again, as a prologue to the acquisition of the rational knowledge of simple remedy and of the logic in creating and assessing compound medicines. Compared to the inner complexity of chapters II and III, the fourth chapter, De operationibus medicinarum singularium, is much easier to understand and to summarize. In this chapter, as declared in the rubrication, Avicenna lists and describes the various therapeutic actions, powers, and effects of drugs, which he divides into universales and particulares: in other words, what we have already defined as secondary and tertiary qualities (the issues related to the composition, and therefore to primary qualities, being the object of the first three chapters of Book II). Most of this chapter is constituted by the description of 49 operationes, some of which we have already seen listed in the first areola that opens the second treatise of Book II51. Despite the similarities between the areola and the content of the fourth chapter of Book I, I think that the main interest of the latter is not in the presence or the absence of the operationes mentioned in the former. In other words, it is not a comparison between the two passages that should attract our attention. More important are the sequence and the principles of organization that determine it. In order to make my point more understandable, let us start with the passage introducing the sequence of descriptions of the various operationes. Here, Avicenna affirms: Medicinarum autem operationes famosas oportet non numerare secundum conditiones predictas, deinde consequenter ponamus descriptiones et expositiones nominum secundum ordinem
51 For the text of this areola, see above, 139 f.
Classification Systems and Pharmacological Theory
153
unum. Dicitur ergo medicina calefactiva, subtiliativa, resolutiva, abstersiva, carminativa vel exsiccativa, exaspersiva, aperitiva, mollificativa, maturativa, digestiva, incisiva, disrumpens, frangitiva ventositatis, attractiva, mordicativa, rubrificativa, pruritum faciens, ulcerativa, corrosiva, adustiva, frangens, disrumpens, putrefactiva, cauterizativa, excoriativa. Iste igitur sunt viginti duo operationes. Et est ordo alius: infrigidativa, confortativa, repercussiva, ingrossativa, crudum faciens, stupefactiva; et iste sunt sex operationes. Et est ordo alius: humefactiva, inflativa, lavativa, sordidativa, lubrificans, planativa. Iste itaque sunt sex operationes. Et est ordo alius: desiccativa, expressiva, contractiva, opilativa, glutinativa, consolidativa, gravativa, carnis sigillativa. Iste ergo sunt octo operationes. Et est aliud genus proprietatum secundum operationes earum perniciosa, venenosa, tyriaca, bezahar. Et etiam est solutiva, provocativa sudoris, evacuativa. Omnes ergo iste sunt XLIX operationes, quas elegimus de summa operationum proportionatarum vel pertinentium. Et nos quidem narrabimus unamquamque harum operationum sua descriptione (Ibid.: II,1,4, f. 85ra).
This passage deserves our attention, because of its considerable meaning for our overview of the criteria of perception, description, and classification of the properties of simple remedy, and for our little history of ‘rational pharmacology’. As far as I can see, it seems that Avicenna does not put together a simple list of the 49 operationes attributed to simple remedies, but that he organizes them into clusters, each of which is constructed around a ‘common feature’. As for the first four groups/clusters, the ‘common feature’ is a potentially good, or better, not necessarily dangerous one, that is, a primary quality, warm (the first, opened by the calefactiva), cold (the second, introduced by the infrigidativa), humid (the third, represented at first place by the humefactiva), and dry( the fourth, opened by the desiccativa), respectively. On the other hand, the fifth and the sixth group include dangerous, or at least potentially dangerous operationes, the ones performed by poisons and purgatives ( perniciosa, solutiva). By arranging thus, Avicenna not only provides a complete list of the secondary quality depending on complexiones and primary ones, but a rationally structured one that helps readers to place each operatio into a coherent cluster and to identify all operationes belonging to it. In other words, Avicenna does not simply offer a list of properties and effects, but a sequence of regroupments characterized by an associative principle which allows us to recognize the ‘common feature’ and to connect, in turn, any of the properties connected by it. To my knowledge, Avicenna is the only one to apply such an associative system, and to provide a rationally understandable way of classifying therapeutic properties and effects, and not a simple sequence. The list of 49 operationes is not, however, the sole system of interpretation and classification of the therapeutical properties of drugs. The beginning of the chapter presents, in fact, a much more interesting way of distinguishing and structuring secondary and tertiary qualities. Here, Avicenna reconfigures the regroupment of secondary vs. tertiary qualities and properties into a threefold system, differentiating between operationes universales, operationes particulares, and operationes similes universalibus. The operationes particulares, whose definition corresponds to our idea
154
Iolanda Ventura
of tertiary qualities acting upon a specific part of the body, are programmatically left aside by Avicenna, who does not consider it necessary to dwell upon them: Et operationes particulares sunt sicut iuvamentum in cancro et iuvamentum in emorroydibus et iuvamentum in ictericia et similia his. […] Nos vero non rememoramur hic nisi operationum earum universalium et similium universalibus (Ibid.: II,1,4, f. 84va).
More interesting for him is, on the other hand, to define the operationes universales and those similes universalibus. By defining the first ones, Avicenna goes beyond their simple connection with the primary qualities and their direct derivation from the primary ones and the main qualities (warm, cold, dry, and humid), by distinguishing two different levels of universal effects, a first (operationes universales prime) that derives directly from the main qualities, and includes their direct effects, such as warming, cooling, hydrating, and drying (calefactio, infrigidatio, humectatio, exsiccatio), and a second one (operationes universales secunde), defined by Avicenna in the following way: Secundarum vero alie sunt que sunt iste eedem operationes, sed sunt mensurate aut comparate cum termino additionis aut diminutionis sicut adustio et putrefactio et congelatio. Nam iste eedem sunt calefacientes et infrigidantes, sed sunt mensurate et comparate. Et alie sunt que sunt operationes alie, sed sunt precedentes ab istis, sicut stupefactio et sigillatio et attractio et lubrificatio, et operatio et conglutinatio et que sunt similia illis (Ibid.).
With these words, Avicenna abandons a simple method of gathering secondary qualities around the primary ones and/or to derive the former directly from the latter, and builds a more sophisticated system of progressive derivation which we could translate again into a geometrical image, this time one recalling concentric circles where the center is represented by the primary qualities forming the composition (warm, cold, dry, humid), the next circle by the properties and effects directly connected with one specific quality (warming, cooling, hydrating, drying), the next one by the powers and effects that cannot be directly identified with one, but are, as it were, ‘in between’, for they can be defined in comparison with the previous ones and/or as stronger or weaker in comparison with them (such as in the case of burning – adustio – in comparison with warming – calefactio –), and the last one as proceeding from them. This last category, that includes attraction or solidification, can be easily perceived as a further step toward a stronger and more precise, specific action, as a further ‘empowerment’, both in quantitative and qualitative terms. One cannot overlook or fail to appreciate the subtleness of Avicenna’s strategy to structure and classify secondary qualities, that overcomes the basic definition of ‘secondary quality’ and a simple list of them by creating a more sophisticated system moving from simple primary qualities to their extreme perfection. The same sophisticated way of distinguishing between operationes of different types and levels, and classifying them, becomes evident in Avicenna’s treatment of
Classification Systems and Pharmacological Theory
155
the operationes similes universalibus. Again, let us Avicenna speak and offer a definition: Universalibus vero similes sunt sicut ventris solutio et sudoris evacuatio. Et ante quidem quod loquamur de operationibus earum, loquimur de proprietatibus quas in seipsis habent. Dicimus ergo quod proprietates quas medicine in seipsis habent alie sunt ipse qualitates quatuor note, et alie sunt odores et colores, et alie sunt proprietates famose, quarum sunt iste subtilitas et spissitudo, et pulverizandi proprietas et viscositas et congelatio et mucillaginositas, et oleoginositas et spongiositas et levitas et gravitas. Medicina autem subtilis est cuius proprietas est ut cum a virtute patitur naturali que in nobis est dividatur in corporibus nostris in partes minimas, sicut crocus et cinamomo. Et hec quidem medicina magis confert in omnibus suis impressionibus quam omnes medicine, ita ut licet in eius desiccatione non sit mordicatio, tamen consequitur desiccationem rei fortis mordicantis. Et intelligo per spissam cuius non est illud de suis proprietatibus sicut cucurbita et gypsum (Ibid.: f. 84vab).
Avicenna’s definition of the operationes similes universalibus cannot be reduced to a simple mechanism of derivation of these properties from the ‘universal’ ones I have just described. The physician’s reasoning is more complex and sophisticated, and his goal is of great meaning: we can define it as an attempt to go beyond Galenic qualities, and to relate main features, ways of acting upon (or within) the human body, and therapeutic powers in a close system. This reasoning has an important consequence, viz.: what could have been identified (and, indeed, was identified), in a general and sometimes vague way, as a ‘tertiary quality’, becomes here a category in itself, and a category of therapeutic power and effects that does not just depend on Galenic main qualities, but on some sort of essential, substantial characteristics that determine the main effect of a natural element and of the simple remedies derived from it (and, consequentially, of similar drugs sharing the same substantial characteristics). In other words: if a natural element (e.g., a plant) and the drug(s) derived from it have a main characteristic (e.g., they are ‘subtle’, or, in modern terms, easy to dissolve and absorb), they will operate in the body in a specific, easily perceivable and identifiable way (e.g., they will dissolve rapidly and completely). Because of that, they will display a specific power, and act against a specific condition. By constructing the link between main characteristics, action, and effects, Avicenna ‘rationalizes’ tertiary qualities, making these effects easy to explain and to connect with internal characteristics that, otherwise, would have been left out of a purely Galenic therapeutical system, and transform what could have been considered an empirically established property into something that can be rationally explained and connected with an internal characteristic. Moreover, the operationes similes universalibus are not perceived as separated from, or opposed to, the ‘universal ones’, but are a further step emanating from them, a last, external, all-encompassing concentric circle. With this rich, sophisticated, and solidly structured theoretical background, the 49 operationes listed in the chapter become more than a simple, albeit typologically ordered, list: they become the last
156
Iolanda Ventura
consequence of a rational process of interpretation and classification of properties and effects that started with the essence of each natural substance or drug. With my journey through the first and the second treatise of Avicenna’s LC, I hope to have presented my readers with the most illustrious representative of the ‘theoretical turn’ in 12th century pharmacology. Compared to the previous authors, Avicenna does not simply offer a multi-layered form of structuring, classifying, and organizing pharmacology both in its content (drugs) and in its principles (such as what we found in al-Majusi’s/Constantine the African’s PP II), but also provides his readers with a sophisticated theoretical background and a profound discussion of concepts of complexion, qualities, and effects. This background can be defined as one that provides pharmacology with an overall, all-encompassing, rationalized, and, above all, solid system which, so to say, does not admit doubt and incertitude. This system does not just define, describe, and classify in a rational and coherent way, but structures, creates hierarchies and clusters, and connects. Moreover, this system is not just a medical or a pharmacological, but also a philosophical one, for its principles are rooted in natural philosophy as well as in medicine. Last but not least, the ‘new pharmacological system’ created by Avicenna revolutionized the background and reshaped the form of all of pharmacological theory and practice, not just single elements or aspects. Particularly important, in this sense, is the innovation provided by Avicenna of the idea of experimentum and experience as a strategy to define the power of medicine rationally, or of the connection between primary, secondary, and tertiary quality. In particular, it must be stressed that this connection becomes, in Avicenna’s reasoning, not just a matter of fact, but is the object of rational explanation and the basis of a rational system characterized not just by links, but by relation of cause and effect (or causal relation). On the other hand, Avicenna’s definition of the role played by practical experience and rational approach contributes toward an integration of, and not an opposition between, these two poles of the theory of acquisition of (general, not just medical) knowledge. Finally, we should recall the decisive contribution provided by Avicenna to the idea of composition, and of ‘specific form’. Later pharmacological culture and thinking, in the West as well as in the East, in Latin academic treatises as well as in Arabic manuals, will have to deal with Avicenna’s heritage, both with his description of the nature and the properties of simple remedies and with his theoretical premises. Every history of Late Medieval pharmacology should take Avicenna’s heritage into account. In particular, each analysis of ‘rational pharmacology’ should deal with him and his lore. Such an analysis cannot be carried out here, unfortunately, but must be reserved to future studies aiming to picture the reception of Avicenna’s pharmacology in Latin and Arabic culture, from the Middle Ages down to Early Modern times.
Classification Systems and Pharmacological Theory
157
3 Conclusions Our journey across the criteria of identification and description of drugs, the classification systems of collections of materia medica, and the development of a theoretical background in pharmacology has hopefully shown that we cannot focus exclusively on the historical development of the external structures and on the various types of organization systems exemplified by pharmacological works, but must acknowledge that behind those structures and systems, their choice, and their use by Ancient and Medieval authors, there is a consistent debate and reflection on the nature and the properties of drugs and the ways in which both nature and properties can be perceived, rationally interpreted, and converted into practice. Therefore, we cannot simply speak of ‘classification systems’ in pharmacology, but of ‘pharmacological theories and their connections with classification systems’. The ‘pharmacological theories’ or the ‘pharmacological theoretical debate’ addressed, during Antiquity and the Middle Ages, various subjects and issues, such as the connection between the nature and the properties, powers, and effects of simple remedies, the role played by experience and rationality in the perception, identification, and interpretation of the nature and properties of drugs, the perception and classification of the intensity of their effects, and the type of relationship between primary, secondary, and tertiary qualities and their organization and classification. These issues and the different answers provided by medical and pharmacological authors cannot be separated from the attempts to order and structure the universe of the materia medica and its elements. Our journey has also shown that the pharmacological literature of Antiquity and the Middle Ages evolved toward a more evident ‘rationalization’ of the portrait of simple remedies and of the structures of the entries. The affirmation of a ‘pharmacological Galenism’ during Antiquity and the Middle Ages had various consequences, among which we can stress the progressive disappearance of the external description of the natural element the drug was derived from, as well as the tendency to reduce and to organize in a more coherent way the multiplicity of properties, effects, and therapeutical powers, to provide a more coherent definition of the intensity of effects, and to organize the entries in a homogenous fashion, where therapeutical employments are presented as the direct consequence of the composition and the general properties and powers. The connection with the ‘reality’ of the perception of the powers of the natural elements and of the effects of the drugs derived from them does not get lost, of course, but is included in a rational and coherent process of interpretation. This ‘rationalized’ way of structuring the representation of nature and powers of drugs has a practical consequence, too, viz. the constant employment of coherent structures, in which any element has its place (and, actually, the same place within the entry), and is described in the same form, and the development of a specific vocabulary e.g., in the definition of the secondary qualities). We cannot speak, however,
158
Iolanda Ventura
of constant uses of the same structures, but of disappearances and ‘comebacks’ of systems of organization such as the alphabetical one, or the one based on properties and effects. The presence or the absence of systems of organization and the structure of drugs cannot be connected, we must stress, only to the esprit du temps, and the consequent preference for one structure instead of another, but also (and, perhaps), above all, to the function needed from, and requested of, pharmacological works. A good example of that can be found in the tradition of the Quid pro quo, a type of text that does not completely disappear after Late Antiquity, but can often be found in medical miscellaneous manuscripts, for example in connection with antidotaria and collections of compound medicines. And this, even after newer, and perhaps better, writings have been produced, because they are useful tools to find substitute drugs. In the same way, we may notice that alphabetical order is often employed when the compiler faces an ‘information overload’, and therefore needs a reliable, albeit not really practice-oriented, system of organization of his abundant material. Finally, we have seen that, even if it is possible to perceive and sketch a history of rational pharmacology, it is less easy to reconstruct a coherent development of rational pharmacological theories. Even the so-called ‘Galenic’ pharmacology is the result, in the Latin Middle Ages, of the mediation of Arabic sources, rather than of the Fortleben of the original treatises written by Galen himself. The consequence of this status quo is that, even if we manage to connect the authors, we struggle to link theories, and to put them on a chronological line. Further researches are not only necessary on that matter, but they must have a solid ground in terms of number and types of texts, and of applied methodologies. To give just one example: we have seen how the vocabulary of the secondary and tertiary qualities developed during Antiquity and the Middle Ages, for instance in the passage from Greek to Latin, or from Arabic to Latin, within medical culture. A linguistic study based on a solid corpus and a coherent methodology is therefore necessary to discover how, and to which extent, the vocabulary of properties and effects evolved through medical history, and how far it was stable or experienced changes. This linguistic study must later be compared to, and connected with, the historical perspective. Then, I cannot but conclude by recalling the preliminary character of this état des lieux, and with the necessity of further studies that will allow us to have a better picture of the process of creation and development of a ‘rational pharmacology’ through the centuries.
Classification Systems and Pharmacological Theory
159
References Primary Sources: Alphabetum Galeni. In: Everett, Nicolas (ed.). 2012. The Alphabet of Galen: Pharmacy from Antiquity to the Middle Ages. A Critical Edition of the Latin Text with English Translation and Commentary by Nicholas Everett. Toronto: University of Toronto Press. Alphita. In: García González, Alejandro (ed.). 2008. Alphita. (Edizione Nazionale “La Scuola Medica Salernitana”, 2) Firenze: SISMEL. Arnaldus de Villanova, Aphorismi de gradibus. In: McVaugh, Michael (ed.). 1975. Arnaldus de Villanova, Aphorismi de gradibus. (AVOMO 2) Granada/Barcelona: Publicaciones de la Universidad de Barcelona. Auslasser, Vitus, Herbarius. In: Mayer, Johannes Gottfried and Konrad Goehl (ed.). 2001. Höhepunkte der Klostermedizin. Der „Macer floridus“ und das Herbar des Vitus Auslasser, edited by Johannes Gottfried Mayer and Konrad Goehl. Holzminden: Reprint Verlag. Avicenna, Liber canonis. Venetia 1507: Paganino de Paganinis. Avicenna, Avicennae medicorum principis Canonum liber una cum lucidissima Gentilis Fulginensis expositione. Venetiis 1520: Octavianus Scotus. Bartholomaeus Salernitanus, Practica. In: De Renzi, Salvatore (ed.). 1856. Collectio Salernitana. Napoli: Filiatre Sebezio (repr. Napoli: D’Auria, 2001). Vol. IV, 322–406. Bernardus Provincialis, Commentum super Tabulas Salerni. In: De Renzi, Salvatore (ed.). 1859. Collectio Salernitana. Napoli: Filiatre Sebezio (repr. Napoli: D’Auria, 2001). Vol. V, 269–328. Constantinus Africanus, Liber de gradibus. In: Constantinus Africanus, Opera, 342–387. Basel 1536: Henricus Petrus. Constantinus Africanus, Pantegni. In: Opera Ysaac. Lugduni 1515: Bartholomaeus Trot. Dinus del Garbo, Expositio super quarta fen primi Canonis Avicennae. Accedit ejusdem Expositio super canones generales de virtutibus medicinarum simplicium secundi canonum Avicennae. Venetia 1514. Pedanii Dioscoridis Anarzabaei opera quae extant omnia. Lugduni 1598: Héritiers d’André Wechel. Pedanii Dioscuridis Anarzabei de materia medica libri quinque, edited by Wellmann, Max. 1907–1914. Berlin: Apud Weidmannos. Dioscorides langobardus. In: Hoffmann, Konrad and Theodor M. Auracher (eds.). 1883. “Der Longobardische Dioskorides des Marcellus Virgilius.“ Romanische Forschungen 1: 49–105; Rönsch, Hermann. 1883. „Textkritische Bemerkungen zum Longobardischen Dioskorides.“ Romanische Forschungen 1: 413–414; Stadler, Hermann. „Dioscorides longobardus (Cod. Lat. Monacensis 337).” Romanische Forschungen, 10 (1897), 181–247 and 369–466; 11 (1899), 1–121; 13 (1902), 161–243; 14 (1903), 601–637 (= Hoffmann et al., 1883–1903). Dioscorides langobardus (edition of book I). In: Mihăescu, Haralambie (ed.). 1938. Dioscorides latino, Materia medica, libro primo. Iasi: Terek. Dioscorides alphabeticus: Dioscorides alphabeticus. Colle 1478: Johannes de Medemblick. Galenus, Opera. In: Kühn, Carolus Gottlob (ed.). 1821–1833. Galeni opera omnia. Lipsiae: Carolus Cnobloch (22 vols.; the ones used in this article are quoted ad locum). Galenus, De purgantium medicamentorum facultate. In: Ehlert, Jürgen (ed.). 1959. Galeni de purgantium medicamentorum facultate. Diss. Göttingen. Galenus, Quos, quibus catharticis medicamentis et quando purgare debeant. In: Raeder, Iohannes (ed.). 1928. Oribasii collectionum medicarum reliquiae, 221–227 (Corpus Medicorum Graecorum, VI,1,1) Leipzig: Teubner.
160
Iolanda Ventura
Henricus Huntingdonensis (Henry of Huntingdon), Anglicanus ortus, edited by and translated by Winston Black. 2012. London: British Library; Toronto: Toronto University Press. (= Black 2012). Iohannes de Sancto Amando, Areolae. In: Pagel, Julius Leopold. 1893. Die Areolae des Johannes de Sancto Amando, nach Handschriften der Königlichen Bibliotheken zu Berlin und Erfurt zum ersten Male herausgegeben. Ein Beitrag zur Literaturgeschichte der Arzneimittellehre im Mittelalter. Berlin: Georg Reimer. Iohannes de Sancto Paulo, De medicinis simplicibus. In: Opera Ysaac, f. 186va–189rb. Lugduni 1515: Bartholomaeus Trot (here printed as Constantini de virtutibus simplicium medicinarum). Modern edition in: Kroemer, Georg-Heinrich (ed.). 1919. Johannes von Sancto Paulo: Liber de simplicium medicinarum virtutibus und ein anderer Salernitaner Traktat: Quae medicinae pro quibus morbis donandae sunt; nach dem Breslauer Codex herausgegeben. Diss. Leipzig. Isaac Iudaeus, De diaetis universalibus sive particularibus. In: Opera Ysaac, Lugduni 1515: Bartholomaeus Trot. Lorscher Arzneibuch. 1992. In: Stoll, Ulrich (ed.). 1992. Das ‘Lorscher Arzneibuch’. Ein medizinisches Kompendium des 8. Jahrhunderts. (Codex Bambergensis Medicinalis, 1) Text, Übersetzung und Fachglossar. (Sudhoffs Archiv, Beihefte 28) Stuttgart: Steiner Verlag. Odo Magdunensis (Macer Floridus), De viribus herbarum, edited by Ludwig Choulant. 1832. Leipzig: Leopold Voss. Petrus de Sancto Floro, Concordantiae. In: Pagel, Julius Leopold Pagel (ed.). 1896. Neue litterarische Beiträge zur mittelalterlichen Medizin, 68–71. Berlin: Georg Reimer. Platearius, Circa instans. In: Wölfel, Hans (ed.). 1939. Das Arzneidrogenbuch Circa instans in einer Fassung des 13. Jahrhunderts aus der Universitätsbibliothek Erlangen. Text und Kommentar als Beitrag zur Pflanzen- und Drogenkunde des Mittelalters. Diss. Berlin. Pseudo-Apuleius, Herbarius. In: Howald Ernst and Henri E. Sigerist (eds.). 1927. Antonii Musae de herba vettonica Liber-Pseudo-Apulei Herbarius-Anonymi de taxone Liber-Sexti Placiti Liber medicinae ex animalibus, 13–225. (Corpus Medicorum Latinorum, IV) Lipsiae-Berolini: Teubner. Salernus (magister), Tabulae. In: De Renzi, Salvatore (ed.). 1859. Collectio Salernitana. Napoli: Filiatre Sebezio (repr. Napoli 2001: D’Auria). Vol. V, 233–253.
Studies: Adamson, Peter (ed.). 2013. Interpreting Avicenna: Critical Essays. Cambridge: Cambridge University Press. Ausécache, Mireille. 2007. “Magister Salernus dans la Collectio Salernitana et au-delà.” In La Collectio Salernitana di Salvatore De Renzi, edited by Danielle Jacquart and Agostino Paravicini Bagliani, 191–226. (Edizione Nazionale “La Scuola Medica Salernitana”, 3) Firenze: SISMEL. Ayari-Lassueur, Sylvie. 2012. “Les propriétés des remèdes simples selon Avicenne (980–1037); analyse de quelques passages du Canon.” Gesnerus 69: 207–246. Bénézet, Jean-Pierre. 1999. Pharmacie et médicament en Méditerranée occidentale (XIIIe–XVIe siècles). (Sciences, techniques et civilisations du Moyen Age à l’aube des Lumières, 3) Paris: Champion. Black, Winston. 2015. Review of “Spuria Macri”: Ein Anhang zu “Macer floridus, De viribus herbarum”. Einleitung, Übersetzung, Kommentar by Ulrike Jansen. (Beiträge zur Altertumskunde, 314) Berlin: de Gruyter, 2013. Speculum 90/1: 260–263. Bottiglieri, Corinna. 2007. “Appunti per un’edizione critica del Liber pandectarum medicinae di Matteo Silvatico.” In La Scuola Medica Salernitana. Gli autori e i testi. Convegno Internazionale (Università degli Studi di Salerno, 3–5 novembre 2004), edited by Danielle Jacquart and Agostino Paravicini Bagliani, 31–58. (Edizione Nazionale “La Scuola Medica Salernitana”, 1) Firenze: SISMEL.
Classification Systems and Pharmacological Theory
161
Burnett, Charles. 2001. “The Coherence of the Arabic-Latin Translation Program in Toledo in the Twelfth Century.” Science in Context 14: 249–288. Burnett, Charles. 2005. “Arabic into Latin: The Reception of Arabic Philosophy into Western Europe.” In Cambridge Companion to Arabic Philosophy, edited by Peter Adamson and Richard Taylor, 370–404. Cambridge: Cambridge University Press. Burnett, Charles. 2009. Arabic into Latin in the Middle Ages: The Translators and Their Intellectual and Social Context. (CSS 939) Aldershot: Ashgate. Burnett, Charles. 2013. “The Legend of Constantine the African.” Micrologus. Nature, Sciences and Medieval Societies, XXI: The Medieval Legends of Philosophers and Scholars: 277–294. Chandelier, Joël. 2017. Avicenne et la médecine en Italie. Le Canon dans les universités. (Sciences, techniques et civilisations du Moyen Âge à l’aube des Lumières, 18) Paris: Honoré Champion. Chandelier, Joël. 2007. La réception du Canon d’Avicenne. Médecine arabe et milieu universitaite en Italie avant la peste noire. Diss. Paris EPHE. Chandelier, Joël, and Aurélien Robert. 2013. “Nature humaine et complexion du corps chez les médecins italiens de la fin du Moyen Âge.” Revue de synthèse 134/4 (ser. 6): 473–510. Cronier, Marie. 2006. “Quelques aspects de l’histoire du texte du De materia medica de Dioscoride: forme d’origine, remaniements et révisions à Constantinople au Xe et au XIe siècles.” In Ecdotica e ricezione dei testi medici greci. Atti del V Convegno Internazionale, Napoli, 1–2 Ottobre 2004, edited by Véronique Boudon-Millot, Antonio Garzya, Jacques Jouanna, and Amneris Roselli, 43–65. (Collectanea) Napoli: D’Auria. Cronier, Marie. 2010a. “La production des manuscrits scientifiques dans l’atelier de Michel Apostolis: l’exemple du De materia medica de Dioscoride.” In The Legacy of Bernard de Montfaucon: Three Hundred Years of Studies on Greek Handwriting. Proceedings of the Seventh International Colloquium of Greek Paleography (Madrid – Salamanca, 15–20 September 2008), edited by António Bravo García and Immaculada Pérez Martin, 462–472 and 879–884. (Bibliologia, 31; 2 vols.) Turnhout: Brepols. Cronier, Marie. 2010b. “Le Dioscoride alphabétique latin et les traductions latines du De materia medica.” In Body, Disease and Treatment in a Changing World. Latin texts and contexts in ancient and medieval medicine, edited by David Langslow and Brigitte Maire, 189–200. Lausanne: Editions BHMS. Cronier, Marie. 2013. “Dioscorides’ Excerpts in Simon of Genova’s Clavis sanationis.” In Simon of Genoa’s Medical Lexicon, edited by Barbara Zipser. (published online on www.simonofgenoa. org) Berlin/New York: de Gruyter. Debru, Armelle (ed.). 1997. Galen on Pharmacology. (Studies in Ancient Medicine, 16) (= Debru 1997a) Leiden/Boston/Köln: Brill. Debru, Armelle. 1997. “Philosophie et pharmacologie: la dynamique des substances leptomères chez Galien.” In Galen on Pharmacology, edited by Armelle Debru, 85–102. Leiden/Boston/ Köln: Brill (Studies in Ancient Medicine, 16 (= Debru 1997b)). De Vos, Paula. 2013. “The ‘Prince of Medicine’. Yūḥannā ibn Māsawayh and the Foundations of the Western Pharmaceutical Tradition.” Isis 104: 667–712. Egmond, Florike et al. (eds.). 2006. Carolus Clusius: Towards a Cultural History of a Renaissance Naturalist. Amsterdam: Edita Publishers, Royal Dutch Academy. Egmond, Florike. 2010. The World of Carolus Clusius. Natural History in the Making. London: Pickering and Chatto. Ferraces Rodriguez, Arsenio. 1999. Estudio sobre textos latinos de fitoterapia entre la Antigüedad y la Alta Edad Media. A Coruña: Universidade de Coruña. Ferraces Rodriguez, Arsenio (ed.). 2007. Tradición griega y textos médicos latinos en el periodo presalernitano. Actas del VIII Coloquio Internacional ‘Textos médicos latinos antiguos’. A Coruña: Universidade de Coruña.
162
Iolanda Ventura
Ferraces Rodriguez, Arsenio (ed.). 2009. Fito-zooterapia antigua y altomedieval: textos y doctrinas, edited by Arsenio Ferraces Rodriguez. A Coruña: Universidade de Coruña. Fischer, Klaus-Dietrich. 2011. “Antidotum, cui nomen est Acharistum.” In Between Text and Patient: The Medical Enterprise in Medieval and Early Modern Europe, edited by Florence Eliza Glaze and Brian K. Nance, 173–200. (Micrologus’ Library, 30) Firenze: SISMEL. Fischer, Klaus-Dietrich. 2013. “Die vorsalernitanischen lateinischen Galenübersetzungen.” Medicina nei secoli 25/3: 673–714. Fischer, Klaus-Dietrich. 2015. “Starch and the Alphabet of Galen.” Studia Philologica Valentina 17: 113–138. Fischer, Klaus-Dietrich. 2017. “Drugs to declare. Two works with pharmaceutical content attributed to Galen.” In Pseudo-Galenic Texts and the Formation of the Galenic Corpus, edited by Charles Burnett, Klaus-Dietrich Fischer and Caroline Petit, forthcoming. Fortuna, Stefania, and Anna Maria Urso. 2010. “Tradizione latina dell’ars medica di Galeno: La translatio antiqua e il completamento di Burgundio.” In Sulla tradizione indiretta dei testi medici greci: le traduzioni. Atti del III seminario internazionale di siena certosa di pontignano 18–19 settembre 2009, edited by Ivan Garofalo, Stefania Fortuna, Alessandro Lami and Amneris Roselli, 137–168. (Biblioteca di “Galenos”, 3) Pisa/Roma: Fabrizio Serra editore. Frati, Lodovico. 1908. “Indice dei codici latini conservati nella R. Biblioteca di Bologna” (parte I). Studi italiani di filologia classica 16: 103–432. French, Roger. 2001. Canonical Medicine. Gentile da Foligno and Scholasticism, Leiden/Boston/ Köln: Brill. Frippiat, Hélène. 2006–2007. Les traités médicaux de Jean de Saint Paul (XIIe siècle). Université Catholique de Louvain, MA thesis. [= DMS] Green, Monica. 1994. “The Recreation of Pantegni, Practica, book VIII.” In Constantine the African and Ali ibn al Abbas al-Magusi. The “Pantegni” and its Related Texts, edited by Charles S. F. Burnett and Danielle Jacquart, 121–160. (Studies in Ancient Medecine, 10) Leiden/Köln/New York: Brill. Gutas, Dimitri. 2014. Orientations of Avicenna’s Philosophy: Essays on Life, Method, Heritage. (CSS, 1050) Aldershot: Ashgate. Hardy, Frederick G., and Laurence M. V. Totelin. 2016. Ancient Botany. London: Routledge. Heuken, Ulrike. 1990. Der achte, neunte und zehnte Abschnitt des Antidotarium Mesue in der Druckfassung Venedig 1561 (Trocisci, Pulver, Suffuf, Pillen). (Quellen und Studien zur Geschichte der Pharmazie, 58) Stuttgart: Deutscher Apotheker-Verlag. Jacquart, Danielle. 1990. “Principales étapes dans la transmission des textes de médecine (XIe – XIVe siècle).” In Rencontres de cultures dans la philosophie médiévale, edited by Marta Fattori and Jacqueline Hamesse, 251–271. (Textes, Etudes, Congrès, 1) Louvain-la-Neuve and Cassino: Institut d’Etudes Médiévales de l’Université Catholique de Louvain. Jacquart, Danielle. 1992. “Les traductiones médicales de Gérard de Crémone.” In Gerardo da Cremona, edited by Pierluigi Pizzamiglio, 57–70. (Annali della Biblioteca Statale e Libreria Civica di Cremona, XLI) Cremona: Libreria del Convegno. Jacquart, Danielle. 1994. “Le sens donné par Constantin l’Africain à son oeuvre: les chapitres introductifs en arabe et en latin.” In Constantine the African and Ali ibn al Abbas al-Magusi. The “Pantegni” and its Related Texts, edited by Charles S. F. Burnett and Danielle Jacquart, 71–89. (Studies in Ancient Medecine, 10) Leiden/Köln/New York: Brill. Jacquart, Danielle. 1997. La science médicale occidentale entre deux renaissances (XIIe s.–XVe s.). (CSS 567) Aldershot: Ashgate. Jacquart, Danielle. 1998. La médécine médiévale dans le cadre parisien (XIVe–XVe siècle). Paris: Fayard.
Classification Systems and Pharmacological Theory
163
Jacquart, Danielle. 2002. “Lectures universitaires du Canon d’Avicenne.” In Avicenna and His Heritage. Acts of the International Colloquium, Leuven – Louvain-la-Neuve, September 8 – September 11, 1999, edited by Jules Janssens and Daniel de Smet, 313–324. Leuven: Peeters. Jacquart, Danielle. 2008. “Islamic Pharmacology in the Middle Ages: Theories and Substances.” European Review 16/2: 219–227. Jacquart, Danielle, and Françoise Micheau. 1996. La médecine arabe et l’Occident médiéval. Paris: Edition Maisonneuve et Larose. Jacques, Jean-Marie. 1997. “La méthode de Galien pharmacologue dans les deux traités sur les médicaments composés (Περὶ συνθέσεως φαρμάκων).” In Galen on Pharmacology, edited by Armelle Debru, 103–129. (Studies in Ancient Medicine, 16) Leiden/Boston/Köln: Brill. Jansen, Ulrike. 2013. “Spuria Macri”: Ein Anhang zu “Macer floridus, De viribus herbarum”. Einleitung, Übersetzung, Kommentar. (Beiträge zur Altertumskunde, 314) Berlin: de Gruyter. Keil, Gundolf. 2002. “Standardwerke mittelalterlicher Drogenkunde und ihre Repräsentation in der Enzyklopädik.” In Die Enzyklopädie im Wandel vom Hochmittelalter bis zur frühen Neuzeit. Akten des Kolloquiums des Projekts D im Sonderforschungsbereich 231 (29.11.–1.12.1996), edited by Christel Meier, 343–389. (Munstersche Mittelalterschriften, 78) München: Fink Verlag. Klimaschewski-Bock, Ingrid. 1987. Die “Distinctio Sexta” des Antidotarium des Pseudo-Mesue in der Druckfassung Venedig 1561 (Sirupe und Robub). (Quellen und Studien zur Geschichte der Pharmazie 40) Stuttgart: Deutscher Apotheker-Verlag. Kristeller, Paul-Oskar. 1986. Studi sulla Scuola Medica Salernitana. Napoli: Istituto Italiano per gli Studi Filosofici. Kroemer, Georg-Heinrich. 1919. Johannes von Sancto Paulo: Liber de simplicium medicinarum virtutibus und ein anderer Salernitaner Traktat: Quae medicinae pro quibus morbis donandae sunt; nach dem Breslauer Codex herausgegeben. Diss. Leipzig. Langermann, Y. Tzivi. 2003. “Another Andalusian Revolt? Ibn Rushd’s Critique of Al-Kindi’s Pharmacological Computus.” In The Enterprise of Science in Islam. New Perspectives, edited by Jan P. Hogendijk and Abdelhamid I. Sabra, 351–372. Cambridge, Mass.: The MIT Press. Lieberknecht, Sieglinde. 1995. Die Canones des Pseudo-Mesue. Eine mittelalterliche PurgantienLehre. (Quellen und Studien zur Geschichte der Pharmazie, 71) Stuttgart: Deutscher Apotheker-Verlag. MacKinney, Loren C. 1936. “‘Dynamidia’ in Medieval Medical Literature.” Isis 24/2: 400–414. Maire, Brigitte. 2007. Se soigner par les plantes: les “Remèdes” de Gargile Martial, avec un avant-propos de Kurt Hostettmann et un dossier iconographique par Michel Fuchs. (Sources en Perspective) Lausanne: Editions BHMS. Maire, Brigitte. 2009.“Gargilio Marziale.” In Dizionario della scienza e della tecnica in Grecia e Roma: autori e testi, Realien, saperi alle radici della cultura europea, edited by Paola Radici Colace, 613. Pisa/Roma: Fabrizio Serra. Maire, Brigitte (ed.). 2014.‘Greek’ and ‘Roman’ in Latin Medical Texts. Studies in Cultural Change and Exchange in Ancient Medicine. (Studies in Ancient Medicine, 42) Leiden/Boston: Brill. Martín Ferreira, Ana Isabel and Alejandro García González. 2010. “La tradición manuscrita del Breviarium de Johannes de Sancto Paulo.” Exemplaria Classica 14: 227–248. Martín Ferreira. 2013. “Génesis, structura y pervivencia de un compendio médico salernitano: el Breviarium de Johannes de Sancto Paulo.” In La compilación del saber en la Edad Media/La compilation du savoir au Moyen Age/The Compilation of Knowledge in the Middle Ages, edited by Maria José Muñoz, Patricia Cañizares and Cristina Martin, 367–381. (Textes et Études du Moyen Age, 69) Porto: Fédération Internationale des Instituts d’Études Médiévales. Mazzini, Innocenzo. 1992. “Dynamidia Hippocratis: esempio di problematiche ecdotiche tardoantiche.” In Tradizione e ecdotica dei testi medici tardoantichi e bizantini. Atti del
164
Iolanda Ventura
Convegno Internazionale, Anacapri, 29–31 ottobre 1990, edited by Antonio Garzya, 257–270. (Collectanea) Napoli: D’Auria. McVaugh, Michael R. 2011. “Medicine and Arts in Thirteenth-Century Paris.” In Crossing Boundaries at Medieval Universities, edited by Spencer E. Young, 189–211. (Education and Society in the Middle Ages and Renaissance, 36) Leiden/Boston: Brill. Newton, Francis. 1994. “Constantine the African and Monte Cassino: New Elements and the Text of the Isagoge.” In Constantine the African and Ali ibn al Abbas al-Magusi. The “Pantegni” and its Related Texts, edited by Charles S. F. Burnett and Danielle Jacquart, 16–47. (Studies in Ancient Medecine, 10) Leiden, Köln, New York: Brill. Newton, Francis. 1999. The Scriptorium and Library at Monte Cassino, 1058–1109. (Cambridge Studies in Palaeography and Codicology, 7) Cambridge: Cambridge University Press. Newton, Francis. 2011. “Arabic Medicine and Other Arabic Cultural Influences in Southern Italy in the Time of Constantinus Africanus (saec. XI²).” In Between Text and Patient: The Medical Enterprise in Medieval and Early Modern Europe, edited by Florence Eliza Glaze and Brian K. Nance, 25–55. (Micrologus’ Library, 30) Firenze: SISMEL. Nicoud, Marilyn. 2007. Les Regimes de santé au Moyen Age: naissance et diffusion d’une écriture médicale (XIIIe–XVe siècle). (Bibliothèque des écoles françaises d’Athènes et Rome, 333; 2 vols) Roma: Ecole Française de Rome Ogilvie, Brian W. 2006. The Science of Describing: Natural History in Renaissance Europe. Chicago: University of Chicago Press. Paavilainen, Helena. 2009. Medieval Pharmacoterapy: Continuity and Change. Case Studies from Ibn Sina and some of his Late Medieval Commentators. (Studies in Ancient Medicine, 38) Leiden: Brill. Paravicini Bagliani, Agostino. 1987. Medicina e scienze della natura alla corte dei Papi nel Duecento. (Biblioteca di Medioevo Latino, 4) Spoleto: CISAM. Petit, Caroline. 2013. “La tradition latine du Traité des Simples de Galien.” Medicina nei secoli 25/3: 1063–1090. Pormann, Peter E. 2001. “The Formation of Arabic Pharmacology between Tradition and Innovation.” Annals of Science, 68/4: 493–515. Pormann, Peter E, and Emilie Savage Smith. 2007. Medieval Islamic Medicine. Edinburgh: Edinburgh University Press. Santamaría Hernández, María Teresa (ed.). 2012. Textos médicos grecolatinos antiguos y medievales: estudios sobre composición y fuentes. Cuenca: Universidad de Castilla-La-Mancha. Scarborough, John. 2010. Pharmacy and Drug Lore in Antiquity. Greece, Rome, Byzantium. (CSS, 904) Farnham: Ashgate. Schalick, Walton O. 1997. Add One Part Pharmacy to One Part Surgery and One Part Medicine : Jean de Saint-Amand and the Development of medical Pharmacology in Thirteenth-Century Paris. (UMI Microform 9730782) Diss. Baltimore. Schmitz, Rudolf. 1998. Geschichte der Pharmazie. Bd. I: Von den Anfängen bis zum Ausgang des Mittelalters. Eschborn: GOVI-Verlag. Schmitz, Rudolf. 2005. Geschichte der Pharmazie. Bd. II: Von der Frühen Neuzeit bis zur Gegenwart, edited by Christoph Friedrich and Wolf-Dieter Müller-Jahncke. Eschborn: GOVI-Verlag. Schulze, Christian. 2003. Die pharmakologische Fachliteratur des Antike. Eine Einführung. (Beihefte zum Göttinger Forum für Altertumswissenschaft, 10) Göttingen: Edition Ruprecht. Siraisi, Nancy G. 1981. Taddeo Alderotti and His Pupils. Two Generations of Italian Medical Learning. Princeton: Princeton University Press. Siraisi, Nancy G. 1987. Avicenna in Renaissance Italy. The Canon and Medical Teaching in Italian Universities after 1500. Princeton: Princeton University Press. Siraisi, Nancy G. 2001. Medicine and Italian Universities, 1250–1600. (Education and Society in the Middle Ages and the Renaissance, 12) Leiden/Boston/Köln: Brill.
Classification Systems and Pharmacological Theory
165
Touwaide, Alain. 1997. “La thérapeutique médicamenteuse de Dioscoride à Galien: du pharmacocentrisme au médico-centrisme.” In Galen on Pharmacology, edited by Armelle Debru, 255–282. (Studies in Ancient Medicine, 16) Leiden/Boston/Köln: Brill. Touwaide, Alain. 2006. “The Development of Paleologan Renaissance: An Analysis based on Dioscorides’ De materia medica.” In Philosophie et sciences à Byzance de 1204 à 1453. Les textes, les doctrines et leur transmission. Actes de la Table Ronde organisée au XXe Congrès International d’Études Byzantines (Paris, 2001), edited by Michel Cacouros and Marie-Hélène Congourdeau, 189–22. (Orientalia Lovaniensia. Analecta, 146) Leuven/Paris/Dudley: Peeters. Touwaide, Alain. 2012. “Quid pro Quo: Revisiting the Practice of Substitution in Ancient Pharmacy.” In Herbs and Healers from the Ancient Mediterranean through the Medieval West. Essays in Honor of John M. Riddle, edited by Anne Van Arsdall and Timothy Graham, 19–61. (Medicine in the Medieval Mediterranean, 1) Farnham: Ashgate. Ullmann, Manfred. 1970. Die Medizin im Islam. (Handbuch der Orientalistik; Ergänzungsband 6/1) Leiden/Köln: Brill. Van’t Land, Karine. 2014. “Long Life, Natural Death. The Learned Ideal of Dying in Late Medieval Commentaries on Avicenna’s Canon.” Early Science and Medicine 19: 558–583. Veit, Raphaela. 2003a. Das Buch der Fieber des Isaac Israeli und seine Bedeutung im lateinischen Westen: ein Beitrag zur Rezeption arabischer Wissenschaft im Abendland. (Sudhoffs Archiv, Beihefte 51) Stuttgart: Steiner Verlag. Veit, Raphaela. 2003b. “Quellenkundliches zu Constantinus Africanus.” Deutsches Archiv für Erforschung des Mittelalters 59/1: 121–152. Veit, Raphaela. 2006. “Al-Majûsî’s Kitâb al-Malakî and its Latin Translation ascribed to Constantine the African: The Reconstruction of Pantegni, Practica, Liber III.” Arabic Science and Philosophy 16: 133–168. Veit, Raphaela, and Lola Ferre. 2009. “The Textual Traditions of Isaac Israeli’s Book on Fevers in Arabic, Latin, Hebrew, and Spanish.” Aleph: Historical Studies in Science and Judaism 9, no. 2: 309–334. Veit, Raphaela. 2011. “Greek Roots, Arab Authoring, Latin Overlay: Reflections on the Sources for Avicenna’s Canon.” In Vehicles of Transmission, Translation, and Transformation in Medieval Textual Culture, edited by Robert Wisnovsky et al., 353–369. (Cursor mundi 4) Turnhout: Brepols. Veit, Raphaela. 2015a. “Les Diètes universelles et particulières d’Isaac Israëli: Traduction et reception dans le monde latin.” Revue d’histoire des textes, n.s. 10: 229–249. Veit, Raphaela. 2015b. “Isaac Israeli: His Treatise on Urine (De Urinis) and Its Reception in the Latin World.” In Isaac Israeli – The Philosopher Physician, edited by Kenneth E. Collins, Samuel Kottek, and Helena Paavilainen, 77–113. Jerusalem: Muriel & Phillip Berman Medical Library. Ventura, Iolanda. 2006. “Il De materia medica di Dioscoride nel Medioevo: mediazione araba e ricezione occidentale.” In Miscellanea Mediaevalia 34. Arabisches Wissen und lateinisches Mittelalter. Akten der 34. Kölner Mediävistentagung (6–9 Sept. 2004), edited by Andreas Speer and Lydia Wegener, 317–339. (Miscellanea Mediaevalia, 34) Berlin/New York: de Gruyter. Ventura, Iolanda. 2007. “Un manuale di farmacologia medievale ed i suoi lettori: il Circa instans, la sua diffusione, la sua ricezione dal XIII al XV secolo.” In La Scuola Medica Salernitana. Gli autori e i testi. Convegno Internazionale (Università degli Studi di Salerno, 3–5 novembre 2004), edited by Danielle Jacquart and Agostino Paravicini Bagliani, 465–533. (Edizione Nazionale “La Scuola Medica Salernitana”, 1) Firenze: SISMEL. Ventura, Iolanda. 2011. “Un’opera araba e la sua traduzione: il Liber fiduciae de simplicibus medicinis di Stefano di Saragozza.” In Il mare e la medicina. Atti del XLVII Convegno Nazionale della Società Italiana di Storia della Medicina, Salerno, 3–5 Settembre 2009, 83–108. Salerno: Laveglia Editore.
166
Iolanda Ventura
Ventura, Iolanda. 2013a. “Il Dioscorides alphabeticus : un esempio di farmacopea arabo-latina?” In Circolazione dei saperi nel Mediterraneo : filosofia e scienze, secoli IX–XVII. / Circulation des savoirs autour de la Méditerranée : philosophie et sciences, IXe–XVIIe siècle, edited by Ahmad Hasnawi and Graziella Federici Vescovini, 159–166. Firenze: Cadmo Edizioni. Ventura, Iolanda. 2013b. “Sulla diffusione del Circa instans nei manoscritti e nelle biblioteche del Tardo Medioevo: ricezione e lettura di un’opera medica.” In La produzione tecnica e scientifica nel Medioevo: libro e documento tra scuole e professioni, edited by Giuseppe de Gregorio and Maria Galante, 465–549. (Studi e Ricerche, 5) Spoleto: Fondazione Centro Italiano di Studi sull’Alto Medioevo. Ventura, Iolanda. 2015. “Il Circa instans attribuito a Platearius: trasmissione manoscritta, redazioni, criteri di costruzione di un’edizione critica.” Revue d’Histoire des Textes, n.s. 10: 249–362. Ventura, Iolanda. 2016. “Medieval Pharmacy and Arabic Heritage: the Salernitan Collection ‘Circa instans’.” Micrologus XXIV: The Impact of Arabic Sources on Divination and the Practical Sciences in Europe and Asia (Erlangen, 21–23 janvier 2014): 339–401. Vogt, Sabine. 2008. “Drugs and Pharmacology.” In The Cambridge Companion to Galen, edited by Robert J. Hankinson, 304–322. Cambridge: Cambridge University Press. Voigts, Linda E. 2008. “Plants and Planets: Linking the Vegetable with the Celestial in Late Medieval Texts.” In Health and Healing from the Medieval Garden, edited by Peter Dendle and Alain Touwaide, 29–46. Woodbridge: The Boydell Press. Wack, Mary Frances. 1994. “Ali ibn al-Abbas al-Magusi and Constantine on Love, and the Evolution of the Practica Pantegni.” In Constantine the African and Ali ibn al Abbas al-Magusi. The “Pantegni” and its Related Texts, edited by Charles S. F. Burnett and Danielle Jacquart, 161–202. (Studies in Ancient Medecine, 10) Leiden/Köln/New York: Brill. Wallis, Faith. 2007. “The Articella Commentaries of Bartholomaeus of Salerno.” In La Scuola Medica Salernitana. Gli autori e i testi. Convegno Internazionale (Università degli Studi di Salerno, 3–5 novembre 2004), edited by Danielle Jacquart and Agostino Paravicini Bagliani, 125–164. (Edizione Nazionale “La Scuola Medica Salernitana”, 1) Firenze: SISMEL. Wallis, Faith. 2011. “Why Was the Aphorisms of Hippocrates Retranslated in the Eleventh Century?” In Vehicles of Transmission, Translation, and Transformation in Medieval Textual Culture, edited by Robert Wisnovsky, Faith Wallis, Jamie C. Fumo and Carlos Fraenkel, 173–193. (Cursor Mundi, 4) Turnhout: Brepols. Wallis, Faith. 2012. “The Ghost in the Articella: A Twelfth-Century Commentary on the Constantinian Liber Graduum.” In Herbs and Healers from the Ancient Mediterranean through the Medieval West. Essays in Honor of John M. Riddle, edited by Anne Van Arsdall and Timothy Graham, 107–151. (Medicine in the Medieval Mediterranean, 1) Farnham: Ashgate. Weill-Parot, Nicolas. 2013. Points aveugles de la nature. La rationalité scientifique médiévale face à l’occulte, l’attraction magnétique et l’horreur du vide. (Histoire, 120) Paris: Les Belles Lettres.
Abbreviations DMM = De materia medica DMS = De medicinis simplicibus DPM = De probanda medicina DSF = De simplicium medicamentorum temperamentis et facultatibus DSM = De simplici medicina LC = Liber canonis LDG = Liber de gradibus
MF = Macer floridus PAH= Herbarius attributed to Apuleius Platonicus PP = Pantegni Practica SAP = De simplicibus ad Paternianum TS = Tabulae Salerni
Tanja Pommerening
Classification in Ancient Egyptian Medical Formulae and its Role in Re-Discovering Comprehensive and Specific Concepts of Drugs and Effects Summary: This essay is intended to embark upon a path of method, in which the classifications contained in Ancient Egyptian formulae, or prescriptions, are more closely characterized; their potential for recovering comprehensive and specific concepts of drugs and effects will be discussed. In medical papyri there are four different kinds of classification, depending on focus: 1) classifiers at the end of individual words – a general phenomenon in the Ancient Egyptian language; 2) the increased occurrence of nominal compounds, especially in relation to the names of drugs and illnesses, compounds that are unavoidably connected to the creation of subordinate level categories; 3) systematic list-type sequences of the formula content and the correlations appearing therein; 4) the configuration of chosen formulae within a papyrus. These different means of classification will be more closely determined with respect to their specificities (i.e. type of source, medium, the circle of users and the duration of use), motivations, types (exclusive or flexible class limits, hierarchizations) and criteria. It can be shown that the classification types 3) and 4) are directly connected to concepts of drugs and effects, but that ascertaining the individual specificities of 1) and 2) in relation to formula texts can reveal the concepts of drugs and effects behind them.
1 Introduction This paper refers to a limited text corpus of Ancient Egyptian medicine, the medical prescriptions. The author has entered some 2000 prescriptions in a database, with their respective positions in papyrus documents, with all the drug names and their sequence within the individual prescriptions and with indications. The types of classification occurring in the formulae can be followed, with their changing arrangements, throughout the entire time period being considered.1
1 For this contribution, prescriptions written on papyri for the period from about 1850 until about 300 BC have been used. The abbreviations follow Westendorf 1999: 808–831; on the history of prescriptions, see Pommerening 2012. DOI 10.1515/9783110538779-006
168
Tanja Pommerening
The aim of the paper is to document what informative value the classification phenomena, which are individually discussed, may have with regard to recovering concepts of drugs2 and effects. In addition, comprehensive systems will be contrasted with specific ones; in this way, the dependencies of the classificational limits and criteria on the context will be more clearly emphasized. This essay is based on the following methodic and definitory considerations with regard to concept and classification:3 A concept is here understood as culturally and cognitively formed and organized conglomerates of properties or (mentally anchored) knowledge components.4 They may be transmitted by various media (images, texts, objects) subject to different processes of formation. The media may express different properties or knowledge components, depending on context (discourse, type of source, space, time, individual). Concepts in historical sources are hardly ever transmitted fully formulated and without being subject to certain changes. For the purpose of reconstruction, so-called meta-concepts are constructed as an aid, superordinated to the existing concepts; all properties and knowledge components of the object, issue, or type of action being considered are united in this. Conceptualization can be seen as a device to divide semantic space into subdomains. Important for this article is the observation that the mechanisms of forming concepts and separating them from others resemble the mechanisms used for classification. The model of prototypes can be applied to both. I.e. a concept and a class can both be defined by linking it to and delimiting it from other concepts / classes in a network of semantic and prototypical properties, some of which it shares with other concepts / classes and some of which it does not. In a model of prototype-based definitions, the borderlines between concepts / classes can be notoriously fuzzy. Often, the specific properties involved for assigning a certain object to a certain concept / class depends on the specific aspect from which it is looked at. For instance, the use of an object for certain healing purposes may lead to subsuming it under a specific concept / class that is relevant for that purpose. In this way classification highlights certain properties of an object. Prescriptions and formulae do not serve to transmit concepts / classes of drugs or effects. They only allow indirect access to the underlying concepts through their content, their structures, and their arrangement within the greater context of a papyrus.
2 All the ingredients (simplicia) of a particular medicine (compositum) are subsumed under this. 3 Cf. the introduction to this volume. 4 Lakoff 1987; Blank 2001; Murphy 2004; Schwarz 2008.
Classification in Ancient Egyptian Medical Formulae
169
2 Classification phenomena in Ancient Egyptian prescriptions Phenomena of classification can be observed at various levels: 1. in the pictorial writing system through classifiers that occur at the end of individual words; 2. in nominal compounds for the creation of subordinate-level categories; 3. in the internal structure of the prescription through principles of selection and arrangement of individual ingredients; 4. in the selection and arrangement of the entire prescription within the papyrus as a whole, according to the principles of ordering that determine the composition of prescriptions within a manuscript. In addition, there are purely linguistic criteria of classification associated with the gender system of Egyptian and aspects such as naming patterns at the level of the root, collective plurals and gender-specific properties (e.g. the use of the feminine gender to express diminutives). These aspects will not be discussed in the present paper, although they could have some relevance to concepts of drugs or effects. As to naming, for example, one has to ask whether the njꜢjꜢ-/njwjw-plant and the njꜢ-symptom might derive from the same root nj ‘to rebuff, to ‘drive away’5, and / or whether the njꜢjꜢ-plant was used against the njꜢ-symptom because of sound equivalence. In both cases we get an impression of underlying healing concepts. We will take a look at the different systems 1) to 4) mentioned above.
2.1 Classifiers
ḳꜢꜢ
n
šnd.t
ḳꜢꜢ-parts
of
thorn acacia
Drug 1
ḏꜢr.t
bnr
carob
date
Drug 2
Drug 3
Table 1: Drugs in the prescription of Papyrus Ebers XCIII.6–8 (Eb 783)
5 Wb II, 201, 4–6; Allen 1984: 548.
170
Tanja Pommerening
Classifiers (once called ‘determinatives’) are a sign class without any phonetic value in the Egyptian writing system; they always follow a lexeme and function as an aid to reading. They assign a semantic category to the particular lexeme (according to Goldwasser 2002: 15). By being determed with the pictogram involved, each and every element carrying this feature is connected with the category symbolized with the sign. This marking was needed because the writing did not show vowels and empty spaces so as to distinguish different expressions and to delimit words from each other. By way of illustration, examples of the drugs in the prescription from Papyrus Ebers XCIII.6–8 (Eb 783) are listed here (Tab. 1): the classifier after the ḳꜢꜢ ( ) indicates small parts (schematic relation, dimension). The tree ( ) behind the ‘thorny acacia’ assigns the word to the group of trees (taxonomic relation). The pictogramme after the ‘carob fruit’ ( ) indicates something long and oval (schematic relation, shape), and that after the ‘date’ ( ) refers to a bulk product that was measured (meronymic relation, portion-mass). The classifiers found after natural objects are thus bound to the previous lexemes in differing ways. Goldwasser (2002: 34) and Kammerzell (2011: 153) were able, within the framework of a typology project, to establish the following relations: taxonomic relations (superordinate – basic level – subordinate level); schematic relations (dimension, shape, consistency, colour, material); meronymic relations (component – integral object, member – collection, portion – mass, stuff – object, feature – activity, place – area); and, in particular, thematic relations for verbs and adjectives.
2.1.1 Specificity Classifiers can be shown to have existed from the beginning of writing (Kahl 1994: 105–115). In the Middle Kingdom, their use had become systematic, and we have for the most part a defined group of the usual symbols. The number of signs in use decreases in the New Kingdom, which has more abstract categories for superordinate concepts (Goldwasser 2002). Research by Goldwasser (2002) and Lincke (2011: 116–118) has shown that lexemes are mostly, but not always, classified in a uniform manner. This can be demonstrated in the area of the prescriptions, too. Proper investigation is still needed of the conditions for choosing a particular classifier (grammatical and status phenomena, diachronic development, author, school, scribe, type of text, medium, type of writing, text surroundings, nuances of meaning, polysemy, etc.).6
6 For the Pyramid Texts, there is the work by Lincke (2011). She was unable to show any dependency of the choice of classifier on the content of the surrounding text in the case of variable classifiers following lexemes. She also excluded polysemy. But she recognized context dependencies in many areas from the earliest times, and pleads for this research to be continued (Lincke 2011: 117), which can only be underscored here.
Classification in Ancient Egyptian Medical Formulae
171
Variability may also quite possibly be due to the ‘fuzzy edges’ of the classification system. With regard to the medical texts, the observations made by Westendorf (GdM VIII: 5 f.) are relevant. He established that words of daily life are given special determiners in medical texts. He names specifically the classifier for body parts F51 ( ) and the classifier for disease symptoms Aa2/Aa3 ( / ). This already makes clear that classification by means of determinatives (classifiers) in the hieroglyphic and the hieratic writing systems represent a superordinated phenomenon, but that the choice of a classifier can indicate a system dependent on discourse, which in turn can be connected with the conceptualization of words. The system seems to be open and variable. If some of this variation can be exclusively assigned to contexts in medicine, the determinatives provide a key to obtaining specific medical concepts. In this case, this would be the recovery of a superordinated concept of body and one of illness, at least with reference to the medical texts. The words with their determiners need to be collected and their delimitation from otherwise classified words investigated – a field for more detailed and comprehensive studies.
2.1.2 Motivation The motivation for using classifiers in the writing system is the simplification of reading (Goldwasser 2002: 13; Kammerzell 2011: 156).
2.1.3 Classification type and criteria As Goldwasser (2002: 13–38; cf. also Kammerzell 2011: 154) was able to show, the classification in the area of determination can be described very well with the theory of prototype semantics. An object is sorted according to its similarity with one representative that is typical of the group, rather than according to presupposed, fixed, and indispensible properties. For instance, a living creature classified as a bird need not necessarily be able to fly, but must be similar to the prototype of the community in many other ways (frequently cited example: the penguin, cf. Murphy 2004: 31 f.). The more difficult it is to assign a word to a definite class, the more often different determinatives appear after the word. Here we can see the fuzzy edges of a category in its broader context and the flexibility of the system as described in prototype models.
172
Tanja Pommerening
The prototype approach in turn is based on experimental investigations that served to understand concepts as mental representations, which correspond to categories (Murphy 2004). The concepts of classification that can be recovered through the Ancient Egyptian system of writing would presumably follow the same sort of organization, if we assume that this theory is right. The classifier comprises the entire category and thus the entire concept. Table 2 is a list of those classifiers whose occurrence is attested with more than five different drug names. The arrangement approximately reflects frequency of occurrence in different words, starting with the most common classifiers in prescriptions. 7
Sign
Gardiner (1957)
Piktogramm
Usage
Relation
highlighted Number feature in recipe
N33
sand grain
after minerals, pulverized components and fruits
schematic relation: dimension, shape
small grains (morphology)
213a)
M2
herb
after plant-like objects
taxonomic herbs relation: super- (morphology) ordinate level
155b)
W24, W22
bulbous/ round objects, pitcher/ jug
after vessels; liquid, doughy or ointmentlike objects (solvents, suspensions, emulsions)
schematic relation: shape (vessel) consistency (drug)
F27
skin of animal with a tail
after mammals taxonomic mammal-like and quadrupe- relation: super- (morphology) des ordinate level
36d)
M1
tree
after trees
taxonomic rela- tree-like objects tion: superordi- (morphology) nate level
28e)
H8
egg
after oval objects
schematic rela- oval objects tion: shape (morphology)
27f)
7
liquid, viscous or salve-like products (morphology)
76c)
7 In addition, there are determinatives that show the animal quite concretely. These and the names belonging to them have been left out here, but they can be found in GdM VI, ch. 7.
Sign
Classification in Ancient Egyptian Medical Formulae
Relation
173
Gardiner (1957)
Piktogramm
Usage
highlighted Number feature in recipe
G38
bird
after things that taxonomic bird-like objects can fly relation: super- (morphology) ordinate level
21g)
Aa2, Aa3
pustule?
after disgusting schematic things relation: shape, consistency, colour
morbid bodily changes (morphology, smell?)
21h)
O39
stone
after hard objects
schematic relation: consistency
hard objects (morphology)
21i)
F51
piece of meat
after body parts schematic and organs relation: consistency
body part (morphology)
20j)
X4
bread
After doughy schematic drugs and pastry relation: consistency
bread-like objects (morphology)
16k)
M3
branch
after wood-like schematic rela- branch-like objects tion: material objects (morphology)
13l)
K1
fish
after fish-like objects
taxonomic rela- fish-like objects tion: superordi- (morphology) nate level
N35A
lines of water
after liquids
schematic relation: consistency
U10
bushel for things after products to stored in be measured granaries
liquid things (morphology)
meronymic objects conrelation: nected to grain portion – mass farming or taxonomic relation: superordinate level
13m)
118
109
89
8 iꜤj.t, pꜢw-Ꜥ, pꜢdw, mwj.t, mhwi, mstꜢ, ḥbb.t, ḥrw.t, ḥsꜢ (combined with W22), šꜢmj.t, tꜢḥ.t. 9 inj.t (‘part of date’): Bln 2, in all other cases N33; ꜤmꜤꜤ (‘kernel of grain’): Bln, H, also N33, and M2 attested; bi (‘groats’): Eb, also H8 and N33 attested; bnr (‘dates’): mostly U10, sometimes N33; bšꜢ (‘barley’) always U10; bd.t (‘emmer’) always U10; mimj (‘a cereal’): mostly N33; ns.tj (‘part of bšꜢ’) always U10; sw.t (‘kind of wheat’): U10, also H8, N33; skj (‘flour’): Bln 45, in all other cases N33.
174 Sign
Tanja Pommerening
Gardiner (1957)
Piktogramm
Usage
Relation
highlighted Number feature in recipe
I14
wriggling animals
after amphibia and reptiles
taxonomic relation: superordinate level
amphibia and reptiles (morphology)
N23
irrigation canal after geologically defined objects (inundation)
thematic relation to mud / relation: source, inundation water location (morphology)
7n)
610
Table 2: The most frequent classifiers following the names of drugs 10
a)–n): See Appendix. If we look at the criteria for distinguishing categories, there are the following observations: 1. Nearly all the criteria used for classification in the field of drug terminology reflect morphological properties of the objects to be classified. The only exeption is the use of the bushel U10 ( ) as a classifier, which relates its objects to a process of administration (measuring) 11. As the sign U10 is also used as a logogram for the word it ‘barley’/‘grain’ it is quite possible that the relation here is based on morphology, as well. 2. The relation between object and class is often a taxonomic one (or based on schemes). The only exeption here is sign N23 which occurs usually behind objects in connection with inundations; thus it implies a thematic relation. 3. We can assume that these criteria overlapped with classificational criteria that were common at their time. This perspective, with its fixation on forms and observation of the animal and plant world (morphology integrates the whole habitus in this interpretation) basically occurs in the concepts of drugs and effects already recognized (the theory of signatures) (Pommerening 2017). The criteria can be roughly divided into the morphological and behavioural properties of plants (tree, herb, branch), animals (quadruped, bird, fish, reptiles/ amphibia), minerals/small fruits/seeds, liquid, doughy or ointment-like products (including oils and resins), pastry products (X4), grains (U10) and Body parts (F51). Another remarkable criterion is the shape-property of ovality (H8) and the association with Disease Symptoms (Aa2, Aa3). The division into drugs derived from plants or from animals, liquids/oils/salves and minerals as it results here from the classifiers, is relevant in later pharmacological
10 iwtn (‘earth’), itrw (‘nile water’), ꜤmꜤ.t (‘mud’), pns (‘earth’), ṯꜢ (‘lump [of earth]’), ḏꜤ.t (?). 11 Cf. Footnote 9.
Classification in Ancient Egyptian Medical Formulae
175
teaching texts (Dioskurides12, Galen, and cf. Ventura in this volume). Of particular interest are all those drugs that deviate in their classification from what we would expect today. Thus, one finds the classifiers N33 ( ) and W24 ( ) and not M1 ( ) or M2 ( ) after certain plant products imported from foreign countries, such as Ꜥntjw ( )13 or šnj-fruits ( )14 , which indicates that the drug is specified in its form as a small grain object, and not according to its botanical origins. This emphasizes yet again the primary categorization according to forms (consisting of small parts). N33 ( ) is the most frequent classifier. It occurs after minerals, pulverized components, and fruits. It seems to be the default determiner that is generally compatible with any ingredient. This is also obvious from its use after compound nouns, for instance, after er ẖsꜢj.t-balm from Nubia (ẖsꜢj.t n.t mḏꜢ15 ) or faeces from a crocodile (ḥs msḥ16 ). Classifier selection is not uniquely determined. For instance, the ‘tadpole’ (‘polliwog’: ḥfrn) can be classified as fish or worm. Many drugs change their classifiers between the herb (M2 ) pictogram and the Small Grains (N33 ) pictogram. If there are no diachronic changes or synchronic polysemies, one can say that several properties are treated as hierarchically equal in this context. 4. In some rare instances, we find two classifiers in a sequence. Thus, Ꜥrw-tree can be classified by M1 ( ) and N33 ( ), or by Aa2 ( ) and M1 ( ).17 What needs to be investigated in more detail is the comparison of the use of the same classifiers and class limits in non-medical texts. In doing so, as already mentioned, the media (papyrus, ostraca, tomb wall, temple wall) and the writing system (hieratic / hieroglyphic) must be taken into consideration.
12 De materia medica, 5 vols. Here is the translation for the introduction to Book V in the edition of Julius Berendes 1902: ‘Nachdem ich in den vorstehenden vier Büchern, theuerster Areios, von den Gewürzen, Oelen, Salben, und den Bäumen, sowie von ihren Früchten und Thränen, ebenso auch von den Thieren, dem Honig, der Milch, den Fetten, den Getreidearten und Gemüsen und überhaupt von den Wurzeln, Pflanzen, Säften und Samen geredet habe, werde ich in diesem als dem letzten des ganzen Werkes von den Weinen und den sogen. Metallen handeln, beginnend mit dem Vortrage über den Weinstock.’ (After I have talked, my dearest Areios, in the previous four books about spices, oils, salves, and the trees, as well as their fruits and exudences, and of animals, of honey, of milk, of fats, kinds of grain and vegetables, indeed of roots, plants, saps and seeds, I shall talk in this last book of the work about the wines and the so-called metals, beginning with an essay on the vine.) 13 GdM VI: 99 (Eb; Bln 51; 95). 14 Vgl. GdM VI: 494 (Eb 486). 15 Eb 164. 16 Eb 344. 17 GdM VI: 105.
176
Tanja Pommerening
With regard to classifiers for small grains (N33 ), body parts (F51 ), Disease Symptoms (Aa2 ), and oily or salve-like products (W24 ), it seems that there was a specific usage pattern of classifiers within medical texts, which had become necessary owing to the special requirements of specification of the materia medica.
2.1.4 Hierarchizations As mentioned above, the connection between lexeme and classifier can be described quite well with the prototype approach. In the area of allocating living creatures, the use of prototypes as classifiers can automatically yield taxonomic relations (see Table 2 above). The concept of taxonomy should not be understood here to mean the modern sense of generating a clear, biunique terminology (a goal, which cannot be reached in the natural sciences themselves); rather, the hierarchization in classes emerges from the present system and becomes visible to us. It is flexible and open for the adoption of new members, and it offers no sharply-defined, clear definitions and limits. As demonstrated above, ‘polliwog’ or ‘tadpole’ (ḥfrn) can be classified as fish or worm. Tab. 1 shows the word šnḏ.t ‘thorn acacia’ with M1 ( ), the sycamore pictogramme, as determiner. Thus the sycamore functions at the superordinate level and all lexemes classified with it, are the basic level of the category (cf. Goldwasser 2012: 46–48). The same applies to quadrupeds, fish, birds, reptiles and amphibia, and herbs. Finally, the use of F51 as a classifier for body parts and Aa2 and Aa3 as a classifier for Disease Symptoms in the medical context seems to establish taxonomic relations at the level of medical concepts as well. The hierarchizations in the above examples differ from what is generally understood by the term, since the determinative denotes both, the basic concept and the superordinate concept. Thus, the word behind M1 ‘sycomore’, eg. nh.t, has a double meaning at the language level. As early as in the period of the Old Kingdom, the word nh.t included both the superordinate concept for tree and the specific designation of ‘sycamore’ (Pommerening 2015a: 128–131).
Classification in Ancient Egyptian Medical Formulae
177
2.2 Nominal compounds If we look at drugs expressed by nominal compounds, we can discover other types of relevant classifications, which are crucial for the understanding of concepts of drugs and efficacy. Ancient Egyptian drug names are often combinations of nomina + attributes. Attributes are often adjectives or so-called genitive attributes, which leads to combinations such as ‘boiled barley’ (it ps) vs. ‘crushed barley’ (it sḥm), or ‘donkey’s milk’ (irṯ.t iꜤꜢ) vs. ‘cow’s milk’ (irṯ.t iwꜢ), or ‘ḳꜢꜢ-parts of sycamore’ (ḳꜢꜢ n nh.t) vs. ‘ḳꜢꜢ-parts of thorn acacia’ (ḳꜢꜢ n šnd.t).
2.2.1 Specificity The modification of nouns by means of adjectives or so-called genitive attributes is a common phenomenon not only in Egyptian (cf. Schenkel 1967: 71–83). In this way, a classification results automatically, for the noun is thus divided into two classes. Thus, when boiled barley is mentioned, one expects there to be unboiled barley as well. The narrower definition of ḳꜢꜢ-parts as ḳꜢꜢ-parts of the acacia gives us an idea that there was at least one more kind of leaf. According to Hansjakob Seiler (1960: 19), the attribute in general has ‘die Funktion, aus der Klasse einen bestimmten Vertreter auszuwählen und zu den übrigen, andersgearteten Vertretern derselben Klasse in Opposition zu stellen’ (cited after Schenkel 1967: 74). In this way modification reduces the set of potential referents of a noun to a smaller set that covers the intersecting set of items that have the properties of the noun plus the properties specified by the modifier. By means of modification altogether, a specification is attained. It has been shown (Pommerening 2016a: 84–88) that the word combinations occurring in the formulae or recipes in the area of materia medica are endowed for the greater part with attributes, and therefore show high specificity in this sense. For example, outside of medicine there exist three expressions in particular, as far as barley is more narrowly specified: ‘north barley’ (it mḥ), ‘white barley’ (it ḥḏ), and ‘red barley’ (it dšr). In contrast, the word ‘barley’ in the recipes is linguistically more narrowly specified by 30 different attributes (in extenso: Pommerening 2016a: 85 f.). One can conclude from this that specification in the choice of drugs to prepare medicines played a large role. It is not only remarkable that more was specified, but also that the subdivision was different. Thus, the three expressions mentioned above from a non-medical context are not found in the terminology of Ancient Egyptian materia medica. The classifications occurring within the drug terminology thus reflect concrete concepts of drugs. Beyond that, with regard to the phenomenon of modification, the possibility of a sequence of attributes is to be mentioned, for example, in the form of epithets after the names of Ancient Egyptian deities.
178
Tanja Pommerening
2.2.2 Motivation Modification in general leads to specification. In the case of attribute sequences, such as the occurence with divine epithets, the motivation was surely to cover and present all the facets of the deity concerned. Looking at the drugs named in the formulae, the source of the classification illustrated for us in the linguistic tradition was quite certainly the actual use of corresponding substances and their precise description. With that, there is a direct correlation between the attributes and the concepts of drugs.
2.2.3 Classification criteria and type The classificational criteria expressed mostly by noun+attribute compounds are processing status18, age19, origin20, quality21, gender22, taste/smell23, and colour24. These criteria are directly derived from the pharmaceutical practice and are tightly related to the intended healing effects (Pommerening 2017; Pommerening 2016a). A cooked drug was considered to have other healing effects than an uncooked drug. Similarly, excrements from a red bull must have had different healing effects than excrements from a black bull. If one takes the criteria of processing status, age, origin, quality, gender, taste/ smell, and colour as the starting point of classifications in the materia medica, then one can establish that none of the drugs has all attributes assigned to it. As a rule there is only one attribute, often there are two, even more frequently, however, none. A consistent system based on several properties therefore does not appear.
18 E.g. ‘chufa’ (tiger nut, Cyperus esculentus), ‘boiled’ (wꜤḥ ps): 4 records: Eb 261, 483, 497, L34; ‘chufa, crushed’ (wꜤḥ nꜤgw): 4 records: Eb 49, H 27, 62, 93. 19 E.g. ‘ben oil’ (bꜢḳ): 45 records, i.a. Bln 20, Brk § 43c/2, Eb 106, H 164, Ram IV C 11–12, Ram V Nr. 3, Zag 3; ‘fresh ben oil’ (bꜢḳ wꜢḏ): 38 records, i.a. Beatty VIII Vs. V.1–3, Bln 64, Brk § 46d, Bt 15, Eb 41, H 162, Papyrus Cairo 58027, IV 1–2/2. 20 E.g. ‘fat of the lion’ (mrḥ.t mꜢj), 1 record: Ram V Nr. 3; ‘fat of the wild lion’ (mrḥ. mꜢj ḥsꜢ), 1 record: Eb 465; ‘fat of the cat’ (mrḥ.t.t mjw), 4 records: Eb 465, 627, 658, 847. 21 E.g. ꜤmꜢw-plant / ‘fruit, best quality’, 1 record: Eb 6; ꜤmꜢw-plant/fruit (ꜤmꜢw), 12 records: Bln 35, Bln 157, Eb 78, 79, 82, 83, 94, 122, 217, 320, 328, 860. 22 E.g. ‘plant slime, male’ (ḥsꜢ ṯꜢj): 5 records: Bln 149, Brk § 55, Eb 199, Eb 565, Eb 650. ‘Plant slime’ (ḥsꜢ), without specification: 54 records. 23 E.g. ‘ben oil, sweet’ (bꜢḳ nḏm), 4 records: Bln 51, 84, 98, Eb 687; e.g. ‘fat (of cattle)’, ‘pleasantsmelling’ (Ꜥḏ nḏm), 4 records: Bln 97, 130, Bt 10, Bt Rs. I.7–II.2; cf. ‘fat (of cattle)’ (Ꜥḏ), 43 records, i.a.: Bln 20, Brk § 76, Eb 75, H 45, L 38, Lvr Rs. I.4–5, Sm Fall 41/3. 24 See Fig. 2.
Classification in Ancient Egyptian Medical Formulae
179
Rather, the reduced choice of properties listed here points apparently to a specific concept of efficacy connected with the particular characteristic, this being relevant to the particular context of application. This interpretation agrees with recent experimental examinations and thoughts on conceptual combinations, as they are present here. Conceptual combinations are understood by Murphy (2004: 475) in the debate with current theories as a ‘knowledge-intensive process that involves coordinating rich conceptual representations’. Causal reasoning plays an important role in this. The concepts at the base of the classification can be apprised by a combination of prototype models and knowledgebased approaches (Murphy 2004). Beyond that, the concepts of drugs also includes the handling of drugs, and other practical aspects.
2.2.4 Hierarchizations Hierarchizations are nicely reflected by the use of different attributes. What remains unclear is whether the determinatives only stood for the superordinate concept in the strict sense of hierarchization or whether they included both the superordinate concept as well as the basic concept (cf. 2.1.4). The following example of ‘salt’ will show that there was a superordinate term ‘salt’ which was further specified by one of its relevant attributes. In the prescriptions, we find ‘salt’ (ḥmꜢ.t)25, ‘salt from Lower Egypt’ (ḥmꜢ.t mḥ.t)26, ‘salt from the East’ (ḥmꜢ.t iꜢb.t)27, and ‘salt from the Mountain Area’ (ḥmꜢ.t n.t ḫꜢs.t)28. What is not mentioned (neither here nor in non-medical texts) is ‘salt from Upper Egypt’ (ḥmꜢ.t šmꜤ). It seems that this was no possible extension of the simplex word ‘salt’ (Fig. 2). It is known that the main sources for salt were situated in Lower Egypt, in Wadi Natrun. Thus, the main concept behind ‘salt’ (ḥmꜢj.t) was, in my view, Lower Egyptian salt. The more explicit term ḥmꜢj.t mḥ.t (‘salt of Lower Egypt’) became necessary in prescriptions at a time when salt from other sources (the East, the Mountain Area) became available. These were now subsumed under the prototype term of ‘salt’ (ḥmꜢj.t). ‘Salt’ (ḥmꜢj.t) generated the superordinate expression when subtypes came into view. The new superordinate term for salt in the realm of drugs covers all possible subtypes. But still, if we find the simple term ‘salt’ in the prescriptions; its meaning
25 16 records, e.g. Bln 18, Brk § 48c, Eb 96, H 197, LL XXIX.32, Ram III B 12–14. 26 175 records, e.g. Bln 16, Brk § 45a, Bt 2, Eb 13, H 71, L 7, Leiden Vs. XXVI.11–12, Lvr Rs. I.10–12, Ram V Nr. 4, Rub 16, Sm Fall 41/4, Sm Rs. XXI.3–6, Zag 2. 27 2 records: (Bt 2), Bt 7. 28 1 record: Eb 663.
180
Tanja Pommerening
is the specific meaning of the ‘salt of Lower Egypt’, not the all-encompassing, more general meaning of ‘salt’. This insight can only be derived if all drugs are compared and the materia medica is reconstructed in this fashion. Thus, it proves that in the course of transcribing formulae ‘salt’ and ‘salt from Lower Egypt’ are interchangeable (GdM VI: 343). It is therefore necessary, when concepts of drugs and efficacy are to be understood, that the class delimitations of nominal compounds are carefully reconstructed.
Fig. 1: Hierarchization of ḥmꜢ.t ‘salt’
Types of excrement represent another good example for this kind of binominal structure that was widely used in the context of the materia medica. There were some 46 types of excrement, which were classified by the animal/man by which or whom they were produced, and by some other attributes (a choice: GdM VI: 358–363). In this context, it is additionally possible to use attributes such as ‘excrement from a red bull’ (ḥs n iwꜢ dšr.t)29, ‘excrement from a black bull’ (ḥs n iwꜢ km)30 as opposed to the less specific ‘excrement of a bull’ (ḥs kꜢ), which occurs in different papyri than the other two31.
Fig. 2: Hierarchization of ḥs kꜢ ‘Excrement of a bull’
29 Brk § 47c, Brk § 54d. 30 Brk § 54e. 31 Bt Rs. I.5–7.
Classification in Ancient Egyptian Medical Formulae
181
I am quite convinced that these specifications were associated with different healing effects. As far as the assignment of colour to the steers is concerned, this is probably some kind of taxonomy. The division of cattle according to the colour of their hides can be shown in other contexts, too (Gerke 2016: 60–77). The system behind the classifications of excrement and salt cannot be taxonomic in the narrower sense, because that would imply that unspecified terms for salt or ointments would still have been specified for their taxonomic status by the editors of the texts.
2.3 Internal structure of the prescription To recover the concepts of drugs and efficacy, the structure of a recipe or formula plays a decisive role.
2.3.1 Specificity Formula texts, no matter from what time, display a special formal structure designed to transmit prepared knowledge. They are kept short, as a rule clearly sequenced, clearly ordered internally and therefore quickly comprehensible as to content. The list-like structure, which permits a sketch of the ingredients and what to do below a keyword-like title, can be found in many cultures, and not only in medicine. Cooking recipes and instructions for other manufacturing processes show a similar construction. As a rule, one can find a clear sequence of things such as titles, ingredients, quantities, short production instructions and information about use.
2.3.2 Motivation The similar text structures, which clearly occur in many places independently of each other, refer to existing basic knowledge on the part of the author and the circle of addresses; they ensure the error-free realization of the instructions contained in them in the perception and practice of the target group. They ensure a priori, too, the associations needed for all this: appearance, origin, form, storage, working properties of the ingredients, tools, times, durations, and places for manufacture; uses, clientele, means of application, places of application, duration of application, degree of success of application, kind of effect – all of this is not mentioned, as a rule. If one reads such instructions from a modern point of view, it is clear that a good deal of knowledge is taken for granted and thus in general not explained.
182
Tanja Pommerening
Here is an example formula: Another (pharmakon), [made for one who suffers from concretions in his urine]: Linseed 1/8 [Dja]; inner part of the fruit of the carob 1/32 [Dja]; crushed sw.t-grain 1/4 [Dja]; goose fat 1/8 [Dja]; honey 1/8 [Dja]; iḥw-plant 1/8 [Dja]; water 1/32 [Oipe]. To be boiled. To be filtered. To be drunk four days long. [Papyrus Ebers text 268 (XLIX.12 f.)]
The knowledgeable medic knew what the ingredients looked like, where they could be obtained, how they were to be measured and prepared, and how they came to be effective. He had the knowledge at his disposal of a great many properties of the ingredients and the medicines made of them. This knowledge was (one can certainly presume this) known to both the author and to the target group envisaged by him – a knowledge that was transmitted, but not to us in a connected form.
2.3.3 Criteria of classification Ancient Egyptian formulae follow a uniform scheme. The title says what the symptoms were, for which the drugs needed to be processed into medicine. Then all of the ingredients are given, with or without measurements, how to manufacture or produce them, and the duration of application. The classification lies in the targeted allocation of the drugs to the illness. The few detailed formulae show that the sequence of ingredients in the short form does not have to correspond to the sequence used in manufacture. For reasons of similarity of the ingredients used with Eb 268, Bln 185 shall be given here as an example.32 Red oil, made as a requirement for the heart’s need (ib), after its heat has come. Linseed 1/8 [Dja]; groats of sw.t-grain 1/64 [Oipe (= 1 Dja)]; inside of the carob fruit 1/32 [Dja]; honey 1/8 [Dja]; oil 1/8 [Dja]; water 1/4 1/8 1/16 1/32 [Oipe]. Let its water and its sw.t-grain fall into the kettle (wḥꜢ.t). Concerning its first bubbling up after it has been heated: let the linseed fall in. Concerning its second heating, let the inside of the carob fruit fall into it. Concerning its next heating: let its oil fall into it, until it has boiled and it has boiled. Before it is taken, let honey fall onto it. Thicken it until it is only 6 [Dja]. Expose it in the night to the dew. Filter it through cloths. It is to be drunk by the man four days long. [Papyrus Berlin text 185 (XIX.11–XX.4)]
Especially in the case of potions, the liquid drug is named at last place in the short form of the formula, whereas the ‘main active ingredient’ appears frequently at first place. The detailed instructions then show how it was actually done.
32 The manufacturing process described in detail in Bln 185 cannot simply be transferred to Eb 268, despite Westendorf 1999: 521, for the indications, individual steps of manufacture, and measurements differ considerably.
Classification in Ancient Egyptian Medical Formulae
183
As described elsewhere (most recently in Pommerening 2017; Pommerening 2016b: 263–265), the manner of manufacture can correlate with the recreation of a course of illness and recovery, so that readings may result that place the concepts of drugs and illness in relation. Along with the strictly synchronic classification, we have here a diachronic course, which is primarily reflected in the sequence of drug manipulations and the technique of manufacturing. Thus, one could hypothetically assume that this formula retraces a process of illness in the area of the heart. The ib-heart has received heat and this heating is simulated with the repeated ebullition of a formula in a special kettle. By the thickening of the drugs for mucous the disquiet and pressure from below are symbolically stopped. The process of the illness can thus be characterized by ebullition33. The first three ingredients possess strong properties of swelling, and in combination with water free up mucous substances. The red colour of the oil could also be interpreted as a metaphor of warmth.34 Only after the clinical picture has been completely characterized and established does the process of filtering take place (Egyptian sẖꜢk – lit. ‘make deceitful’, cf. Pommerening 2016b: 263). The signature of the process of illness is invisibly contained in the potion, which effects healing according to the principle simila similibus, which may also be deduced here. Assuming such associations, an exact classification of the drugs and their preparation with regard to clinical pictures then follows. The detailed formulae themselves mediate the process of illness by metaphorical imitation. We cannot read these aspects from a short formulae; the knowledge was presumed. Formula and recipe structures, like lists, seem to be a phenomenon that overarches cultures. The allocation of medicines to the clinical picture, in contrast, is carried out on the basis of different criteria. Today, the primary role for the desired effective and safe application is played by clinical, medical, pharmaceutical, and – in the sector of traditional medicines – the historical evidence. The choice of drugs is first determined by concepts, however. For the Ancient Egyptian formulae these may be reconstructed today from indications and frequently from the sequence in which the drugs are mentioned, as far as the process of production is described in detail.
33 Theoretically one could immediately obtain a healing drink by means of filtering, but the instruction first prescribes a reduction of the boiling suspension. In this manner the mass is thickened. From today’s point of view, particularly with an eye to the ingredients used, one could think of heartburn. 34 Cf. Köhler 2016.
184
Tanja Pommerening
2.3.4 System type and hierarchizations It has become clear that the classification of the structures present in a formula, and here especially the total number of drugs with regard to the illness, is connected with concepts of body, sickness, and healing. The instructions for manufacture and application in the formulae, too, refer back to concepts of drugs, body, illness, and efficacy.35 The classifications here, that are based on analogies from all areas of perception, are highly complex and can hardly be reconstructed, owing to the poor state of transmission. Besides they differ from case to case. Seemingly clearly delimited illnesses are treated in manifold ways, and the richness of associations seems infinite. Empirically won knowledge enters into reciprocal interaction and influences the formation of classes. The sequence of the information transmitted in the formulae is process-orientated; to this extent, a hierarchical structure is present which orientates itself towards the course of formula seeking, obtaining the ingredients, using them in their preparation, and the application of the resulting medicine.
2.4 Arrangement of formulae within one source If we look at the sequence of formulae within individual papyri, the hierarchies can be rather divergent. This can be seen from the fact that a group of prescriptions, which occurs on different papyri can be arranged in a sometimes similar, sometimes different way in individual papyri.36 Grapow (GdM II: 97 f.) thus concludes that an extensive collection of medical notes must have served as the template for the medical papyri, and the scribes joined the texts together according to certain criteria upon being commanded to do so. This possibly previous command, and the resulting arrangement, which is what we have today, is what permits us to reconstruct concepts of drugs and efficacy, using the classification yielded by the arrangement.
2.4.1 Specificity The medical papyri that have come down to us diverge noticeably from each other in their texts. Comprehensive statements on specific and less specific classifications
35 Their reconstruction is described elsewhere in more detail. For more on this, cf. Pommerening 2015a and Pommerening 2017. 36 On parallel individual texts and common text groups on different papyri, see the overview in GdM II: 105 f.
Classification in Ancient Egyptian Medical Formulae
185
are hardly possible, as until now only a small number of medical papyri with larger collections of formulae have been edited and published.37 In those that have, we see an array of ways in which formulae could be classified. A prime example is the Ebers Papyrus, with a length of 20 m, in which the order is clearly recognisable, for which reason it shall be reproduced here in more detail.38 The criteria of classification, which are also valid for the following formulae, are printed in blue or orange. The indentations document the list of prescriptions after a marked new section (by the word ‘beginning’) (ḥꜢty-Ꜥ); deviant classification criteria are marked by in green. Lines are used to show the way of arranging groups together. Very often the last prescription of one group offers a link to the next one. Eb 4–103: Beginning of the collected manuscripts of the pharmaceuticals (made for) removing suffering in the abdomen39 Eb 7: Opening the abdomen Eb 8: Another reason that one excretes Eb 9–15: Remedy for excretion Eb 16–17: Another for opening the abdomen Eb 18–20: Another for the emptying of the abdomen Eb 21: Another for treating the lungs Eb 22–34: Another for the emptying of the abdomen Eb 35: Another for removing any suffering from the abdomen and treating the lungs Eb 36–38: Another for the emptying of the abdomen
Eb 39: Another to remove swelling (šf.wt) from the abdomen Eb 40: Another for the removal of suffering from half of the abdomen (each half itself) Eb 41–43: Removing suffering in the abdomen
Eb 44–48: Pharmakon to end excretions Eb 49–52: Another to stop the excretion of blood
Eb 53–85: Removal of the worm from the abdomen
Eb 86–89: Pharmakon to break the wḫd.w-substances in the abdomen
37 See the overview in Pommerening 2012: 34, Fig. 1. 38 Cf. also Westendorf 1999: 23–34, who does not, however, elaborate further on the criteria of ordering. 39 As long as the title of the formula agrees with the main title, and refers only to illnesses of the belly or to the belly itself, the formulae are not regarded as further categorized.
186
Tanja Pommerening
Eb 90–92: Another to remove the wḥꜢ.w-illness from the abdomen40 Eb 93: Another to remove srf-illness Eb 94–95: Another to remove the wḥꜢ.w-illness from the abdomen Eb 96: Another to remove srf-illness Eb 97–103: Pharmakon to break the wḫd.w-substances in the abdomen Eb 104–120: Beginning of the salves to remove the wḥꜢ-illness Eb 121–124: Another (salve) to remove the wḫd.w-substances Eb 125: Another to remove the wḥꜢ-illness: potion Eb 126–127: Another to remove the seeds (?) of the wḫd.w-substances: potion Eb 128: Another (salve) for the skin Eb 129–130: Another (salve) to remove the wḫd.w-substances Eb 131: Incantation of the wḫd.w-substances Eb 132–138: Pharmakon to treat the abdomen and the treatment of the rectal anal complex (potion) Eb 139: Another to remove heat in the rectal-anal complex Eb 140: Cooling of the rectal-anal complex Eb 141: Getting rid of wḫd.w-substances (suppository/enema) in the rectal-anal complex Eb 142: Getting rid of heat in the rectal-anal complex Eb 143: Cooling of the rectal-anal complex Eb 144: Improve the rectal-anal complex Eb 145–146: Another (suppository) against prolapsus of the rectal-anal complex Eb 147–152: Pharmakon to treat the rectal-anal complex (potion) Eb 153–154: What to do when there is a blockage of heat in the rectal-anal complex (potion) Eb 155–159: Getting rid of heat in the rectal-anal complex Eb 160: Another to cool the rectal-anal complex (potion) Eb 161: Another to make the vessels of the rectal-anal complex more comfortable (dressing) Eb 162: Healing the rectal-anal complex (suppository) Eb 163: Cooling the rectal-anal complex (suppository)
Eb 164: Making the rectal-anal complex more comfortable, making comfortable the lower abdomen (pill in the rectal-anal complex)
40 From this it follows that wḥꜢ and wḫd.w are conceptually connected; see, too, further below.
Classification in Ancient Egyptian Medical Formulae
187
Eb 165–174: Removal of magic in the abdomen Eb 175–181: Pharmakon to remove heat in the lower abdomen Eb 182: Another to remove the influence of a dead in the abdomen
Eb 183–187: Pharmakon to treat the breast
Eb 188–21641: Experience of illness in the area of the stomach and heart (rꜢ ib) Eb 217–218: Another to remove illness of the heart (ib) Eb 219–220: Pharmakon to remove an agglomeration of heat on the heart (ḥꜢ.tj) Eb 221–241: Beginning of the pharmaceuticals to remove the ꜤꜢꜤ-poison seeds in the abdomen and in the heart (ḥꜢ.tj) Eb 225–226: Another to remove the ꜤꜢꜤ-poison seeds of a god and a dead person in the abdomen of the man Eb 227–228: Another to remove the ꜤꜢꜤ-poison seeds on the heart (ḥꜢ.tj)42 Eb 229–232: Medicine with immediate effect to remove the ꜤꜢꜤ-poison seeds Eb 233–234: Another to treat the heart (ḥꜢ.tj) and to remove the wḫd.w-sub stances Eb 235: Medicine with immediate effect to remove the ꜤꜢꜤ-poison seeds Eb 236–237: Another to remove the ꜤꜢꜤ-poison seeds on the heart (ḥꜢ.tj) Eb 238–240: Medicine with immediate effect to remove the ꜤꜢꜤ-poison seeds
Eb 241–247: Beginning of the pharmaceuticals made for Re himself [by a deity]
Eb 248–260: Pharmakon to remove actions in the head
Eb 261–283: Beginning of the pharmaceuticals to remove a blockage of urine
Eb 284–293: Beginning of the pharmaceuticals to cause the heart (ib) to take up nourishment
Eb 294–300: Beginning of the pharmaceuticals to cause mucous to descend Eb 301: Another pharmakon to remove the illness on any part of the body of the man (ingredients are mucous) Eb 302: Another to remove the ‘bitterness’ illness (ingredients are mucous) Eb 303–304: Another to defend against heat (šmm.t) (ingredients are mucous) Eb 305–306: Beginning of the pharmaceuticals for the removal of coughing
41 Eb 204 does not name rꜢ ib; instead ‘blockage in his left half’; Eb 210 ditto, instead ‘blockage in his right half’. 42 The two last-named formulae are referred to the right place within the papyrus, namely here, using a marking.
188
Tanja Pommerening
Eb 326–335: Beginning of the pharmaceuticals to kill the gḥw-illness
Eb 336–34143: Beginning of the pharmaceuticals from the manuscript collection for both eyes
Eb 432–435: Another for a bite from a human being Eb 436: What to do about the bite of a crocodile
Eb 437–450: Beginning of pharmaceuticals to remove the ḫnsj.t-illness on the head Eb 451–463: Beginning of the pharmaceuticals to remove gray hair and to treat hair Eb 464–473: Beginning of the pharmaceuticals to make fast the hair Eb 474–476: Another to cause the hair to disappear
Eb 477–481: Beginning of the pharmaceuticals to treat the liver
Eb 482–509: Beginning of the pharmaceuticals for a burn Eb 510–514: Pharmakon to remove the weals of a blow Eb 515–550: Beginning of the pharmaceuticals to heal a wound in the flesh Eb 551–555: Beginning of the pharmaceuticals to remove ulcers Eb 556–591: Beginning of (the pharmaceuticals to) remove a swelling and stop gorging Eb 592–602: Beginning of the pharmaceuticals for gorging on blood and stilling the gorging
Eb 603–615: Beginning of the pharmaceuticals to soften the knee Eb 616–626: Beginning of the pharmaceuticals for the finger when it hurts, or the toe Eb 627–696: Beginning of the salve to strengthen the vessels and the pharmaceuticals to improve the vessels
Eb 697–704: Beginning of the pharmaceuticals to remove an illness of the tongue, when it hurt Eb 705–707: Beginning of the pharmaceuticals to remove the špn-illness Eb 708–711: Pharmakon to remove smells Eb 712: Pharmakon to remove tpꜢw from the head (smells?)
Eb 713–738: Pharmakon to open the body (cosmetic improvements)
Eb 739–749: Beginning of the pharmaceuticals to treat a tooth Eb 750–756: Beginning of the pharmaceuticals to remove the spoilt bitterness
43 A more exact investigation of which rule the connected formulae follow must be made.
Classification in Ancient Egyptian Medical Formulae
189
Eb 757–769: Beginning of pharmaceuticals to treat the right half by drainage Eb 761–763: Beginning of pharmaceuticals for the cold Eb 764–770: Beginning of the pharmaceuticals for the ear, when hearing is slight Eb 771–777: Pharmakon to remove ‘cuts’ Eb 778+779: Pharmakon to remove the mḥs-illness Eb 780+781: Another pharmakon for the zr-illness Eb 782: Pharmakon to get rid of much shrieking of children Eb 783–807: Beginning of pharmaceuticals made for women Eb 808–811: Beginning of pharmaceuticals for the breast, when it hurts Eb 812–827: Pharmakon to remove mucous in the uterine complex Eb 828–830: Pharmakon to pull out the blood in a woman
Eb 840–851: Beginning of pharmaceuticals made to remove [fleas]44 in the house Eb 852–853: Incense, made to make the smell of the house and the clothing pleasant
Eb 854–855: Beginning of the secret knowledge of the physician: The Knowledge of the Course of the Heart (ḥꜢ.tj), the knowledge of the heart (ḥꜢ.tj) Eb 856: Beginning of the Book of the Itinerancy of the wḫd.w-substances
Eb 857–862: Experience of a ḥnḥn.t-ulcer on the throat of the man Eb 863–875: Experience of a ꜤꜢ.t-ulcer Eb 876: Experience of a sfṯ-ulcer Eb 877: Experience of a Ꜥnw.t-ulcer
It appears that, in concentrating the topics, in the majority of cases the locations of the illness are in the foreground. Here the following superordinate groupings can be made out: abdomen (bis Eb 335), eyes (Eb 336–431), hair (Eb 437–476), liver (Eb 477– 481), changes in the skin (Eb 482–602) and extremities, and vessels (Eb 603–696). In the further division the particular symptoms of illness play a role. The circle of users is different in the case of the formulae for gynaecology (Eb 783–839), followed by substances against vermin and unpleasant smells (Eb 840–853). The conclusion of the papyrus is formed by two comprehensive theoretical works (Eb 854–856) and finally a collection of teaching texts, which primarily recommend surgical measures (against all kinds of ulcers and abscesses). Structural notes, such as ‘Beginning of (...)’, are hardly present in other papyri. Yet groups of formulae can be recognized there, too.
44 According to the following animals (snake, fly, mosquito, mouse, harrier) in the sense of vermin.
190
Tanja Pommerening
To obtain significance with regard to specificity, here are the structural criteria of division in Papyrus Berlin 3038 (key points), written some 300 years later, as a comparison: Bln 1–10: Killing the worm in the body Bln 11–13: (destroyed) Bln 13 ends with breast Bln 14–18: suffering breast Bln 19–20: fnṯ-worm (illness of the nails) Bln 21–28: painful places (dressing; beginning of the relevant formula destroyed) Bln 29–47: coughing Bln 48: mucous in both sides, potion Bln 49–57: swellings Bln 58–79: incense Bln 80–113: salves (exception in Bln 107/08: potion to be sipped) Bln 114–117: medicine against ꜤꜢꜤ-poison seeds Bln 118: ꜤꜢ.t-ulcer, treatment with smoke Bln 119–120: pains in arms/legs, potion Bln 121–135: pains (arms)/legs, dressing Bln 136–147: mucous Bln 148–150: belly, potion Bln 151–152: destroying the spread of blood on the heart (ib) Bln 153–156: wḫd.w-substances places of body Bln 157–160: body engorgement Bln 161–163: wḫd.w-substances Bln 164–183: emetic medicines, partly against wḫd.w-substances Bln 184–188: potion Bln 189–191: sayings Bln 192–199: gynaecology (esp. prognoses) Bln 200–203: ear Bln 204: potion Here, the locality of the diseased phenomenon is not so much in the foreground as the cause and the connected applications by using smoke, salves, potions or suppositories. The theory for this is delivered by the Book of the Itinerancy of the wḫd.w-Substances, already transmitted in the Ebers Papyrus (Bln 163a–h; Eb 856). The arrangement according to forms of application such as suppositories and potions in combination is based on the knowledge that a vessel in the rectal-anal complex is connected to the entire system of vessels in the body. That is to say that the structure here is obviously influenced by theoretical knowledge. A direct derivation of the concepts of efficacy from categorizations is thus present here, too.
Classification in Ancient Egyptian Medical Formulae
191
Other collection manuscripts show their own kinds of division. Naturally, they are closely connected with the practical use of the manuscript concerned, which results from the total transmission of the texts. Noticeable are manuscripts on limited subjects, such as gynaecology (Papyrus Kahun), snakebites and their causes (Papyrus Brooklyn 47.218.48/.85), treating injuries to the head down to those of the abdomen (Papyrus Edwin Smith), arranged a capite a calcem and within that according to severity (cessation in the area of the abdomen), or a papyrus that contains only emetic medicines (Papyrus Chester Beatty VI). The choice of single texts can thus have been based on larger, topic-orientated works.
2.4.2 Motivation The compilation of the prescriptions is determined by strictly practical purposes.
2.4.3 Criteria of classification The compilation of the prescriptions differs across the different papyri. The papyrus may focus on the place of application, the target group (women), the specific diseases (wounds, injuries), the origin of the prescription (divine) or its pharmaceutical form (salve, ointment). This clearly illustrates that there were no uniquely fixed structures, that allowed only for one categorization. Nevertheless, it is clear that no papyrus contains the same text twice, apart from an obvious scribal error in the Ebers Papyrus. In the area of the transmission of formulae, a strict classification was thus undertaken within the particular manuscript. That means that it was sorted out of a pool of formulae into only one category. A desideratum is the investigation of which criteria of ordering possessed the greatest relevance in cases of multiple possible assignments.
2.4.4 System type and hierarchizations The various different ways to order the formulae and their choice in the papyri offer a colourful picture of the possibilities at the disposal of the Egyptians. It appears that every papyrus must be investigated individually with regard to its system of ordering. Associations in changing from one group of formulae to another can frequently be traced. In the end, here, too, the model of prototype semantics can be applied. Formulae at the end of a group frequently appear to be further distant from the typical representative and, for this reason, can form the connection to the most prototypical representative of the next group more easily. If one looks at the collection manuscripts in comparison, no over-arching system of topics can be shown to exist (e.g. first abdomen, then head, then limbs, as in the Ebers Papyrus). It is also not
192
Tanja Pommerening
always comprehensible as to what could have led to the topic areas and the internal ordering of the formulae in one topical area. For immediate practical use, the rapid finding of information will have been relevant.
3 Conclusion We can observe different types of classification in Egyptian medical texts, particularly in prescriptions. The most specific type of classification shows up in the nominal compounds of drug. Compared to non-medical texts, there is a strong tendency in Egyptian medicine to specify the plant, mineral and animal components of a drug. This is arranged by the use of adjective attributes or genitive attributes. The distinctions are clearly practically motivated. Individual plants, animals and minerals are most importantly classified according to their healing properties. Those distinctions concerning processing states, properties, origins, etc. seem to be of less importance in non-medical texts. In the area of the classifiers we find three frequently used hieroglyphs F 51 and Aa2 / Aa3 in the formulae, which determine lemmata that probably had to be newly conceptualized in the context of medicine: the two large groups of body terminology and disease terminology. This clearly illustrates, that there were no uniquely fixed structures in the system. The classification systems are open and flexible. Concepts of drugs and their efficacy can best be recontructed with the help of the arrangement of the individual prescription in the larger context of the papyrus. If one observes the order of the formulae within the manuscripts, different structures can be established, which will have been orientated to the practical application of the texts (ex. Papyrus Berlin 3038). Particularly complex knowledge builds the basis for the assignment of drugs to a disease complex and their manipulation during manufacture (internal structure of the prescription, selection and arrangement). In this area, in contrast, over-arching concepts can be found. Because of the uniform purpose of the formula texts as presumed in this investigation, and the assumption of a common horizon of expertise obtaining between the historical author and the historical user (agreement of concept), one may assume generally known concepts of drugs and their efficacy in defined periods of time. Nonetheless, the arrangement within the manuscripts displays individual traits in each case. By having analyzed the four levels which show up with classification processes, namely pictorial, nominal compounds, internal structure of the prescription, selection and arrangement of the individual prescription within the larger context of a papyrus, it becomes obvious that every structure is able to provide information on the underlying concepts of healing. Methodologically it is most important to consider whether a phenomenon is specific with regard to the source, topic, time, etc. This
Classification in Ancient Egyptian Medical Formulae
193
presumes a detailed examination including big data of texts45 for comparison. In particular, for the pictorial levels those studies still have to be done. Hierarchizations in terms of taxonomies as they are known today cannot be found in the data. In particular, the Egyptian way of using a typical species of a genus to build a generic name, and to use this term still as a species name at the same time differs from other systems.
Appendix a) Ꜣw.t-ib, Ꜣbw, iꜢw, iꜢrr.t, iꜢgw.t, iwrj.t, iwḥw, ibw, ibnw, ibr, ibsꜢ, ipšnn, imꜢ, ins.t, inj.t, inr-spdw, inhmn, inktwn, ir.t-p.t, ir.tj, irtjw, ihm.t, iḥw, išd, išd.t, idw, ꜤꜢmw, ꜤꜤꜢm, ꜤfꜢ, ꜤmꜢw, ꜤmꜤꜤ, Ꜥnnw, Ꜥntjw, Ꜥrw, Ꜥš, Ꜥḏ, Ꜥḏrn, wꜢm, wꜢḏw, wꜤn, wꜤḥ, wnšj, wš.t, wšꜤ, wšb.t, wtj.t, wḏꜤjt, bꜢj, bj, bj.t, bjꜢ, ḥwj.t n.t bnw.t, bnn.t, bnr, bnjw, bsn, bsbs, bsn, pꜢj.t, pꜢ Ꜥr.t, pr.t, pr.t-šnj, prš, pḫ.t-ꜤꜢ, psḏ, pšn.t, pr.t mꜢꜢ, mꜢt, mꜢd, mjm.t, mjmj, mjkꜢ.t, mw.t, mnj, mnḥ, mnš.t, mhwj, mstꜢ, msdm.t, mkj, nbj.t, nbs, nnjb, nḥꜢ-s-Ꜥwj, nḥš, nḥd.t, ns-š, ns-šm.t, nḳꜤw.t, nṯrj.t, pr.t nḏm, nḏḥꜤḏḥꜤ.t, rj.t, rwḏ, hnꜢj.t, ḥwj.t, ḥwr.w, ḥm.t, ḥmꜢ.t, ḥmꜢj.t, ḥmj.t, ḥmꜢr.t, ḥmw.t, ḥnj-tꜢ, ḥnw.t, ḥrj.t, ḥrs.t, ḥsmn, ḥknw, ḥtm, ḥḏw, ḥḏ.t, ḫpr-wr, ḫprj, ḫnf.t, ḫrš.w, ḫs, ḫšw, ḫsbd, ḫt-ꜤwꜢ, ḫt-ds, ḫt-dšr, ẖꜢw, ẖꜢw nw bjꜢ, ẖꜤ Ꜥ.w, ẖnn.w, ẖsꜢ, ẖsꜢj.t n.t mḏꜢ, sw.t, swḥ.t, sꜢ-wr, sjꜢ, sin, sꜤꜢm, sm.t, snn, snṯr, sr, shr.t, sḫp.t, sẖ.t, sskꜢ, sšw, skj, stj, šꜢw, šb.t, špnn, šꜤw, šꜢšꜢ, šꜤj, šꜤ.t, šb.t, šps, šfšf.t, šnj-tꜢ, šnf.t, šs, šsꜢj.t, škr, ḳꜢꜢ, ḳꜢꜢ.t, ḳꜢ.w, ḳbw, ḳmj.t, ḳnj.t, ḳsntj, ḳtḳt.w, kꜢ.w, kꜢ.t-šw.t, kmw, krk, ksb.t, kšw, gngn.t, gsfn, tꜢ-msḥ, tjꜢ, tjꜤm, tj-šps, tbn, tpꜢ.w, tpꜢw.t, tpnn, tpnn.t, tf, tntj, tntm, ṯwṯkn, ṯr.t, ṯrw, ṯḥn.t, dꜢb, djdj, dbn, dbḥ.t.t, dḥꜢ, dḥtj, dšr, dšr.w, dḳw, dgm, dḏw, ḏꜢj, ḏꜤr.t, ḏꜤs, ḏꜤb.t, ḏꜤm.w, ḏmtn, ḏrn.t, ḏḥtj. b) Ꜣgg.t, ꜤꜢmw, ꜤmꜢw, i, iꜢr, iꜢḳ.t, ib.t, ibw, ibsꜢ, ip.t, imꜢ, ims.t, inj.t, inwn, inb, innk, ins.t, iḥj, iḥw, isw, išd, išd.t, ꜤꜢb, ꜤꜤꜢm, ꜤfꜢ, ꜤmꜤꜤ, Ꜥnḫ-imy, Ꜥḫ, Ꜥẖm.w, Ꜥgj.t, wꜢb, wꜢm, wꜢnb, wꜢdw, bꜢgs.w, bb.t, bḥḥ, bsbs, bt.t, bddw-kꜢ, pꜢḫ, pꜢḫ-srj.t, pp.t, pḫḫ, mꜢ.t, mꜢt.t, ḥs mjw, mnj.t, mnwḥ, mnḥ, mnḏj, mḥj, mḥj.t, ḥs msḥ, mstꜢ, msḏr-ḥḏr.t, mḳr.t, mgꜢ, njꜢjꜢ, nwꜢ, nwꜢn, nb, nbj.t, nfr.t, nḫtjw, nšꜢw, ns.tjw, nšꜢ.w, nṯr, rkrk, hrj, hdn, ḥb.t, ḥpꜤpꜤ.t, ḥmj.t, ḥmꜢw, ḥmw, ḥmm, ḥnj, ḥnj-tꜢ, ḥnw, ḥḏw, ḥḏ.t, ḫꜢ.w, ḫꜢsj.t, ḫb.w, ḫfꜤ Ꜥ=j Ꜣm Ꜥ=j, ḫnš, ḫsꜢw, ḫt-ds, sꜢ-Is.t, sꜢj.t, sꜢp.t, sꜢr, sꜤꜢm, sꜤm, sw.t, sbttj.t, smw, sn-wt.t, sẖ.t, sšn, sd-pnw, šꜢw, šꜢwj.t, šꜤb.t, šꜢms, šꜢs, šꜢšꜢ, šw.t-ḏḥwtj, šww, šbb, špn, šps, šmꜢ.w, šmšm.t, šnꜤ.w, šnḏ.t, šsp.t, ḳꜢꜢ, ḳꜢd.t, ḳbw, ḳn.t, ḳnj, ḳst.t, kꜢkꜢ, ktkt, gꜢb.t, gj.t, gjw, gngn.t, grš, tjꜢ, tjꜤm, tjw, tj-šps, twr, tbt.t, tḥwꜢ, tḫw, ṯꜢ, ṯwn, ṯhw, ṯḥf.tj, dwꜢ.t, dbj.t, dr-nkn, dḥꜢ, dḥꜤꜤ, dšr, dgm, ḏꜤr.t, ḏꜤs, ḏꜤꜤ, ḏsr.t. c) ꜢꜤꜤ, Ꜣbḫ.t, Ꜣḥ, iꜢrr.t, iꜤj.t, ibw, ibr, ibḫj, if.t, inj, irp, irṯ.t, išf, ꜤꜢ.t, ꜤꜢgj.t, Ꜥwj.t, ꜤꜤꜢm, Ꜥntjw, Ꜥnḏw, Ꜥš, Ꜥḏ, wꜢḏw, bꜢḳ, bj.t, bnjw, pꜢ-ib, pꜢ-wr, mrḥ.t, mhwj, mstꜢ, msdm.t, nwd.w
45 The growing text source data base of Thesaurus linguae aegyptiae in connection with the ‘Zettelkasten’-archive provides a good amount of different sources.
194
d)
e) f) g) h) i) j) k) l) m) n)
Tanja Pommerening
šps.w, nḥbw.t, nḥḥ, rhd.t, hnw, ḥꜢ.t.t, ḥmw, ḥrw.t, ḥsꜢ, ḥnḳ.t, ḫꜢwj, ḫprj, ḫmt.nj, ḫt-ds, ḫḏw, ẖꜤꜤ.w, sjn, sfṯ, smj, srm.t, sḫp.t, sẖ.t, sḏr, šꜢmj.t, šꜤj.t, šb.t, šꜢmw, šb.t, šfw, šnf.t, ḳmj.t, gꜢš.w, tꜢ, tꜢḥ.t, tj-šps, twjꜢ, tp.t, tpꜢw.t, tḫw, dbj, dbj.t, dbn, ḏwjw, ḏsr.t, ḏdm.t. Ꜣbj, ib, ibṯrsw, ꜤꜢ, Ꜥw.t, Ꜥbḫn, Ꜥpnn.t, ꜤpšꜢj.t, ꜤmꜤmw, Ꜥnḫ, wꜤ.tj, wnšj, pnw, pgg.t, mꜢj, mꜢj-ḥsꜢ, mjw, mj.t, nrꜢw, hnn, ḥntꜢ, ḥḏr.t, ḥtr, ẖn.t, sm.t, skꜢ, šꜢj, šsꜢ.w, štw, ḳꜢdj.t, ḳrr, gḥs, ṯsm, ṯsm.t, db, dḥr. imꜢ, inhmn, isr, išd.t, iḳrw, Ꜥwn.t, Ꜥnnw, Ꜥrw, Ꜥš, wꜤn, bꜢḳ, bḫb, mꜢꜢ, mꜢf.t, mrj, mnḳ, nbs, nnjb, nh.t, nḏm, hbnj, šb.t, šnḏ.t, ksb.t, gꜢb.t, tj-šps, ṯr.t, dgm. Ꜣjs, ꜤꜢ.t, iṯꜢ, im, ꜤꜢgj.t, Ꜥmm, wḏꜤjt, bj, bnf, bd.t, bd.t-ḥꜢwr.t, mw.t, ḫt-ds, sw.t, swj.t, swḥ.t, smꜢ.tj, sḫt, sẖ.t, šꜤw, šnf.t, šsp.t, ḳꜢꜢ, ksb.t, gnn, ḏꜤr.t, ḏꜤb.t. irḥnn.t, Ꜥff, wjꜢ.t, bjbjw, pꜤr.t, mn.t, mnj.t, mšꜤ, njw, nr.t, rꜢ, ḥwr, ḥn.t, s.t, smn, sḥjḥ.t, gꜢbgw, gnw, ṯrp, dꜢgj.t, ḏrj.t. Ꜣjs, Ꜥwj.t, Ꜥrw, Ꜥš, wꜤn, wsš.t, bnf, bd.t-ḥꜢwr.t, rj.t, ḥs, ḥs msḥ, ḥnb.t, ẖpꜢ, ẖrj pddw, šꜢw, špnn, šꜤmw, kꜢj.t, tbn, tpꜢ.w. Ꜣs.t, imrw, inr, ihm.t, Ꜥnw, Ꜥḏ, wꜢḥ-nḥb.t, wꜢḏw, wḏꜤjt, bnw.t, pꜢḳj.t, mꜢt, mꜢd, msḳ, ḥḏ.t, ḫtm, šs, ḳꜢḥ, dbn, ds, ḏb.t. iꜢ.t, iwf, ins.t, wdd, bnf, bḳsw, mjs.t, mw.t-rmṯ, mnḏr, mẖtw, nnšm, ns, ḥꜢ.tj, ḫnd, ḫs, ẖ.t, swg, sḫw, sṯr.t, kꜢ.t. Ꜣḥ, iꜤj.t, Ꜥwj.t, wtjw, bj.t, prsn, ḥsꜢ, ḥkn, ḫnm.t, ẖꜤꜤ.w, šꜤ.t.t, šꜤj.t, šb.t, šns, šd.t, t. iḳrw, ꜤꜢgj.t, wꜢb, mnj.t, rn.t, hbnj, ḫpr-wr, ḫs, sꜢw, smꜢ.w, tj-šps, ḏꜤ, ḏꜤꜤ. Ꜣbḏw, in.t, ꜤḥꜢ, Ꜥḏw, wḥꜤ.w, njs, nꜤr, rm.w, ḥfrn, ẖꜢ.t, tm.t, dšr.w, ḏdb. Ꜥpnn.t, ꜤnꜤr.t, wꜤwj.t, pḳrw, ḥfꜢ.t, ḥfꜢ.w, ḥfrn.
References Allen, James P. 1984. The inflection of the verb in the Pyramid Texts. Malibu: Undena Publications. Gardiner, Alan Henderson. 31957. Grammar: being an introduction to the study of hieroglyphs. London: Oxford University Press. Gerke, Sonja 2016. Hirtenbilder – Untersuchungen zur kulturimmanenten Sicht auf eine altägyptische Personengruppe. Unpublizierte Dissertation. Mainz. Goldwasser, Orly. 2002. Prophets, lovers and giraffes. Wor(l)d classification in Ancient Egypt. (Classification and Categorization in Ancient Egypt 3; Göttinger Orientforschungen, 4. Reihe: Ägypten 38,3) Wiesbaden: Harrassowitz. GdM = Grapow, Hermann, Hildegard von Deines and Wolfhart Westendorf. 1954–1973. Grundriß der Medizin der Alten Ägypter. 9 vols. Berlin: Akademie-Verlag. II: Grapow, Hermann. 1955. Von den medizinischen Texten. VI: Deines, Hildegard von, Hermann Grapow. 1959. Wörterbuch der ägyptischen Drogennamen. VIII: Westendorf, Wolfhart. 1962. Grammatik der medizinischen Texte. Kahl, Jochem. 1994. Das System der ägyptischen Hieroglyphenschrift in der 0.–3. Dynastie. (Göttinger Orientforschungen, 4. Reihe: Ägypten, Bd. 29) Wiesbaden: Harrassowitz. Kammerzell, Frank. 2011. “Auszug aus dem Abschlußbericht des deutsch-israelischen Kooperationsprojekts ‘Typologie und Gebrauch der ägyptischen Hieroglyphenschrift’.” In Eliese-Sophia Lincke 2011, 151–159.
Classification in Ancient Egyptian Medical Formulae
195
Köhler, Ines. 2016. Rage like an Egyptian. Die Möglichkeiten eines kognitiv-semantischen Zugangs zum altägyptischen Wortschatz am Beispiel des Wortfeldes [wut]. (SÄK-Beihefte 18) Hamburg: Buske. Lakoff, George. 1987. Women, fire, and dangerous things. What categories reveal about the mind. Chicago: University of Chicago Press. Lincke, Eliese-Sophia. 2011. Die Prinzipien der Klassifizierung im Altägyptischen. (Classification and Categorization in Ancient Egypt 6; Göttinger Orientforschungen, 4. Reihe: Ägypten, Bd. 38,6). Wiesbaden: Harrassowitz. Murphy, Gregory L. 2004. The big book of concepts. Massachusetts: MIT. Pommerening, Tanja. 2012. “Altägyptische Rezepte – Eine diachrone Betrachtung.” Geschichte der Pharmazie 64 [Heft 3]: 33–38. Pommerening, Tanja. 2015a. “Bäume, Sträucher und Früchte in altägyptischen Listen – eine Betrachtung zur Kategorisierung und Ordnung.” In Die Liste. Ordnungen von Dingen und Menschen in Ägypten, edited by Susanne Deicher and Erik Maroko, 125–166. (Ancient Egyptian Design, Contemporary Design History and Anthropology of Design, 1) Berlin: Kadmos. Pommerening, Tanja. 2016a. “Wege zur Identifikation altägyptischer Drogennamen – eine kritische Betrachtung.” In Zwischen Philologie und Lexikographie des Ägyptisch-Koptischen. Akten der Leipziger Abschlusstagung des Akademienprojekts ‚Altägyptisches Wörterbuch‘, edited by Peter Dils and Lutz Popko, 82–111. (Abhandlungen/Sächsische Akademie der Wissenschaften zu Leipzig. Philologisch-historische Klasse 84, Heft 3) Stuttgart/Leipzig: S. Hirzel. Pommerening, Tanja. 2016b. “Heilkundliche Texte aus dem Alten Ägypten: Vorschläge zur Kommentierung und Deutung.” In Translating Writings of Early Scholars in the Ancient Near East, Egypt, Greece and Rome. Methodological Aspects with Examples, edited by Annette Imhausen and Tanja Pommerening, 175–279. (Beiträge zur Altertumskunde, 344) Berlin/Boston: de Gruyter. Pommerening, Tanja. 2017. “Medical Re-enactments: Ancient Egyptian Prescriptions from an Emic View.” In Proceedings of the XIth Congress of Egyptologists, Florence, Italy 23.–30.08.2015, edited by M. Cristina Guidotti and Gloria Rosati. Oxford: Archaeopress Archaeology. Schenkel, Wolfgang. 1967. “Adversarien zu Attribut, Apposition und Genitiv-Relation des Ägyptischen.” Mitteilungen des Deutschen Archäologischen Instituts, Abteilung Kairo 22: 71–83. Schwarz, Monika. 32008. Einführung in die Kognitive Linguistik. Tübingen/Basel: A. Francke. Seiler, Hansjakob. 1960. Relativsatz, Attribut und Apposition. Wiesbaden: Harrassowitz. Wb = Erman, Adolf, and Hermann Grapow. 41982. Wörterbuch der aegyptischen Sprache. 6 Bde. Berlin: Akademie-Verlag. Westendorf, Wolfhart. 1999. Handbuch der altägyptischen Medizin. 2 vols. (Handbuch der Orientalistik, Abt. I, 36) Leiden et al.: Brill.
Figures Fig. 1: Tanja Pommerening Fig. 2: Tanja Pommerening
Walter Bisang
Classification between Grammar and Culture – a Cross-Linguistic Perspective Summary: Classification is involved in a multitude of grammatical categories and manifests itself in a variety of different types of classifier systems. The present paper will offer a survey of the categories and the classifier types across the world’s languages. Moreover, it will address the following three questions: (i) Why does classification cover such a large number of grammatical categories? (ii) To what extent are the criteria of classification universal and to what extent are they language-specific and culture-specific? (iii) What is the relation between classification and conceptualization in the grammar of human languages?
1 Introduction Speakers and learners of Indo-European languages are familiar with the phenomenon of classification through their experience with gender. They may be less familiar with the fact that gender is by no means limited to Indo-European languages (cf. Corbett 1991 for a survey) and they may be even less aware that gender is only one type of classification among many others. Aikhenvald (2000a) presents seven types of classifier systems in her survey and shows additionally that many languages have more than one type (different classifier sets with different classificational criteria in one and the same language) or that a single set of markers covers the functions of more than one type of classification (multiple classifier languages). It is needless to say that this broad typological variation also comes with a multitude of grammatical functions that can be expressed by classifier systems, among them agreement, pronominal reference to concepts activated in discourse, definiteness, possession and the formation of relative clauses. Such a wealth of cross-linguistic variation calls for a systematic overview of classification in grammar and this is in fact the first aim of this paper. It will start with a short sketch of the relevant semantic or cognitive criteria of classification in section 2 and the different types of classifiers and the grammatical functions expressed by them in section 3. Based on this information, the paper will then address the following three questions concerning the relevance of classification from a broader perspective: i. What motivates the use of classification across a large number of different grammatical categories? (section 4) ii. What are the universal features of classification and what is language-specific and culture-specific? (section 5) DOI 10.1515/9783110538779-007
200
Walter Bisang
iii. What is the relation between classification and conceptualization in the grammar of human languages? (section 6) Questions (i) and (ii) will be seen in the context of human cognition. More specifically, it will be argued that classification is employed for highlighting conceptual boundaries of objects and for optimal parsing, i.e. the recognition of syntactic units as early as possible in the course of speech production. The first function crucially contributes to the cognitive processes of the individuation and the identification of objects in space and discourse, two processes that are shared by the many grammatical categories in which classification is involved (Bisang 2002). The second function is associated with agreement as it is found by definition in systems of gender or noun class. Question (ii) on universal vs. specific features of classification directly focuses on two types of features, those that are based on deeper cognitive foundations and those that depend on culture-specific distinctions. Finally, question (iii) shows that classification and conceptualization are strongly intertwined. This can be seen from the potential of creating specific concepts out of more general concepts in languages with flexible classifier systems and from the historical development of classifier systems.
2 Criteria of classification Systems of nominal classification divide the nominal inventory of a language into different sets of objects. Thus, classification can be defined as ‘[a] mental operation that causes an object or a multitude of objects to fall under a concept X’ (Seiler 1986: 95). The semantic criteria for defining the concept X as they are relevant in grammatical systems have been described from various perspectives. The three most important ones are conceptual classification (Allan 1977), social-interactive classification (Denny 1976) and taxonomic classification (Berlin, Breedlove and Raven 1968; 1973). Allan’s (1977) classical conceptual approach is based on a cross-linguistic comparative study of classifier systems. As a result, he presents a list of seven criteria which occur recurrently in classifier systems: i. material ii. shape iii. consistency iv. size v. location vi. arrangement vii. quanta The first criterion addresses what is ‘the essence of the entities referred to by nouns’ (Allan 1977: 229) and is divided into the three subcategories of animate object, abs-
Classification between Grammar and Culture
201
tract notions (mostly verbal nouns, e.g. sing > his singing) and inanimate objects. The second criterion of shape is defined by spatial dimensionality. It includes the three criteria of long for one-dimensional objects, flat for two-dimensional objects and round for three-dimensional objects. Criterion (iii) is characterized by the properties of flexibility, hardness/rigidity and discreteness. The criterion of (iv) size refers to the smallness or bigness of the concept to be classified. Many languages have specific classifiers for (v) locations, i.e. for plots of land, countries, gardens, buildings, institutions, etc. Criterion (vi) points out the spatial structure of the object to be classified. In English, this type of arrangement is reflected by words like strand in a strand of hair or coil in a coil of wire. Finally, criterion (vii) refers to measure words that quantify the concept expressed by the noun as in a bottle of water or a cup of tea. Denny’s (1976) classification is based on the way in which humans interact with objects. He distinguishes three types of interaction: Three main kinds of human interactions seem to be conveyed by noun classifiers, physical interaction such as handling, functional interaction such as using an object as a vehicle, and social interaction such as interacting appropriately with a human compared to an animal, or a high status person compared to a low status one. (Denny 1976: 125)
Physical interaction is characterized by spatial properties and strength of material, while functional interaction classifies the way in which a given object is made use of. The third criterion of social interaction assigns animate beings to social groups arranged in a hierarchical scale. Classification in terms of taxonomy is based on the observation that humans have a universal habit of splitting up the world of living organisms into categories and subcategories. Berlin, Breedlove and Raven (1968, 1973) suggest the following five levels: (1)
Five levels of abstraction (Berlin, Breedlove and Raven 1968; 1973) 1 unique beginner: plant, animal 2 life form: tree, bush, flower 3 generic name: pine, oak, maple, elm 4 specific name: Ponderosa pine, white pine, jack pine 5 varietal name: northern Ponderosa pine, western Ponderosa pine
Even though it seems to be straightforward to classify objects in terms of sets of subclasses at different levels, Croft and Cruse (2004: 147) rightly point out that ‘[t]axonomizing is clearly not simply a matter of dividing a larger class into smaller classes’. As they show, there are ‘logically impeccable subdivisions’ (Croft and Cruse 2004: 147) that do not form good taxonomies: (2)
? A stallion/mare/foal is a kind/type of horse. (A stallion is a horse.)
(3)
? A blonde/queen/actress is a kind of woman. (An actress is a woman.)
202
Walter Bisang
To account for this fact, Cruse (1994) introduces the notion of perspective – taxonym and hyperonym must share the same perspective. (2’)
stallion is not a good taxonym of horse because it has a sexual perspective that horse has not
(3’)
blonde is not a good taxonym of woman because it takes the perspective of ‘hair-colour’, and woman does not.
Later, Croft and Cruse (2004: 149) introduce the notion of focal orientation: ‘a good taxonomy must have at its core a specification of the core of the hyperonym’ (Croft and Cruse 2004: 150). As the authors admit themselves, this solution ‘can at present only be offered tentatively’ (Croft and Cruse 2004: 150). In spite of these difficulties in providing a semantically convincing definition, the rough definition in terms of subcategories is met quite frequently in systems of nominal classification. However, the extent to which these criteria exactly match with the biological criteria suggested in (1) remains to be seen from detailed anthropological studies on how objects are classified in different cultures (cf. the two papers of Ellen, this volume).
3 Classifier systems 3.1 Introduction Classification is represented in various types of noun categorization devices. Aikhenvald (2000a) suggests the following seven types: i. gender and noun-class systems ii. noun classifiers iii. numeral classifiers iv. classifiers in possessive constructions v. verbal classifiers vi. locative classifiers vii. deictic classifiers Due to limitations of space, the present paper will only discuss the first four types. Moreover, types (vi) and (vii) are cross-linguistically rare. Gender/noun-class systems will be discussed in subsection 3.2, noun classifiers in subsection 3.3, numeral classifiers in subsection 3.4 and possessive classifiers in subsection 3.5. Finally, subsection 3.6 will address the issues of languages with more than one classifier system and problems of clearly defining different classifier types.
Classification between Grammar and Culture
203
3.2 Gender and noun-class systems Systems of gender or noun class differ from the other noun categorization devices by agreement (Corbett 1991). Thus, the nominal class to which a noun belongs must be marked on other syntactic units like demonstratives, articles, numerals, adjectives, personal pronouns, relative pronouns and on the verb in the case of subject (or object) agreement. In the following example from German, the indefinite article and the adjective agree with the gender of the head noun (Mann ‘man [M]’, Frau ‘woman [F]’ and Mädchen ‘girl [N]’): (4) a.
ein-ø ART:INDEF:M:SG ‘a clever man’
klug-er clever-M:SG
Mann man:M:SG
(4) b.
ein-e ART:INDEF:F:SG ‘a clever woman’
klug-e clever-F:SG
Frau woman:F:SG
(4) c.
ein-ø ART:INDEF:N:SG ‘a clever girl’
klug-es clever-N:SG
Mädchen girl:N:SG
As the name implies, gender systems are based on the distinction between male and female and they are often supplemented by a neutral class for objects that belong to neither of these two classes. The semantic transparency of gender assignment is subject to enormous cross-linguistic variation. Tamil (Dravidian) has a very transparent system with a male class for nouns denoting males and gods, a female class for women and goddesses and a neutral class for all other nouns (Corbett 1991: 8–10). In most systems, gender is assigned additionally by morphological and phonological criteria. A good example of a language whose classification is based extensively on phonology is French (Tucker, Lambert and Rigault 1977; Corbett 1991: 57–62). The term ‘noun class’ is used for agreement-based classification that is not determined by gender. The proto-typical example is found in Swahili and many other Bantu languages. In Swahili, nominal class is expressed by prefixes on the noun1 and by agreement within the nominal phrase (demonstratives, numerals, adjectives, possessive pronouns, relative clauses) as well as on the verb (subject and object agreement). For the purpose of comparative work on Bantu noun-class systems, the individual classes are numbered (Guthrie 1967–1971, Welmers 1973: 159–183, Katamba 2006).
1 Notice that there is a considerable number of nouns with no class prefix. In these cases, the noun class is assigned lexically.
204
Walter Bisang
Thus, class 1 refers to humans in the singular, while class 2 refers to humans in the plural. In most cases, the classes have a singular form represented by odd numbers and a plural form represented by an even number. Since the number of classes varies significantly across the Bantu family, not every number is represented in every language. In Swahili, we find the classes given in Table 1. In some cases, the semantic criteria for class assignment are pretty consistent. Good examples are the classes 1 and 2 for humans and 7 and 8 for things. The classes 16 to 18 represent a characteristic property of Bantu languages. While locatives are expressed by prepositions like at, in, etc. in English, Bantu locatives are expressed by the noun-class system (cf. Bresnan 1994, Demuth 2000). In the remaining classes, the criteria are more complex. The descriptions given in Table 1 are meant as a rough survey on the classes found in Swahili with the class prefixes as they occur on the noun:
Class number Prefix
Criterion
Examples
1/2
m-; wa-
Human
m-tu ‘man’, wa-tu ‘men’
3/4
m-; mi-
Trees, plants
m-ti ‘tree’, mi-ti ‘trees’
5/6
ø/ji-; ma-
Various
ji-we ‘stone’, ma-we ‘stones’ ø-tunda ‘fruit’, ma-tunda ‘fruits’
7/8
ki-; vi-
Things
ki-tu ‘thing’, vi-tu ‘things’ ki-tanda ‘bed’, vi-tanda ‘beds’
9 / 10
N-/ø-; N-/ø-
Various
n-dege ‘bird(s)’, m-bu ‘mosquitos’ rafiki ‘friend(s)’
u-
Abstracts, a collection of very s mall objects, Concrete objects
u-kubwa ‘size’, u-dogo ‘smallness’ u-nga ‘flour’, u-dongo ‘soil’ u-bao ‘blank’, u-fagio ‘broom’
16 17
paku-
18
mu-
Position at a place, Movement to/from a place, Position in a place
11 / 14
Table 1: Noun classes of Swahili The following example illustrates the agreement system of Swahili. It is needless to say that this is a constructed example that shows extensively how noun class is marked on the verb and on different constituents of the noun phrase:
(5)
Classification between Grammar and Culture
205
Swahili (Allan 2001: 114): vi-su vi-dogo vi-wili hi-vi amba-vy-o ni-li-vi-nunua ni 8-knife 8-small 8-two this-8 REL-8-REL 1SG-PST-8-buy be vi-kali sana. 8-sharp very ‘The two small knives which I bought are very small.’
Gender and noun-class systems are often also involved in evaluating the nouns they mark. In many Bantu languages, this is expressed by a change from the default class prefix to another class. Thus, swapping from the regular noun class to classes 7 / 8 (ki-; vi-) in Swahili is associated with smallness (diminutives), while changes to the classes 3 / 4 (m-; mi-) and 5 / 6 (ø/ji; ma-) induce interpretations in terms of bigness (augmentatives) (cf. Shepardson 1982: 57). Good surveys of gender/noun class and evaluatives are given in Körtvélyessi (2015), Grandi and Körtvélyessi (2015) and Orozco (2016).
3.3 Noun classifiers Noun classifiers are the second type of classifiers to be discussed in this section. They co-occur with a noun and have scope over the whole noun phrase. In contrast to the other types of classifiers, their occurrence does not depend on the presence of another constituent (Aikhenvald 2000a: 81). They typically express the grammatical functions of (i) definiteness, (ii) reference tracking2 by personal pronouns of the third person and (iii) relative-clause marking3. Noun classifiers are found very frequently in Australian languages, in some Mesoamerican languages (Maya family) and in some subfamilies of Austronesian (West Austronesian, Oceanic). Their function will be illustrated by examples from Jacaltec, a Maya language spoken in Guatemala, whose system of 24 noun classifiers is excellently described by Craig (1986a, b, 1987; for some details on how this system functions, cf. section 5). In (6), the noun classifiers ix for ‘female non-kins’ and ixim for ‘corns’ express definiteness. As the reader will see immediately, the classifier for female non-kins is the same as the noun ix ‘girl’. This shows that at least some noun classifiers are derived from common nouns.4 (6)
Jacaltec (Craig 1986a: 264): swatx’ ix ix ixim made CL girl CL ‘The girl made the tamales.’4
b’itx. tamale
2 On the term ‘reference tracking’, cf. section 4.3. 3 They also occur in some specific contexts of nominal possession, for details, cf. Craig 1986b. 4 Tamales are a Mexican dish.
206
Walter Bisang
The next example shows the use of noun classifiers in the function of pronouns. In (7a), the proper name xuwan ‘John’ and the noun lab’a ‘snake’ are first introduced by their corresponding classifiers naj and no7, respectively. In (7b), the bare classifiers are used as pronouns: (7) a.
Jacaltec (Craig 1986a: 264; Grinevald 2000: 65): xil naj xuwan no7 lab’a. saw CL John CL snake ‘John saw the snake.’
(7) b.
xil naj saw CL ‘He saw it.’
no7. CL
3.4 Numeral classifiers In languages with numeral classifiers, numerals cannot be counted directly as in English (e.g. three letters, five cars, seven apples, etc.). As in the case of the following example from Chinese, they need an additional marker, which is the numeral classifier feng in the context of the noun xin ‘letter’: (8)
Mandarin Chinese: san feng three CL ‘three letters’
xin letter
Numeral classifiers are found very frequently in the languages of East and mainland Southeast Asia (Sinitic [Chinese], Japanese, Korean, Ainu, Tai-Kadai [Thai], MonKhmer, Hmong-Mien; cf. Bisang 1999) and in the languages of Oceania. Moreover, we find numeral classifiers in numerous languages of Mexico and Central America as well as in languages of South America (for a more detailed account, cf. Aikhenvald 2000a: 121–124). Most languages with numeral classifiers have significantly more than one classifier. The main criteria for selecting a classifier are animacy (humans, animals, plants), shape (dimensionality), size, instrument or man-made object, location and additional classifiers with highly specific functions. The following list of some of the most important classifiers in Mandarin Chinese illustrates these functions. As one can also see from the translations given in this list, some classifiers also have a concrete meaning or are historically derived from a word with such a meaning. Moreover,
Classification between Grammar and Culture
207
one and the same classifier may have additional functions that go beyond the general functions of animacy, shape etc. mentioned in the list. (9)
List of some classifiers in Mandarin Chinese:5 Animacy: Animals: zhi ‘animals and moving objects’, tou ‘head’ for ‘hearded cattle and garlic’, kou ‘mouth’ for ‘cattle but also wells and languages’, wei ‘tail’ for ‘fish’. ke ‘trunk, stem’ for ‘trees and plants’, duo ‘flower’ for ‘flowers Plants: and clouds’. Shape: One-dimensional: tiao ‘string, strip’ for ‘long, flexible objects’, zhi ‘branch, stylus, rod’ for short objects like ‘fountain pens’ and ‘candles’. Two-dimensional: zhang ‘extend’ for objects with a surface ‘table, painting, sheat of paper’. Three-dimensional: kuai ‘lump, chunk’ for objects like ‘earth/land, soap, wristwatch, money/currency units’. Size: Small objects: li ‘grain’ for ‘grain-shaped objects’, mei ‘stalk, stem’ for ‘coins and other small objects’. Instrument (man-made objects): ba ‘handle, handful’ for ‘objects with a handle but also for chopsticks and a handful of small objects’, jian for ‘garments above the waist and for social matters and occasions’, jia ‘frame, construction’ for ‘large engines, e.g. airplanes’, tai ‘platform’ for ‘electronic equipment (e.g. computers, tv sets) and theatrical performances’. Place: suo ‘place’ for buildings, jian ‘space, interval’ for ‘rooms’, zuo ‘seat’ for ‘big locations like mountains, palaces and large Buddha statues’. Special classifiers: feng ‘seal’ for ‘letters, postal items’, ben for ‘books’, zun ‘for pieces of artillery and Buddha statues’, pi for ‘horses and other hoofed animals’, liang for ‘vehicles’, shou ‘head’ for ‘poems’, zhan ‘bowl’ for ‘lamps’.
In addition to these classifiers, Chinese also has a general classifier ( ge) which can replace most of the classifiers listed in (9) in spoken language. In written or more formal language, there is a more elaborate set of classifiers. The general classifier is also the default classifier for nouns denoting humans ( san ge ren [three CL man] ‘three men’). The rigidity of classifier assignment to individual nouns varies cross-linguistically. A particularly rigid system is found in Thai (Hundius and Kölver 1983). Chinese is more flexible at least with some nouns. In most cases, flexibility points out a specific property of the object to be classified that matters for the discourse situation (Ahrens 1994). Thus, a noun like snake can be marked by the classifier for animals ( zhi) or for one-dimensional objects ( tiao) if it moves and thus appears in its full length.
208
Walter Bisang
For nouns denoting humans, it is often possible to use quite a number of classifiers for highlighting specific properties as is illustrated by the following example with jiangjun ‘general’ and the classifiers ge ( ), wei ( ), ming ( ) and yuan ( ).5 (10) a.
Mandarin Chinese (Zhang 2007: 53): yi ge jiangjun one CLgeneral general ‘a general’
b.
yi wei jiangjun one CLpolite general ‘an [honorable] general’
c.
yi ming jiangjun one CLname general ‘a general [as an item in a list of generals]’
d.
yi yuan jiangjun one CLrank general ‘a general [highlighting his military rank]’
3.5 Classifiers in possessive constructions Nominal possession is associated with no less than three different types of classifiers which are defined by the three components that constitute a nominal possessive construction, i.e. a possessor (the owner), a possessed (the thing owned) and a relation of possession in a more or less metaphorical sense between these two components (Aikhenvald 2000a: 125 f.): (11)
POSSESSOR John
Relation ’s
POSSESSED car
In the case of (11), the possessor is John and the possessed is car. The relation of possession is expressed by the genitive ’s (of course, there is also the other option of genitive formation with the preposition of as in the car of John). Depending on the component on which the classification operates, we get the following three types of possessive classifiers:
5 This list summarizes descriptions found in the grammars written by Chao 1968; Li and Thompson 1985; Wiedenhof 2015.
Classification between Grammar and Culture
209
–– Relational (possessive) classifiers –– Possessed classifiers –– Possessor classifiers While relational possessive classifiers and possessed classifiers are relatively frequent, possessor classifiers are very rare in the population of the world’s languages. For that reason, the present paper only introduces the first two classifier systems.6 Relational classifiers minimally distinguish between the two types of alienable vs. inalienable possession (Lévy-Bruhl 1914, Bally 1926, Seiler 1983, Lichtenberk 1983, Chappell and McGregor 1996). Inalienable possession is characterized by its permanence as it typically shows up in kinship relations or body-part relations. Thus, the relations expressed in his mother or her head are inherently related to the possessor even in the unlucky case that the possessor ceases to have the thing possessed through death or an accident. In contrast, alienable possession is not permanent and can be transferred or disposed of as in the case of John’s car. In Fiji (Austronesian; Schütz 1985: 445–466 and Dixon 1988: 136–139 on Boumaa Fiji), inalienable possession is expressed either by a suffix (pronominal possession) or by a possessive marker i (nominal possession): (12) a.
Fiji: Inalienable possession (Schütz 1985: 445): na ulu-na b. na ulu i Jone DEF head-POSS:3.SG DEF head POSS John ‘his/her head’ ‘John’s head’
Alienable possession is expressed by an additional set of four classifiers which specify the semantic relation between the possessor and the possessed. The first two classifiers express a relation of incorporation between possessor and the thing possessed either in terms of drinking (classifier me-) or eating (classifier ke1-). The classifier ke2-7 is used in cases in which the possessor has no control over initiating or terminating the relationship. Finally, there is the classifier no- which is employed for other types of relations which are characterized by the fact that the possessor can control the thing possessed:
6 On possessor classifiers, cf. Aikhenvald 2000a: 139 f. and the literature mentioned there on the two languages of Dâw and Hupda (Makú family, spoken on the Brazil/Colombia border). 7 Geraghty 1983: 249 argues that ke1- and ke2- are derived from two different lexical meanings of kana. Kana1 means ‘eat’ and generates ke1- for expressing an eating relation between possessor and possessed, while kana2 with the meaning of ‘suffer, undergo’ developed into ke2- for marking the absence of control by the possessor.
210 (13)
Walter Bisang
Relational classifiers in Fiji: methe possessed can be drunk (i.e., possessor drinks possessed) ke1the possessed can be eaten (i.e., possesssor eats possessed) ke2the possessor has no control over the possessed noall other relations between possessor and possessed which are [+control]
The pronominal possessive marker as well as the possessive marker i are suffixed to the classifier. This is illustrated with the classifier ke1- (possessor eats the possessed) in the following example: (14) a.
Fiji: Inalienable possession (Schütz 1985: 450): na ke-na uvi b. na dalo ke-i Toomasi DEF CL-POSS:3.SG yam DEF taro CL-POSS Thomas ‘her/his yam (for eating)’ ‘the taro of Thomas’
The next three examples illustrate the use of the other relational classifiers. Example (15) illustrates a drinking relationship. The difference between ke2- for relations with no control and no- for relations with control is illustrated by the noun vacu ‘punch’8. If the possessor is the agent who gives the punch to somebody else, the classifier no- is used (16a). The classifier ke2- is used if the possessor received the punch as a patient (16b): (15)
Fiji (Schütz 1985: 450): na tii me-i DEF tea CL-POSS ‘Paula’s tea’
(16) a.
Fiji (Schütz 1985: 462): no-mu i-vacu CL-2.SG punch ‘your punch (you give)’
Paula Paula
b.
ke2-mu i-vacu CL-2.SG punch ‘your punch (you receive)’
In Kipeá-Karirí, a North Arawak language spoken in the Amazonian Area of Brazil, the relation between possessor and possessed is defined by the way in which the possessor treats the possessed. Aikhenvald (2000a) lists following twelve classifiers:
8 Notice, that the details can be much more complex, cf. Schütz 1985: 445–464.
(17)
Classification between Grammar and Culture
211
List of relational possessive classifiers in Kipeá-Karirí (Aikhenvald 2000a: 135): Classifier Semantics uaprú gathering of wild plants enkí raising of animals uanhí cultivation of Manioc udjé cultivation of other plants udé boiling upodó roasting ubó maturation at home uitó finding ukisí sharing ubá gift from outsiders boronumú booty e transportation, carried goods
Depending on the noun, the possessive classifier may vary in accordance with the activity that is relevant in a given discourse situation. Thus, the noun sabuka ‘fowl’ can be found with the following classifiers: (18)
Kipeá-Karirí (Aikhenvald 2000a: 135): a. dz-upodó do sabuka b. 1.SG-CL:roast POSS fowl ‘my fowl (roasted)’ c.
dz-ukisí do sabuka 1.SG-CL:share POSS fowl ‘my fowl (that was my share)’
d.
dz-udé do sabuka 1.SG-CLboil POSS fowl ‘my fowl (boiled)’ dz-ubá do sabuka 1.SG-CL:gift POSS fowl ‘my fowl (that was given to me)’
In systems of possessed classifiers, the selection of the classifier is determined by the semantic properties of the possessed noun. This system is extensively developed in Panare, a Carib language spoken in southern Venezuela (Mattéi-Muller 1974; Carlson and Payne 1989; quoted from Aikhenvald 2000a: 127 f.). There are no less than 21 classifiers which characterize the possessed according to the criteria of shape, structure and consistency. Three classifiers are combined only with a single noun (single-item classes) and another four classifiers have the same form as the noun they occur with (repeaters).
212 (19)
Walter Bisang
List of possessed classifiers in Panare (Aikhenvald 2000a: 128): Semantics Classifier Examples General iyu Soap, gasoline, hammer, paper, table, etc. Edible 1 yung Food items in paste form: mango, corn (mash), manioc Edible 2
tä’ma
Edible 3 Edible 4 Drinkable/liquid Animal
empa yo’/are uku yiki
Vehicle Hunting arms Musical instrument Body paint Clothing Container Artificial light Single item classes
kanowa ko ntyën yanoë po’ mara’pi uyung ëwi´ tipi’ ichi’ pata chistë wata wa’to
Repeaters
Something softer than yung: egg, rice, soup Fruit with its pulp, not squeezed out Meat (cooked or raw): fish, chicken, beef Liquids: coffee, milk, blood Live animals, including domesticated ones Canoe, truck, helicopter, car Spear, arrow, harpoon Flute, guitar, violin Body paint Necklace, shirt Gourd, etc. Lantern, flashlight House Swidden garden Hammock Village, community Hatchet Blowgun Fire
The following example illustrates the use of the classifier -uku- for liquid entities: (20)
Panare (Aikhenvald 2000a: 128): y-uku-n wane 1.SG-CL:liquid-GEN honey ‘my honey (mixed with water for drinking)’
Classification between Grammar and Culture
213
3.6 Languages with more than one classifier system and questions of how to distinguish different classifier types The description of classifier systems as they are discussed today in linguistics does not stop at the description of its different types. One additional topic in the typology of classifiers starts out from the observation that many languages have more than one classifier system. Another observation is that it is often difficult to define clear-cut criteria for distinguishing the different types. The presence of more than one system in a language manifests itself in two different ways. There are multiple classifier languages in which a single set of markers covers the functions of more than one type of classification and languages with different types/sets of classifiers (the terminology is from Aikhenvald 2000: 184–241). Hmong, a Hmong-Mien language spoken in the North of Thailand, is a good example of a multiple classifier language. This language has only one set of classifiers which can be used in the functions of noun classifiers as markers of definiteness (21), numeral classifiers (22) and possessed classifiers (23). Finally, they also occur with demonstratives (24) (Bisang 1993)9: (21)
Hmong: mas tus tsov txawm ya then CL tiger then fly ‘Then the tiger arrived swiftly, (…)’
(22)
Hmong (Bisang 1993: 27, from Mottin 1980): rau tus ntxhais six CL girl/daughter ‘six girls/daughters’
(23)
Hmong (Bisang 1993: 27, from Mottin 1980): nws (tus) ntxhais he CL girl/daughter ‘his daughter’
(24)
Hmong: mus xyuas lub tsev ub, go CL CL house this ‘to visit this house and that house’
ceev ceev swiftly
lub CL
los … come
tsev no house that
9 The classifier tus occurs with animate nouns, including humans, while the classifier lub is used with inanimate nouns and with three-dimensional objects.
214
Walter Bisang
Tariana, an endangered language of the Maipurean family (Northern Arawak) spoken along the Vaupés river in the Amazonian area of Brazil, can be situated somewhere between a multiple classifier language and a language with different types/sets of classifiers. Classifiers are used in the following grammatical environments: noun class/gender agreement on adjectives, numerals (between one and four), possessed nouns, verbs, demonstratives, articles and a few other modifiers which form closed classes. Moreover, they are used as derivational suffixes. These environments are filled by three types of classifiers: (i) a set of classifiers that uses different forms for each of these functions and thus may be treated as different sets, (ii) a set with the same form that is limited to the functions of noun classes and derivational suffixes, and (iii) a set of classifiers with the same form in each environment (cf. Aikhenvald 2000a: 235; also cf. Aikhenvald 1994; 2000b). To provide the reader with at least one example, (25) illustrates the use of the type (iii) classifier -dapana for habitats in combination with a demonstrative, a numeral, a possessive, an adjective and the verb:1011 (25)
Tariana (Aikhenvald 2000a: 204): ha-dapana pa-dapana na-tape-dapana na-ya-dapana DEM:INAN-CL one-CL 3.PL-medicine-CL 3.PL-POSS-CL hanu-dapana heku na-ni-na-dapana-mahka. big-CL wood 3.PL-make-VOICE10-CL-RECENT.PST11 ‘This one big hospital of theirs has been made of wood.’
The language that is reported to have the most elaborate system of different classifier types is Palikur, another member of the Maipurean family. It is described as follows by Aikhenvald (2000a: 192 f.):12 Palikur has genders (with two or three agreement forms depending on the construction type), and the following classifier types: numeral classifiers; verbal classifiers with two subtypes— the ones used with stative verbs and the ones used with transitive verbs …; locative classifiers …; and possessive classifiers (which combine properties of relational and of possessed classifiers …). Only numeral classifiers, the two subtypes of verbal classifiers, and locative classifiers show some similarities—i.e. the same forms are used in some instances.
The increasing number of data on more complex systems in languages with more than one classifier type shows that the distinction of the individual types of classifiers is not always clear. This is particularly evident if we compare multiple-classifier systems as the one presented in example (25) from Tariana with highly developed systems of
10 The suffix -ni is described by Aikhenvald 2000a: 204 as a topic-advancing voice marker. 11 The suffix -mahka marks an event that has taken place in the recent past and was not seen by the speaker her/himself. 12 For further details, the reader is referred to Aikhenvald and Green 1998.
Classification between Grammar and Culture
215
gender/noun-class agreement as we find them in example (5) from Swahili. Even if there may be some criteria for distinguishing the two types of systems13, the surface form may often look rather similar. In the case of (5) and (25), different components of the noun phrase take the same or almost the same marker (vi-, -vi and -vy- in Swahili, -dapana in Tariana) with different types of noun modifiers, irrespective of whether this is in an agreement system as in Swahili or in a language that combines different classifier systems as in Tariana. In a more recent paper, Corbett and Fedden (2016) have tried to resolve the problem of distinguishing gender/noun class systems from other classifier systems from the perspective of canonical morphosyntactic features (Corbett 2012). With this approach, it is possible to measure the degree to which a given classifier system deviates from a canonical or prototypical gender/noun class system. To what extent this approach will be able to come up with typologically relevant solutions depends on the quality of data from a larger number of languages like Tariana. At the moment, there do not seem to be enough data (and maybe enough languages) for implementing this model. Another approach developed by Passer (2016) has come up with a remarkable difference between gender/noun class systems and the other classifier systems (noun classifiers, numeral classifiers, possessive classifiers, multiple classifiers). Passer (2016) distinguishes two types of criteria for measuring the degree of grammaticalization, i.e. criteria of linguistic form and criteria of semantic transparency. Based on a sample of some 40 languages, he presents convincing evidence that gender/noun class systems are characterized by a uniformly high degree of morphosyntactic integration on the form side, while their semantic side reaches from high arbitrariness to full transparency. In contrast, the other classifier types are semantically more transparent and they show less formal integration. As will be discussed in subsection 4.4, this has to do with the use of classification for marking agreement.
13 Some criteria are mentioned in Bisang 2002. Among them: (i) In languages with more than one system, the selectional criteria tend to be more semantic than form-based (morphological, phonological), (ii) in languages with more than one classifier system, there may be nouns which do not undergo classification and (iii) gender/noun-class agreement is obligatory, while other types of classifiers may be optional.
216
Walter Bisang
4 Motivations of using classifiers in grammar 4.1 Introduction The description of the different types of classifier systems shows that classification is used for an impressive number of different grammatical functions. The following list summarizes the functions discussed in section 3: –– Counting (with numerals and other quantifying expressions): numeral classifiers –– Definiteness: Noun classifiers, numeral classifiers –– Reference tracking (personal pronouns): noun classifiers –– Possession: Possessive classifiers, numeral classifiers –– Relative clauses: noun classifiers, numeral classifiers –– Agreement: gender/noun class markers As will be argued in this section, these functions are not accidental, they are motivated by three cognitive tasks for which classification can be used, i.e. individuation/ atomization, identification and processing (Bisang 2002). Looked at from the perspective of these cognitive purposes, these three functions are involved with the following functions expressed by classifiers: –– Individuation: Counting –– Identification: Definiteness, reference tracking, relative clauses, possessed classifiers –– Processing: Agreement Each of these tasks will be described individually in the remainder of this section. Subsection 4.2 will be on individuation, subsection 4.3 on identification and subsection 4.4 on processing.
4.2 Individuation The function of individuation is based on the notion of transnumerality (Greenberg 1972)14, i.e. the absence of a number distinction (singular vs. plural) in the lexical meaning of nouns in languages with numeral classifiers. A noun like Chinese xin only refers to the concept of ‘letter(s)’ with no distinction between the singular ‘letter’ or the plural ‘letters’. For that reason, the concept expressed by a noun must be individuated before it can be counted. This is the function of the numeral classifier, which provides the concept expressed by the noun with the necessary conceptual boundary
14 The notion of transnumerality is taken up by a large number of typologically oriented papers (among them Hundius and Kölver 1983, Bisang 1999).
Classification between Grammar and Culture
217
by which it can be perceived as a countable entity. In Chierchia’s (1998) formal semantic approach, nouns in numeral classifier languages are all mass nouns like English furniture. Nouns of this type cannot be counted directly in English (*two furnitures is ungrammatical), they need a classifier like piece for that purpose (cf. two pieces of furniture). While nouns like furniture are rare in English, Chinese nouns are supposed to behave generally like the English word furniture (for a different view, cf. Cheng and Sybesma 1999). In Chierchia’s (1998) approach, the function of numeral classifiers is atomization. In his analysis, mass nouns like furniture are inherently plural and do not need additional plural marking. If they have to be counted they have to be atomized, i.e., they have to be singled out as individual atoms by a numeral classifier. The process of singling out individual entities is supported by classification. Irrespective of whether one calls this process individuation or atomization, numeral classifiers make nouns accessible to counting by highlighting some of their inherent properties or conceptual boundaries as they are reflected in their classificational semantics (Bisang 1999). In other words, the classificational properties of an individual classifier point out the conceptual boundaries of a concept by which it can be perceived as a discrete entity that can be counted.
4.3 Identification The relation between classification and the task of identification is defined by what I would like to call ‘restriction of the search domain’ in a given context (for an earlier version, cf. Bisang 1999). More specifically, the classificational semantics of a classifier (shape, animacy, etc., cf. section 2) limit the kind of objects a speaker may refer to in her or his utterance to those objects whose corresponding nouns are compatible with the classifier. If in the process of speech production the hearer reaches the position of the classifier in an utterance s/he will already be able to reduce the number of potentially relevant nouns even before the relevant noun itself will be mentioned. Thus, if a Chinese speaker hears the sequence san zhang piao [three CL: two-dimensional ticket] ‘three tickets’, s/he will anticipate that what follows after zhang must belong to the set of entities that are compatible with that classifier. If a noun can occur with two or more classifiers, classifier selection even supports the identification of an object in a given situation. Thus, the Chinese noun for snake can occur either with the classifier zhi ( ) for animals or with the classifier tiao ( ) for onedimensional objects. Depending on the concrete form of the snake in the discourse situation, the speaker may select the classifier tiao if the snake moves and thus appears as a one-dimensional object, or s/he may select the classifier for animals zhi if the snake presents itself in a less salient form (Ahrens 1994). From a more general perspective, the occurrence of the classifier increases the predictability of a nominal concept in ongoing speech production—it reduces the search domain and thus enhances the identifiability of a nominal concept.
218
Walter Bisang
The relevance of classification for restricting the search domain is confirmed by various psycholinguistic studies that address issues of the cognitive organization of information and its retrievement from memory. They show that classification in fact does delimit the number of potential candidates in experiments of naming or recognizing objects. In that sense, classification reduces the search domain for candidate objects and thus contributes to their identification. Most of these studies focus on grammatical gender (Friederici and Jacobsen 1999) and its effects on picture naming (Jacobsen 1999, Bentrovato et al. 1999), the recognition of words (Grosjean et al. 1994) and the identification of an object among a set of objects depending on whether French words are preceded by a gender-marked article (le [masc.] vs. la [fem.]) or a non-gender-marked plural article (les [masc/fem]) (Dahan 2000). In addition, gender marking shows effects in the resolution of pronominal reference (he, she, it) (Garnham et al. 1995). While the above studies focus on gender, there are other studies dealing with the cognitive consequences of speaking a language with numeral classifiers (Zhang and Schmitt 1998; Gao and Malt 2009; Kuo and Sera 2009; Sera et al. 2013). Of particular interest is the study by Srinivasan (2010), which shows how classification is used for deciding whether an object belongs to a specific search domain. The search domain was defined by a visual search task that consisted in identifying different objects within a given search array. Participants had to count target objects that were presented on a display of pictures that included distractor objects. There were 80 displays arranged in two types of 40 sets. In the first set of displays, the distractor objects had the same classifier as the target object (same classifier trials), in the other set, the distractors belonged to different classifier sets (different classifier trials). Each display contained between eight and eleven targets plus eight to eleven distractors. The participants were native speakers of Chinese, Russian or English who had to decide whether the number of target objects on the display indicated by the experimenters was correct. The result showed that there was a greater difference in reaction time for Chinese speakers between the same and different classifier trails than for Russian and English speakers (Srinivasan 2010: 184). Moreover, the same-classifier effect was greater for Chinese speakers than for Russian and English speakers (Srinivasan 2010: 185). On the whole, ‘classifier membership predicted Mandarin speakers’ performance more than it did the search performance of Russian and English speakers’ (Srinivasan 2010: 186). Srinivasan’s (2010) experiment thus provides some good evidence that the task of identifying objects by restricting them to the domain that needs to be counted really depends on classification as it operates with numeral classifiers. Identification is basic for many grammatical categories, among them definiteness, reference tracking, possession and relative clauses (Bisang 2002). Even though the definition of definiteness is a controversial issue (cf. Lyons 1999 for an excellent survey), it is uncontroversial that identification plays a major role in its definition. For some languages like Chinese, the possibility of identifying a concept is the crucial factor that determines the marking of definiteness by a numeral classifier
Classification between Grammar and Culture
219
(Li and Bisang 2012). In more general terms, identification is associated with the hearer’s assessment of the extent to which a given concept is activated in the hearer’s mind (Lambrecht 1994). Thus, the speaker selects a definite article in a context in which she/he assumes that a given concept can be identified by the hearer. This is even true in examples in which a concept like beer was not mentioned before but can be identified by the previous context as the beer that belongs to the party: (26)
The party was nice but the beer was bad.
Reference tracking is another important domain in which classification serves the purpose of reducing the search domain. The term of reference tracking covers all the linguistic strategies which make sure that a given noun with its referent can be clearly identified through extended passages of discourse. Very frequently, a noun will be taken up in later strings of discourse by pronouns as in the following example: (27)
John and Mary went to the shopping mall. She bought new shoes and he bought a new t-shirt.
This example shows how the gender of the pronoun contributes to the unambiguous identification of who bought what. The pronoun she reduces the contextually given search domain of [John and Mary] to Mary, who is the only referent belonging to the class of feminine elements in (27) and thus qualifies as the subject of bought new shoes. Simple examples like (27) may give an idea of how important the class membership of a noun is in more complex scenarios of reference tracking. Relative clauses are associated with classification in two ways. The first way is tightly linked to reference-tracking. Relative clauses do not only modify their head noun, they are also characterized by the fact that the head noun must play a semantic role within them (cf. e.g. Lehmann 1984). This second condition is important for reference tracking because grammar must offer a device by which the head noun gets assigned the right semantic role within the relative clause and class membership can be an important criterion for making clear which role the head noun plays in the relative clause. In various Indo-European languages, this function of classification is expressed by gender agreement on the relative pronoun (often in combination with number and case). The second way has to do with the fact that the modification of a noun automatically reduces the search domain for its potential referents. Thus, the number of potential referents of the noun book gets significantly reduced if it is modified by a relative clause like which I bought. It is this reduction that may call for the clarification of the identifiability status of a noun modified by a relative clause in languages with noun classifiers and numeral classifiers. In this context, the function of the classifier is to identify the head noun for making it accessible to further modification.
220
Walter Bisang
Possession is not as prominently associated with identifiability as definiteness, reference tracking and relative clauses. Relational classifiers cannot contribute to the identification of their nominal components of possessed or possessor because they focus on the property of the relation between them. This is different with possessed classifiers. In their case, the motivation is the same as with modification in the case of relative clauses. The set of potential referents of the possessed noun is reduced by the modifying possessor noun and this invites identification by a classifier. Finally, possessor classifiers are cross-linguistically rare and need further research.
4.4 Processing Gender and noun-class systems are defined by agreement (cf. subsection 3.2) and many of them lack semantic transparency (cf. subsections 3.6 and 4.4). As I argued in Bisang (2002), the use of gender/noun classes with agreement has nothing to do with the cognitive-semantic properties of classification but rather with the purpose of agreement in language processing. It is of vital importance for the hearer to analyze the input of linguistic structures as quickly as possible for being able to follow track of an utterance and for understanding it within a short time window of a few seconds (short-time memory). For that reason, the human parser, i.e. the cognitive mechanism responsible for the detection of syntactically related components, aims at recognizing larger syntactic structures and their components as early as possible (Hawkins 1994). In an example like these two small knives which I bought from Swahili (cf. (5)), the parser will have to recognize that the components vi-dogo ‘small’, vi-wili ‘two’, hi-vi ‘these’ and the relative clause amba-vy-o ni-li-vi-nunua ‘which I bought’ form a coherent syntactic unit with their head vi-su ‘knives’. If each of the elements depending on the head noun (e.g. vi-su ‘knives’) is marked with a class marker that reflects the class of that head noun (vi-, -vi or -vy-), this significantly supports the speed and the accuracy with which the parser can decode the syntactic structure of an utterance and that is the advantage of agreement. Agreement is semantically arbitrary, i.e. it does not care about the meaning of its markers. What matters is only that the markers are reliably present for pointing out the relevant syntactic structures as coherent units. For that reason, agreement is not limited to classification. Other grammatical categories that are equally reliable and are used for the same purpose are number, case and person. In fact, these categories are more frequent in the agreement systems of the languages of the world than gender/noun class and they are often accumulated within a single marker. A good example are Latin nominal agreement markers. They combine number, case and gender in a single nominal marker (cf. amic-um bon-um [friend-SG.ACC.MASC good-SG.ACC.MASC] ‘a good friend’). This shows that there is no specific semantic motivation for the use of gender/noun class in agreement. What matters is that the gender/noun class system of a language includes the whole inventory of its nouns
Classification between Grammar and Culture
221
(exhaustiveness) and that agreement is marked obligatorily because only under this condition can gender/noun class markers operate as reliable indicators of a syntactic unit with its nominal head. The fact that the semantics of the marker is irrelevant for agreement markers under these circumstances also explains why the values for transparency vary so widely in languages with gender/noun class (cf. Passer 2016 as discussed in subsection 3.6). The classifier needs to be a reliable indicator of the class of the head noun on its modifiers and semantic transparency does not contribute to that property. No matter whether the gender/system is semantically transparent or arbitrary, it only needs to be exhaustive and obligatory.
5 Classification between universality and specificity A cross-linguistic view on grammatical systems shows that the general criteria of classification are recurrent in the population of the world’s languages. The criteria of animacy (animate, inanimate), shape (dimensionality), consistency, size and location (Allan 1977) are found very frequently in a large number of language families and classifier systems. Similarly, criteria of social function and physical or functional interaction (Denny 1976) are widespread. Given the general frequency of these criteria across cultures, language families, classifier systems and geographic areas, it is reasonable to assume that they are universal in the sense that they reflect deeper cognitive foundations of the human brain. Other criteria like alienability and the eating-relation or the drinking-relation are limited to possessive constructions (cf. subsection 3.5). In spite of this, their cross-linguistic and cross-cultural frequency indicates that they transcend individual cultural boundaries and reflect a more general/universal way of determining relations of possession. At a lower level, the details of how the universal distinctions are split up into more fine-grained categories leaves a lot of room for culture-specific criteria. This will be illustrated by the noun-classifier system of Jacaltec (cf. subsection 3.3) as it is described by Craig (1986a, b, 1987). Jacaltec has 24 noun classifiers whose selection criteria are determined by the way in which humans interact with objects in terms of Denny (1976). As Table 2 shows, twelve classifiers are concerned with social interaction (subsystem I) and another twelve are based on physical and functional interaction (subsystem II): Subsystem I: Social interaction Subsystem II: Physical/Functional interaction
222
Walter Bisang
1.
cumam
male deity
13.
no7
animal
2.
cumi7
female deity
14.
metx’
dog
3.
ya7
respected human
15.
te7
plant
4.
naj
male non-kin
16.
ixim
corn
5.
ix
female non-kin
17.
tx’al
thread
6.
naj ni7an
young male non-kin
18.
tx’ang
twine
7.
ix ni7an
young female non-kin
19.
k’ap
cloth
8.
ho7
male kin
20.
tx’otx’
soil/dirt
9.
xo7
female kin
21.
ch’en
rock
10.
ho7 ni7an
young male kin
22.
atz’am
salt
11.
xo7 ni7an
young female kin
23.
ha7
water
12.
unin
infant
24.
k’a7
fire
Table 2: Noun classifiers in Jacaltec (Craig 1986a: 266 f.)
The twelve classifiers of subsystem I reflect a remarkably complex system of social organization. At the top of the system (classes 1 and 2), we find the male high deities represented by the Christian God, the sun, lightning, measles and whooping cough (cu-man ‘our father’) and the female high deities represented by the Virgin Mary, the moon and the corn (cu-mi7 ‘our mother’). The classifier naj ‘male non-kin’ also includes ‘lower deities or spirits expressing human contingency and living among humans’ (Craig 1986a: 271). These male spirits stand for notions of disease, fight/war, sadness, poverty, fear and wind. The question of kinship and gender is of central importance in the Jacaltec classifier system and reaches from classifier 4 to classifier 11. It is irrelevant only at the highest level of social status (classifier 3 ya7 ‘respected human’) and at the lowest level (classifier 12 unin ‘infant’). The classifier ya7 stands for individuals who have earned personal high status in their community, while the classifier unin is used with infants whose status as full members of the community is not confirmed. The other classifiers of the domain of social interaction are defined by combinations of the criteria kinship, gender and age. On the whole, the twelve classifiers of subsystem I can be described as follows: One sees in the choices of classifiers the recognition of higher powers of god-like nature, and of the powers of both god-like and human nature, in a mixture of Christian and native worship typical of the Mayan communities. And at the level of the social organization of the community the noun classifier system appears to encode two important characteristics of the Jacaltec community, one being that the nuclear family is the basic social unit of the community, and the other that the social status of a person is a function of his or her personal worth and not of social class status. (Craig 1986a: 272)
Classification between Grammar and Culture
223
The twelve classifiers of subsystem II can be divided into two subtypes. There is a basic set of five general classifiers which cover large natural domains of human interaction: no7 ‘animal’, te7 ‘plant’, ch’en ‘rock’, ha7 ‘water’ and k’a7 ‘fire’. The remaining seven classifiers are specific classifiers. They cover specific objects such as metx’ ‘dog’, ixim ‘corn’, tx’al ‘thread’, tx’ang ‘twine’ and k’ab ‘cloth’. The semantic relation between the head noun and the noun classifier is either taxonomic or what Craig (1986a: 279) calls ‘chaining’, i.e. a relationship characterized by the primary substance of which the head noun is made of. The five general classifiers are in a taxonomic relation to the noun they classify. On a lower biological rank, the specific classifiers metx’ ‘dog’ or ixim ‘corn’ are also associated with taxonomic classification. Other specific classifiers such as tx’al ‘thread’, tx’ang ‘twine’ clearly highlight the material properties of the head noun and thus establish a chaining relation between the classifier and its head noun (tx’ang ‘twine’ refers to a product of the agave plant, tx’al ‘thread’ to a product of the cotton plant). The interactional aspect of classification shows up in both types of subsystem II classifiers. While the general classifiers refer to ‘ecologically and culturally important and common objects manipulated by the Jacaltecs, the specific classifiers identify objects of particular cultural significance to the Jacaltec community of speakers’ (Craig 1986a: 279). Of particular importance for the Jacaltec culture are material classifiers. As is shown very convincingly by Craig (1986a), the specific classifiers expressing a chaining relation classify exactly those objects of Jacaltec agricultural life, which are essential for the subsistence and the identity of the members of this culture. The classifier ixim ‘corn’ is a good example for the case of subsistence. It marks maize and wheat, the main foods of this culture. The classifiers tx’al ‘thread’ and k’ab ‘cloth’ are associated with identity. They refer to weaving as an activity that is typically performed by women. Objects marked by tx’al are produced by unmarried women and include head-bands and sashes, while k’ab is used with basic clothing for the community that is produced by mature women.
6 Classification and conceptualization Classification and conceptualization are strongly intertwined. This can be seen very clearly in the following two cases: –– Flexible classifier systems (in contrast to rigid systems) –– The historical/diachronic development of some classifier systems The fluidity between classification and conceptualization can be described from both sides. Classification can be synchronically used for conceptualization in the case of flexible systems and, in turn, conceptualization is often the diachronic source of classifier systems. Both sides will be briefly discussed in the remainder of this section.
224
Walter Bisang
In a rigid classifier system the lexicon fully determines the class-membership of a noun. While rigid systems are limited to classification because each noun is clearly assigned to only one classifier out of the set of available classifiers, the possibility to select a classifier from a set of options for an individual noun in flexible systems adds more specific information to the noun and thus may contribute to conceptualization. This can be illustrated by an example from Miraña, a Witotoan language spoken in the Colombian Amazon region (Seifart 2005). This language has four types of nouns. The one that is important for this paper are optionally classified nouns, which express a rather general meaning that is further specified by the classifier. A good example is the noun ‘banana’, which refers to the concept of ‘banana’ in a very broad sense. More specific concepts are generated by classifier selection. The classifier for three-dimensional oblong objects limits the reference of the noun to banana fruits (28a). The selection of other classifiers yields other more specific conceptualizations (28b–28f): (28)
Miraña (Seifart 2005: 114): a. b. banana-CL:oblong banana fruit’ d.
banana-CL:drinkable ‘banana drink’
e.
banana1D:pointed ‘banana plant’
banana-CL:bag ‘bag of bananas’
c.
banana-CL:bunch ‘bunch of bananas’
f.
banana-CL:powder ‘pulverized dry banana’
The above example from Miraña shows how the use of a particular classifier out of a set of possible classifiers in a flexible system specifies the concrete concept to which a given noun is meant to refer in a given situation of discourse. The diachronic perspective from conceptualization to classification is based on taxonomic word formation. In a large number of East and mainland Southeast Asian languages, at least some classifiers are clearly related to nouns referring to concepts of higher taxonomic levels (cf. (1); for a more detailed discussion, cf. Bisang 1999). Such nouns are used in word formation for creating compound nouns. In Thai, the noun ‘flower’ is the lexical basis for the formation of a large number of nouns denoting individual species of flowers or flower-like objects like [flowerrose] ‘rose’, [flower-lotus] ‘lotus’, [flower-poppy] ‘poppy’ and [flower-mushroom] ‘mushroom’. In contexts in which a classifier is needed the first element of these compound nouns is repeated in the classifier position (29). Moreover, the independent status of as a classifier is confirmed by the fact that it is also used with words in which it does not occur (cf. ‘key’ in (30)):
(29)
Classification between Grammar and Culture
Thai:
sǎam three
rose ‘three roses’ (30)
Thai: key ‘three keys’
sǎam Three
225
CL
CL
In Thai, this process of classifier formation is no longer productive. In fact, Thai has one of the most rigid classifier systems in which each noun is clearly associated with a single classifier. For that reason, the use of conceptualization for classification can only be reconstructed historically. In less rigid systems in which numeral classifiers are not fully developed, one can observe how both ends, classification and conceptualization meet somewhere in between. A good case in point is Vietnamese as described by Löbel (1996, 2000). Vietnamese has a set of ten well-developed classifiers: (31)
List of Vietnamese classifiers (Löbel 1996): cái ‘a non-living thing’ cây ‘a tree or plant, a stickshaped or plantlike object’ chiếc ‘individual item of floating object, vehicles’ con ‘a living thing that is not human’ hòn ‘a stone or stonelike thing’ quả ‘a fruit, a round, globular object’ quyển ‘a volume’ sợ ‘a hair, thraid, cord, etc.’ tấm ‘a flat piece of material’ tờ ‘a sheet of paper, document’
In addition to these classifiers, basically any noun can take on the function of a classifier if the speaker wants to be specific enough in a given context. An important factor that drives this system is taxonomy15. In Vietnamese, a head noun (N1) can be modified by another noun that modifies it (N2). Thus, a head noun N1 like cây ‘tree’ can be specified by another noun N2 like táo ‘apple’ in the form of cây táo ‘apple tree’. More generally, any head noun N1 that is taxonomically higher than another noun N2 following it can be analyzed as a classifier under specific pragmatic conditions in which N2 is profiled from the perspective of its superordinate concept. In the following
15 Another important factor is the part-whole relationship. In this case, the element denoting the part (i.e. the meronym) will be reanalyzed as a classifier, Löbel 1996; 2000.
226
Walter Bisang
examples, the noun cây ‘tree/plant’ in (32) is analyzed as a classifier of rau ‘vegetable’ in (33). In turn, the noun rau ‘vegetable’ in (33) is analyzed as a classifier marking cần ‘celery’ in (34): (32)
Vietnamese: một cái one CL ‘a tree/plant’
(33)
Vietnamese: một cây rau one CL vegetable ‘a vegetable (as a plant)’
(34)
Vietnamese: một rau one CL ‘one celery’
cây tree/plant
cần celery
In Vietnamese we can observe how conceptualization contributes to the development of a classification system but at the same time we can also see how that emerging system of classification is employed for conceptualization, i.e. the creation of semantically more specific concepts out of more general ones.
7 Conclusion Classification manifests itself in various types of classifier systems and is associated with an impressive number of grammatical categories (sections 2 and 3). As the present paper has shown, there is no unified motivation for the wide functional range of classification across the grammars of individual languages (section 4). The motivation for numeral classifiers is the need of singling out countable units in languages in which that information is not encoded in the lexicon. In functional-typological approaches, this is explained by the individuation of transnumeral nouns (Greenberg 1972; Bisang 1999). In Chierchia’s (1998) formal semantic view, nouns are inherently plural in numeral-classifier languages and must be atomized by the classifier for making them countable. The use of classification in the context of definiteness, reference tracking, possession and relative clauses is based on the restriction of the search domain and its relevance for identification. Finally, the use of classification in gender and noun-class systems has no semantic motivation. It is rather due to the positive effects of agreement for the human parser.
Classification between Grammar and Culture
227
The semantics of the criteria for classifying objects for grammatical purposes seem to be partly universal and partly language-specific and culture-specific. The higher levels of conceptual classification (Allen 1977) and social-interactive classification (Denny 1976) are universal in the sense that they operate through different types of classifier systems and across different language families and different geographic areas. Culture-specific features make their way into the grammatical system at the lower levels of classification. Interactions between classification and conceptualization show up in languages with flexible classifier systems and in the diachronic development of classifier systems. Of particular interest are languages like Vietnamese, which are at the crossroads of a diachronic development from conceptualization to classification and at a synchronic application of classification for the purpose of conceptualization.
References Ahrens, Kathleen. 1994. “Classifier production in normals and aphasics.” Journal of Chinese Linguistics 22.2: 202–247. Aikhenvald, Alexandra Y. 1994. “Classifiers in Tariana.” Anthropological Linguistics 36: 407–465. Aikhenvald, Alexandra Y. 2000a. Classifiers. A Typology of Noun Categorization Devices. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. 2000b. “Unusual classifiers in Tariana.” In Systems of Nominal Classification, edited by Gunter Senft, 93–113. Cambridge: Cambridge University Press. Aikhenvald, Alexandra Y., and Diana Green. 1998. “Palikur and the typology of classifiers.” Anthropological Linguistics 40.3: 429–480. Allan, Keith. 1977. “Classifiers.” Language 53: 285–311. Allan, Keith. 2001. “Review of: Aikhenvald, Alexandra Y. 2000: Classifiers. A typology of noun categorization devices.” Linguistic Typology 5.1: 111–116. Bally, Charles. 1926. “L’expression des idées de sphère personnelle et de solidarité dans les langues indo-européennes.” In Festschrift Louis Gauchat, edited by F. Frankhauser and J. Jud, 68–78. Aarau: Sauerländer. Bentrovato, S., Devescovi, A, D’Amico, S. and Bates, E. 1999. Journal of Psycholinguistic Research 28: 677–693. Berlin, Brent, Denise A. Breedlove, and Peter H. Raven. 1968. “Covert categories and folk taxonomies.” American Anthropologist 70: 290–299. Berlin, Brent, Denise A. Breedlove, and Peter H. Raven. 1973. “General principles of classification and nomenclature in folk biology.” American Anthropologist 75: 214–242. Bisang, Walter. 1993. “Classifiers, quantifiers and class nouns in Hmong.” Studies in Language 17: 1–51. Bisang, Walter. 1999. “Classifiers in East and Southeast Asian languages: Counting and beyond.” In Numeral Types and Changes Worldwide, edited by Jadranka Gvozdanovic, 113–185. Berlin: Mouton de Gruyter. Bisang, Walter. 2002. “Classification and the evolution of grammatical structures: a universal perspective.” Sprachtypologie und Universalienforschung (STUF), 55.3: 289–308. Bresnan, Joan. 1994. “Locative inversion and the architecture of universal grammar.” Language 70: 72–131.
228
Walter Bisang
Carlson, Robert, and Doris Payne. 1989. “Genitive classifiers.” In Proceedings of the Fourth Annual Pacific Linguistics Conference, edited by Robert Carlsson, Scott DeLancey and Gildea, Spike and Payne, Doris and Saxena, Anju, 89–119. Eugene: University of Oregon. Chao, Yüan Ren. 1968. A Grammar of Spoken Chinese. Berkeley/Los Angeles: University of California press. Chappell, Hilary, and William McGregor. 1996. The Grammar of Inalienability: A Typological Perspective on Body Part Terms and the Part-Whole Relation. Berlin: Mouton de Gruyter. Cheng, Lai-Shen Lisa, and Rint Sybesma. 1999. “Bare and not-so-bare nouns and the structure of NP.” Linguistic Inquiry 30: 509–542. Chierchia, Gennaro. 1998. “Reference to kinds across languages.” Natural Language Semantics 6: 339–405. Corbett, Greville G. 1991. Gender. Cambridge: Cambridge University Press. Corbett, Greville G. 2012. “Canonical morphosyntactic features.” In Canonical Morphology and Syntax, edited by Dunstan Brown, Martina Chumakina and Greville G. Corbett, 48–65. Oxford: Oxford University Press. Corbett, Greville G. and Fedden, Sebastian. 2016. “Canonical gender.” Journal of Linguistics 52.3: 495–531. Craig, Colette. 1986a. “Jacaltec noun classifiers.” In Noun Classes and Categorization, edited by Craig, Colette G., 263–293. Amsterdam/Philadelphia: Benjamins. Craig, Colette. 1986b. “Jacaltec noun classifiers.” Lingua 70: 241–284. Craig, Colette. G. 1987. “Jacaltec noun classifiers, a study in grammaticalization.” Lingua 71: 241–284. Croft, William A. and Cruse, Alan D. 2004. Cognitive Linguistics. Cambridge: Cambridge University Press. Cruse, Alan D. 1994. “Prototype theory and lexical relations.” Rivista di linguistica 6.2: 167–188. Dahan, D., Swingley, D., Tannenhaus, M.K. and Magnuson, J. S. 2000. “Linguistic gender and spoken-word recognition in French.” Journal of Memory and Language 42: 465–480. Demuth, Katherine. 2000. “Bantu noun class systems: loanword and acquisition evidence of semantic productivity.” In Systems of Nominal Classification, editd by Gunter Senft, 270–292. Cambridge: Cambridge University Press. Denny. Peter J. 1976. “What are noun classifiers good for?” Chicago Linguistic Soiety: 122–132. Dixon, R. M. W. 1988. A grammar of Boumaa Fijian. Chicago/London: The University of Chicago Press. Friederici, Angela D. and Jacobsen, Thomas. 1999. “Processing grammatical gender during language comprehension.” Journal of Psycholinguistic Research 28: 467–484. Gao, Ming Y. and Malt, Barbara C. 2009. “Mental represenations and cognitive consequences of Chinese individual classifiers.” Language and Cognitive Processes 24: 1124–1179. Garnham, A., Oakhill, J., Ehrlich, M.-F., and M. Carreiras. 1995. “Representations and processes in the interpretation of pronouns: New evidence from Spanish and French.” Journal of Memory and Language 34: 41–62. Geraghty, Paul. 1983. The history of the Fijian languages. (Oceanic Linguistics Special Publication no. 19) Honolulu: University of Hawaii Press. Grandi, Nicola and Körtvélyessy, Livia. 2015. Edinbourgh Handbook of Evaluative Morphology. Edinburgh: Edinburgh University Press. Greenberg, Joseph H. 1972. “Numerical classifiers and substantival number: problems in the genesis of a linguistic type.” Working Papers on Language Universals 9: 1–39. (Stanford University) Grinevald, Colette. 2000. “A morphosyntactic typology of classifiers.” In Systems of Nominal Classification, edited by Gunter Senft, 93–113. Cambridge: Cambridge University Press. Grosjean, F.; Dommergues, J.-Y.; Cornu, E. and Guillelmon, D. 1994. “The gender-marking effect in spoken word recognition.” Perception and Psychophysics 56: 590–598.
Classification between Grammar and Culture
229
Guthrie, Malcolm 1967–1971. Comparative Bantu. Farnborough: Gregg. Hawkins, John A. 1994. A Performance Theory of Order and Constituency. Cambridge: Cambridge University Press. Hundius, Harald and Kölver, Ulrike. 1983. “Syntax and semantics of numeral classifiers in Thai.” Studies in Language 7.2: 165–214. Jacobsen, Thomas. 1999. “Effects of grammatical gender on picture and word naming: Evidence from German.” Journal of Psycholinguistic Research 28: 499–514. Katamba, Francis. 2006. “Bantu nominal morphology.” In The Bantu Languages, edited by Nurse, Derek and Philippson, Gérard, 103–120. London: Routledge. Körtvélyessy, Lívia. 2015. Evaluative Morphology from a Cross-Linguistic Perspective. Cambridge: Cambridge Scholars Publishing. Kuo, Jenny Yi-chun and Sera, Maria D. 2009. “Classifier effects on human categorization: the role of shape classifiers in Mandarin.” Journal of East Asian Linguistics 18: 1–19. Lambrecht, Knut. 1994. Information Structure and Sentence Form. Cambridge: Cambridge University Press. Lehmann, Christian. 1984. Der Relativsatz. Tübingen: Narr. Lévy-Bruhl, Lucien. 1914. “L’expression de la possession dans les langues mélanésiennes.” Mémoires de la Société de Linguistique de Paris 19.2: 96–104. Li, Xuping and Bisang, Walter. “Classifiers in Sinitic languages: From individuation to definitenessmarking.” Lingua 122: 335–355. Lichtenberk, Frantisek. 1983. “Relational classifiers.” Lingua 60: 147–176. Löbel, Elisabeth. 1996. Klassifikatoren. Eine Fallstudie am Beispiel des Vietnamesischen. Universität Köln: Habilitationsschrift. Löbel, Elisabeth. 2000. “Classifiers vs. genders and noun classes: a case study of Vietnamese.” In Gender in Grammar and Cognition, edited by Unterbeck, Barbara and Rissanen, Matti, 259–319. Berlin: Mouton de Gruyter. Lyons, Christopher. 1999. Definiteness. Cambridge: Cambridge University Press. Mattéi-Muller, Marie-Claude. 1974. “El sistema de posesión en la lengua panare.” Antropologica 38: 3–14. Mottin, Jean. 1980. Contes et légendes Hmong Blanc. Bangkok: Don Bosco. Orozco, Inti Aedo. 2016. Evaluatives in Grammar. Munich: LINCOM. Passer, Matthias Benjamin. 2016. The Typology and Diachrony of Nominal Classification. PhD dissertation, University of Amsterdam. Published in Utrecht: Netherlands Graduate School in Linguistics (LOT). Schütz, Albert J. 1985. The Fijian language. Honolulu: University of Hawaii Press. Seifart, Frank. 2005. The Structure and Use of Shape-Based Noun Classes in Miraña (North West Amazon). (MPI Series in Psycholinguistics) Wageningen: Ponsen and Looijen. Seiler, Hansjakob. 1983. Possession as an Operational Domain of Language. Tübingen: Narr. Sera, Maria D., Johnson, Kaitlin R., and Kuo, Jenny Yichun. 2013. “Classifiers augment and maintainance of shape-based categorization in Mandarin speakers.” Language and Cognition 5: 1–23. Shepardson, Kenneth N. 1982. “An integrated analysis of Swahili Augmentatives-Dimunutives.” Studies in African Linguistics 13.1: 53–76. Srinivasan, Mahesh. 2010. “Do classifiers predict differences in cognitive processing? A study of nominal classification in Mandarin Chinese.” Language and Cognition 2: 177–190. Tucker, G. R., Lambert, W. E., and Rigault, A. A. 1977. The French Speaker’s Skill with Grammatical Gender: An Example of Rule-Governed Behavior. The Hague: Mouton. Welmers, William A. 1973. African Language Structures. Berkeley/Los Angeles/London: University of California Press.
230
Walter Bisang
Wiedenhof, Jeroen. 2015. A Grammar of Mandarin. Amsterdam and Philadelphia: John Benjamins. Zhang, Hong. 2007. “Numeral classifiers in Mandarin Chinese.” Journal of East Asian Linguistics 16:1: 43–59. Zhang, Shi. and Schmitt, Bernd H. 1998. “Language dependent classification: The mental representation of classifiers in cognition, memory, and ad evaluations.” Journal of Experimental Psychology: Applied 4:s 375–385.
Joachim W. Kadereit
Classification and Naming of Living Objects – a Biologist’s Perspective (Extended Abstract) The usage of so-called binomials for naming organisms – a species name is composed of a genus name (Nicotiana) preceding a specific epithet (tabacum; i.e., Nicotiana tabacum L. for the tobacco plant) – was first consistently applied by Linnaeus (1753) in his Species Plantarum and has not changed since. In contrast to this, his classification (system) of plants has long been abandoned. He grouped the plants known to him into 24 classes based almost entirely on the number and arrangement of stamens, those organs of a flower which produce pollen. This ‘sexual system’ represents what is now called a ‘special purpose’ or ‘artificial’ classification. Although Linnaeus may have been ‘led to it [his sexual system] by an a priori belief that parts so fundamental for the reproduction of plants would be equally so for their classification’ (Stearn 1957: 24 f.), its major purpose was ‘to enable botanists to determine plants without having a teacher at hand’ (Stearn 1957: 34) and thus to create a ‘very convenient filing system’ (Davis and Heywood 1973: 16). Already before Linnaeus, John Ray (1682) ‘formulated the principle (often neglected by later botanists) that all parts of the plant should be used for classification’ (Davis and Heywood 1973: 14) and thus advocated the use of general purpose or natural classifications. Even though a general purpose classification may seem superior to a special purpose classification because ‘(Gilmour notes that) the high information content of such a classification allows many predictions to be made about the properties of members of its constituent groupings’ (Briggs and Walters 1997: 380), it should be clear that the quality of a classification can only be assessed in the context of its stated purpose. Thus, a system of plants arranged by flower colour is appropriate for a gardener who intends to plant a flower bed. After Linnaeus, various efforts were made to construct natural systems, e.g., by Michel Adanson. In his Familles des Plantes (1763), Adanson produced 65 special, single-character systems, in which the position of one genus relative to another varied greatly. Nevertheless, as he found that the same genera came together in several systems, he believed that the genera which were found side by side in the greatest number of systems should be classified together because they showed the greatest number of similarities (Davis and Heywood 1973: 23).
In the 20th century, this astonishingly modern Adansonian approach to the determination of similarity was further quantified (and eventually computerized) and resulted in what became known as ‘Numerical Taxonomy’ (Sokal and Sneath 1963).
DOI 10.1515/9783110538779-008
232
Joachim W. Kadereit
The whole course of biological thought was radically changed by the publication of Darwin’s The Origin of Species in 1859. It influenced taxonomy by the enunciation of two important principles; (...) (1) Species are not unchanging creations but have evolved from one another in the long history of life... (Davis and Heywood 1973: 31).
This should have resulted in equally radically changed approaches to classification, because ‘if organisms have reached their present state by a process of evolution, it follows that they have a built-in classification, and man’s problem is to find it. This is quite a different problem from that presented to the classifier of man-made objects’ (Holttum 1967: 37). However, ‘Post-Darwinian systems have differed little in content, though they have differed in arrangement, from those of the later pre-Darwinian taxonomists’ (Davis and Heywood 1973: 31). The reason for this may be that similarity was regarded as the best proxy for, and the only means to detect evolutionary (phylogenetic) relationship, and that ‘the systems of de Candolle and others were, in the content of their groups, very largely natural’ (Davis and Heywood 1973: 31). B
C
relationship
A
similarity
Fig. 1: Incongruence of overall similarity and relationship. Relationship is defined by the relative recency of a last common ancestor (B and C are more closely related to each other than either is to A), and overall similarity by lateral distance (B is more similar to A than to C).
Similarity is critical in two respects. First, similarity need not reflect evolutionary relationship. This is illustrated in fig. 1 , which shows that, although species B is more closely related to species C, it is more similar to species A. Evolutionary relationship is here defined by the relative recency of the last common ancestor, and species B and C have a more recent common ancestor (at the branching point of the lines leading to B and C) than species B and A. One reason for incongruence of overall similarity and relationship as seen in fig. 1 is difference in speed of change (evolution). Species C has evolved faster than species A and B, a situation often encountered when species, by pure chance, are displaced into entirely new environmental conditions with entirely new selective conditions. Second, similarity cannot be objectively determined, and
Classification and Naming of Living Objects
233
the outcome of a similarity analysis will depend on which properties were considered (and subjectively chosen), which algorithms were used (and subjectively chosen), etc. Both methodological (Hillis et al. 1996) and conceptual (Felsenstein 2004) advances lead us to believe that our ability to reconstruct the course of evolution and to formulate well-supported hypotheses about relationships has greatly increased. Thus, the use of molecular genetic information (DNA-sequences) in ever-increasing quantity and its analysis with increasingly refined model-based bioinformatic tools have resulted in a multitude of phylogenetic trees. Although the relationships that were found on the basis of such approaches were supported by similarity in many cases, this was not so in many other cases. In the following, I want to use an example to illustrate such incongruence between similarity and relationship and its implication for classification and naming.
Fig. 2: Fruits (capsules) of representative species of the traditionally recognized genera Meconopsis (A M. villosa, B M. discigera, C M. cambrica), Papaver (D P. alpinum, E P. argemone, F P. aculeatum, G P. californicum, H P. rhoeas), Roemeria (I R. hybrida) and Stylomecon (J S. heterophylla). Scale: 1 cm. For further explanation see text.
234
Joachim W. Kadereit
Papaver L., the Poppies, with the Opium Poppy (P. somniferum L.) and the Common Poppy (P. rhoeas L.) as the probably best known of altogether about 70 species, is part of a group of four genera distributed almost entirely in the Old World (Schwarzbach and Kadereit 1995). The other three genera are Meconopsis Vig. with over 50 species, Roemeria Medik. with two species, and Stylomecon G. Taylor with one species. Subdivision into these four genera is based largely on fruit morphology (fig. 2). Papaver has capsules which open and release seeds by apical pores, and are crowned by a flat stigmatic disc. Meconopsis has capsules with an apical extension – the style of the flower at fruiting time – which open by valves well below the apex, and have no stigmatic disc. Roemeria has capsules without apical extension which open by valves, and have no stigmatic disc, and Stylomecon has capsules similar to those in Meconopsis, but is an annual (completing its life cycle within one year) and grows in western North America. In contrast to this, Meconopsis is perennial (with a life cycle of more than one year) and largely Himalayan in distribution, with one species (M. cambrica (L.) Vig. – Welsh Poppy) in westernmost Europe. Various analyzes of these four genera (Kadereit et al. 1997, Carolan et al. 2006, Kadereit et al. 2011), all using DNA sequences, have revealed that patterns of relationship cut across traditional generic delimitations. Thus (fig. 3), three subgroups of Papaver (Papaver s.str., P. californicum A. Gray, P. aculeatum Thunb.) are most closely related to Meconopsis cambrica and Stylomecon heterophylla (Benth.) G. Taylor. Another subgroup of Papaver (Papaver sect. Argemonidium Spach) is most closely related to Roemeria, and a last subgroup of Papaver (Papaver sect. Meconella Spach) is most closely related to one of the three subgroups of Meconopsis. This is the best hypothesis about relationships in this group of plants currently available – one should be aware of the possibility that examination of additional species or the use of other DNA sequences may lead to different results in future studies. Should these results affect our classification of these plants and, if yes, how would this affect naming? In order to illustrate the range of issues relevant when answering these questions, I want to discuss four (of many more) possible options. 1. The first option would be to leave classification and names unchanged (1 in fig. 3), based on the argument that the four genera as traditionally recognized are well-characterized by their morphology and that one should classify according to similarity and not according to relationship. The difficulties with similarity have been discussed above. It is quite evident that this option would result in unnatural groups: Papaver would contain five subgroups (Papaver s.str., P. californicum, P. aculeatum, P. sect. Argemonidium, P. sect. Meconella) which are not closest relatives to each other, and Meconopsis would contain three subgroups (M. cambrica, Meconopsis 1 Asia, Meconopsis 2 Asia) which again are not closest relatives to each other. 2. A second option (2 in fig. 3) could be to limit Papaver to Papaver s.str., P. californicum and P. aculeatum, to maintain M. cambrica and S. heterophylla, and to include P. sect. Argemonidium in Roemeria and P. sect. Meconella in Meconopsis. Such an approach may be supported by considerations of both relationships and similarity.
Classification and Naming of Living Objects
235
For example, Papaver sect. Argemonidium resembles Roemeria in having bristly capsules, and Papaver sect. Meconella, distributed across the Arctic and in European high mountains, resembles some species of Meconopsis, which by growing in the Himalayas also prefers similar habitats. On the other hand, the capsules of M. cambrica and S. heterophylla are clearly distinct from those of Papaver. Such a classification would create a paraphyletic instead of a monophyletic Papaver. The principle of monophyly was introduced to systematics by Hennig (1950), who defined a group as monophyletic if all members of this group descended from a most recent common ancestor, and if all descendants of this ancestor are included in that group. When Papaver is defined as NOT containing M. cambrica and S. heterophylla, as has been done in this second option, Papaver is paraphyletic, because two of the descendants of the last common ancestor of Papaver, i.e., M. cambrica and S. heterophylla, are not included in Papaver. With this option, all species of Papaver sect. Argemonidium and Papaver sect. Meconella require new (or already available) names in Roemeria and Meconopsis, respectively. This is an implication of the hierarchical nature of the Linnaean classification, where a species belongs to a genus, a genus to a family, a family to an order, etc. For example, Papaver argemone L. as one species of Papaver sect. Argemonidium would need to be called Roemeria argemone (L.) C. Morales, R. Mend & Romero Garcia. Apart from the binomial species name introduced above, a scientific name also includes its author, Linnaeus (abbreviated as L.) in the case of Papaver argemone L. The string of authors for Roemeria argemone indicates that the species was first described by Linnaeus in a different genus, which is why his name is put into brackets, and that the three authors after the bracket combined this species into another genus, here Roemeria. Thus, scientific names have a memory and contain some information about past classifications. 3. A third option (3 in fig. 3) would be to include Meconopsis, Roemeria and Stylomecon in Papaver. This would constitute a monophyletic group, because all species in this group descend from a most recent common ancestor and all descendants of this ancestor are included in that group. Of the four genus names available, Papaver has to be chosen because it is the oldest validly published name, following the principle of priority. In spite of its monophyletic character, this group would be morphologically highly heterogeneous and, moreover, a rather large number of species would require name changes. The last point raises the issue of name stability. Not only is ‘A plant’s name (is) the key to its literature’ (van Steenis 1957: clxxiv), but users of plant names outside botany, e.g., horticulturalists, ethnobotanists, etc., have an understandable interest in name stability. 4. Considering all issues raised above, i.e. similarity and relationship, monophyly and paraphyly, morphological homogeneity and name stability, I prefer a fourth option (and have followed this in recent publications: Kadereit and Baldwin 2011, Kadereit et al. 2011), where Papaver including M. cambrica and S. heterophylla, Roemeria incl. Papaver sect. Argemonidium, Meconopsis incl. Papaver sect. Meconella and Cathcartia for the second part of Meconopsis (M. Asia 1) are recognized (4 in
236
Joachim W. Kadereit
fig. 3). These groups reflect relationship and are each monophyletic. They are morphologically fairly homogeneous (Papaver in such circumscription admittedly less so) and distinct from each other, i.e. diagnosable. In addition, many names which will have to be given to species, changing genus allocation, are already available, i.e. have already been used at some point in the past. For example, M. cambrica was originally described as Papaver cambricum by Linnaeus. A final note on the stability of names. In order to avoid name changes as a consequence of changes in a hierarchical classification system, some authors (for references see Cantino and de Queiroz 2010) have suggested abandoning this hierarchy and instead naming clades, where a clade is a branch of the phylogenetic tree which needs no change of name when it changes its position in the phylogenetic tree. A set of rules for this type of approach has been published under the so-called PhyloCode (Cantino and de Queiroz 2010). However, this approach has so far failed to gain momentum for various reasons (Stuessy 2009: 101–103; 134).
Fig. 3: Phylogenetic relationships among Meconopsis, Papaver, Roemeria and Stylomecon (left) and four options for their classification (right). For further explanation see text.
Classification and Naming of Living Objects
237
References Adanson, Michel. 1763. Familles des Plantes. Paris: Vincent. Briggs, David, and S. Max Walters. 31997. Plant Variation and Evolution. Cambridge: Cambridge University Press. Cantino, Philip D., and Kevin de Queiroz. 2010. International Code of Phylogenetic Nomenclature. Version 4c. http://www.ohio.edu/phylocode/toc.html. [Accessed 1/2010] Carolan, James C., Ingrid L. J. Hook, Mark W. Chase, Joachim W. Kadereit and Trevor R. Hodkinson. 2006. “Phylogenetics of Papaver and related genera based on DNA sequences from ITS nuclear ribosomal DNA and plastid trnL intron and trnL-F intergenic spacers.” Annals of Botany 98: 141–155. Darwin, Charles. 1859. On the Origin of Species by Means of Natural Selection. London: John Murray. Davis, Peter H., and Vernon H. Heywood. 1973. Principles of Angiosperm Taxonomy. Huntington: Robert E. Krieger. Felsenstein, Joseph. 2004. Inferring Phylogenies. Sunderland, MA: Sinauer Associates. Hennig, Willi. 1950. Grundzüge einer Theorie der phylogenetischen Systematik. Berlin: Deutscher Zentralverlag. Hillis, David M., Craig Moritz and Barbara K. Mable (eds.). 21996. Molecular Systematics. Sunderland, MA: Sinauer Associates. Holttum, Richard E. 1967. “Comparative morphology, taxonomy and evolution.” Phytomorphology 17: 36–41. Kadereit, Joachim W., and Bruce G. Baldwin. 2011. “Systematics, phylogeny, and evolution of Papaver californicum and Stylomecon heterophylla (Papaveraceae).” Madroño 58: 92–100. Kadereit, Joachim W., and Claudia Erbar. 2011. “Evolution of gynoecium morphology in Old World Papaveroideae: a combined phylogenetic/ontogenetic approach.” American Journal of Botany 98: 1243–1251. Kadereit, Joachim W., Kirstin B. Jork and Andrea Schwarzbach. 1997. “The phylogeny of Papaver s.l.: Polyphyly or monophyly?” Plant Systematics and Evolution 204: 75–98. Kadereit, Joachim W., Christopher D. Preston and Francesco J. Valtueña. 2011. “Is Welsh Poppy, Meconopsis cambrica (L.) Vig. (Papaveraceae), truly a Meconopsis?” New Journal of Botany 1: 80–87. Linnaeus, Carl. 1753. Species Plantarum. Stockholm: Laurentii Salvii. Ray, John. 1682. Methodus Plantarum Nova. London: Henrici Fairborne, and Joannis Kersey. Ridley, Mark. 1986. Evolution and Classification. The Reformation of Cladism. London/New York: Longman. Schwarzbach, Andrea, and Joachim W. Kadereit. 1995. “Rapid radiation of North American desert genera of the Papaveraceae: evidence from restriction site mapping of PCR-amplified chloroplast DNA fragments.” In Systematics and Evolution of the Ranunculiflorae, edited by Uwe Jensen and Joachim W. Kadereit, 159–170. (Plant Systematics and Evolution Supplement, 9) Wien/New York: Springer. Sokal, Robert R., and Peter H. A. Sneath. 1963. Principles of Numerical Taxonomy. San Francisco: Freeman. Stearn, William T. 1957. An Introduction to the Species Plantarum and Cognate Botanical Works of Carl Linnaeus. London: Unwin. Stuessy, Tod F. 22009. Plant Taxonomy. The Systematic Evaluation of Comparative Data. New York: Columbia University Press. van Steenis, Cornelius G. G. J. 1957. “Specific and intraspecific delimitation.” In Flora Malesiana, edited by Cornelius G. G. J. van Steenis, Series 1, vol. 5: clxvii–ccxxix. Djakarta: Nordhoff-Kolff N. V.
238
Joachim W. Kadereit
Figures Fig. 1: Modified from Ridley 1986: 5. Fig. 2: Original drawings by the author. Fig. 3: Original, modified from Kadereit and Erbar 2011: 1244.
Roy Ellen
Tools, Agency and the Category of Living Things Summary: Humans and other animals attribute the qualities of living matter and agency to what we call tools and other cultural objects. In both cases a paradox may arise when autonomy is attributed to the object at the same time that it is recognized that its life-like characteristics are motivated by human actions. Nuaulu people in eastern Indonesia describe many kinds of objects as having the qualities we might otherwise reserve for biological organisms. They also distinguish entities that have many of the qualities of life but which ordinarily have no corporeal existence (spirits). While all cultural objects are potentially regarded in this way, in practice some objects are more alive and have more agency than others. I argue that part of the problem with existing anthropological treatments of the category living things is that they are either logical extrapolations through polythetic extension or based on formal taxonomic deduction/induction (ethnoscience). Using examples of meat-skewers, outboard motors, coconut graters, and sago-processing devices, together with certain peripheral forms of biological life, I demonstrate how Nuaulu ideas of what is animate and agentive are always fuzzy and contingent, and that by combining data from different kinds of ethnographic context, using different elicitation procedures, a more complex picture emerges. Human cognition, it appears, is handily eclectic. (Atran 1990: 78)
1 Introduction: thinking like a dog I never cease to be amazed and intrigued by the way in which certain animals respond to humans using tools. Let me give two very specific examples from my own domestic life. When I am brushing away ashes in the fire hearth our border collie will snap at the brush, but only when I am using it. As soon as the brush is put down, and my hand withdrawn, and the brush is completely inert, the snapping and barking ceases. Similarly, when I am outside – let us say – raking autumnal leaves, the collie will snap and bark at the rake, but as soon as the rake is put down it becomes completely disinterested and wanders off. The dog also reacts to wheelbarrows and tractor mowers in a similar way, and bites at the wheels, consistent with the widely reported canine ‘grab-bite’ predatory behaviour pattern (Coppinger and Schneider 1995: 27).
DOI 10.1515/9783110538779-009
240
Roy Ellen
What is interesting about this observation is not simply the phenomenon itself (a characteristic of the breed), but a paradox that lies at the heart of our wonderment at the dog’s perception and comprehension. It is a relatively uninteresting inference to say that when the tool is moving the dog responds to it as if the tool were animate. What is more interesting is to note that the tool becomes animate because it is attached to me. The dog knows me well enough and does not bark and snap when I am not raking or brushing, or whatever – only when the activity involves certain kinds of tool that interact with the ground. Thus, it does not happen when I am, say, using a hammer, or working at a bench. So, why, when the dog does not react aggressively to me normally, does it do so when I am using certain kinds of tools in a particular way? One possible explanation is that the dog responds to the rake or brush because it thinks it is an independent organism, and because of its close proximity must be threatening me. The collie has reverted to guard dog mode to protect me, responding to the rake or brush as if it were a recalcitrant sheep. So although it can see that I am the agent moving the rake it responds to the rake as though it was independent and itself had agency. Can we therefore conclude that the dog is confused? Of course, not all animals respond in this way in such situations, certainly not all species and or most breeds of dog. Collies are trained as guard dogs to herd animals, and their behavioural features have been selected by humans, and cultivated to encourage such behaviour. However, quite apart from the puzzle of canine perception and understanding that this raises, both the general observation that animals other than humans can attribute the qualities of living matter and agency to what we call tools or affordances, and the specific paradox concerning the simultaneous attribution of autonomy to the tool and the visual evidence that it is motivated by a human, are, I think, relevant. We can conclude that both humans and other animals attribute the qualities of living matter and agency to what we call tools and other cultural objects. And in both cases a paradox may arise when autonomy is attributed to the object at the same time that it is recognized that its life-like characteristics are motivated by human actions.1 My account of the interaction between dogs and human tools is relevant to anthropological debates about the differences between human cognition of living and non-living things (e.g. Atran 1990). Nuaulu people in eastern Indonesia describe many kinds of object as having the qualities we might otherwise reserve for biological organisms. They also distinguish entities that have many of the qualities of life
1 It may be that dogs behaving in this way are responding to a key stimulus, an old ethological concept that refers to specific stimuli that have the potential to release a specific behaviour (modal action pattern). The rapidly moving stimuli of the kind described here release the predatory behaviour, more effectively in collies and in particular dogs. In some cases this behaviour can become stereotypical (obsessive-compulsive) and therefore abnormal. In such cases what we see has little to do with seeing/perceiving directly, but rather reflects the brain’s sensitivity to certain stimulus configurations.
Tools, Agency and the Category of Living Things
241
but which ordinarily have no corporeal existence (spirits). While all cultural objects are potentially regarded in this way, in practice some objects are more alive and have more agency than others. I argue here that part of the problem with existing anthropological treatments of the category living things is that they are either logical extrapolations through polythetic extension, or based on formal taxonomic deduction/induction. Using examples of meat-skewers, outboard motors, coconut graters, and sago-processing devices, together with certain forms of peripheral biological life such as slime fungi and algae, I shall try to demonstrate how Nuaulu ideas of what is animate and agentive are always fuzzy and contingent, and that by combining data from different kinds of ethnographic context, using different elicitation procedures, a more complex picture emerges.
2 Tools, machines, and engines Nuaulu people of eastern Indonesia provide us with plenty of examples of tools that might be said to have agency, but not all tools are the same. They attribute to many kinds of object the qualities that we might otherwise reserve for biological organisms, and although I have continued unashamedly to describe Nuaulu as animists despite the nuanced theorising of the ontological turn in anthropology, I do not have in mind here simply practices that we might formally understand as animism. Nuaulu do not animate every inanimate thing in their environment, all of the time. Neither by describing Nuaulu as animist do I mean that they exclusively resort to what Descola (2005) would understand by an animist ontology. Totemic, analogical, naturalist and animist ontologies are all available to Nuaulu, who like many people resort to each and all as and when the context makes them appropriate. I think we need to resist turning these important cognitive and representational distinctions into a mechanical typology. However, in examining the case of tools it is impossible to avoid notions of animation wherever we find the property of physical motion. Tools have been classified in many ways, but here it helps to distinguish between: (a) tools that once made do not move (e.g. a fence or a stake in a pit trap), and (b) tools that move when activated by the human body and are often characterised ergonomically or in terms of physiological mechanics as an extension of a body part. These would include a hammer, sago pounder, or the fire hearth brush that featured in my opening example of dog-human interaction. Type (c) tools are those having simple moving parts that relay energy released by initial human motion, such as the Nuaulu sago flour extraction device, a treadle-operated coconut grater, or a hand-held string-making apparatus. These are effectively ‘machines’, or ‘tended facilities’ (Oswalt 1976) using mechanical power, having several parts for performing a particular task. Finally, there are (d) tools that are also machines in the sense outlined but that run on stored energy, and do not require continuous human inputs, only
242
Roy Ellen
initially to prime them or to add fuel. In the Nuaulu material universe these might include clocks, trucks, chainsaws, or outboard motors. We would usually describe these as engines, in that they each convert energy into useful mechanical motion. Let us now examine Nuaulu examples of each of these. A common type (a) tool would be, as I have indicated, a fence or a pit trap stake. Although Nuaulu might in theory attribute some independent agency to either of these under certain circumstances, I have no examples of this ever being the case. A more interesting example would be a meat-skewer or asunaete. These are made of wood or bamboo and used in the preparation and transport of meat, but they are also a material instrument facilitating communication between the living and ancestral spirits. Thus, after a wild pig has been killed it will first be singed to remove most body hair near to the place where it was caught and before returning to the village. A fire is lit and a wooden stake cut to serve as a skewer to manipulate the carcase. A chip (kakomate) from the skewer – representing its soul or spirit – is placed inside the belly of the pig, removed after singing and re-united with the asunaete. The pig is butchered by first stripping out the lower jaw bone (penesite), throat and lungs. After the pig has been butchered, the skewer will be stuck in the ground and the throat and lungs (the organs of breathe) attached to the top as an offering to ancestral spirits, before being taken back to the village. A similar skewer (asunaete marane) is used for marsupial cuscus of the genera Phalanger and Spilocuscus that are the most frequent species hunted by Nuaulu (Ellen 1996). In this case, however, the skewers are not used to remove the fur through singing and do not have the organs of breathe tied to them. Instead, having served their purpose in cooking and for transport they are stuck into the ground and the chip initially removed to make the spike is re-attached to the skewer. The soul is thus re-united with the body, and it is believed that the spirit of the cuscus will return in another body, which will be eaten again. Thus, the asunaete has a simultaneous purpose of skewering the meat, roasting it or carrying it, but by virtue of this role also ensures that the spirit of the animal returns to the cosmos to sustain the population of cuscus for further hunting (fig. 1). Asunaete for the four different kinds of cuscus are differentiated by the presence or absence of notches, or the number of notches on the skewer. The skewer for mara kokowe (male Phalanger orientalis) has no notches, while that for mara osu (female Phalanger orientalis) has one or two notches. Thus the asunaete works in a similar way to the Huarani blowpipe, being a regulatory instrument inserted in webs of systemic relations through which the reproduction of society (Rival 2013: 97)
is effected through the reproduction of the cuscus population.
Tools, Agency and the Category of Living Things
243
Fig. 1: Series of asunaete: cuscus (Phalanger) skewers planted as an offering to ancestral spirits; June 1970
Type (b) tools conform to our prototypical notion of what a tool should be, and readily fit the idea of tool as prosthesis, as an extension of the living human anatomy, moving with the body and sometimes even resembling human body parts: for example a sago pounder. But while tools become animated by virtue of being extensions of the human body, this – unlike the case of dog-brush interaction in my opening example – is in itself insufficient to make them animate in the sense understood by animism. Most Nuaulu tools, most of the time, are not attributed with animacy. However, some are more likely to be, and we need to identify what the conditions might be to satisfy this condition. In the first place, these conditions relate to the kind of subsistence activity with which they are connected. Thus, hunting (and in a previous era warfare) are highrisk activities in which supernatural support can make a difference. Secondly, some technologies are less under the continuous control of an operator. A bamboo pounder never leaves the hand that holds it during the actions entailed in pounding sago pith. However, arrows once they leave the projectile release mechanism we call a bow are subject to a variety of hazards over which the archer has no direct control: wind, movement of the prey, or of objects intervening between the hunter and his prey. In this situation, to attribute an arrow with agency and to make a preliminary offering to its embodied spirit to ensure its effectiveness in flight makes sense. However, in addition to such technological contexts, there is a category of tools that are always regarded as fully animate because of their simultaneous status as ritual objects. In the first place, any tool or other artefact that has been used in ritual becomes sacred and cannot be disposed of, being stored until it disintegrates in the smokey lofts of clan houses. But there are other ritual objects, such as shields nowadays used only in ceremonial contexts (Ellen 1990). Most of these are not subject to special ritual
244
Roy Ellen
attention, but others are produced and stored with anthropomorphic regard under controlled ritual circumstances. They well illustrate Gell’s (1998: 230–231) notion of ‘objectified personhood’ through deferred or abducted human agency, existing both ‘objectively’, and subjectively as persons. But not only are there sacred shields but also sacred spears and baskets (Ellen 2012) and other objects, all of which are exemplary instances of each kind of tool. To have all baskets, spears and shields accorded special respect would be inconvenient, but to have just some objects in a given category with sacred agency is sufficient to make the point. The logic here is similar to the totemic logic prohibiting one species within a category of potential food animals or plants, the rest of which can be eaten (Ellen 1998: 254–255). The third category of tool (type c) contains those with moving parts, and those that I will discuss here are the sago flour extraction apparatus and the treadle-operated coconut grater. Although like type (b) tools they are in a sense, like any tool, an extension of the body, part of the extended phenotype, they are so in a less obvious way and rather than simply being a specialized extension to a limb, such as a sago pounder, the relationship between tool and person is more complex, the body and apparatus entwining in each other holistically. This appears to be the case for many tools that we describe as machines, and we need to examine whether it makes a difference when attributing agency or animacy. The first of these two cases (fig. 2), what Nuaulu call the aha, is a device of great antiquity and of wide distribution in the Moluccas and New Guinea, used for separating starch granules from pith residue. The apparatus consists, basically, of two troughs made from sago leafstalk resting on a frame, in which one – the input trough or sihane – overlaps with and is slightly higher than the lower (solo-solo) output trough. Attached to the overlapping end of the sihane is a filter mesh (nunte), usually of stitched coconut fibre matting, which acts as a semi-permeable membrane. The upper part of the membrane is attached by string to a flexible sapling that acts as a spring-loaded assembly (hehune hatai). As water flows into the upper trough, the resulting mulch is pressed against the filter with one hand and the string pulled down with the other. Because of the flexibility of the sapling the membrane automatically retracts, enabling further compression of the mulch, and is ready for the operator to press once more. Thus, the latent energy produced by the bent sapling assists the process of efficient filtering. There is a process of complex bodily engagement between the individual and the apparatus. Here the tool is not simply motivated by the arm, but in a process initiated by the operator, actually collaborates in the process of filtering in a multifunctional kind of way (Ellen 2004). This is such a common and culturally embedded technique that Nuaulu seldom reflect on its mode of function. As far as I know, it is never described in terms that would lead to an understanding that it is anthropomorphized, though the success of the processing activity is routinely ensured by making a small offering to the ancestral spirits who own the sago, consisting of tobacco, Areca fruit and betel pepper fruits wrapped in a leaf, and tucked into a convenient joint in the apparatus.
Tools, Agency and the Category of Living Things
245
Fig. 2: Aha, Nuaulu sago-processing apparatus
By contrast, the rotary action treadle-operated coconut grater (fig. 3) is of relatively recent introduction, and is perhaps no older than the late nineteenth century, part of a new technological repertoire associated with an increase in the planting of coconut palms for commercial copra production, and an increase in oil production for local consumption. The operator (usually male) presses the treadle with one foot, which pulls a string that rotates a cutting head, first in one direction and then in the other, using successive actions on the treadle. The head is made of wood in which metal cutting blades are embedded, and this assembly is attached to a spindle at waist height. As can be seen from fig. 3, the action – like the sago flour processing apparatus – requires engagement with the whole body. To my knowledge, there are no associated ritual practices, though in several technical respects the treadle-operated coconut grater is comparable to the much older sago-washing device.
Fig. 3: Treadle-operated coconut grater, Rouhua village, Seram; December 1970
246
Roy Ellen
Finally, there are type (d) tools, machines driven by a non-human source of power, which in the historical experience of the Nuaulu consist of devices reliant on steam power, diesel, or petrol. These have been familiar parts of the Nuaulu world for over a century, with steam and diesel driven boats, and power-driven ban saws used in timber yards. The internal combustion engine found in trucks and generators, mainly became familiar along with aircraft during the Second World War, and through increased travel to Ambon and other islands. With independence the road system deteriorated and most devices with engines were out of the reach of most people in rural areas of Seram. As the economy grew under the New Order, so Nuaulu became more familiar with road vehicles, tools such as chainsaws, and especially the outboard motor, the ubiquitous and eponymous motor jonson. Outboard motors became common with local fisherman and small scale traders during the 1980s, though being essentially forest-oriented Nuaulu have acquired them only recently. What is relevant to our discussion here is that Nuaulu treat outboard motors as essentially animate and motivated by a spirit even though some have a good knowledge of the practicalities of how they work. Nevertheless, when they go wrong, as they often do, Nuaulu will, after physical inspection and adjustment, make offerings and invocations to get them moving. These offerings are basically no different from those that they might offer before hunting, or tuck into the joints of a sago processing apparatus. I was first introduced to this practice by a chance observation of Rosemary Bolton of Nuaulu ‘feeding an outboard motor’ in 1996. Up until that time no Nuaulu of my acquaintance had owned an outboard motor and I was determined to follow through on the observation. To some extent this fits in with existing ritual practices relating to outrigger canoes and other sailing craft found in central Moluccan waters, that are anthropomorphized and subjected to life-cycle rituals and rituals at various stages in their manufacture (Ellen 2003: 161 f.). However, the outboard motor confers new properties to the vessel, namely a source of propulsion independent of human bodily action. While this has obvious technical advantages, it has the disadvantage that the operation of the vessel is less under the control of the crew. In a sailing boat, the crew can appeal to the spirits of the wind and may encourage them to blow by supplication in the form of banging gongs or similar metal objects. This is not possible with a jonson, and it is hardly surprizing that the crew seek to maximize the forces working in their favour when the risk is increased. In making ontological sense of these machines and the practices that accompany them it is also relevant that outboard motors and sago-grating machines, move, whirr, hum and get hot when they are used. Moreover, tools that have a quasi-independent existence in not needing to rely on continuous human manual power also resemble biological life in their capacity to die. Paul Taylor (1990: 49) in his componential analysis of biotic forms claims that Tobelo define life as anything that has the capacity to die, and there is something paradoxically potent in this definition. Nuaulu speaking of an outboard motor that has just spluttered to a standstill having run out of fuel will say mataenya – ‘it is dead’, just as we would use the same word to describe a malfunc-
Tools, Agency and the Category of Living Things
247
tioning vehicle or power tool. Interestingly, Bloch (1998: 48) notes that Zafimaniry speak of life in relation to entities as diverse as clouds, quartz and motor engines, and use maty (an Austronesian cognate of Nuaulu matae) for almost anything that breaks down. So, the attribution of the qualities of life to engines is hardly unique to the Nuaulu, and in Western societies too these same properties encourage modes of behaviour, emotion and linguistic expression that are wholly comparable. The point I want to make is that in these various attitudes concerning the degree to which tools have animacy, agency or intentionality reveals an underlying pan-human tendency. We all anthropomorphize tools (especially Western males) to the extent that some, such as cars, might be said to be fetishized (Miller 1987: 85–108). We develop an intimate relationship with them that can involve by turns love and anger when they do not perform as expected. Who has not shouted at a printer for breaking-down when racing to meet a deadline – an example that Alfred Gell (1998) has felicitously explored – notwithstanding a clear distinction between intentional persons and inanimate things in dominant Euro-American ontology. Many of you will know the famous scene in Gourmet Night, an episode of the British comedy show Faulty Towers, in which Basil (John Cleese) attacks his mini car that has mechanically failed while delivering food from a local restaurant to the hotel. Here is the dialogue to remind you: Come on! Start! Start, you vicious bastard! Come on! Oh, my God! I’m warning you! If you don’t start I’ll count to three. … One two three right! That’s it! I’ve had enough! You stalled just once too often! Right! Well, don’t say I haven’t warned you! I’ve laid it on the line to you time and time again! Right! Well, this is it! I’m gonna give you a damn good thrashing!
We laugh at this because we see in Basil’s behaviour an exaggeration and mocking of our own attempts to interact with tools, where there is a clear dissonance between our rational techno-scientific selves and the more intuitive compulsion to treat an object as if it were biological and willful.
3 Life as a taxonomic category I think I can fairly claim that Nuaulu ideas about animacy, are – like our own – generally fuzzy and contingent. However, if we look at scientific definitions there have been repeated attempts to define the boundaries of living matter and what this might imply. These approaches come from philosophy, developmental/cognitive psychology, twentieth century anthropology and particularly ethnobiology or ethnoscience. Let us start with the ethnoscience model. Brent Berlin (1992) describes the concept of ‘living things’ as a ‘unique beginner’ in a general taxonomy of biological organisms. For example, Taylor (1990) in his work on Tobelo ethnobiological classification uses a strict componential form of semantic analysis. Quite reasonably, he begins by indicating that Tobelo distinguish living from non-living, defining the former (as we have
248
Roy Ellen
seen) as entities with the characteristic of being able to die. But rather than talk of living things he uses the term ‘biotic forms’, which he then says are divided-up by Tobelo into ‘sexual biotic forms’ that contrast with four other groups at the same taxonomic level: coral, sponges, fungi, and (a single category) ‘moss, mould, bryozoans, small algae’. ‘Sexual biotic forms’ are in turn divided into ‘breathers’ and ‘non-breathers’, the latter exemplified by seaweed and black coral, and the former by fauna and flora (fig. 4). Biotic form
Sexual biotic form
Breather
Fauna
Flora
Coral
Sponges
Fungi
Non-breather
Seewead
Black coral
Moss, mould, bryozoa, small algae
Fig. 4: Tobelo taxonomy of ‘biotic forms’ based on semantic componential analysis
Edmund Leach (1964) approached the issue in a different way, but with much the same result. Leach makes no reference to ethnoscience literature, and would probably be scathing in his dismissal of it if he had. Instead, he adopts a kind of logical formalism derived from Lévi-Strauss and Mary Douglas. The main problem with Leach, however, is not his underpinning theory, but his methods, or lack of them. The diagram in fig. 5 is drawn from Leach’s imagination, augmented by his knowledge of the role of certain oppositions and notions from English popular culture and sacred texts. Leach uses the category ‘nature’ in a way that he has presumably derived from Lévi-Strauss, though he is sceptical of naive dualism and ambiguous on the universal proclivity of the nature-culture opposition (e.g. Leach 1972; 2000). Nature is, therefore, not the same as ‘life’ or ‘living organisms’, or even less Taylor’s ‘biotic forms’. He divides ‘nature’ into animate and inanimate, though we are left to speculate whether this means that the category ‘inanimate’ includes some living organisms, most obviously plants. His category ‘animate’ (lexico-logically) applies only to animals, given that it is then sub-divided into warm-blooded and cold-blooded. However, although I have often used Leach’s folk-English scheme as a useful teaching example, it has always struck me as highly problematic. Not only does it suspiciously provide a convenient taxonomy of all living things that fits the logical conventions of a taxonomic approach, but it conflates discourses that we know to be separate, and flies in the face of ethnographic evidence. Unfortunately, ‘living things’ is a category that is easily
Tools, Agency and the Category of Living Things
249
yielded using formal elicitation techniques, but it is less obvious once you aggregate data from different ethnographic contexts using different research procedures. ‘NATURE’ animate
inanimate
warm-blooded (meat)
cold-blooded (not meat)
LAND CREATURES BEASTS
(ambiguous creatures)
BIRDS
INSECTS
(sometimes edible) (tame)
PETS
LIVESTOCK
HOUSE (inedible) SELF-dog*-cat-horse ass goat
REPTILES
FISH
(inedible) (ambiguous)
(inedible)
WATER CREATURES
GAME
(edible subject to rules)
FARM (edible if castrated) pig-ox-sheep
(edible)
(wild)
WILD ANIMALS (inedible)
OPEN FIELD
(near)
(far)
REMOTE
(edible if hunted at correct season)
(inedible)
rabbit-hare-deer-fox cony
zoo animals
Fig. 5: Edmund Leach’s (1964) version of the English classification of nature
Possibly because of the pitfalls in defining it, or in order to avoid the problems of using it as a unique beginner in a taxonomic description, many accounts of ethnobiological classification simply avoid considering what is meant by a ‘living thing’, and simply get on with the business of describing categories within the separate plant and animal domains. This approach is adopted by Hunn (2008) in his recent Zapotec natural history, and not even Berlin (1992), who while adopting the over-riding rubric of ‘ethnobiological classification’ and dwelling on the formal criteria for establishing ‘folk-kingdoms’ of (often nomenclaturally covert) plants and animals, is interested in considering the unique beginner that might define the domain of ‘living things’ overall. Such approaches, however, logically (that is inferentially or syllogistically) assume that ‘living things’ must exist as a categorical phenomenon. Thus, if plants are living things and animals are living things, then there must be a superordinate or more encompassing category that both belong to. But this definition assumes that we are only dealing with what might be conventionally understood as ‘biological taxa’. For an example of an early ethnoscience analysis of the category ‘living thing’ which, unlike the example imagined by Leach, is based on real ethnographic data, we can do no better than look at a paper published by Mary Black in 1969 on the Ojibwa category ‘/bema.diziwa.d/’. This Black describes as the ‘head-term of the taxonomic universe’ and it indicates just how complex the concept can be in a particular local manifestation. Black uses a classic lexicographic distinctive feature
250
Roy Ellen
approach. In her table 1 (1969: 186) ‘living things’ are divided into ‘indians’, ‘white people’, ‘negroes’ and ‘asiatics’; and in another context into ‘large animals’, ‘insects’ and ‘other’. In her table 2 (1969: 177) ‘living things’ are divided into ‘human’, ‘large animals’, ‘small animals’, ‘birds’, ‘fish’, and ‘spirits’. And we might recall here that Hallowell (1960) sees Ojibwa spirits as ontologically ‘nonhuman persons’. Moreover, ‘/bema.diziwa.d/’, depending on informants, sometimes also includes ‘trees’, ‘stones’, ‘leaves’, ‘berries’, ‘shells’, ‘sun’ and the ‘moon’ (1969: 178). Well aware of the problems of confusing category with label, Black explains that choice of label here depends on ‘context and type of contrast or relation’. They are taxonomically homonymous – sometimes just people, sometimes all living things. So, in Black’s understanding of Ojibwa ontology, ‘living things’ may be biological taxa, human ‘racial’ groups, spirits or astronomical entities.
4 Comparing biological and non-biological classifications Although we might now dismiss some of the formal approaches discussed above as methodologically naïve, they do seem to point at some interesting similarities in the way in which the cultured mind makes sense of domains as diverse as biodiversity, human groups, the spirit realm, and other natural kinds such as minerals. Sixteenth and seventeenth century European natural history routinely incorporated minerals as natural kinds. Early attempts to systematically describe the ‘mineral kingdom’ and related entities ‘dug out of the earth’ often followed closely the organization of local floras (Cooper 2007: 87–115). This approach was hardly novel even then. Rumphius in his Ambonese Curiosity Cabinet (1705 [1999]) follows Pliny who discussed minerals in his Natural History. Rumphius juxtaposes descriptions of species of crustacea, echinoderms, coelenterates, seaworms, molluscs, cephalopods, stones, metals, minerals, gemstones, concretions, and objects that come from animals and plants often used medicinally. Thus, a kind of crystal found in Ambon is named Crystallus ambonica, while Amianthus ambonicus is a variety of asbestos (1705 [1999]: 538). Also included are curiosities such as Mestica sontong (cuttlefish stones, 1705 [1999]: 531) and Dendrites metallica (small pieces of iron found in trees), or Ambra grysea (ambergris, the intestinal secretion of the sperm whale, 1705 [1999]: 499). Many of these objects are described using a version of the Latin binomial system. Then there are fossils such as Cancri lapidescentes (said to live under water but petrify when removed), and prehistoric tools and weapons. In such schemes fossils provided both a link with living biota and a problem. The idea that stones might reproduce, as Theophrastus had suggested in his On Stones, had still not been entirely repudiated. Following Rumphius, Linnaeus too attempted a ‘taxonomy’ of mineral ‘species’, and in the tenth edition (1758–1759) of the Systema Naturae proposed 4 ‘classes’: Petræ, Mineræ, Fossilia and
Tools, Agency and the Category of Living Things
251
Vitamentra. Such classifications and attempts to integrate minerals in a more general taxonomic approach to natural history were eventually abandoned, superseded in global science by different arrangements based on chemical composition. But as both the history of science and the comparative ethnographic record attest, there have been repeated attempts since to integrate various domains of natural history knowledge (Atran 1990: 78), but generally resulting in the confrontation between irreconcilably different organising principles. The same thinking has been extended to the humanly-made world of artefacts. Drawing his inspiration from Berlin’s work on the universal features of folk ethnobiological classifications, Brown (Brown et al. 1976) claimed to find the same formal features of taxonomy in the organization of other domains, including both human artefacts and spirits, referencing also the work of Frake (1961, on disease categories) and Spradley (1970, on social categories). Brown concludes that ‘principles originally attributed solely to biological classification (1976: 73) extend beyond biological and perhaps even beyond taxonomic classification’ (1970: 84). He has in mind hierarchic depth, class contrast and inclusion, partonymy, and how nomenclature relates to category level. We can see some of these features in his rendering of American English tool taxonomy (fig. 6).2
Fig. 6: American English tool taxonomy
2 In considering the ranks in non-biological domains Brown is content to employ the term ‘life form’ for level 1, though without flagging-up any sense of irony. Of course the generic level, like many concepts employed in Western scientific and other folk classifications of living things, frequently reflect social categories (genus, family, tribe …), echoing another broadly understood principle, that is how – of necessity – we use social categories to make sense of the natural world and non-social categories to make sense of nature, e.g. Lévi-Strauss 1966. The relationship between tool and organism displays the same conceptual mutuality, one domain being used to explain the other. Consider, for example, Aristotle’s use of ‘organon’ in his natural history, and the long post-Aristotelian tradition of explaining animal function using mechanical analogies, e.g. Smith 2011.
252
Roy Ellen
In addition to treating certain tools in a life-like way, Nuaulu distinguish entities that have many of the qualities of life, but which in the ordinary way have no corporeal existence. Thus, Nuaulu – like many peoples – think of spirits as a kind of pseudoorganism and represent them as species-like entities that can be ordered in terms of family resemblances (Ellen 1993: 176–190), even if, as Boyer (1994: 97) argues, their conception as such violates the intuitive principles that some psychologists (e.g. Keil 1979) have demonstrated for the domain-specificity of living kinds. Because spirits have these qualities they are sometimes also attributed with corporeal manifestations: they become birds, or lizards, or indeed certain animals may always be regarded as the physical manifestations of spirits, such as scarab and long-horned beetles, or the death adder (Acanthophis antarcticus). In a way this is not surprizing, since we can only imagine the spirit world through our experience of the physical and social world. It was difficulties of this kind – both ontological and epistemological – that pre-occupied Scott Atran in much of his Cognitive Foundations of Natural History (1990). For Atran there is only a superficial similarity between biological and nonbiological domains, ‘living things [… being …] everywhere ranked into transitively structured taxonomies, with no other natural-object domain so structured’ (1990: 47). In concluding thus he draws on the support of field experiments in child psychology (1990: 50) and the work of Keil (1979). Keil has argued that children possess an ontological category of living things that includes animals and plants, allowing Atran (1990: 73 f.) to claim that young children ‘categorically distinguish artefacts from living things’ and come to presume that only the latter constitute ‘natural kinds’ with underlying essences, while limiting certain concepts (such as growth) to living things. For Atran (1990: 5) there is a significant contrast both in regard to the ‘ordinary categorization of artefacts and the extraordinary scientific classification of living kinds’. Thus science is based on a kind of universal ethnoscience in which ‘our universally held conception of the living world is both historically prior to, and psychologically necessary for, any scientific – or symbolic – elaboration of that world‘ (1990: 13). As well as using experimental data, Atran (1990: 55–57) argues his case on formal logical grounds assuming this to be a universal grammar. Thus, furniture cannot be part of the definition of chair, although animal is part of the definition of cat. Transitive hierarchy, he therefore reasons, works for living kinds but not for artefacts, because the domain of artefacts fails to meet the inductive and deductive requirements of ranked taxonomies. Biota, like artefacts are often placed in different categories (whether we call these taxa or not), and although he recognizes that people often ‘confound’ artefacts with living kinds and ‘confuse’ plants with things made from plants, the underlying field structures are quite different. But one reason that artefacts are not conveniently organized taxonomically is because Atran has defined taxonomy in such a way that it can only be used for biota (1990: 57). If we are to argue on the basis of logical formalism, then we might forgive ordinary people from deviating from it in the practices of their ordinary lives. No wonder Atran (1990: 71) dismisses that work in developmental psychology (e.g. Carey 1985; see also Gelman 2003)
Tools, Agency and the Category of Living Things
253
that shows that children will spontaneously attribute a common invented property, a kind of underlying nature, to dogs, flowers and inanimate objects. Such theories and data tend to support the idea of the integration of domains of knowledge, and of fundamental ambiguity in the concept of natural kind.
5 The body, social cognition and the limits of modularity While we may still wish to argue that there are special features in ethnobiological classifications that distinguish them from the organization of other domains, I would argue that this has less to do with the cognitive apparatus brought to bear on them than the pattern of empirical discontinuities found in a particular ecological context. Similarly, the mind cognizes tools in the same way as living organisms partly because of a shared cognitive architecture, but also fundamentally because our models of apprehending and thinking about the world derive from our own bodily experience, and from the social worlds of which we are part. It has long been recognized that at the core of human cognition is a necessary duality and tension whereby humans understand the natural world through their experience of society, and the social world through their experience of nature (e.g. Lévi-Strauss 1962). This is why despite repeated attempts to counter naive dualism and challenge the culture-nature divide, the divide keeps on re-emerging (Leach 2000: 340; Astuti 2001). Thus, there is a general tendency in human relations with the inanimate world to attribute and represent that world in organic terms, and to attribute inanimate objects with the properties of living things (Ellen 1988). It happens because we are bound to model our world directly on our experience of our own body (Mauss 1934) and we employ this selfsame model as a source of labels and concepts to interpret the world outside the body. The lexicon of animal parts is, after all, for the most part that of human anatomy. Botanical nomenclature is less so, and that of inanimate objects less still, but body terms – or at least terms that appear concurrently in anatomical lexicons – are still crucial (Ellen 1977). More than this, if we thingify or entify parts of a living system, and then observe that the things move, so to speak, it logically follows that the things may well be regarded, or spoken of, as if they were sentient beings; they will appear as though they were indeed animated. Thus, phenomena that have life are turned into objects, only be re-animated in turn (Ellen 1988: 223). The organic models we use vary along a continuum from general organic analogies (organomorphs), plant analogies (phytomorphs), animal analogies (zoomorphs), general human analogies (anthropomorphs), and the attribution of particular personalities (personification) (1988: 224). It is as if the mind progressively enlivens non-living entities, and humanizes other entities in which it recognizes life. I have elsewhere illustrated this point in relation to the personification of sacred shields that Nuaulu attribute with soul (nemati). The
254
Roy Ellen
shields and other sacred valuables (such as sacred spears or barkcloth beaters) are treated with reverence, anthropomorphically, granted personhood and have complex biographical histories, as reported for other parts of Indonesia (Ellen 1990, Hoskins 1998). When Viveiros de Castro (2004: 465) tells us that Amazonian beliefs attributing perceptual perspectives of the world to different species are connected with their belief that animals are ex-humans, this is not only an example of mythic legitimation but of how objects are treated as subjects and species as persons. We can see the same process at work when Carrithers, Bracken and Emery (2011) refer to ‘the person-like character of a species’ in conservation and taxonomic discourse and show how the ‘axiom of amity’ finds an afterlife in the same. At this point too even Atran seems to concede the point that ‘because humans and animals are adjacent overlapping domains, then one might expect children to borrow from knowledge of humans to organize animals and plants ’(Atran 1990: 74). Because the role of the social world in the way we organize and understand nature is so entrenched in ordinary thought and discourse, this has been taken as good evidence that the faculties of social cognition have evolved in humans to eclipse other modules of knowledge (Mithen 1996); at some point in our evolutionary history the barriers between hitherto specialized intelligences distributed as quasi-separate neural networks in the brain – natural history cognition, intuitive physics and others – breakdown and merge with the hegemonic module of social cognition. Thus, if we are to understand those cultural behaviours we describe as animistic we must take into account a fundamental fact of human cognition: that we use social intelligence to make sense of the natural world. What we call animism is in a way the reification of multiple instances of such thinking, of repeated observations, reinforced by cultural elaboration into something approaching a coherent set of connected beliefs.
6 Organomorphism, motion, agency, intention Thus, through the working out of the recent evolutionary history of the embodied brain, and polythetic linkage between different cognitive and semantic domains in any particular cultural population, we will find a group of things that are regarded as like life, or life-like. But in attributing the notion of life and more narrowly animacy, what features does the mind latch on to? The evidence of the previous section would suggest that we start with physical resemblance, especially since given our dominant sense of vision it is visually-salient attributes that first register and that are most easily encoded in memory. In other species it might be other senses that occupy this position, but even in humans these other senses can be critical in determining whether something is likely to be living. Imagine a world in which life is mainly apprehended and registered through the senses of sound or tactility (which we can perhaps imagine, especially if we are blind) or through smell (which we can only imagine with
Tools, Agency and the Category of Living Things
255
difficulty). In reality, of course, our cognition of life is multi-sensorial and synaesthetic, but we must start with physical form. Some of the classic boundary problems of the category living matter are raised through an engagement with forms of life that do not fit easily into the cognitively universal (but not lexically-universal) prototypical categories of plant or animal, that are peripheral to biological life as most ordinary people experience it. Our different senses register different qualities on the life/ non-life continuum, but combined offer a sufficiently discriminating instrument. In my work on Nuaulu ethnobiological classification fungi and algae are classic liminal forms (Ellen 2008). In the classifying and naming strategies that Nuaulu employ there is a tension between placing fungi with plants and according them their own separate kingdom, while seaweed is nomenclaturally aligned with fungi. The term unate refers to all visible fungi, with the exception of lichen-forming Ascomycetes; and is also applied to sponges (unate nau moti, literally, ‘mushrooms of the reef’), and to all algal seaweed (una nuae), despite the first being phylogenetically Animalia and the latter either Plantae or Protista. All unate are marked lexically by the contracted prefix unafollowed by an adjectival or other qualifier: for example, una msinae (red mushroom) for Pycnoporus sanguineus, or una pate (fig-tree mushroom) for Trametes corrugata. The broadly inclusive character of unate is in itself interesting, given that mushrooms as a phylogenetic grouping are extremely varied in shape, structure, colour, habit and reproduction. Since there can be considerable morphological difference, say between rigid shelf mushrooms or bracket fungi, fragile fleshy mushrooms and the highly salient coral mushrooms, we must assume that placing them together in a single clearly labelled category (whether ‘mushrooms’ or unate) must reflect some combination of cognitive prototype and common distinctive features. Not subsumed under unate are forms such as Usnea, a fruticose lichen which Nuaulu call ahane, and molds and mildew (which they call rekunai). Slime molds, which are not fungi, are given their own category as well: sona, literally ‘sago jelly’ which they are said to resemble. Freshwater and terrestrial algae are described as mapunua by Nuaulu and seen as related forms e.g. lumu-lumu (Chaetomorpha javanica) found commonly on trees. For Nuaulu all these forms are undeniably living, even though some seem to lack movement, others much evidence of growth, and others still a capacity for reproduction. Movement in the narrow sense is easily observable, growth less so, but reproduction in many cases has to be inferred. In most folk biologies while sexual biotic forms clearly reproduce, entities such as sponges merely endure. So, by deliberately selecting liminal forms we can test the extent to which the notion of life applies, and if that test is passed whether such entities conform best to animal or to plant prototypes. But we also need to recognize that both these prototypes are in semantic tension, simultaneously sharing features and contrasting them. All living matter could be said to have both a vegetal and animalistic aspect, there being a vegetal quality especially found in plants, and an animal quality especially found in animals. This is why where certain peripheral organisms are placed in classificatory space varies between cultures, compared with the greater regulari-
256
Roy Ellen
ties reported for core plants and animals. But it is also reflected in the evolutionary convergence of forms that we might otherwise have no problem is assigning to one or other of the taxonomic kingdoms. Thus, for Nuaulu, certain plants have attributes of animals, such as the insectivorous pitcher plants (koitipi: Nepenthes spp.) found in the Manusela National Park, and certain animals the attributes of plants, as in various phasmid stick and leaf insects (kau ai otoe, e.g. Platycrana viridana). But in addition, we are constantly also primed to look out for what is potentially human in other parts of the living world, as we have seen from the previous section. We try to make things anthropomorphic, but while the basic attributes of morphology make this more promising in the case of, say, the Nuaulu marsupial cuscus, with its simian anatomy and multiple symbolic resonances (Ellen 1972), conservationists struggle to convince us of the humanity in the hapless pearl mussel (Carrithers et al. 2011). In the attribution of life, and even more so of animacy in the sense of ‘animality’ (Reed 1988), morphological resemblance is not enough. As indispensible is motion, in all or any of its manifestations. Everywhere that liminal biological forms are attributed with animacy it is because in some way or another they display a characteristic that is semantically rooted in movement or its metaphorical extension, such as locomotion, growth, reproduction, fission, fusion, fragmentation. Even eruption and erosion are recognizably for all humans, kinds of motion. When we talk about the living landscape we have in mind the idea that it is dynamic, whether vegetally, animalistically or geomorphologically. Time, change and cause are all described in various contexts as if they were like motion. But to qualify for the condition of life, the source of motion has to be independent of any interlocutor. So, as we move between the different types of Nuaulu tool in the order that we considered them above, the cognitive stimuli amenable to the attribution of life seem to increase. In the transition from skewer to sago pounder, and from sago pounder to the sago flour extraction apparatus, from the treadle-operated coconut grater to the power-assisted sago grater and the outboard motor, there is a gradual shift in the source of the motion – and therefore seemingly of agency: from the using subject to the tool itself. We now come to agency in the strict sense. As Gell (1998: 72) puts it, ‘agency implies the possession of a mind which intends actions prior to performing them’ and ‘what matters is where an animated object stands in a network of social relations’ (1998: 123). For Gell (1998: 132) there is only a slight dividing line between the intentionality of humans and of anthropomorphic objects. The problem with this is that, conceptually, it mixes up not only agency with intention, but with animacy and sociality as well. The term agency has been much stretched in recent anthropological discussions of animism and technology, but in this context I think we need to shrink it back to its core meaning: an intervention to produce a particular result, or something with ‘the faculty of an agent’ (Little, Onions and Friedrichsen 1973: 37). Thus, a sagoprocessing device has agency in the sense that it can produce unique outcomes in the context of human-object interaction. This neither makes it animate nor gives it intentionality. So, looking at our selected Nuaulu tools, on material grounds neither the
Tools, Agency and the Category of Living Things
257
meat skewer nor the sago pounder can be said to have agency, though more complex machines with many parts which form a system with emergent properties not entirely under the control of the operator might indeed be said to possess agency. Thus, this would be the case for the sago-processing device, the coconut grater and the outboard motor. A quality that Nuaulu commonly associate with the essentials of animal and human life is breathe (nahai), and breathe too is a kind of motion. At the birth of a child the moment at which a child breathes independently, when the chest begins to move and the lungs ventilate, is the beginning of autonomous human life, while death is the expiration of breathe and the end of movement, and therefore is (paradoxically) also a characteristic of life. Breathe as a concept is recurrent in ritual and attributed to physical entities that are anthropomorphized, such as sacred houses or ritual shields. Ritual shields, like the meat skewer, retain their organs of breathe in the first chip of wood to be cut, and which is thereafter kept in the loft of a sacred house. Most salient of all, however, is the literal extraction of the organs of breath (the lungs and trachea) of large game animals and their offering as a sacrifice to the clan ancestors in the form of a gift to the clan head. In the asunaete ritual these same organs are attached to a wooden skewer as a form of repayment for a life taken and as to ensure that life will thereby be replenished. In humans and large animals the physicality of breathe is clear enough, but in other biological organisms and non-biological entities it is not, and here we find that sound (more specifically vocalization) may sometimes serve as a proxy, for sound is only possible where there is breathe or motion to produce it. So, when cicadas sing it is evidence of breathe, or even when the wind blows through certain rocks to produce eery sounds it may be taken as evidence of life, recalling somewhat the remarks of Herodotus on the Colossi of Memnon at Thebes. Finally, we come to intentionality. In the Nuaulu world, humans, some animals, spirits, the godhead, animals and some plants that have become coterminous with spirits might be said to act intentionally. Their actions in the world do not simply have consequences independent of the humans that interact with them, but they have minds that permit them to make plans and act in the world in ways that are deliberate responses to human and other behaviour, and which are often contrary.
7 Life as a matter of degree, while animacy is not animism I am hardly the first to argue that the attribution of life is necessarily gradual, contextual and, from a biological perspective, sometimes inconveniently deviant. Bloch (1998: 53) calls this the ‘more-or-less’ character of life. Ingold (2006) argues that we can make anything seem alive, but we do not always chose to do so. Moreover, there is much ethnographic evidence that people do not agree about what life is and no uni-
258
Roy Ellen
versal distinctions as to what is alive and what not. Ingold (2011: 29) further speaks of ‘bringing things to life’, of ‘things in life, rather than life in things’, suggesting that life is less a property of individual entities than a phenomenon of which those entities are part. In this model, the animacy of the lifeworld is ‘ontologically prior to their differentiation‘ (Ingold 2006: 10). What I have tried to do here is to unpack those conditions for the recognition of life and to show how they might constitute a series of progressive cognitive steps that when aggregated are more likely to prompt the attribution of life than not. These are: recognition of morphological resemblance, recognition of motion, and recognition of independent motion.3 As we have seen, these are each reflected in a pattern of conceptual attribution that follows a broadly phylogenetic progression: we apprehend physical entities as being plant-like, animal-like, humanlike and – ultimately – like individual human persons (Ellen 1988). Part of the problem in the literature is that the discourses on biocognition, life and animism begin from different starting points and have different intellectual histories. While analytically we need to separate these and certainly not confuse and conflate the concepts and terms employed, in particular ethnographic cases it is unlikely that we will ever discover a convenient congruence. It may be that we need to separate vital from symbolic forces – that we need both ‘vital (biological), and symbolic (spiritual) ontologies’ – just as Mauss suggested we separate technical from symbolic classifications, and Berlin general-purpose ‘natural’ schemes from special-purpose schemes. Unfortunately real-life examples do not give us much hope that this is possible. In some cases we may find evidence of two life forces (spiritual and biological) as Rival (2012) suggests for Makushi cassava, but my informed bet is that because ‘vital and technical processes [are] already situated in a complex relational complex when we experience them’, in practical everyday life organisms and things are treated as though they are motivated by a single underlying force, unless otherwise prompted to re-consider. This would be my reading of the Nuaulu data. While all cultural objects are potentially regarded in this way, in practice some objects are more alive, are more likely to evoke the characteristics of agency or indeed intentionality than others. Part of the problem with existing anthropological treatments of the category living things is that they are either logical extrapolations through polythetic extension, based on formal ethnotaxonomic deduction/induction, rooted in observations of how children under artificial conditions perform in field experiments of a very abstract kind, or (completely differently) assume that life is recognized phenomenologically prior to its differentiation. Nuaulu ideas of what is animate and agentive are always fuzzy and contingent. By combining data from different kinds
3 Compare this with a similar progression noted by Rival 2012: 71 in the contributions from Ingold 2006 and others in the animism debate regarding the logical implicatory relationships between communication, intentionality, consciousness, life, and movement. The problem, as Rival observes, is in a ‘hazardous slippage’ between concepts.
Tools, Agency and the Category of Living Things
259
of ethnographic context, using different elicitation procedures, a more complex picture emerges. The same applies to animism, with which the attribution of life is too readily conflated. If, following Descola (2005: 183 f.), animism is ‘the granting by humans to non-humans of an interiority identical to theirs, an attribution that humanizes animals (and plants)‘ (Rival 2012: 70), I would say that it is virtually impossible to separate ethno-theories of life processes from what is sometimes described as animism. However, not every object in nature is animated, and animism is not totalizing. Recent work (including my own) suggests much selectivity with which it is evoked ontologically, and limits to the extension of personhood as a humanlike category. Praet (2014) even argues that the outstanding feature of animism is its peculiar restrictiveness. Moreover, regardless of the arguments surrounding how we might best understand animism as a specific worldview, whether we are considering humans or other species (including dogs), we can observe a tendency for the mind to use an understanding of living bodies to interpret experience of artefacts, including tools. When they exhibit evidence of motion, and especially independent motion, objects prompt responses and interactions that suggest recognition of animacy, agency and intentionality to different degrees. If you do not know whether something is alive then, as Ingold (2006) suggests, it is better to assume that it is. We have evolved a tendency to attribute the characteristics of life to parts of the world and to the world as a whole, since our experience of existence is how we must represent, model, understand and act in the world. Life in its most generalized phenomenological sense must emerge, in terms of our experience, from the aggregation of lives in particular. When Nuaulu refer to mahai (life), they are first and foremost thinking of human life, but this does not mean that they do not also readily extend the notion to other biota and non-biota as necessary. A final word on motion as a condition of life. Motion is often necessary, but is seldom sufficient. Motion is often accompanied by multi-sensorial – often synaesthetic – characteristics that are shared with biological life, for example the expenditure of heat and the emission of sound. But while none of the clanking, whirring and buzzing of the treadle-operated coconut grater, or even the sloshing and slapping of a sago-processing device are sufficient for Nuaulu to conceive of them as living entities, the same features in an outboard motor or power-assisted sago grater are intrinsic to recognition of its animate status. For while the treadle-operated coconut grater and sago-processing device have the technical characteristics of a machine, they do not have the autonomy of an engine. Once primed and fueled the engine will run by itself until the fuel runs out or malfunctions. These technical processes are fully understood and Nuaulu have wondrous ways of fixing malfunctioning engines, but the combined features that give them quasi-autonomy also give them the vitality that is more than just the combination of the parts and crosses a boundary that places them with other vital biological and quasi-biological entities such as spirits. This is so despite their not sharing other characteristics that are often focal to our defi-
260
Roy Ellen
nitions of life, such as growth and reproduction. Humanly-operated machines may have agency, but engines can also appear to act intentionally. When Basil Fawlty is thrashing the car he is exacting revenge through punishment from an entity that has ‘stalled just once too often’. It has willfully disobeyed its owner and driver. We laugh because we recognize that all of us, while fully accepting the technical reasons for mechanical failure, insist on treating the vehicle as if it were a sentient person who is deliberately contrary.
References Astuti, Rita. 2001. “Are We All Natural Dualists? A Cognitive Developmental Approach.” Journal of the Royal Anthropological Institute 7: 429–447. Atran, Scott. 1990. Cognitive Foundations of Natural History. Towards an Anthropology of Science. Cambridge: Cambridge University Press. Berlin, Brent. 1992. Ethnobiological Classification: Principles of Categorization of Plants and Animals in Traditional Societies. Princeton: Princeton University Press. Black, Mary B. 1969. “Eliciting Folk Taxonomy in Ojibwa.” In Cognitive Anthropology, edited by Stephen A. Tyler, 165–189. New York: Holt, Rinehart and Winston. Bloch, Maurice. 1998. “Why Trees, Too, are Good to Think With: Towards an Anthropology of the Meaning of Life.” In The Social Life of Trees: Anthropological Perspectives on Tree Symbolism, edited by Laura Rival, 39–56. Oxford: Berg. Boyer, Pascal. 1994. The Naturalness of Religious Ideas; A Cognitive Theory of Religion. London: University of California Press. Brown, Cecil H., John Kolar, Barbara J. Torrey, Tipawan Truong-Quang and Phillip Volkman. 1976. “Some General Principles of Biological and Non-biological Folk Classification.” American Ethnologist 3: 73–85. Carey, Susan. 1985. Conceptual Change in Childhood. Cambridge, Mass: MIT Press. Carrithers, Michael, Louise J. Bracken, and Steven Emery. 2011. “Can a Species be a Person? A Trope and its Entanglements in the Anthropocene.” Current Anthropology 52: 661–685. Cooper, Alix. 2007. Inventing the Indigenous: Local Knowledge and Natural History in Nearly Modern Europe. Cambridge: Cambridge University Press. Coppinger, Raymond, and Richard Schneider. 1995. “Evolution of Working Dogs.” In The Domestic Dog: Its Evolution, Behaviour and Interactions with People, edited by James Serpell, 21–47. Cambridge: Cambridge University Press. Descola, Philippe. 2005. Par-delá Nature et Culture. Paris: Gallimard. Ellen, Roy. 1972. “The Marsupial in Nuaulu Ritual Behaviour.” Man 7: 223–238. Ellen, Roy. 1977. “Anatomical Classification and the Semiotics of the Body.” In The Anthropology of the Body, edited by John Blacking, 343–373. London: Academic Press. Ellen, Roy. 1988. “Fetishism.” Man 23: 213–235. Ellen, Roy. 1990. “Nuaulu Sacred Shields: The Reproduction of Things or the Reproduction of Images.” Etnofoor 3: 5–25. Ellen, Roy. 1993. The Cultural Relations of Classification: An Analysis of Nuaulu Animal Categories from Central Seram. Cambridge: Cambridge University Press. Ellen, Roy. 1996. “Individual Strategy and Cultural Regulation in Nuaulu Hunting.” In Redefining Nature: Ecology, Culture and Domestication, edited by Roy Ellen and Katsuyoshi Fukui, 597–635. Oxford: Berg.
Tools, Agency and the Category of Living Things
261
Ellen, Roy. 1998. “The Inedible and the Uneatable: Totemic and Other Restrictions on the Use of Biological Species Among the Nuaulu.” In Old World Places, New World Problems: Exploring Issues of Resource Management in Eastern Indonesia, edited by Sandra Pannell and Franz von Benda-Beckman, 243–266. Canberra: Australian National University. Ellen, Roy. 2003. On the Edge of the Banda Zone: Past and Present in the Social Organization of a Moluccan Trading Network. Honolulu: University of Hawaii Press. Ellen, Roy. 2004. “Processing Metroxylon sagu (ARECACEAE) as a Technological Complex: A Case Study From South Central Seram, Indonesia.” Economic Botany 58: 601–625. Ellen, Roy. 2008. “Ethnomycology Among the Nuaulu of the Moluccas: Putting Berlin’s ‘General principles’ to the test.” Economic Botany 62: 483–496. Ellen, Roy. 2012. Nuaulu Religious Practices: The Frequency and Reproduction of Rituals in a Moluccan Society. Leiden: KITLV Press. Frake, Charles O. 1961. “The Diagnosis of Disease Among the Subanun of Mindanao.” American Anthropologist 63: 113–132. Gell, Alfred. 1998. Art and Agency: An Anthropological Theory. Oxford/New York: Clarendon Press. Gelman, Susan. 2003. The Essential Child: Origins of Essentialism in Everyday Thought. Oxford: Oxford University Press. Hallowell, A. Irving. 1960. “Ojibwa Ontology, Behavior and World View.” In Culture in History: Essays in Honour of Paul Radin, edited by Stanley Diamond, 19–52. New York: Columbia University Press. Hoskins, Janet. 1998. Biographical Objects: How Things Tell the Stories of People’s Lives. New York: Routledge. Hunn, Eugene S. 2008. A Zapotec Natural History: Trees, Herbs and Flowers, Birds, Beasts and Bugs in the Life of San Juan Gbëë. Tucson: University of Arizona Press. Ingold, Tim. 2006. “Rethinking the Animate, Re-animating Thought.” Ethnos 71: 9–20. Ingold, Tim. 2011. Being Alive: Essays on Movement, Knowledge and Description. London: Routledge. Keil, Frank. 1979. Semantic and Conceptual Development: An Ontological Perspective. Cambridge, Mass: Harvard University Press. Leach, Edmund. 1964. “Anthropological Aspects of Language: Animal Categories and Verbal Abuse.” In New Directions in the Study of Language, edited by Eric H. Lenneberg, 23–63. Cambridge, Mass.: MIT Press. Leach, Edmund. 1972. Humanity and Animality. (Fifth Conway Memorial Lecture) London: South Place Ethical Society. Leach, Edmund. 2000 [1977]. “Anthropos.” In The Essential Edmund Leach, edited by Stephen Hugh-Jones and James Laidlaw, 324–380. New Haven, CT: Yale University Press. Lévi-Strauss, Claude. 1962. Le Totémisme Aujourd’hui. Paris: Presses Universitaires de France. Lévi-Strauss, Claude. 1966. The Savage Mind. London: Weidenfeld and Nicholson. Linnaeus, Karolus. 1758–1759 [1735]. Systema Naturae (10th edition). Holmiae: Impensis Direct. Laurentii Salvi. Little, William, Charles T. Onions, and George W. Friedrichsen. 1973. The Shorter Oxford English Dictionary. Oxford: Clarendon Press. Mauss, Marcel. 1973 [1934]. “Techniques of the Body.” Translated by Ben Brewster. Economy and Society 2: 70–88. Miller, Daniel. 1987. Material Culture and Mass Consumption. Oxford: Blackwell. Mithen, Steven. 1996. The Prehistory of the Mind: A Search for the Origins of Art, Religion and Science. London: Thames and Hudson. Oswalt, Wendell H. 1976. An Anthropological Analysis of Food-getting Technology. New York: Wiley. Praet, Istvan. 2014. Animism and the Question of Life. London: Routledge.
262
Roy Ellen
Reed, Edward S. 1988. “The Affordances of the Animate Environment: Social Science From the Ecological Point of View.” In What is an Animal?, edited by Tim Ingold, 110–126. London: Unwin Hyman. Rival, Laura. 2012. “Animism and the Meaning of Life: Towards an Understanding of Manioc Domestication.” In Animism in Rainforest and Tundra: Personhood, Animals, Plants and Things in Contemporary Amazonia and Siberia, edited by Olga Ulturgasheva, Marc Brightman and Vanessa Elisa Grotti, 119–141. Oxford: Berghahn. Rival, Laura. 2013. “The Materiality of Life: Revisiting the Anthropology of Nature in Amazonia.” In The Handbook of Contemporary Animism, edited by Graham Harvey, 92–100. Durham: Acumen. Rumphius, Georgius E. 1999 [1705]. The Ambonese Curiosity Cabinet. Translated by Eric M. Beekman. New Haven: Yale University Press. Smith, Justin. 2011. Divine Machines: Leibnitz and the Sciences of Life. Princeton, NJ: Princeton University Press. Spradley, James P. 1970. You Owe Yourself a Drunk. Boston: Little, Brown and Company. Taylor, Paul M. 1990. The Folk Biology of the Tobelo People: A Study in Folk Classification. Washington, DC: Smithsonian Institution. Viveiros de Castro, Eduardo. 2004. “Exchanging Perspectives: The Transformation of Objects into Subjects.” Common Knowledge 10, 463–484.
Figures Fig. 1: Photo Ellen 70-06-25. Fig. 2: After Ellen 2004: 612, fig. 10. Fig. 3: Photo Ellen 70-15-02. Fig. 4: Taylor 1990: 48. Fig. 5: Leach 1964: 41, table 1. Fig. 6: After Brown et al. 1974: 78.
Acknowledgments I would like to thank Walter Bisang and Tanja Pommerening for the invitation to Mainz that has permitted me to develop the ideas presented in this chapter. An earlier version was given at the colloquium Des Êtres Vivants et des Artefacts held April 2014 at the Musée du Quai Branly in Paris. I have benefitted from comments from Laura Rival, Perig Pitrou, Ludovic Coupaye, Adam Miklosi, Garry Marvin and (in Mainz) Joachim Kadereit and Jochen Althoff.
Roy Ellen
Categorizing Natural Objects: Some Issues Arising from Recent Work in Cognitive Anthropology and Ethnobiological Classification Summary: Classification is not logically, developmentally or historically prior to conceptualization, but rather consubstantial. Using examples of folk categories for plants and animals (including of humans), I show that classifications are always imminent, tentative and just sufficiently good to achieve their purpose. While there is evidence for some evolved regularities in the way classifying is cognized, it is always expressed culturally and contextually. This limits our ability to generalize about classifications as entities. As social phenomena, classifications are variable and dynamic, but under certain circumstances their shared elements can be stabilized and enforced. This happens through the conventions of language, where recognized authorities can enforce their will, and in the interests of effective communication within increasingly larger groups. Writing and pictures are powerful means of simplifying and sharing classifications, but they also transform them and introduce new properties.
1 Introduction My original brief for the workshop on which this volume is based was to respond to some of the issues in the position statement circulated by Tanja Pommerening and Walter Bisang. This is a tall order as these issues have all been the subject of much reflection in cognitive anthropology and allied subjects for the last half century. I fully realize that in an attempt to say something that is of general relevance regarding the organization of folk categories for plants and animals (including humans), I risk oversimplification in some specific areas. That said, I will focus here on just seven issues: classification as a situated process, the relationship between words and categories, hierarchies and taxonomy, cultural universals, domain specificity, the position of humans within folk classifications, and the impact of writing and depiction technologies. Nothing here is particularly original, and my main purpose is to provide an overview that may help connect papers in this volume coming from different disciplinary perspectives.
DOI 10.1515/9783110538779-010
264
Roy Ellen
2 Classification as situated practice The first thing to emphasize is that classifications as they are encountered in ordinary life are seldom fixed schemes to which we assign semantic content, but rather endproducts of a process. Put differently, classifications emerge consubstantially with conceptualization. There has been a problem in the past in attributing to them more stability than they deserve. Indeed, to speak of classifications is to run the risk of reifying schemes as permanent cultural artefacts or mentally stored old knowledge, when they are often more properly understood as the spontaneous and often transient consequence of underlying processes. The assumption to the contrary is what we might call ‘the classificatory fallacy’ (Ellen 2006: 27; 31). This is understandable when we think of the grand schemes that have arisen in the history of science (perhaps most obviously those of Linnaeus and Mendeleev), or when as students of ancient writing systems we are restricted to limited physical evidence (e.g. Goldwasser and Grinevald 2012). But when we look at classifications in the context of oral culture and social interaction, it becomes clear that they are generally imminent works in progress, just sufficiently good for their purpose, always tentative and plural. There is no one major scheme that all members of a particular cultural population share in its entirety, though there may be pressure to achieve as much commonality as possible. Consider an example, of cultural rather than natural objects. My father used to store screws in old tobacco tins that he aligned in shelves in his garden shed. Each tin had a label. These included the following: –– ‘Posidrive screws’ –– ‘Medium flathead screws’ –– ‘Short dumpy flathead screws’ –– ‘Smallish flathead screws’ –– ‘Largish woodscrews’ –– ‘Long large flathead screws’ –– ‘Short fat round head screws’ On first inspection this list is baffling. It contains trademark names (‘Posidrive’ TM); what Lakoff (1973) calls hedges: vague terms that defy the determination of boundaries (‘smallish’, ‘largish’ …), idiosyncratic allusive terms (such as ‘dumpy’), and a mixture of morphonyms describing the physical properties of the fixing agent (‘large’, ‘round’, ‘flathead’ …), as well as functional properties (what the screw is used for: e.g. ‘woodscrew’). Systematic contrast is difficult to establish across the range of different types because the same attribute does not appear in all names. The best we can achieve is a partial polythetic set (Ellen 2006: 33 f.). Thus, although this series made perfect sense to my father and allowed him to always find what he was looking for, I would defy anyone to reconstruct the list of labels as – say – a taxonomic hierarchy, a challenge I have routinely proposed to students in my Kent cognitive anthropology classes. A taxonomy is impossible to establish because there are so many cross-cutting
Categorizing Natural Objects
265
features, not all of which are present in each case, and because boundaries between categories are generally fuzzy. The information my father supplies is neither exhaustive nor logical, only sufficiently detailed for him to achieve his practical purpose. Of course, classifications have been the subject of interest in science for much of the twentieth century. As an object of anthropological scrutiny they came to prominence during the 1960s, exemplified in the British (constructionist) tradition by the writings of Mary Douglas, and in the American ethnosemantics (cognitive) tradition by the likes of Harold Conklin and Brent Berlin.1 My own work has always tried to bridge this divide and develop a more embedded approach. In particular, I have used the analysis of people’s categorization of natural kinds as a means of understanding how classifying more generally works. In this view it is no longer necessary, when speaking of cognition, to choose between the chicken position (an approach which assumes a world with pre-given qualities) and the egg position (where reality is simply the reflection of internal laws of the system).
3 Words and categories Most studies of classification begin with language. This is because, most data acquired in natural fieldwork settings are generated through interviews and by hearing people talk about what they think, perceive and experience because this is how most people themselves share classificatory knowledge; and because many classifications are stabilized in literate cultures and by researchers alike by committing them to written language. However, we have long known that words are not always a perfect indicator for the existence of categories: several words may label the same category, and the same word can be used for quite different ideas. What are often described as ‘classifiers’ by linguists in both spoken and written languages of various kinds (e.g. Grinevald 2000, Goldwasser and Grinevald 2012) – grammatical entities embedded in utterance or text – are often unreliable indicators of how people assign – say plants and animals – to groups. Moreover, some categories may be inferred without being labelled – they are covert but no less real (e.g. Berlin et al. 1974: 415). It is nevertheless true that the nomenclature for labelling categories tells us something both about classificatory knowledge and about the attributes which people find important in distinguishing different entities, attributes and phenomena. Words provide evidence of more inclusive categorizing strategies: for example, plant binomials often indicate two categories linked by a kind of relationship. Thus, for Nuaulu in eastern Indonesia, the term sinsin msinae, red sinsinte (a kind Codieum
1 Background to these debates will be found in several of the essays included in Ellen 2006. Note also that I draw on some of the content of that book, as well as Ellen 2011, for the present chapter.
266
Roy Ellen
variegatum, a croton widely available in Europe as a houseplant), is a binomial. Local linguistic conventions have to be carefully observed, and it is important to note, for example, that kasipi sinsinte is not a kind of sinsinte, but rather a variety of cassava. In this context sinsinte becomes, instead, an adjectival qualifier. What is important is not that something has a fixed name, but that the association is registered through continual and repeated perceptual events, reinforced over the longer term and transmitted between individuals. Early attempts to understand how categories within the cultural domain of natural kinds are established and used employed a distinctive feature model, in which category A was distinguished from category B by a number of key characteristics. For example, birds have wings, feathers, beaks and fly, in contrast to fish, which swim and have fins. This approach was largely based on lexicography and logic (e.g. Conklin 1962). However, it was early noted that the condition of contrast required for the model to work was not always evident. Thus, category A might be linked to category B by one common attribute, and category B linked to category C through a different common attribute, thus linking categories A and C even though they had nothing in common. Such a ‘polythetic classification’ (e.g. Ellen 1993: 128 f.) works on the same principle as my father’s idiosyncratic labelling system for screws; while a similar critique of distinctive feature analysis coming from a different disciplinary angle is offered by Barsalou (1992). As work on ethnobiological classification expanded it became obvious that the distinctive feature model was inadequate, and that a better way of modelling the cognition of basic and more inclusive categories might be in analogue terms, as cognitive prototypes. In this model – for example as exemplified in the work of Eleanor Rosch (e.g. 1977) – the brain has an image of, say, ‘birdness’ or ‘treeness’ to which incoming perceptual images are matched. The presence or absence of particular features is not an overriding consideration, only closeness of match. In this model an image could be a close match or a marginal match. Thus, in British English classification of birds a sparrow would configure closely the core prototype, but an emu would be marginal. Of course in practice, both the notion of contrasting features and cognitive prototypes are necessary to understand how classifications work in detail.
4 Hierarchies and taxonomies Early anthropological work on cognitive domains modelled their subdivisions largely in terms of taxonomy (e.g. Tyler 1969), meaning a hierarchical model of contrast and class inclusion. This is partly because taxonomies had become so dominant in the literary and scientific tradition of the West, and particularly because of the powerful precedent of Linnaeus. The recursive relationship between the properties of classifications developed in the literate mode of science and our interpretation of pre-literate
Categorizing Natural Objects
267
or non-literate classifying behaviour has proved to be a major complicating factor in their analysis, an issue that I shall return to below. However, the work of Brent Berlin and his associates (e.g. Berlin et al. 1973, Berlin 1992) has put forward a strong claim for taxonomies as the general way in which at least ethnobiological classification works cross-culturally, hypothesising a series of ranks reflecting the Linnaean scheme: unique beginners, life-forms, intermediates, generics, specifics and varietals. The principle of taxonomy is persuasive, as people undoubtedly classify living things into increasingly inclusive groups, and the idea of taxonomy provides a strong inductive framework for making systematic inferences about the properties of organisms. But this need not imply taxonomy in the formal or domain-specific sense. Systematic contrast and class inclusion are present across a number of domains. It is particularly striking in plants and animals because of their characteristic materiality or thinginess, and because they are the outcome of an evolutionary process reflected in patterned physical and behavioural resemblance. In the domain of living kinds these tendencies converge in a special way, not obviously because of the character of the mind which does the classifying, but because of regularities in the objective world so classified and to which the mind responds. However, there is also evidence that taxonomic thinking is encouraged more in some cultural contexts than in others, such as in the performance of formal pedagogic operations in classroom contexts, that it is – at least in its strong version – a consequence of patterns of cognitive practice that emerge with writing (e.g. Goody 1977, Lancy and Strathern 1981). Because of the propensity of most researchers to rely heavily on taxonomic thinking embedded in their own protocols, it is easy to yield taxonomies in patterns of data collected from research subjects who may be simply being helpful within the constraints imposed by methodologies such as frame elicitation and sorting tests. Moreover, if we accept the centrality of prototypical thinking and polythesis in classifying activity, it is not at all surprising that it is often difficult to establish systematic and consistent hierarchical relationships between superordinate and subordinate categories.
5 Cultural universals Claims for the universality of certain patterns of classification were easier to sustain in the past. Studies of intra-cultural variability (e.g. Romney et al. 1986) are now common and provide evidence for the role of social and situational factors, which challenge the validity of some key assumptions regarding the extent of sharing. Moreover, continuous exposure to even locally-distributed classificatory routines can result in a kind of fixing of culturally-derived habitual practices through neuroplasticity (Ellen 2008: 19–22). One universal, however, the existence of which few would
268
Roy Ellen
now deny, is that the classifications (and especially folk biological classifications) of all cultural groups display some concept of basic category, the segregates of which are then either aggregated or disaggregated (e.g. Rosch 1977, Atran 1990, Berlin 1992: 70–78). These may refer to natural kinds, people, social groups or other types of object or entity. But there is less agreement as to how consistent such a level of basic categorizing is cross-culturally. When applied to natural kinds, it is the phylogenetic genus that generally gives us the basic level for plants, species only obtaining priority with Linnaeus. More problematic, as we have seen, has been the notion of taxonomy. Brent Berlin has always argued in favour of its universality for ethnobiological schemes, but this only really works if we also assert the clear separation of general-purpose from special-purpose schemes; that is, those that are logical and natural from those that arise to meet particular cultural requirements – say medicinal plants or domesticated livestock. Indeed, the effective demonstration of the primacy of taxonomy depends on the extent to which categories are linked in a particular way, although we know that they are often flexibly connected in numerous different ways, ways that undermine implicit taxonomic levels and contrasts and the general-purpose/special-purpose distinction. It also depends upon the reliability with which we can elicit transitivity statements: i.e. if A is a kind of B, and B is a kind of C, then A must be a kind of C. Scott Atran (1998), who has followed Berlin in being a powerful advocate of the universality of the basic natural features of ethnobiological classification, suggests that his Itzaj Maya findings do not uphold the customary distinction between general-purpose and special-purpose classifications, which would violate their primary concern with ‘ecological and morpho-behavioural relationships’ in favour of abstract properties. I believe that one of the problems central to the methodology that we use to generate much of our ethnobiological data is not knowing quite how independent the system of ranks that we discover is from the kinds of concepts with which we start. On the whole, it is my experience that data from long-term ethnographic research are more consistent with the notion of a holistic and dynamic conception of the relations between categories, one which allows for the generation of particular classifications depending on context. For example (Ellen 1998), some cultures clearly place palms within trees whereas others assign them to separate categories. The variable position of palms in comparative ethnobotanical schemes and the nebulousness of its position as a life-form, intermediate or unaffiliated generic, is an excellent example of the preeminence of local ecological and cultural considerations. On balance, the more dense our knowledge of a particular domain the more we deviate from general models, such that in a very real sense taxonomies only become possible by simplifying experiential complexity in ways which make knowledge less useful.
Categorizing Natural Objects
269
Having urged caution, there are certain discontinuities that are so much part of the lives of so many human populations, that they can be said to be universal. We can show this to be true for natural kinds as a phenomenal type, and also for unique beginners, such as plant or animal. Since hominids have evolved in environments which display a particular phylogenetic and phenomenal discontinuity, it is not surprising that they should demonstrate a capacity to utilize a notion of natural kind which assists the management of diversity. Similarly, living matter, animacy and animality are not simply an end-product of classification based on multiple cognitive discriminations, but relate to a fundamental ability of the human mind to distinguish an organic form that registers a particular kind of saliency matching objective phylogenetic features. And there can be little doubt about the neurobiological organization which governs what Rosch (1977) calls ‘cognitive economy’: the propensity to store information in ways which make best use of the perceptual and cultural resources available and which provide templates with which to model fuzzy concepts.
6 Category formation and domain specificity Many aspects of rule-governed category formation and classification work in the same way irrespective of cognitive or semantic domain, but there are also significant differences that we must note which suggest domain specificity, some of which have major theoretical and methodological implications. This is so even if we do not accept hardcore modularity of the kind which typologizes generic cognition into, for example, intuitive physics, natural intelligence, and social cognition. If we consider just four different semantic domains – colour, kinship, biota and the body – we can see that each involve different premises. Colours are not really things at all, but rather properties of things, measurable along the dimensions of hue, saturation and brightness; while kinship classes are part of social deixis (those aspects of language which vary with the occasion, time and location of utterance, and with the identity of speaker and hearer), and refer to the properties of the relations between things. Bodies are clearly bounded entities, but the way we divide them up into parts – through partonymy – involves some degree of arbitrary grouping despite a large degree of crosscultural conformity. Of the four domains, only natural kinds map directly on to real, discrete objects in an objective world. But even with biota, some gaps between purportedly discrete kinds and objects are bigger and more salient than others, in most environments, and therefore serve as more widespread (even universal) markers in classifying behaviour.
270
Roy Ellen
When we classify as humans, not only do we use codes established in one domain to make sense of another in ways which distort aspects of experience, we systematically repress or forget or ignore certain characteristics and associations of particular natural things, and exaggerate and foreground others. Any one species, entity, idea or percept presents too complex an aggregation of traits to take into account in routine practical memory storage and information handling. Sometimes this simplification results in more naturalistic classifications, sometimes it results in more symbolic ones, or a combination of the two. This is very clear when we look at graphic icons for natural species in different aesthetic and writing traditions. Thus in Britain a child is likely to see a picture of a teddy-bear, or a teddy bear toy, before it sees a real bear.
7 Humans as animals One of intellectual legacies of dualism, as reflected in certain kinds of structuralism and the analysis of cosmologies, is the problem of where to place humans in classifications of the natural world. In developmental psychology this has translated as experimental work on the extent to which children in different cultures experience the similarities between people and animals in the narrow sense (Atran and Medin 2008: 130–132). On the whole, I think too much is made of this. All cultures simultaneously situate humans within both their ethnobiological classifications (emphasizing similarities) and emphasize the singularity of humans in contrast to other kinds of organism, most usually in theological and philosophical contexts. In the various diagrams of plant and animal classifications found in works of ethnobiology, humans are the necessary other and the main point of reference – absent from the picture but nevertheless informing its interpretation: a bit like the puzzle that Velázquez presents us with in his Las Meninas (fig. 1). We can only interpret these classifications because we know that they involve a series of steps in which other species are being compared with humans. In those ethnobiological classifications that I am aware of, anthropomorphism is a template by which to structure knowledge of the natural world: we make sense of other organisms by projecting on to them – as far as we can – human characteristics. In most folk classifications, while the default position is that humans are outside the frame, this evidence of constant comparison indicates that many peoples do understand that humans are a kind of animal – even where there is no word for the folk kingdom animal. Some of these issues are explored in my other contribution to this volume.
Categorizing Natural Objects
271
Fig. 1: Las Meninas, or The Family of Felipe IV, by Diego Rodríguez de Silva y Velázquez, Ca. 1656
In some local ecologies the presence of non-human primates of close phylogenetic proximity forces people to confront the intellectual issue more explicitly, as in cultures with elaborate beliefs about orang hutan or other large primates (e.g. Richards 1993). And where such evidence for transitional forms exists it is not unusual for local people to invent mythical forms that – as it were – close the gap. Consider, for example, the plethora of hybrid humanoid forms that have inhabited the Western imagination from the Nuremburg Chronicle (fig. 2) to modern popular science fiction. In Gregory Forth’s (2008) book on Images of the Wildman in Southeast Asia he is even able to plausibly suggest that the concept of Ebu gogo on the Indonesian island of Flores is a folk memory of the non-sapient hominid Homo floresiensis that we now know existed there until some 10,000 years ago.
272
Roy Ellen
Fig. 2: Image from the Nuremberg Chronicle, first published in 1493
8 Writing, depiction and social control In these remarks I have mainly been concerned with what we have discovered about the process of classifying in oral cultures, though I have already flagged-up the ways in which the interpretation of this process has been influenced by how we practice classification in the literary traditions of science. For as long as classifications were oral and shared they were constantly being reinforced by the cognitive limitations of the mind-brain and the body. In such situations storing knowledge as causal hypotheses (or models) was more efficient than making the right responses by induction alone, simply because humans relying on oral culture and low levels of division of labour do not have sufficient memory to make induction alone reliable. While oral language and graphic depiction seem to have common evolutionary roots (Mithen 1996), for most human populations over the last 10,000 years the reflection of classi-
Categorizing Natural Objects
273
fication in writing and art has been demographically of minor significance, becoming important mainly with the development of elite writing cultures, and latterly with innovations in the technology of writing that made it more accessible. But what happens when we put classifications in writing or attempt to physically depict them? In fact, we find that physical depiction has permitted degrees of complexity in the arrangement of categories that individual minds could not cope with. Specialization, the creation of visual images, and the written word allowed for the long-term storage and manipulation of classifications, in ways which were not limited by (even distributed) memory. The social distribution of knowledge and increasing specialization led to specific classifications having semi-autonomous histories of their own, and to the generation of emergent categories within a completely cultural framework unconstrained by ecological experience and cognitive limitations. Writing reduces the scope for variation and plurality, increases sharing and can therefore become a powerful force for social control. Writing also makes knowledge more portable and permanent, reinforcing the dislocation that arises when knowledge rooted in a particular place and set of experiences (i.e. local or indigenous), and generated by people living in those places, is transferred to other places. Thus, lion could be imagined as a category and transmitted between generations in Medieval Europe even where lions had never existed; and people could agree on categories even where there was apparent disagreement over content of those categories, such as with the basilisk in medieval and early modern writings. Such transformations in the technology of knowledge management have had dramatic practical consequences in certain spheres of human activity, for example at the point where orally-transmitted medicinal plant knowledge interfaces with the great scholarly traditions (Leonti 2011). Thus, illustrated Tibetan medical texts (fig. 3) provided a powerful means of aggregating and arranging knowledge and passing it on in a summarized form to a new generation of practitioners over a wide area, but at the same time uprooted it from local contexts, simplified the detail and posed challenges for those interpreting its meaning outside the geographic situation of its creation. Similar issues occur in committing classificatory knowledge to scripts more generally (Goody 1968; 1977; 1986; Goody and Watt 1968).
274
Roy Ellen
Fig. 3: One of a set of 77 medical paintings copied in the time of the Thirteenth Dalai Lama in the 1920s for the training of physicians in Buryiatia. The originals are late seventeenth century, and accompany a medical text compiled by the then Regent of Tibet, Desi Sangye Gyatso (sde srid sangs rgyas rgya mtsho). See Parfionovitch and Meyer 1992.
What is more, the written mode can create logical totalizing synthetic classifications that bring together the mundane and the symbolic in ways that in oral culture are flexible and fuzzy (Ellen 2006: 56–58). The Tibetan texts to which I have just been referring are an excellent example of this, but a good contender for the most problematic in the Western tradition is the much-discussed tree of Porphyry (fig. 4). Such attempts to compress classificatory relationships into a single specialized domain of technical practice (in this case philosophy) have given rise to many of the logical problems that we see today reflected in the modern study of classification. And so – in a sense – we come full circle: from the ad hoc but highly pragmatic and successful attempt of my father to solve a limited technical problem in the storage and retrieval
Categorizing Natural Objects
275
of fixing devices, through the regularities of shared cultural schemes for organizing domain-specific knowledge through the oral mode, to the consequences of a technology of sharing (writing) that while undoubtedly increasing the evolutionary fitness of those that have access to it, has led to all manner of profound problems – including paradoxically – shaping the way we understand the classificatory process in nonliterate contexts, including the position occupied by humans in classifications of the natural world.
Fig. 4: The Tree of Porphyry, translated from a version by Peter of Spain (1239), after Sowa 1999
References Atran, Scott. 1990. Cognitive Foundations of Natural History: Toward an Anthropology of Science. Cambridge: Cambridge University Press. Atran, Scott. 1998. “Folk Biology and the Anthropology of Science: Cognitive Universals and Cultural Particulars.” Behavioural and Brain Sciences 21: 547–609. Atran, Scott, and Douglas L. Medin. 2008. The Native Mind and the Cultural Construction of Nature. Cambridge, MA: The MIT Press. Barsalou, Lawrence W. 1992. “Frames, Concepts and Conceptual Fields.” In Frames, Fields, and Contrasts: New Essays in Semantic and Lexical Organization, edited by Adrienne Lehrer and Eva F. Kittay, 21–74. Hillsdale, NJ: Lawrence Erlbaum. Berlin, Brent. 1992. Ethnobiological Classification: Principles of Categorization of Plants and Animals in Traditional Societies. Princeton: Princeton University Press. Berlin, Brent, Dennis E. Breedlove, and Peter H. Raven. 1973. “General Principles of Classification and Nomenclature in Folk Biology.” American Anthropologist 75: 214–242.
276
Roy Ellen
Berlin, Brent, Dennis E. Breedlove, and Peter H. Raven. 1974. Principles of Tzeltal Plant Classification: An Introduction to the Botanical Ethnography of a Mayan-speaking Community in highland Chiapas. New York: Academic Press. Conklin, Harold C. 1962. “Lexicographical Treatment of Folk Taxonomies.” In Problems in Lexicography, edited by Fred W. Householder and Sol Saporta, 119–441. (International Journal of American Linguistics 28) Bloomington: Indiana University Research Center in Anthropology, Folklore and Linguistics. Ellen, Roy. 1993. The Cultural Relations of Classification: An Analysis of Nuaulu Animal Categories from Central Seram. Cambridge: Cambridge University Press. Ellen, Roy. 1998. “Palms and the Prototypicality of Trees: Some Questions Concerning Assumptions in the Comparative Study of Categories and Labels.” In The Social Life of Trees, edited by Laura Rival, 57–79. Oxford: Berg. Ellen, Roy. 2006. The Categorical Impulse: Essays in the Anthropology of Classifying Behaviour. New York/Oxford: Berghahn. Ellen, Roy. 2011. “‘Indigenous Knowledge’ and the Understanding of Cultural Cognition: The Contribution of Studies of Environmental Knowledge Systems.” In A Companion to Cognitive Anthropology, edited by David B. Kronenfeld, Giovanni Bennardo, Victor C. de Munck and Michael D. Fischer, 290–313. Oxford: Wiley-Blackwell. Ellen, Roy, and Angela Muthana. 2010. “Classifying ‘Eoliths’: How Cultural Cognition Featured in Arguments Surrounding Claims for the Earliest Human Artefacts as these Developed Between 1880 and 1900.” Journal of Cognition and Culture 10: 341–375. Forth, Gregory. 2008. Images of the Wildman in Southeast Asia: An Anthropological Perspective. London and New York: Routledge. Goldwasser, Orly, and Colette Grinevald. 2012. “What are ‘Determinatives’ Good For?” In Lexical Semantics in Ancient Egyptian, edited by Eitan Grossman, Stéphane Polis and Jean Winard, 17–53. (Lingua Aegyptia Studia Monographica 9). Hamburg: Widmaier Verlag. Goody, Jack. 1968. “Introduction.” In Literacy in Traditional Societies, edited by Jack Goody, 1–26. Cambridge: Cambridge University Press. Goody, Jack. 1977. The Domestication of the Savage Mind. Cambridge: Cambridge University Press. Goody, Jack. 1986. The Logic of Writing and the Organisation of Society. Cambridge: Cambridge University Press. Goody, Jack, and Ian Watt. 1968. “The Consequences of Literacy.” In Literacy in Traditional Societies, edited by Jack Goody, 27–68. Cambridge: Cambridge University Press. Grinevald, Colette. 2000. “A Morpho-Syntactic Typology of Classifiers.” In Systems of Nominal Classification, edited by Gunter Senft, 50–92. Cambridge: Cambridge University Press. Lakoff, George. 1973. “Hedges: A Study in Meaning Criteria and the Logic of Fuzzy Concepts.” Journal of Philosophical Logic 2: 458–508. Lancy, David F., and Andrew J. Strathern. 1981. “Making Twos: Pairing as an Alternative to the Taxonomic Mode of Representation.” American Anthropologist 83: 773–795. Leonti, Marco. 2011. “The Future is Written: Impact of Scripts on the Cognition, Selection, Knowledge and Transmission of Medical Plant Use and its Implications for Ethnobotany and Ethnopharmacology.” Journal of Ethnopharmacology 134: 542–555. Mithen, Steven. 1996. The Prehistory of the Mind: A Search for the Origins of Art, Religion and Science. London: Thames and Hudson. Parfionovitch, Yuri, Gyurme Dorje, and Fernand Meyer. 1992. Tibetan Medical Paintings: Illustrations to the Blue Beryl Treatise of Sangye Gyamtso (1653–1705), London: Serindia Publications. Richards, Paul. 1993. “Natural Symbols and Natural History: Chimpanzees, Elephants and Experiments in Mende Thought.” In Environmentalism: The View from Anthropology, edited by Kay Milton, 144–159. (ASA Monographs 32) London/New York: Routledge.
Categorizing Natural Objects
277
Romney, Kimball, Susan Weller, and William Batchelder. 1986. “Culture as Consensus: A Theory of Culture and Informant Accuracy.” American Anthropologist 88: 313–338. Rosch, Eleanor. 1977. “Principles of Categorization.” In Studies in Cross-Cultural Psychology, vol. 1, edited by Neil Warren, 1–49. London: Academic Press. Sowa, John F. 1999. Knowledge Representation: Logical, Philosophical, and Computational Foundations. Pacific Grove, CA: Brooks Cole. Tyler, Stephen A. 1969. Cognitive Anthropology. New York: Holt, Rinehart and Winston.
Figures Fig. 1: Original (Inv. No. P01174) in Museo Nacional de Prado, Madrid. 2015. Fig. 2: Hartmann Schedel (2001) Weltchronik: Kolorierte Gesamtausgabe von 1493: Einleitung und Kommentar von Stephan Füssel. Cologne: Taschen. Fig. 3: The present image (Item 40928) is from a private collection and reproduced courtesy of Himalayan Art and the Shelley and Donald Rubin Foundation. Fig. 4: After Sowa 1999.
Acknowledgments I would like to thank Walter Bisang and Tanja Pommerening for the original invitation to speak at the Mainz seminar, Keith O‘Sullivan of the Canterbury Cathedral Library in relation to fig. 2, and Cathy Cantwell for advice in sourcing fig. 3. I am grateful to the copyright holders of the images reproduced in figs. 1, 2 and 3, as indicated in the captions.
Dietrich Busse
Frames as a Model for the Analysis and Description of Concepts, Conceptual Structures, Conceptual Change and Concept Hierarchies Summary: Frame models are generally understood to be instruments for explaining, analyzing and describing concepts or terms and conceptual structures. They are also suitable, however, for describing conceptual change, or for describing changes in entire conceptual systems and concept hierarchies or systems of classification. The concept of frame inheritance as used by some frame researchers particularly emphasizes the classificatory aspect of conceptual knowledge. Although current frame research has yet to provide complete approaches to analysis and description of conceptual systems, the author of this chapter would like to show how an analysis of conceptual knowledge and concept hierarchies can gain in clarity and provide structuring options from the perspective of frame theory.
1 Introduction Human knowledge, especially abstract knowledge, is generally understood to be conceptual in nature and seen as structured by concepts. This is a truism as long as one accepts that the term concept or its derivation conceptual refer not only (or at least not primarily) to words, that is linguistic signs, but principally to the mental content or knowledge components behind them. Even though serious philosophical and linguistic doubts are often raised about the concept of concept, and have in particular been raised about many current theoretical and philosophical conceptions connected with this term, yet this term and the idea of an analysis of knowledge in the form of a conceptual analysis are still (again?) enjoying a certain popularity and currency. This is true especially in recent and very recent tendencies, in the cognitively based research on cognition, knowledge, and significance, concentrating on the terms term or concept as the core of the theoretical models. More recent cognitive science (which combines cognitive psychology approaches, philosophical and thought theory approaches, linguistic and language-theoretical approaches, and artificial intelligence models) makes prominent use of concepts and conceptual structures (or ontologies) in its central theoretical considerations. More advanced models from this area actually attempt to theoretically throw light on the precarious relation, always present to some extent when using the term concept, between individuality and society, i.e. the individual-psychological disposition and DOI 10.1515/9783110538779-011
282
Dietrich Busse
the supra-individual conventional, social character of the knowledge structures which constitute concepts, as well as between concrete situative realizations referring to individual cases (tokens) and the general knowledge structure transcending the individual case (type). One theory in particular, that of frames, offers a model with which the structure of conceptual (or, if preferred, concept-related) knowledge can not only be explained, but practically described, too. While in conventional research on concepts as practised in all disciplines (mostly, but not always, carried out empirically for, and within the area of the history of concepts) a more hermeneutic – often quite strongly intuitive – form of description of conceptual structures or significance dominates, we need to demand of a systematic conceptual analysis that it be able to serve both to explore and to describe conceptual knowledge with maximal reflection and methodological systematicity (as far as this is possible at all in the fields of knowledge analysis and semantics). This paper presents some thoughts on what such a systematic analysis might look like if frame semantics are applied to it. It is assumed that the approach suggested (or other, similar approaches) is valid as a universal theory and method for various conceptual domains (disciplines), and that it is suited especially (and more so than other, older models of semantic and conceptual analysis) to lay open and describe the structures of very complex abstract concepts and conceptual systems. After a short reflection on some basic theoretical and methodological issues and the discussion of some aspects concerning concepts, concept theory, and conceptual analysis, the conceptual model of frame analysis will be briefly sketched. Against this background, it will be possible to discuss why exactly this model was chosen for the descriptive approach to conceptual analysis. After that, specific aspects of conceptual relationships and systems will be dealt with, so as then to reflect on the possibilities and limits of frame-semantic conceptual analyzes, especially in view of the aims of a history of concepts.
2 The concept of concept The use of the terms concept and meaning is often quite unreflected in many publications dealing with linguistic problems, , both within and outside of the language sciences, and even in language philosophy. It is mainly based on everyday language, or at best on terminology within specialist disciplines, but it lacks linguistic reflection. The fact that the term concept is chosen as a reference point of semantic analysis shows in itself a certain understanding of the function and value of concepts, which ascribes a central role to these in the process of linguistically delimited acquisition (or constitution) of reality (or knowledge of the same). In everyday use, the expressions word and concept are not clearly distinguished. Definitions, such as those often met with in
Frames as a Model for the Analysis and Description of Concepts
283
dictionaries, which explain concept with ‘meaning of a word, ideal content’,1 do not by any means cover the everyday use of this expression. In many, if not most, non-specialist uses, concept is a synonym of word, although not an exact one. This everyday understanding thus reflects the fact that it is not as simple to separate the ideal content from the word(s) carrying it as it might seem. Yet word and concept are not used in an identical way in everyday speech. The main meaning of word clearly has to do with specifically linguistic characteristics, the language sign (Saussure’s signifiant, the external side of the linguistic sign), whereas the main meaning of concept places the aspect of the sign, the external side of the linguistic sign, more in the background, emphasizing the meaning (Saussure’s signifié, the internal side of the linguistic sign, the mental momentum of ‘comprehending’). But in the use of concept in everyday language, the intellectual side of the sign is not completely separated from the specifically linguistic side, the character of the word possessed by the linguistic sign, in contrast to such separation in scientific texts. Daily language, as a rudimentary reservoir of centuries-long theoreticization processes, thus reflects the variance in the definition of concept that has been an integral part of thinking about language since Plato. Two determinants still dictate today what is generally understood by concept: on the one hand, concept as an abstractive synopsis of object characteristics, on the other, concept as a higher-ranking word (in relation to its abstractive function). The aspect of linguistic signs being used in communicative acts of expression, and that this use can be not only meaningful, but also forms the power of language to constitute reality, has, by contrast, been overlooked in the definitions of concept and word and the difference between these notions. It has been overlooked by a perception and conception of language which wanted to see thinking (and thus the mental side of concepts) as independent of language, which was wrongly seen as a purely mediating instrument. The problem for every analysis of concept is how to deal with the subsumption (in a traditional understanding of language) of the concept, perceived as purely mental and therefore individual, in the form of a linguistic sign by the medium of language, which transcends the individual. The perception of concepts as purely cognitive entities leads to any conceptual idea being private, if its constitution (or, put in a more modern way, its cognitive representation) cannot be shown to be a genuinely linguistic process. The thought of a completely private nature of conceptual ideas is a wrong assumption when the analysis of concepts is to be used to represent the analysis of trans-individual (and only in this sense objective) knowledge. Concepts must be intersubjective, objectifiable, by means of language as a process of communication and understanding. According to Wittgenstein, we acquire concepts with the meanings of the words in their usage; they
1 This remark refers to the German term Begriff as defined in a popular dictionary (Wahrig 1979: col. 614 f.). The English term concept seems not to reflect the double meaning of word (of a certain kind) and ‘meaning/content of a word’ as the German Begriff. (For more details cf. Busse 2011, and Haller and Mittelstrass 1971).
284
Dietrich Busse
are not delimitable or finite (Wittgenstein 1971: §§ 67–71). A single word has a meaning for us only when it is used in a concrete context, and only against the background of the systems of significance constituted by a language, called ‘language games’ by Wittgenstein. ‘Wenn sich die Sprachspiele ändern, ändern sich die Begriffe, und mit den Begriffen die Bedeutungen der Wörter (Wittgenstein 1970: § 65).’ (‘When the language games change, the concepts change, and with the concepts the meanings of words.’) And with these, the things we have referred to with the words change for us, too; our image of reality changes.2 What is commonly referred to as conceptuality appears to be a conglomerate of attributions of characteristics, abstractions, cross-references and situational references, all of which are the result of a multitude of communicative acts in which individual linguistic signs can be the trigger or the point of reference, of aspects of meaning (themselves constituted by the context of application), but which can never represent the object (or the concept) as a unit in its totality. Concept words thus induce the synopsis of communicative experiences, which can never be exactly the same for all speaking individuals of a language community. What could be designated a concept is not constituted by individual uses of signs, but by the totality (or by a multitude) of sign uses in a continuum of acts of utterance, texts, and discourses. If one wants to reconstruct concepts analytically, one must look for them at their place of constitution, the acts of sign use within the framework of communicative acts (regardless of whether these occur orally or in the form of texts and text constituents). Since the cognitive turn at the latest, models that regard concepts as purely mental phenomena are en vogue in linguistics and linguistic philosophy. Cognitive scientists, in particular, mostly talk of a level of concepts, which is often wrongly seen as being separable from the level of language. According to the interpretation accepted here, the question of a possibly language-independent or pre-linguistic level of concepts does not arise, the problem of describing the relation of cognition and language with regard to the language and the cognition of the people using language does not pose itself.3
2 Wittgenstein often uses concept and word nearly synonymously, thus making clear that whatever embodies a concept is, for him, explicable and comprehensible only with reference to lexical use (and that means to lexical meaning). The lexical use serves as an example, a paradigm for what is to be grasped as thing, and can have meaning in turn only in the context of the language game belonging to it. 3 From what I can see, cognitive evidence in animals makes it not unlikely that we should assume the possibility and existence of a pre-linguistic level of cognitive category-formation. Whether one should call this a conceptual level is extremely questionable. But one should never forget that there are areas or levels of cognition in the language-cognizant human being which cannot be categorized, and thus conceptualized. Their content, if one can and should use this concept here at all, are thus not linguistically accessible and not directly communicable. See the pertinent discussion by Wittgenstein on expressions of pain.
Frames as a Model for the Analysis and Description of Concepts
285
As far as people have a language and use it, as Wilhelm von Humboldt said so unforgettably, it may be assumed that the word, ... adds a good part of itself to the concept. (The word which first makes the concept an individual in the mental world also adds a considerable part of itself, and, as the idea receives by means of the same determination, it is simultaneously held within certain limits.)4
If one takes concept here to be chiffre for thought, the episteme, the knowledge structures, then the relation between linguistic elements and elements of knowledge (cognition) is as follows, according to Humboldt: Even when the thesis of a complete identity between language (e.g. the semantics inherent to language in their sum) and knowledge (or thought) is not being supported here, yet the fact that epistemic content can only (or mostly) be expressed and thus communicated in linguistic form does considerably influence the structure and content of knowledge itself. (‘The word [...] also adds a considerable part of itself.’) Knowledge elements can only be identified in as much as we have the linguistic means to isolate them and to evoke them. (‘The word [...] makes [...] the concept an individual in the mental world [...]’) Without words (linguistic means of expression) there are no identifiable thoughts (epistemic elements); only through them does what is thought first receive an identity, the ability to be (re)called and repeated; but this also means that only by means of words does thought become changeable and able to acquire a history. (‘The idea attains definiteness through the word.’) At the same time, the linguistic means give structure to the epistemic content and limit it, tying it into the corset of linguistically constituted structures, as it were. (‘The idea is held captive by the word in certain limits.’) By way of a preliminary conclusion, then, the unit concept cannot be considered in complete dissolution from units of the type word (or more exactly: ‘linguistic sign’). Quite apart from what one may think of their relation as a theoretical consideration, it is irrefutably the case that a practical, analytical approach to concepts can only be made through words, texts and text corpora. Words (texts, linguistic-communicative acts) are thus the decisive key to access to the concepts, to the content behind the words, to the knowledge communicated or alluded to in language. The relation between word and concept becomes accessible primarily through the momentum of knowledge, the episteme. Cognitive processes in the language-using human being operate to a greater (and in our context more interesting) extent on and with knowledge which has been constituted and structured in the use of language. The connecting factor is the schematization of knowledge and the architecture resulting from this.5 The concept of frame has been suggested for these formations of schemata. In
4 ‘Ueber das vergleichende Sprachstudium’ (1820), cited after Humboldt 1985: 20. 5 On the aspect of the architecture of knowledge, cf. the considerations in Busse 2005.
286
Dietrich Busse
linguistics this is known as the concept of the semantic frame in the sense of the ‘interpretive semantics’ of Charles Fillmore.6 In the field of general cognitive science, it was Marvin Minsky (1974) who exemplified the frame character of the knowledge required for certain optical perceptions.7 I personally prefer the term knowledge frame, so as to emphasize the general, basic nature of this concept. The formation processes of schemata (or the formation of knowledge frames) are linguistic insofar as only (or, if you want to be more cautious, especially) the active use of the schemata (frames) in acts of linguistic communication stabilizes (gives them continuity) these schemata, enriches them with knowledge, and renders them changeable. ‘The word adds a part of itself to the concept’ (Humboldt), in that the epistemic schemata which form concepts, but also those that go beyond this, are only laden with epistemic material (elements of knowledge) through their use in the context of linguistic utterances/texts. (With an eye to the theory of signification or meaning, it makes some sense in this context to recall Husserl’s distinction – referring to the interpretation of signs and symbols – between ‘sense-giving [mental] acts’ or ‘meaning-conferring [mental] acts’ and ‘meaning-fulfilling [mental] acts’ (Husserl 1913: 38). One may assume that this distinction can also be established in the case of schemata/frames. One could then speak of fully specified frames and distinguish these from frames that are not epistemically/with regard to content fully specified, which however would not permit complete comprehension.) As both the linguistic signs and the frames/schemata forming the basis of their comprehensibility can fulfill their (complete) epistemic function only within a context, one can speak of a process of contextualization with regard to a reasonable comprehension of linguistic signs (sentences, texts).8 Language is, so to speak, the medium in which not only articulation and communication of social knowledge takes place, but in which this knowledge is, at the same time, constituted and structured as such (i.e. as social knowledge). Language (and language events such as texts) is thus by no means the archive of this knowledge. If an archival metaphor is to be applied in any way at all to language, then one could perhaps best characterize language as the index or register of the archive of social knowledge. This index or register contains only references; these references are something that every speaker of the language must first of all epistemically realize and substantiate for him/herself in the course of the process of understanding (more precisely: in the course of the processes of concluding and inference that lead to com-
6 For an overview, cf. Fillmore 1977 and 1982; he speaks there, too, of the ‘semantics of understanding’. 7 Here with a clear reference to the schema theory of Bartlett 1932; on the central role of Bartlett’s schema theory in frame theory, see Busse 2012: 311–331. 8 The concept of contextualization used here is an epistemic contextualization, thus it means some thing like ‘situated in a structured (in respect of content, and thus not accidental, but always contingent) arrangement of frames/schemata’. Cf. on this in more detail Busse 2007.
Frames as a Model for the Analysis and Description of Concepts
287
prehension). One can call this in Husserl’s sense the ‘fulfillment of sense’. Language as the register of knowledge fulfills its task in that the individual signs and their specific combinations evoke (Fillmore)9 knowledge (frames, schemata and frame or schema complexes) in each case. Frame theory, with its foundation in and on cognitive and linguistic-semantic science, is an effective means of making accessible, and describing, the epistemic content of concepts more precisely than was possible with the old models of semantic or conceptual feature-lists or lists of semantic markers (so-called ‘checklist semantics’, as Fillmore called it in his severe criticism of 1975). How this is possible, and how it is done in practice, will be explained in more detail in the following.
3 The conceptual model of frame theory (what is frame semantics?) Generally, frame theories (theories of knowledge frames) understand frames (or knowledge frames) as ‘structures consisting of concepts or ideas’.10 Frame semantics, or – more generally – frame theory as it is presented today (particularly in linguistics) is, nevertheless, not a monolithic block, not a closed model, but rather a heterogeneous set of concepts with the most varying ancestries in different branches of science, each of which shows, to some extent at least, very different goals of knowledge, objects of research, and basic assumptions. The frame semantics of the linguist Charles J. Fillmore (and of the research association founded by him, FrameNet, with its centre in Berkeley) – the only genuinely linguistic frame concept – is rooted in thoughts and theorems, some of which are quite different from those of the frame models in the cognitive sciences, such as the models of Marvin Minsky (1974 and 1986), Schank & Abelson (1977) and Lawrence Barsalou (1992). While Fillmore’s linguistic frame model had and still has roots and theoretical points of reference to a much greater extent in valency grammar and the syntactic theory of ‘case frames’ derived from this in Fillmore (1968), the cognitive frame concepts of Minsky and Barsalou or the closely related script model of Schank & Abelson, have borrowed heavily from the schema concept of the memory psychologist Frederick Bartlett (1932). While Fillmore focuses on ‘frame evoking’ types of words, primarily verbs in their function as semantic and syntactic structural centres of the sentence frame (and evaluates the other types of words, such as nouns, adjectives, adverbs etc.
9 For reasons of space, I will not discuss the difference between evoke and invoke introduced by Fillmore 1982. However, cf. on this Ziem 2008: 231–237 for a wider-ranging, affirmative discussion, and (critically) Busse 2012: 122–131, 203–209, and 644–670. 10 Thus i.a. Fillmore 1992: 40 and 2006: 613, as well as Barsalou 1992: 31.
288
Dietrich Busse
primarily with regard to their function in a semantic or knowledge structure dominated by the verb), Barsalou’s (1992) frame idea – as Minsky’s (1974) – aims primarily at nouns (nominal concepts/ideas). If a frame (knowledge frame) is defined as a structure consisting of concepts, then this means something else for Fillmore than it does for Barsalou or Minsky. A frame according to Fillmore would be a structure consisting in concepts/ideas forming, for example, the basis for the semantics of a sentence (meaning concepts for the verb content, for the subject noun, for the object noun[s], and so forth). In contrast, a frame according to, say, Barsalou would be an epistemically or cognitively viewed (nominal) concept, complex and structured of and for itself, which in turn consists of (sub- or part-) concepts; every frame is itself, in turn, a structure consisting of frames, or, in Barsalou’s terminology, every concept (every idea) must be comprehended as a structure of concepts (ideas). (The idea of the recursivity of all frame structures or conceptual structures or knowledge structures which underlies this thought was, incidentally, borrowed from linguistic syntactic theory by the cognitive scientist Barsalou.)11 What Fillmore’s sentence- or verb-oriented conception has in common with the general cognitive frame model founded by Minsky (1974) is his talking of slots (or, terminals in Minsky 1974), here understood as blank positions for arguments and their fillers. It is this notion of slots that makes up the charm, the distinctiveness, and the essential core of frame theories and that has rendered them so attractive to so many scientists in a multitude of disciplines. Valency theory in linguistics, which deals more with sentence structure, borrowed this basic idea metaphorically from chemistry, or, more exactly, from the conceptual difference between the bonding ability of atoms and the concrete bonds met with in given molecular structures. By the circuitous road of linguistic valency grammar, primarily focused on sentences and the binding ability of central sentence predicates in the form of verbs, and its semantic extension to case frame theory in Fillmore (1968), this model was then transferred to the content structures of concepts. Here a few points should be mentioned which act as the centre of the frame idea being discussed and further developed in many disciplines (such as linguistics, cognitive science, psychology).12 A frame or knowledge frame is a structure of knowledge in which a certain constellation of knowledge elements is grouped, with reference to a structural frame
11 Recursivity in the syntactic sense means the embedding of a sub-structure with a certain construction in the (super-)structure of the same type. Thus, for example, a nominal group such as the friend’s father’s brother’s house itself contains an attributive nominal group the friend’s father’s brother’s, which in turn contains an attributive nominal group the friend’s father’s. According to Barsalou 1992, such recursions are in principle endless when the principle is described within frames, i.e. endlessly divisible or refinable. 12 The following description is based on the thesis-like summary of the core ideas of frame theory/-ies in Busse 2012: 819–827.
Frames as a Model for the Analysis and Description of Concepts
289
core which can be understood as an object or theme of the frame; the constellation of knowledge elements functions in this perspective as a frame-constituting element. These knowledge or frame elements are not epistemic quantities filled out with concrete data; rather, they act as slots, to which, in an epistemic contextualization (embedding, filling), concrete (filling out) knowledge elements (so-called fillers, values, or attributions) are allotted.13 Thus, frames represent knowledge structures (to put it in somewhat simple terms) that attach to a category certain attributes, which in turn can be filled out with certain concrete values. (In other frame theories, the attributes are called ‘blank positions’ or slots and the values are called fillers.) The type and number of attributes of a frame is not laid down once for all, but varies. Thus, new attributes may be added. Frames are then generally understood to be structures made of concepts (concepts understood here as purely epistemic quantities or entities), which, since all concepts in turn are structured in the form of frames, reveal themselves to be structures made of frames. As far as frames essentially specify (epistemic) possibilities and constraints of expansion (of further detailed frame elements), their structure can be described as an arrangement consisting of epistemic relations (to the attached elements and among them). Since frames, in this view, are basic structures (elements) of cognition/knowledge, and thus are to be assumed on all levels of their description, it follows compellingly that different levels and types of frames (and frame analysis) must be assumed. Within the framework of an application of frame theory to semantic or conceptual analysis, the following dichotomies approximately designate level differences that must be taken into consideration in frame theory and frame analysis: individual versus social; shortterm memory versus long-term memory; token versus type; actual (meaning) versus occasional (meaning); concrete versus general; exemplar versus category. In the description of frames (as conceptual structures), then, the description of the slots or attributes or terminals and their relation among each other and to the frame core, has a central function. This can be defined as follows: the slots (terminals, frame elements, attributes) of a frame are the knowledge elements that are connected to a firm set of such elements in a particular frame, and that constitute this frame, and that define the object of reference (the theme) of the frame. These knowledge elements are not fully specified in their epistemic content; rather, they simply establish the conditions that must be fulfilled by concrete and specifying knowledge elements
13 With such structures, linguists immediately think of the valency framework in dependency grammar according to Lucien Tesnière 1959, which are discussed in modern research under the concept of ‘argument structures’, but also of the concept of subcategorization from the linguistics of the 1970s. A valency framework is opened by a verb. Thus, the verb give (as in donate) opens a three-place valency framework (one then says that the valency of give is three-valued) which provides for places for an agent of the verbal action (subject), the object being given (direct object), and the recipient of the gift (indirect object).
290
Dietrich Busse
when they, as constitutive characteristics or components of a frame, are to render said frame an epistemically completely specified (instantiated) arrangement of knowledge. As the slots establish concretising conditions for the epistemic characteristics of the fillers, they can also be characterized as a set of conditions of attributes (or set of conditions for attributability, set of subcategorization conditions).14 In so doing, the following must be taken into consideration: the characteristic of being a slot (a terminal, an attribute) is not attached in an absolute sense to a knowledge element, but only in relation to a higher-ranking frame. In isolation, such knowledge elements form their own frames, with their own slots/attributes in turn subordinated. The slots or attributes that are important for an epistemological analysis are those ascriptions of concepts (functioning in this relation as aspects) for which there exists, in the linguistic/cultural community in which this attribution occurs, an established convention of attribution. Slots determine the relations (and thus also types of relations) which subsist between the frame core and the specified knowledge elements (fillers, fillings, values) which are attached by means of these relations. But the knowledge elements themselves can be characterized as relations between the set of attributive conditions defining them and the frame of reference. In other words, between the slot/attribute and the frame core that is thus specified there exists a relation of allocation, a correlation. A working definition of the fillers or values could be as follows: ascriptions/ fillers/values are those knowledge elements which are attached via slots or terminals to a (general, abstract) frame, so as to make this an epistemically fully specified frame of knowledge (an instantiated frame, an instantiated concept). Ascriptions or fillers or values important for an epistemic analysis are those attributions of concepts (that function in this relation as fillers) to other concepts (that function in this relation as slots) which form expected or possible concretisations/instantiations of the general type conditions of the slot, according to the conditions that define the slot (terminal, attribute) of this frame.15 Here, too, it is important to note that the characteristic of being an attribute (a filler, a value) in such a conceptual structure does not attach absolutely to a knowledge element, but only in relation to a superordinated slot (attribute). In isolation, such knowledge elements form their own frames, with their own slots/attributes, in turn subordinated, and ascriptions/fillers/values. In token frames all ascriptions/
14 In linguistics, the concept of subcategorization designates e.g. that a verb such as bark (of dogs) not only requires an agent as subject (that results already from the valency requirement lexicalized with this verb), but that this agent must belong to a category that is more closely specified by a characteristic such as doggish, canine. 15 (Important terminological notice: Barsalou 1992 uses attribute for slot and value for filler; so the term attribution in this paper has to be understood in the sense of process, i.e. ascription, and not to be confounded with Barsalou’s term attribute.)
Frames as a Model for the Analysis and Description of Concepts
291
fillers/values must be specified (as far as the conditions of filling determined by the slots or attributes require this). As long as slots are not taken up (depending on the situation and context) by concrete, specific ascriptions/fillers/values, they are occupied by standard fillings (default values) added from conventionalized (prototypical) knowledge. Instantiated slots (slots in a substantiated and instantiated frame) can, as a rule, only be taken up by a single ascription/filler (a single value). To illustrate this, and to make things clearer, the figures 1–3 below show a schematic representation of two (albeit relatively simple) concept frames according to Barsalou (1992) and then the representation of a predicative frame according to Fillmore et al. (2002) and FrameNet:
Fig. 1: Attribute frame for companion according to Barsalou 1992: 33; 62
Fig. 2: Representation of tokens/exemplars for bird according to Barsalou 1992: 45
292
Dietrich Busse
Fig. 3: Criminal process frame from Fillmore et al. 2002: 5
A linguistic (semantic) frame analysis, but also a conceptual one, thus registers, with the assumption of frames, structures in knowledge (relevant to comprehension). It is generally agreed by nearly all researchers active in frame research that there is no way to distinguish strictly between linguistic knowledge and so-called encyclopaedic knowledge (or knowledge of the world). An important interaction between the linguistic and the generally epistemic levels is to be seen in the fact that linguistic signs focus knowledge of the world in a specific manner. (See, for instance, perspective according to Fillmore 1977, which he sees realized through the verbs buy, sell, pay, cost in the example of the Commercial Event frame.) But it should be noted here that this interaction between linguistic and generally epistemic levels is strongly influenced by recursivity, undelimitability, flexibility and vagueness (Barsalou 1993 has emphasized this particularly). Using the frame model for the purposes of a conceptual analysis (as is the central focus of this examination), frame elements show up as conceptual elements (attachment positions, slots, attributes of a category). Barsalou’s frame elements, which he calls attributes, are, with reference to the lexeme class of nomina, typically classes of characteristics that can be specified in the reference objects of a category (size, colour, material, etc.). Within the group of characteristical frame elements, one can and should distinguish between so-called structural frame elements and functional frame elements/attributes. Structural frame elements refer typically to attributes such as colour, form, mass in the case of physical entities (things, living things, people); place, time, goal etc. in actions, events, etc. Functional frame elements/attributes have most recently also been subsumed under the name affordances. Affordances are typically assumed in the case of objects and things (usually artefacts). A possible
Frames as a Model for the Analysis and Description of Concepts
293
working definition of affordance would perhaps be: functional characteristics, in relation to people, uses, purposes, of things (e.g. nail, hammer, screwdriver, etc.). Frames can also be regarded as subdivided, organized, in different structural levels. An important type of the internal structure of frames is centred on the pair of concepts type-token. A distinction referring to these might be the distinction between abstract pattern frames and concrete exemplar frames. Strictly speaking, this is not a distinction within a single frame, but a distinction referring to different types or levels of frames. The relationship between the two levels is not only a difference between a structure consisting of empty slots (or slots filled simply with standard values) and a structure consisting of filled (with concrete values) slots. Rather, exemplar frames can augment a pattern frame by adding slots, when these occur more frequently (via a larger number of exemplars, or in especially salient exemplars). Frames (on the level of general social knowledge structures, i.e. patterns or types) are not simple and closed structures. On the contrary, one must reckon with considerable social variance in the degree of ‘granulation’ and differentiation of the frames. Owing to the general principle of recursivity, frames are, in principle, knowledge structures that can be infinitely refined. This shows in that, in social domains with different needs of knowledge, the differentiation of frames varies, too (typically known as the so-called expert/layman divergence). The key words frame systems or frame nets have been used in the literature hitherto mainly for discussions of conceptual taxonomies (so-called ontologies). Beyond the interconnectedness that results in such taxonomic orderings of knowledge, the associative relations between frames and frame elements, based frequently on analogous formation, perception of contiguity, and metaphorical transfer, should not be underestimated with regard to their constitutive and structuring effect on knowledge. Thus far the general outlines of frame theories as the basis for a semantic frame description, or one aiming at conceptual analysis. It has become clear that some representatives (in cognitive disciplines) of frame theory, at least, conceive of frames primarily as conceptual structures (in a double sense: internal structures of concepts and structures consisting of several concepts, conceptual systems). In this way, frame theory has become an important contribution to a theory of concept. Frames or concepts are seen here as knowledge structures, i.e. as complexes consisting of knowledge elements which are arranged internally in a regulated, structured relationship to each other. The advantage of frame theories as opposed to previous concept theories is to be seen primarily in the ability to allow a structural description of the internal semantic, conceptual or epistemic structure of concepts, based on uniform criteria and a unified model of structure.
294
Dietrich Busse
4 Frames, conceptual orders, hierarchies and frame systems In work on frame theory, too, it is often overlooked that an important impulse in developing the frame conception was formed by the hope of being able to better and more systematically explain and describe relations between concepts, up to and including complex conceptual systems (concept hierarchies and so-called ontologies), with this model. It is not well known that the founder of the cognitive variant of frame theory, Marvin Minsky, in his later monograph for the basis of a theory of cognition (Society of Mind, 1986), complained clearly about, and was surprised at the total lack of attention to what he held to be the core and the actual goal of his 1974 text (especially to what was really new and went beyond previous schema theories, for example, that of Bartlett). This core aspect, which he also used as the title of a partial printing of his paper at that time, was Frame system theory. This aspect still plays hardly any role in research even today, although pursuing it further would be of great use in exploiting the frame model as the instrument of a general analysis of knowledge and epistemology (one could say ‘the architecture of knowledge’, cf. Busse 2005). Yet the explanation and description of the relations between frames and the structures within frame systems must remain an important point in any frame analysis, also – and especially – in linguistic semantics, as systematic aspects implicitly result from it again and again, whether in connection with ontologies and taxonomies or in connection with phenomena such as frame inheritance, frame proximity or frame analogy.16
4.1 Frames as structures consisting of frames and relations While Fillmore first conceives of frames in a rather real-world encyclopaedic way, as representations of holistic scenes, Minsky builds the idea into his model from the start that frames themselves are in turn structures made of frames. This becomes clear when he defines the so-called slots17 as terminals (connective positions), and adds that we use these to add frames to frames. In Barsalou, too, it becomes clear that frames, being constellations of concepts and concepts themselves in turn represent frames, are always structures consisting of frames. As is well-known, Barsalou narrows the whole thing down to this: Frames are basically recursive (all components of frames are themselves frames). Since attributes (slots) of a frame, according to Barsalou, are frequently highly complex embedded frames, one could also define a frame as a kind
16 The following description is based strongly on chapter 7.6 in Busse 2012. 17 The frequently used German term for the English term slot, i.e. Leerstelle (‘empty or blank position’) is far less adequate than Minsky’s 1974 original term terminal, that could be translated to German as Anschlusstelle (connective position).
Frames as a Model for the Analysis and Description of Concepts
295
of mini-system consisting of frames. But not only slots or attributes are organized like frames; naturally values, including the standard values themselves, invoke frames, as Ziem (2008) emphasizes. That is, every frame is, per se, an integral component of an extensive conceptual network (which, according to Ziem, is organized hierarchically by means of superordinate relations). It would be wrong to understand frames as mere formations of elements; they are this, too, but that is not all. By means of the concept of slot, in particular, the aspect of relationality is highlighted. In other words, frames are structures consisting of elements and the relations obtaining between them. More precisely, a good part of what goes to make up frames refers more to relations, types of relations, general conditions for relational attachment and so forth, than to information in the sense of static elements. Bartlett had already underlined this thought when he pointed out that any knowledge is always knowledge of relations (interconnectedness). Relationality is already inherent in the concept of schema (as a predecessor of the frame idea) in the form of a basic characteristic. Barsalou, too, understands frames as ‘dynamic relational structures’ whose form is flexible and dependent on context. An important aspect in frame analysis is formed not only by the frame-constituting relations between the frame core and the sub-frames or value frames (or default value frames) attached by attributes (slots). Equally important are the relations, as Minsky was the first to emphasize, that exist between the elements (concepts, frames) attached to a frame. These relations, too, are not random, but specified, or at least they can be. While Minsky and Barsalou focus on relations within frames and between frame elements (core, attributes, values) in the same frame (one could call these intra-frame relations, except that there is a risk of absolute misunderstanding, unless one constantly keeps in mind the basic recursivity and non-rigidity of frames), Fillmore is more concerned with the relations between frames (one could then call these interframe relations), as they exist in hierarchical, taxonomic circumstances (as superand sub-frames) – he calls this frame inheritance. Besides this, Fillmore also mentions frame blending and complex frames as types of inter-frame relations. Frame inheritance and taxonomic levels, epistemic frame connectedness and frame systems, types of relation between frames and frame elements and types of frame systems are thus topics within the context of complex frame structures which need to be further investigated. At this point, however, I can only discuss them very briefly.
4.2 Frame inheritance and taxonomic levels In the development of the frame model by Fillmore and FrameNet, the idea of frame inheritance plays an increasingly important role. This fact is mainly due to experiences from trying to describe the frames for lexematic meanings in linguistic semantics, since this soon raised the question of whether certain slots or attributes that can be determined for a frame-constituting concept should be descriptively registered on
296
Dietrich Busse
a more general or a more concrete level. Let us take, by way of example, the frame for a lexeme such as to bark. For the description of the slot agent, apparently, one has something of a dilemma as to whether this is to be classified as dog (or having one or more characteristics of a dog), or as animal (with the specification sub-category dog), or as living thing (with the specification sub-category animal, and then the sub-specification sub-sub-category dog), and so forth. In other words, the same problems occur as the ones that always occur in conceptual hierarchies or taxonomies. In frame research, Konerding (1993) in particular has drawn the most radical conclusion from this circumstance, in that he has developed a graded scale of categories in conceptual hierarchies for the description of frames (he calls this approach ‘hyperonym type reduction’), which, in the end, leads to extremely abstract so-called matrix frames, of which the concrete frames are always the sets of derivations or concretisations, attained by reducing the number of slots. The closeness to conceptual systems (a speciality of science in the Enlightenment of the 18th century, thus at the apex of encyclopaedism) is quite remarkable. Fillmore started by introducing and explaining the idea of frame inheritance with the example of frames for verbs. Frame inheritance is the elaboration of a more general (and more abstract) parent frame by one or more child frames. In so doing, the child frame inherits all the frame elements and characteristics of the parent frame, but can add its own additional elements and characteristics to these. Examples named include: a general movement frame and travel as its realization. The parent frame and child frame(s) thus behave towards each other like superordinate concepts and subordinate concepts in conceptual hierarchies (taxonomies) and ontologies (in whose image these frames are clearly conceived of, although this similarity is not more closely examined). In contrast to such conceptual and hierarchical frame-to-frame relations, designated frame inheritance, the frame relations designated by Fillmore and FrameNet sub-frame relations (or frame compositions) represent a kind of part-whole relation. A typical example of this are partial actions (or partial course of events) as more comprehensive complex chains of action (or courses of events). Thus, to come back to the example of travel, partial or sub-frames such as depart and arrive show this. Sub-frames share only some frame elements with the superordinated frame (thus, the performer frame element in the superordinated frame travel and in the subframes depart and arrive is identical). It then depends on the specific character of the sub-frame (or the type of embedding of the sub-frame in a superordinate frame) as to what kind of frame elements, and how many, agree in each case. For example, in a complete picnic excursion frame, the performer frame elements are identical in the superordinate frame travel and in the sub-frames depart, arrive, consume food, but need not be so in the sub-frame preparation of food. Larger courses of events or action complexes are, in the FrameNet project, divided into partial actions or events, each of which is then described as a frame (a sub-frame). The analysis does not just touch upon what is recognisable in everyday
Frames as a Model for the Analysis and Description of Concepts
297
life as a partial action or partial aspect (as is the case in a travel frame with the partial aspects depart, drive [vel sim.] and arrive), but upon very subtle part elements, which can only be tapped intellectually by means of an epistemologically/cognitively oriented breaking down. The example arrive in Jack entered the room can be described as a transition between two conditions (A = Jack is not in the room, B = Jack is in the room), each of which shall represent sub-frames of the superordinate frame. The application of the frame relation sub-frame in FrameNet extends from such very subtle partial aspects of events to such large-scale frame compositions as the criminal trial with all its partial events, partial actions, courses of events, etc. In contrast to inheritance relations, sub-frame relations reflect the relations between constitutive elements of a whole and the whole itself. A sub-frame therefore represents a constitutional relation, while inheritance is an abstractive (or specifying) relation. Each is thus on a different theoretical level. They are, however, connected, insofar as the identification of sub-frames may require steps of abstraction which do not always reveal themselves in everyday consciousness from the start. Thus far, at least, the registration of this type of frame relation does demand a decidedly abstractive look at frames of the semantically relevant knowledge. The most important aspect of frame relations according to Fillmore concerns the recurrence of frame elements (in this case: actant frame elements in predicative frames) in several frames, either as identical frame elements or as relations between more general and more substantive versions of frame elements. One could, with the same justification, speak of frame-element relations instead of frame relations. The aim of describing frame relations is primarily to enable semantic generalizations across more than one frame. Referring to the definition of frames, typical for FrameNet, as ‘systems of concepts’, the frame relations are then designated ‘semantic relations between collections of concepts’. With this, the connection between Fillmore’s frame inheritance and conceptual taxonomies becomes clear. But with this, the concept of frame inheritance also inherits all the problems entailed by ‘conceptual hierarchies’. Here, too, we are dealing with a relation between levels of higher abstraction and levels of lower abstraction. Taxonomic inheritance relations (in the sense of superordinate/subordinate concept) are what Minsky, too, has in mind, when he emphasizes that every frame is embedded in superordinated (higher-ranking) frames. Barsalou concisely sums up this aspect with his recursivity axiom. He particularly underlines the special power of the frame model in analyzing taxonomies, conceptual combinations, and conceptual fields. The aspect of attributes and values (slots and fillers) being connected to each other in frames through inheritance relations (i.e. superordinate and subordinate concept relations) leads, then, by the principle of recursivity, to taxonomically graduated inheritance chains. In this sense, Barsalou speaks of recursively graduated attribute taxonomies. He ascribes an important heuristic value to these for the construction of knowledge systems by the single individual. According to him, frames and frame systems help individuals to develop and build taxonomies from existing
298
Dietrich Busse
epistemic material. He then speaks of conceptual fields. Frames and frame inheritance relations can then structure complex conceptual fields in their entirety. Barsalou pleads the case for the strong thesis that ‘Every frame defines an implicit conceptual field’. These conceptual fields are understood in a purely cognitive way and are not to be equated with the lexical fields of older schools. On the basis of the recursivity of frames, conceptual fields, according to Barsalou, can grow in exponential fashion, but most positions of a conceptual field are not lexicalized (only a minority is). The taxonomic conception of (or perspective on) frame inheritance and frame systems leads to frame systems sharing the problems of all taxonomic hierarchies, for instance, when one understands them as systems consisting of inheritance relations. These problems are particularly pertinent in the practical description of meaning in linguistic semantics. The deeper one enters into a frame hierarchy, the more subslots can be established. (This corresponds to the increase in semantic characteristics on the lower levels of a conceptual hierarchy). In contrast, there is the problem that top-level categories in taxonomies are very abstract and can often no longer be expressed by a corresponding lexical term—a fact that sometimes gives rise to serious problems of naming descriptive categories. From a view that understands taxonomies as descriptive methods (cf. Konerding in particular) rather than as theoretical models (as does Barsalou, perhaps), the main problem with taxonomic approaches is, however, that they pretend to reflect a totality which, in fact, cannot actually be attained with encyclopaedically aligned descriptive models, in contradiction to the emphatic claims generally made by these very models.18 Another problem with frame taxonomies comes from the area of type-token. While on the level of the current frame instantiation (a level that Barsalou, for one, primarily, although not exclusively, has in mind) probably inheritance relations can be comparatively precisely determined (and the relation between slots and fillers can also be understood as an inheritance relation), this may turn out to be considerably more difficult on the level of abstract patterns, particularly if the function of signs as either types or tokens is involved as well. One reason for this is that epistemic relations (knowledge relations) can grow rampant in actual everyday knowledge because of analogy, a problem that is known in semantics since the 19th century. As is wellknown, human beings are masters at seeing (sensing, feeling) analogies. These do not simply stop where supposedly systematical conceptual hierarchies and inheritance relations would tell them to do so, and they would not even permit themselves to be limited by these factors (apart from the fact that every conceptual system is always the result of a certain, historically perhaps only accidental, view, idea, ideology, or
18 For example, the practical test to which Lönneker 2003 (according to Konerding 1993) subjected the taxonomic model of hyperonym type reduction is very sobering. Konerding’s top level frames (which he names ‘matrix frames’) covered only 38% of lexemes in one of Lönneker’s corpora in a first step. Only after they were complemented by new, ad-hoc matrix frames did they manage to cover 89%!
Frames as a Model for the Analysis and Description of Concepts
299
everyday theory, which may confront other views with other criteria of classification and inheritance relations). The aspect of frame inheritance and taxonomic relations must, then, in any adequate frame theory and analysis, always be complemented by the view of other, less systematic, epistemic frame connections and systems resulting from unsystematic (or systematically/taxonomically not strictly explicable) cross-references/analogies. In summary, the aspects of frame inheritance and taxonomic orders can be described as follows: Frames are recursive hierarchical structures consisting of knowledge elements that can, in turn, be described as frames. In this sense, frames are always structures made of sub-frames and supra-frames. A central aspect of frame structures is that, as we have seen, attachment positions (slots, terminals, attributes) can determine categorial characteristics of what is attached (fillers, values). In the taxonomic view, this implies that the fillers ‘inherit’ categorial aspects from their slots. This aspect can be called frame inheritance. Inheritance, seen this way, is a typical characteristic in hierarchically graduated conceptual systems i.e. in conceptual taxonomies. Besides the inheritance relation between slots and fillers (attributes and values), there are also inheritance relations between frames and superordinated (more abstract) frames, in regard to which the given frames represent specializations. It is important now to realize that such inheritance relations do not only refer to individual, isolated conceptual characteristics (knowledge elements). An essential effect of the recursivity of frames and frame structures is that certain constellations of frame elements (together with the typified relations existing between them) can also be inherited. This mainly unburdens cognition, since certain constellations of frame elements then only need to be stored once for an entire system of hierarchically graduated frame elements (or only the differences, additions or reductions). This can be observed in actant constellations in actant frame systems. For example, numerous frame elements are typologically identical in the specialized individual frames of a transport frame system (place of departure, destination, route, direction, energy expense, etc.).19 But it can be observed in other categorial frames and their frame elements of characterization as well. Thus, in frames for physical objects, frame elements such as form, colour, size, material are always present (if not always relevant to the same extent). Sub-domains of such inheritance relations can be organized (or described as such) in taxonomic, hierarchically constructed conceptual systems, as far as the material or the subject matter permits. I regard the construction of such hierarchical taxonomies
19 And the specific thing of a ‘Beam me up, Scotty!’ frame is roughly that the frame element Path, provided by the system, is more or less null instantiated, at any rate dissolved from its normal characteristics, irrelevant.
300
Dietrich Busse
(or rather: partial taxonomies)20 as an instrument of human cognition or epistemes, which is better adapted to certain subjects and areas of life, and less so to others. To put this in a different way: one may not misunderstand the taxonomic character of conceptual systems or frame systems as a value in itself (and one most definitely may not hold this to be ubiquitous); rather, one must always keep the instrumental character (sometimes productive, but sometimes misleading, too) in mind. Looked at it this way, the approach set forth by Fillmore and FrameNet is right in saying that taxonomic inheritance relations should only be assumed in the description/analysis where they are unavoidable and have a recognisable additional benefit.21
4.3 Epistemic frame connectedness and frame systems Frame research was drawn early on to the relations between frames and frame connectedness in knowledge which cannot simply be reduced to taxonomic structures. Various kinds of frame connectedness were identified. Fillmore pointed out, comparatively early, that not only the activation of a single frame/schema was needed for lexeme-related knowledge, but also knowledge of the schemata (scenes, frames) with which the word/lexeme itself (or the frames it activates) is connected. Fillmore mentioned structures similar to lexical/semantic fields by way of example, although he regarded the actual theory of semantic fields as inadequate. From a perspective that goes beyond Fillmore, one can also understand verb frames (such as buy, sell, pay, cost) as partial frames of an interconnected frame system which realizes different perspectives of a common total scene called ‘commercial event’—a term they mentioned by him as the starting point for the development of a frame theory. Minsky extended this aspect of perspectivization to perspective in the literal sense in visual perception, and described the different perspective frames of a visually perceived object (e.g. a table whose legs are completely covered and two other ones with partially covered legs), two others with partially covered legs) as elements of a frame system of the total object. This total object cannot be a purely ‘visual’ system. It is rather a cognitive or epistemic system integrated into the memory system, since there can be no total perspective on a visually perceived object.22 Frame systems are not
20 These are called, terribly misleadingly, ontologies in computer linguistic (and to some extent in cognitive) literature. 21 It is in any case known from cognitive lexicology that a median level of taxonomy (not too abstract, not too concrete) is evidently preferred in the mental lexicon. 22 For these frame systems, too, the following is valid: the different frames of a system share the same slots. One can see quite well that this is not just a characteristic in inheritance relations. I would assume rather that such relations, as they are described here by Minsky, enjoy priority to the inheritance relations in taxonomic conceptual systems. Hence, the latter prove to be a special case of a more general phenomenon, for which Minsky’s examples represent something like the original form.
Frames as a Model for the Analysis and Description of Concepts
301
limited to visual perspectives as introduced by Minsky, they can be extended to any epistemic system. Fillmore’s perspective frames on action complexes or events are an example of this. Frame systems as combinations of several perspectives or single aspects are therefore a more general phenomenon of knowledge and its structures. The above remarks show that knowledge as it is relevant to comprehension is characterized by a high degree of complexity and structuredness. Frames are the format for reflecting the organization of such structures, including frame systems and networks. Beyond the interconnectedness which is the result of taxonomic orders of knowledge, the associative relations between frames and frame elements, often based on the formation of analogies, the perception of contiguities, and metaphorical transfers, should not be underestimated with regard to their constitutive, structuring effect on social and individual knowledge. Frame systems can constitute themselves, in that the various individual frames of a system share common basic constituents (slots, attributes) or even groups of constituents as in the case of sets of frames that look at the same thing from various perspectives. In this case, they can be compared to inheritence relations, for which the same holds true as well. Frame networks can also be created by slots in other frames which are filled by more complex frames which get entangled with the epistemic structure of these frame networks through this fact. That is the case, for instance, if the complex legal frame property has a constitutive role in the meaning of the term theft. The description of such structures and networks of frames is a matter for applied epistemology and is of a clearly encyclopaedic character. This circumstance is best recognised in the Fillmore/FrameNet project association, where taxonomic considerations have been minimized in favour of a more descriptive, thesaurian approach to analysis, which is believed by linguists to be the most adequate approach to the objects of linguistic semantics. Frames or frame elements (or groups of frame elements) do not only belong to one and the same frame (e.g. as slot frames and filler frames that are assigned to a category concept as the elements constituting such a frame), they also form connections with different degrees of epistemic solidity. If such connections are of a certain duration and stability, one can speak of frame systems or frame networks. Depending on the types of connections, one can distinguish the following types of frame system (without any claim to systematicity and independence of criteria):
4.3.1 Frames as a frame system Every frame is first of all a frame system in itself (owing to the principle of recursivity), because it consists of subordinate frames (sub-frames, e.g. in the case of an instantiated frame consisting of attribute frames and value frames). Object frames for physical objects are one of the prototypes of such a frame system which is most frequently discussed in the literature. Minsky himself called these frames ‘frame systems’, arguing that completely formed frame systems exist only for the most important objects (not
302
Dietrich Busse
for all objects!). In this case, one may speak of micro-frame systems. The pertinent type of relation for such frame systems is the slot-filler relation.
4.3.2 Taxonomies Taxonomies are complex macro-frame systems that divide into numerous levels and groups of subordinate frame systems. The well-known conceptual hierarchies are a typical case in point. Taxonomies are characterized by hierarchical relations (superordinate or supra-frame, subordinate or sub-frame). Superordinate frames in taxonomies determine the frame elements (slots, default values) of the subordinate frames (either individually, as an inheritance of frame elements, typically however as the inheritance of groups of frame elements).
4.3.3 Congruency networks In contrast to the hierarchical relations in taxonomies, frame systems in the sense of congruency networks are based on matches of individual elements of knowledge. The pertinent type of relation for congruency networks is the relation of parallelity (or more precisely: congruency in the absence of hierarchical relations). Here, neighbouring frames share single frame elements or groups of them. A connection to taxonomy exists in that co-hyponyms in a hierarchy are always also congruency networks in the sense defined. Perhaps congruency networks are a prerequisite for the creation of taxonomies. A special case of such a congruency network is the well-known lexical/ semantic field.23
4.3.4 Seriality networks The individual frames in seriality frame networks may share single frame elements or groups of frame elements. If so, they share this property with congruency networks, even though this is not a necessary condition. What characterizes this type of networks are consequential relations (results) in the broadest sense. Seriality networks occur in the form of event frame systems and action frame systems (or as mixtures of the two). One can (and should) distinguish at least the following sub-types of seriality:
23 Congruence networks in the sense thus defined correspond to what Minsky called ‘similarity networks’.
Frames as a Model for the Analysis and Description of Concepts
303
a. Temporal seriality These are chronological sequences (series in the most narrow sense), with no implication of any logical causal relation. Most forms of such temporally constituted frame networks will be culturally conditioned (if they are not causal) and for that reason they cannot always be clearly distinguished from sub-type (b). An example could be work-out in the gym and then right after that visiting the sauna there. b. Culturally conditioned seriality This type of seriality consists of sequences of action frames, event frames, or both of them. They are based on socially anchored knowledge concerning the seriality of results in terms of conventions or prototypes. Examples are skiing and après-ski, or, more particularly, institutionalized systems of consequences with result relations, such as sermon and blessing, or Fillmore’s criminal trial etc. As far as they are not based on causal relations, the scripts in the sense of Schank and Abelson (1977) (at least to some extent) belong to this sub-type. c. Causally conditioned seriality What is meant here are frame systems that connect frames on the basis of causal relations. Causal relations are a type of such result relations about which there exists solid individual or social knowledge, or both, on the foundation of reliable supra-individual sources of knowledge. At the core of causality, reliable coherence naturally exists; but there is also socially (or culturally) conditioned causality (of the type of results that require no further verbalization: ‘They caught John with a BAC of .15 ...’). Causal seriality always presupposes temporal seriality (a result cannot exist in the world before its cause), and is therefore, in the end, a special case of temporal seriality. The descriptive or backward-directed perspective of causality corresponds to the forward-directed or hypothetical perspective of conditionality. The frame relation types causality and conditionality play an important role in the so-called scripts. Prior to systematization of the script analysis, the script concept should not be determined for causality and conditionality too quickly (as it is the case in Schank and Abelson). Between causal, cultural and temporal seriality relations there are mutual overlaps, so that frame systems (such as, for example, scripts or script systems) are likely to be mixtures of different seriality types. The plans, goals and themes postulated by Schank and Abelson are sub-types of frame systems closely associated with causal and conditional relations. d. Linguistic seriality Linguistic seriality may constitute a further sub-type of frame networks. It occurs, for instance, in all syntagmatic relations, which one might also describe by syntagmatic contiguity. What Barsalou calls ‘sentences’ and ‘texts’ in the context of frame networks qualify as good examples of linguistic seriality. But simple sentences, at any rate, belong to type 1 in the sense of Fillmore’s predication frame model: frames as frame systems. Complex texts (e.g. tales, novels, scientific monographs) are very
304
Dietrich Busse
special and strongly convoluted frame systems requiring special treatment and can hardly be comprehensively accounted for in terms of a pure frame analysis. They would at least strain frame semantics to the point of breakdown.
4.3.5 Associative networks The type of frame systems or networks with the weakest form of relations are the associative networks. The relations pertinent to associative networks occur in different forms. Relevant types of relation are at least the relations of contiguity, similarity, partial congruence (e.g. semantic isotopies in the sense of Greimas 1969), and (procedural) correlations/co-occurrences. a. Contiguity This type of relation is constitutive of frame systems of type (5) only if we are not dealing with contexts already created by another type (for instance, by type (1) frames as frame systems). Thus, for example, the part-whole contiguity is to be counted – as a rule – as type (1). Contiguity always occurs when certain frame connections do not belong to the knowledge constituting concepts, yet are frequently to be expected.24 Contiguity can but does not have to create frame systems or networks, but does not necessarily have to. That is true particularly of the following sub-type. b. Similarity The relation of similarity (in its constitutive form for frame networks) is always strongly dependent on subjective evaluation and perceptions; similarity must be seen.25 There may be numerous words in natural language that, in the end, are based on something very much like similarity networks. Take some fashionable expression, such as chav26, as a label for a class of object frames whose close relation is not strictly defined by element congruence, but rather via associative relations. In such a case, there is no concept in the strict sense of the word but there is an expression which can be taken as a label for an associative network – here in a pejorative sense as a social stereotype.
24 Thus, there were times, for example, when the frames Automobile and Gloves entered a solid connection in the sense of contiguity (occur in common, but without necessity). 25 When Minsky, then, designates the congruence networks, in the sense defined above, ‘similarity networks’, this is misleading, for he means factual agreement in clearly determinable frame elements, which does not apply to the usual understanding of similarity. 26 The original German example was Warmduscher that cannot be well tranlated in English. (Meaning a person who prefers to have a warm shower – contrary to hot or cold, used in a non-literal, pejorative sense for certain traits of personality or behaviour.).
Frames as a Model for the Analysis and Description of Concepts
305
c. Partial congruence In contrast to congruence networks as described under frame systems of type (3) above, the connectedness of this sub-type of (5c) is created by only a few frame elements. A rather well-known case is represented by certain types of the so-called isotopic relation (defined on the basis of individual semantic characteristics, thus not of entire meanings or concepts), as Greimas postulated them within the framework of the structural semantics of the 1960s. Isotopic relations create associative chains and can thus constitute entire networks out of the frames involved. Possibly, affordances are particularly well-suited to creating such associative frame networks , since they go back to only a single element (e.g. the associative frame network, tried and proven a thousand times in party games: ‘what I would take with me to a desert island’). d. (Procedural) correlations/co-occurrences Finally, there may possibly exist a type of associative frame networks that are created by something that one might call procedural contiguity, and that cognitive psychologists might describe like this: whatever is frequently processed together cognitively, can enter into associative concatenations in the memory. Every language system knows such forms of concatenation based on procedural co-occurrences.27 In particular, however, there are probably mixed forms of connection, created by frequent or salient co-processing between the data of various channels of perception (seeing, hearing, smelling, tasting, feeling).28
5 Performance and limits of frame-semantic analyzes The empirical research on frame analysis is still too new and too rarely applied – in the sense of systematic research comprising the entire spectrum of theoretical models and possibilities – as yet to have attained a conclusive evaluation of its potential and its limits. Linguists such as Fillmore and cognitive scientists such as Minsky and Barsalou have made impressive lists of what might be investigated with the help of their rather different types of frame models in the field of concepts and language in the broadest sense. Beginning with verbs, nouns and sentences, through cognitive concepts, texts, morphemes, metaphors, anaphors, presuppositions, to prepositions and conjunctions – nearly everything in the field of linguistics has been named as a pos-
27 Much of what is called ‘constructions’ by Fillmore may belong to this category. 28 This is certainly where the phenomena of so-called synaesthesia belong; these can grow to be genuine illnesses. In particular, certain types of traumatic disturbances are a part of this too. (This sub-type 5d may not be easily separable from sub-type 5a of associative contiguity.).
306
Dietrich Busse
sible field for the application of some kind of frame research. The programmatic part, then, is certainly ambitious. However, there are good reasons to asume that this is not going to happen. It will not be possible to analyze all the phenomena described above equally well. Taking the example of Fillmore’s background frames, it is even unclear to what extent they can be integrated into Barsalou’s model of frames. It is true of many of Fillmore’s prime examples (orphan, widow, bachelor, vegetarian, on land, on the ground) that the knowledge associated with these concepts is so complex and demanding that a suitable paraphrase may require a greater number of additional frame elements (structures, relations), which in turn need to be analyzed and defined with all their elements in a frame-semantic way. In this way, one would rapidly arrive at a rather complex description. This is true too, as we have seen, of analytically complex concepts such as those typical of the world of law,29 but certainly also of other kinds of complex concepts, such as those in philosophy, or maybe the basic historical and social concepts as Koselleck (1972) and others saw them.30 Such complex structures probably will not permit descriptions within a single frame (and certainly not, if this is a graphic), but must be divided into numerous sub-structures, each of which is first investigated for its own sake, before they can be integrated into superordinate structures. Thus, it would surely be mistaken to see a magic genie in frame theory that can solve all theoretical problems concerning concepts and semantics and just about any possible research question. The frame theory is strong where it fills in the gaps left by older theoretical conceptions of meaning (such as feature semantics, logical semantics, word-isolating lexical semantics, logic-based compositional sentence semantics, word- and concept-isolating historical semantics). More precisely, frame theory comes in wherever the scope, the complexity, the subtlety, the power to distinguish, and the epistemic connectedness of the knowledge relevant to comprehension or to concepts has been (sometimes severely) underestimated in older models. In the field of analyzing complex conceptual structures, concept competition, and conceptual change, frame analysis can unfold its special ability to perform, and is, as far as I can see, superior to other approaches. The limits of its possibilities become (and can become) visible only when these possibilities are tested in empirical analyzes of the most various kinds in the most various fields of semantics or concept analysis. The history of concepts and the investigation of conceptual systems could be one of these fields.
29 For more details and examples of frame analyzes of legal terms, cf. Busse 1992; 2008a; 2008b. 30 For more details on ‘Begriffsgeschichte’, i.e. conceptual history as founded by Koselleck 1972, cf. Busse 1987; 2003; Busse and Teubert 1994; and, with relation to frame theory, Busse 2007; 2008c.
Frames as a Model for the Analysis and Description of Concepts
307
References Barsalou, Lawrence W. 1992. “Frames, concepts, and conceptual fields.” In Frames Fields and Contrasts, edited by Adrienne Lehrer and Eva. F. Kittay, 21–71. Hillsdale, N.J.: Erlbaum. Barsalou, Lawrence W. 1993. “Flexibility, Structure, and Linguistic Vagary in Concepts: Manifestations of a Compositional System of Perceptual Symbols.” In Theories of Memory, edited by Alan F. Collins, Susan E. Gathercole, Martin A. Conway and Peter E. Morris, 29–101. Hove, UK/ Hillsdale, N.J.: Lawrence Erlbaum. Bartlett, Frederick C. 1932. Remembering: A Study in Experimental and Social Psychology. Cambridge: UP. Busse, Dietrich. 1987. Historische Semantik. Stuttgart: Klett-Cotta. Busse, Dietrich. 1992. Recht als Text. Linguistische Untersuchungen zur Arbeit mit Sprache in einer gesellschaftlichen Institution. Tübingen: Max Niemeyer Verlag. Busse, Dietrich, Fritz Hermann and Wolfgang Teubert. 1994. Begriffsgeschichte und Diskursgeschichte. Methodenfragen und Forschungsergebnisse der historischen Semantik. Opladen: Westdeutscher Verlag. Busse, Dietrich. 2003. “Begriffsgeschichte oder Diskursgeschichte? Zu theoretischen Grundlagen und Methodenfragen einer historisch-semantischen Epistemologie.” In Herausforderungen der Begriffsgeschichte, edited by Carsten Dutt, 17–38. Heidelberg: Winter. Busse, Dietrich. 2005. “Architekturen des Wissens. Zum Zusammenhang von Semantik und Epistemologie.” In Begriffsgeschichte im Umbruch, edited by Ernst Müller, 85–99. (Archiv für Begriffsgeschichte, Sonderheft) Hamburg: Meiner. Busse, Dietrich. 2007. “Linguistische Epistemologie. Zur Konvergenz von kognitiver und kulturwissenschaftlicher Semantik am Beispiel von Begriffsgeschichte, Diskursanalyse und Frame-Semantik.” In Sprache – Kognition – Kultur. Sprache zwischen mentaler Struktur und kultureller Prägung, edited by Heidrun Kämper, 73–114. (Jahrbuch 2007 des Instituts für deutsche Sprache) Berlin and New York: de Gruyter. Busse, Dietrich. 2008a. “Semantische Rahmenanalyze als Methode der Juristischen Semantik. Das verstehensrelevante Wissen als Gegenstand semantischer Analyze.” In Rechtstheorie in rechtspraktischer Absicht. Freundesgabe zum 70. Geburtstag von Friedrich Müller, edited by Ralph Christensen and Bodo Pieroth, 35–55. (Schriften zur Rechtstheorie 235) Berlin: Duncker & Humblot. Busse, Dietrich. 2008b. “Interpreting law: text understanding – text application – working with texts.” In Law and Language: Theory and Society, edited by Frances Olsen, Alexander Lorz and Dieter Stein, 239–266. Düsseldorf: Düsseldorf University Press. Busse, Dietrich. 2008c. “Begriffsgeschichte – Diskursgeschichte – Linguistische Epistemologie. Bemerkungen zu den theoretischen und methodischen Grundlagen einer Historischen Semantik in philosophischem Interesse anlässlich einer Philosophie der Person.” In Diskurse der Personalität: Die Begriffsgeschichte der ‚Person‘ aus deutscher und russischer Perspektive, edited by Alexander Haardt and Nikolaj Plotnikov, 115–142. München: Fink. Busse, Dietrich. 2009. Semantik. Eine Einführung. München: Fink. Busse, Dietrich. 2011. “Begriff.” In Lexikon der Geisteswissenschaften, edited by Helmut Reinalter and Peter J. Brenner, 31–39. Köln/Weimar and Wien: Böhlau. Busse, Dietrich. 2012. Frame-Semantik – Ein Kompendium. Berlin and Boston: de Gruyter. Fillmore, Charles J. 1968. “The Case for Case.” In Universals in Linguistic Theory, edited by Emmon Bach and Robert T. Harms, 1–88. New York: Holt, Rinehart & Winston. Fillmore, Charles J. 1975. “An alternative to checklist theories of meaning.” In Proceedings of the First Annual Meeting of the Berkeley Linguistics Society, edited by Cathy Cogen et al., 123–129. Berkeley: Berkeley Linguistics Society.
308
Dietrich Busse
Fillmore, Charles J. 1977. “Scenes and Frames Semantics.” In Linguistic Structure Processing, vol. V, edited by A. Zampolli, 55–81. Amsterdam: North Holland. Fillmore, Charles J. 1982. “Frame Semantics.” In Linguistics in the Morning Calm, edited by The Linguistic Society of Korea, 111–137. Seoul: Hanshin Publishing Corp. Fillmore, Charles J. 1992. “‘Corpus linguistics’ vs. ‘computer-aided armchair linguistics’.” In Directions in Corpus Linguistics, edited by Jan Svartvik, 35–60. (Proceedings of 1991 Nobel Symposium on Corpus Linguistics) Berlin and New York: Mouton de Gruyter. Fillmore, Charles J. 2006. “Frame Semantics.” In Encyclopedia of Language and Linguistics (2nd Edition), edited by Keith Brown, 613–620. Amsterdam: Elsevier. Fillmore, Charles J., Srini Narayanan, Collin F. Baker and Miriam R. L. Petruck. 2002. “FrameNet Meets the Semantic Web: A DAML+OIL Frame Representation.” In Proceedings of the Eighteenth National Conference on Artificial Intelligence and the Fourteenth Annual Conference on Innovative Applications of Artificial Intelligence, edited by American Association for Artificial Intelligence. Edmonton, Canada. [http://www1.icsi.berkeley.edu/~snarayan/semweblr.pdf] Greimas, Algirdas Julien. 1966: Sémantique structurale. Paris: Larousse. Haller, Rudolf and Jürgen Mittelstrass. 1971. Art. “Begriff”. In Historisches Wörterbuch der Philosophie, Bd. 1, edited by Joachim Ritter, 780–787. Darmstadt: Wissenschaftliche Buchgesellschaft. Humboldt, Wilhelm von. 1820. “Ueber das vergleichende Sprachstudium.” In Über die Sprache. Ausgewählte Schriften, edited by Jürgen Trabant. München: dtv. Husserl, Edmund. 61980 [1913]. Logische Untersuchungen. Bd. II / 1.Tübingen: Niemeyer. [English translation: Edmund Husserl: Logical Investigations. 2nd edition. Ed. Dermot Moran, translated by Finlay (1970). London/New York: Routledge 2001] Konerding, Klaus-Peter. 1993. Frames und lexikalisches Bedeutungswissen. Untersuchungen zur linguistischen Grundlegung einer Frametheorie und zu ihrer Anwendung in der Lexikographie. Tübingen: Niemeyer. Koselleck, Reinhart. 1972. “Einleitung.” In Geschichtliche Grundbegriffe. Historisches Lexikon zur politisch-sozialen Sprache in Deutschland, Bd. 1, edited by Brunner, Otto/Conze, Werner/ Koselleck, Reinhart, S. XIII–XXVII. Stuttgart: Klett-Cotta. Lönneker, Birte. 2003. Konzeptframes und Relationen. Extraktion, Annotation und Analyze französischer Corpora aus dem World Wide Web. Berlin: Akademische Verlagsgesellschaft AKA. Minsky, Marvin. 1974. “A Framework for Representing Knowledge.” In Artificial lntelligence Memo No. 306. (M.I.T. Artificial Intelligence Laboratory) [Reprint 1975 in: The Psychology of Computer Vision, edited by Patrick H. Winston, 211–277. New York: McGraw-Hill]. Minsky, Marvin. 1986. The Society of Mind. New York: Simon and Schuster. [German translation: Minsky, Marvin. 1990. Mentopolis. Stuttgart: Klett-Cotta]. Schank, Roger C. and Robert P. Abelson. 1977. Scripts, Plans, Goals and Understanding: An Inquiry into Human Knowledge Structures. Hillsdale, N. J.: Lawrence Erlbaum Associates. Tesnière, Lucien. 1959. Eléments de syntaxe structurale. Paris: Klincksieck. [quoted German translation: Tesnière, Lucien. 1980. Grundzüge der strukturalen Syntax. Edited and translated by Ulrich Engel. Stuttgart: Klett-Cotta. English translation: Tesnière, Lucien. Elements of structural syntax. Amsterdam: John Benjamins 2015.] Wahrig, Gerhard. 1979. Deutsches Wörterbuch. Gütersloh: Bertelsmann. Wittgenstein, Ludwig. 1970. Über Gewißheit. Frankfurt a. M.: Suhrkamp. [English Translation: Ludwig Wittgenstein: On Certainty (Uber Gewissheit) ed. Gertrud E.M. Anscombe and Georg H.von Wright. Translated by Denis Paul and Gertrud E. M. Anscombe Oxford: Basil Blackwell 1969–1975.] Wittgenstein, Ludwig. [1953] 1971. Philosophische Untersuchungen. Frankfurt a. M.: Suhrkamp. [English Translation: Ludwig Wittgenstein: Philosophical Investigations. Translated by G.E.M.Anscombe. Oxford: Basil Blackwell, 2001.]
Frames as a Model for the Analysis and Description of Concepts
309
Ziem, Alexander. 2008. Frames und sprachliches Wissen. Kognitive Aspekte der semantischen Kompetenz. Berlin: de Gruyter. [English translation: Alexander Ziem: Frames of Understanding in Text and Discourse. Theoretical foundations and descriptive applications. (Human Cognitive Processing, 48) Amsterdam/Philadelphia: John Benjamins 2014]
Figures Fig.1: After Barsalou 1992: 45. Fig 2: After Barsalou 1992: 45. Fig 3: Fillmore et al. 2002: 5.
Frederik Elwert and Simone Gerhards
Tracing Concepts – Semantic Network Analysis as a Heuristic Device for Classification Summary: Since the digital age is affecting more and more aspects of historical studies, the need to explore and evaluate new computational methods for the constantly growing amount of digital resources has become greater than ever. Due to the huge amount of digitized text, new methods may help to answer innovative research questions, challenge existing theoretical patterns and generate new approaches and perceptions. The research field emerging as a result of these reflections, the so-called digital humanities, might increasingly be incorporated into the established humanities disciplines. The authors present in this paper a heuristic method for the study of textual sources that allows the tracing of ancient Egyptian concepts and classification schemes by uncovering the relationships of words. By using automated network analysis in combination with a close reading of the sources, we can trace the outlines of the semantic field of a word; the ancient Egyptian expression for day serves as an example and will be studied in detail with this method.
1 Introduction: what is semantic network analysis and what is it good for? One growing field in which digital approaches are applied is historical network analysis. Originating from the social sciences, network analysis is a method to analyze relations between units, mostly entities like people or groups, and to reconstruct and visualize underlying patterns. A related field is semantic network analysis, a method that allows relationships between words within one or more textual sources to be analyzed. But the term semantic network analysis does not have the same meaning in every field of research.1 In this paper, we define it as follows: Semantic network analysis understands text as a complex semantic system that consists of different lexical units, such as nouns, verbs, and proper names. These are in a certain relationship which is defined through closeness and distance to each other, not only through frequency. These relationships can be used to extract information about the concepts behind
1 For an overview of different definitions of semantic network analysis, see van Atteveldt 2008: 2–6; 13–61. DOI 10.1515/9783110538779-012
312
Frederik Elwert / Simone Gerhards
them, but also to reconstruct hidden patterns of meaning. It captures information about the context of a term as a network that can be visualized in order to guide interpretation.2 The paper is based on research in the context of the SeNeReKo project. SeNeReKo is a joint research project of the Center for Religious Studies in Bochum (CERES) and the Trier Center for Digital Humanities (TCDH), Germany.3 The project aims to utilize new methodological approaches for the study of digitized historical corpora. This approach was first applied in Egyptology as a method for the reconstruction of the ancient Egypt concepts in a paper on the terms heka and maat (Hofmann and Elwert 2014). This paper describes a method for the study of textual sources that allows the tracing of ancient Egyptian concepts by investigating the relationships of words. Through the use of automated network analysis as a heuristic device in combination with a close reading of the sources, the ancient Egyptian expression for day will be studied in detail. Different implicit conceptualizations connected with the term are detected by looking at its context. We argue that clusters of textual units, detected through analyzing relations between words, can serve as indicators for the classification of typical usage patterns of key terms in historical corpora. The results highlight which conceptualizations are specific for a certain ancient Egyptian time period, and which of them have a broader scope.
2 Theory: distant reading and close reading Traditionally, textual scholarship focuses on extensive reading of relevant source texts. This close reading allows us to grasp details in style or linguistic structure that can only become visible against the backdrop of scholarly expertise in the field under study. While these approaches allow a thorough reconstruction, they are also limited in their scope. In his widely discussed book, Franco Moretti argues for distant reading, and he provides two main arguments. First, he mentions sampling. Close reading depends on the selection of works that are included in the study, while others – often a majority – are left out. He criticizes this kind of canon building that he observes in literary studies: ‘you invest so much in individual texts only if you think that very few of them really matter’ (Moretti 2013: 48). With increasing digitization and digital pro-
2 This definition consists of ideas based on van Atteveldt 2008: 66; Lietz 2007: 1 f.; Rieger 1988: 44–46. 3 The acronym SeNeReKo stands for the German title Semantisch-soziale Netzwerkanalyse als Instrument zur Erforschung von Religionskontakten. The project was funded by the German Federal Ministry of Education and Research (BMBF) as an eHumanities project under the project number 01UG1242A until June 30th 2015. For further information see http://senereko.ceres.rub.de/. The authors are responsible for the content of this chapter.
Tracing Concepts – Semantic Network Analysis
313
duction, canonization becomes more and more a matter of choice, not of necessity. It is possible to study texts on a much larger scale now, also including works outside the classical canon. At the same time, this kind of large-scale work requires computational methods. Even if larger and larger collections are available in digital form, time is still a limiting factor for close reading. Only computer-assisted techniques, which utilize quantitative analysis, allow us to make use of the available source material. This is when text becomes data. This might explain the vivid interest in distant reading methodology in the study of early modern and modern literature. But for ancient history, sampling depends on many other factors besides canonization. In fact, physical preservation might be the most limiting factor in the selection of source material. However, there is a second argument for distant reading, namely theory. In the context of social science research, Strauss and Corbin differentiate theory from description by its use of concepts, and its making statements about the relations of these concepts (Strauss and Corbin 1996: 13). Moretti makes a similar point in favour of distant reading. Only abstraction enables insight on a systematic level, although it comes at the price of losing detail (Moretti 2013: 49). For him, this lack of detail is not a disadvantage, quite the contrary: when it comes to theory, ‘distance […] is a condition of knowledge’ (ibid., 48 [emphasis in original]). But conceptualization is not only a scholarly activity; it is also part of social interaction and expressed through language. Corpus linguistics as an emerging field can be seen as a result of the empirical turn in linguistics. Instead of seeing category building as a theoretical exercise, it aims at describing acts of categorization as an empirical phenomenon: It is the discourse itself, and not a language-external taxonomy of linguistic entities, which will have to provide the categories and classifications that are needed to answer a given research question. This is the corpus-driven approach. (Teubert 2005: 4).
In this regard, network analysis can serve as a bridge between linguistics and cultural studies. Network analysis seeks to uncover the implicit structures hidden in textual data (Bubenhofer 2009: 100). These structures represent meaning. In this sense, meaning is expressed as well as constituted through repetition. Only the previous use of a word informs statements about its meaning, and the continued use of a word constitutes its meaning in the future (Rieger 1988: 44; Teubert 2005: 3).4 Empirical, corpus-driven methods then allow us to trace the acts of conceptualization that are part of the texts we study. As a practice in distant reading, the single text itself disappears, while the
4 For a detailed elaboration of the semiotic theory that informs this approach, see Rieger 1989.
314
Frederik Elwert / Simone Gerhards
patterns of language use start to show, providing hints about the concepts that stand behind the text. The use of network metaphors in the study of language goes back as far as Wittgenstein (2008 [1953]: 57). On a methodological level, the meaning – or better yet, the meanings – of a word can be seen as a network of interwoven dependencies to other words. When operationalizing this basic idea that meaning is constituted through language use, the context in which a word appears in the corpus is an important indicator. The interplay of a given word with the words surrounding it indicates the concepts behind the use of the word. In computational linguistics, a network model has been used as one approach to solve the problem of polysemy. When a word has multiple meanings, the different contexts of its use should become visible in the network structure of interrelated words (Biemann 2006; Dorow and Widdows 2003). Here, we apply a similar approach to trace ancient Egyptian concepts. Still, a corpus-driven approach can only serve as a first step. The patterns that emerge from the computational analysis require interpretation. The interpretation cannot be based on the discovered patterns alone; it requires taking the text back into account (Bubenhofer 2009: 103). In a second step, the network can and should be used as an index to the text. It allows us to spot interesting structural phenomena and trace their origins in the corpus (Elwert 2016). Semantic network analysis as we understand it is merely a heuristic device that guides the work of the scholar, namely: the reconstruction of underlying phenomena and concepts.
3 Definition: what is a network and what does it look like Before we start discussing the method, it is necessary to give a definition of the technical terms used in this paper. The idea of networks is almost ubiquitous in different fields of the humanities and social sciences. Generally, it simply expresses the idea of relationality. Formal network analysis as a specific form of computational analysis, however, is based on a very narrow idea of a network, rooted in mathematical graph theory. A network (also called graph) consists of nodes (also called vertices) and edges (also called links). Since this model is of a completely abstract type, there is no limitation as to the nature of the nodes or edges. Anything that can be described as a set of entities (nodes) and their relations (edges) can be modelled as a network. In our case of semantic networks, words are the nodes, while the edges describe a shared context (for details about network extraction, see below). The network model used here has two additional properties: it is undirected (in contrast to a directed network), which means that all relations between words are treated as symmetrical. And it is weighted, so that the frequency of relations can be taken into account.
Tracing Concepts – Semantic Network Analysis
315
This abstract graph model allows us to mathematically uncover structural properties of the network and of individual nodes. Many network metrics and network analysis algorithms are available that can be used to describe network characteristics. Here, we mainly use two kinds of algorithms. Centrality describes which nodes are more important than others (Freeman 1978: 215–217; Borgatti and Everett 2006: 466–468). There are different possibilities to define importance, so consequently there are several ways to compute centrality. The plainest of the centrality measures is the degree. Simply put: in an undirected network, the degree is the number of edges a node has (Freeman 1978: 218). The more nodes a node is directly connected to, the more central it is, following this definition of centrality. An extension of the degree measure is the weighted degree. It also takes into account the weight of edges, which in our case represent the frequency of a relation. A single edge with a weight of 2 contributes as much to the weighted degree as two edges with a weight of 1. While centrality measures give information about single nodes, they reveal only little about the internal structure of the network as a whole. A more complex set of analytical algorithms deals with the problem of community detection. In network terms, communities are clusters of nodes that are relatively tightly connected, while having only a few connections to nodes outside the community (Blondel et al. 2008: 2). In many cases, there is no single solution to this problem, or it is computationally too complex to be solved in a reasonable time. As a consequence, several different strategies are applied to find an approximate solution in a reasonable time. Additionally, many of these algorithms have a random component, so even running the same algorithms twice does not necessarily return exactly the same communities. While this often seems unsatisfying, in a way it mirrors the nature of humanities research. In contrast to simple statistics, these questions are so complex that there is no single correct answer. Rather, the result of the computation is just a heuristic device that highlights one possible view of the structural properties of the object under study. At the same time, these results are not arbitrary, as they do represent one possible division of the network into groups.
316
Frederik Elwert / Simone Gerhards
4 Material: big data and small data in egyptology Our starting point for the network analysis of ancient Egyptian terms is the Thesaurus Linguae Aegyptiae,5 a database and publication platform by the Ancient Egyptian Dictionary Project at the Berlin-Brandenburg Academy of Sciences and Humanities.6 The Thesaurus contains a digital corpus of Egyptian texts that supports computerassisted research. The TLA kindly provide the SeNeReKo project with their digitized and linguistically annotated data, which contain more than one million lexical units. Every unit is annotated with part of speech information like noun, preposition, etc. For names, additional information about entity classes such as king, god or toponym are given. This very detailed annotation schema is extremely helpful for research methods like network analysis and allows results of a high quality to be obtained. German and partly English translations exist for the ancient Egyptian words and texts, but the networks are generated directly from the Egyptological transliteration of the textual sources. The digitized texts that are part of the TLA collection go back as far as the third millennium BC. The corpus consists of different text genres, such as religious hymns, biographical inscriptions, or literary and medical texts. The length of the texts varies from a few words to more than 8,000. In this regard, it is a very heterogeneous text corpus. Although it consists of more than a million lexical units, it cannot necessarily be classified as big in the context of computational analysis. One million sounds like a rather big amount, but it is a) still only a small excerpt of all ancient Egyptian texts existing and b) still very small compared to big data in the usual sense of data with a size beyond the ability of standard software tools (Snijders, Matzat and Reips 2012: 1). But by Egyptological standards it can be considered a reasonably big database with quite a large number of sources. Therefore, it is a very good starting point for our approach of network analysis of ancient Egyptian terms.
5 For further information see http://aaew.bbaw.de/tla/index.html. Hereafter referred to as TLA. 6 Contributions have been made by project groups at the Berlin-Brandenburg Academy of Sciences and Humanities in Berlin, the Saxonian Academy of Sciences and Humanities in Leipzig, the Demotic Text Database Project of the Academy of Sciences and Literature Mainz, the Book of the Dead Project of the North Rhine-Westphalian Academy of Sciences, Humanities and the Arts (Bonn), the project Digital Heka Leipzig, the Leuven Online Index of Ptolemaic and Roman Hieroglyphic Texts Project of the Katholieke Universiteit Leuven and the Edfu Project.
Tracing Concepts – Semantic Network Analysis
317
5 Use case: the egyptian day represented by the term hrw The ancient Egyptian common term hrw (WB II, 498–500.24) for day has been chosen as example to show the method of semantic network analysis because it is frequently used during all epochs and in different contexts. To date, there is no scholarly publication which examines the semantics and contextual fields of hrw in detail. By offering a diachronic perspective on the context of hrw, network analysis provides a new approach to classifying and categorizing its use (cf. the categorization scheme: day – night; day – month – year; etc.). However, because only a limited sample of texts is available in the TLA, it is not possible to reconstruct all uses of the term. Therefore, the statements in the following examples correlate just to the references that are listed in the TLA database and cannot be generalized for all of ancient Egypt. The analyses presented here should be seen as a starting point and can be extended once larger digital corpora are available.
5.1 Method: strategies for network extraction In order to extract a network representation of the term hrw, one has to specify rules that define nodes and edges with regard to words in a corpus. The basic unit for defining the nodes are words. However, not all words are relevant for semantic analysis, and different surface forms of a single lexeme tend to distort the analysis. So in order to identify the nodes of the network representation, words are first mapped to their lexical base form or lemma. This step is called lemmatization in computational linguistics (Manning and Schütze 1999: 132). The second step is to remove so-called stop words. Stop words are frequent terms that – of themselves – carry only little meaning, like prepositions, conjunctions, auxiliary verbs, and other frequently used, generic words (Ganesan 2014). In certain syntactic arrangements, these words can influence semantics of other words or of larger phrases. But since network analysis removes the words from their syntactic context, these functions of prepositions and other stop words are discarded. Additionally, in the TLA corpus, words are assigned discrete lemmata in case prepositions significantly alter the word sense. They can be removed without affecting the semantic information. Removing these words can help to reduce the amount of noise and usually improves the results of further analyses. For every network of hrw, the same stop word list was used in order to ensure stable conditions for analysis. In addition to a fixed stop word list, words can also be selected based on
318
Frederik Elwert / Simone Gerhards
part-of-speech information. In the case of the analyses presented below, we built the networks only from nouns, verbs and adjectives.7 The edges in the network representation approximately express semantic relations. Corpus linguistics and computational linguistics usually define word associations as co-occurrences.8 If two words regularly appear in the same context, they will likely have a semantic relationship (Manning and Schütze 1999: 185; Bubenhofer 2009: 115). In order to build a network model of these semantic relations, it is necessary to specify a formal criterion for linking two words. A usual procedure is to define a window of a given size: all the words that appear around the target term within the specified distance will be linked to that term (Bullinaria and Levy 2007: 513). Here, we are interested not only in the links of our target term hrw itself, but also in the links between the words surrounding hrw. Thus, also the contexts of all related terms are taken into account when determining the edges in the network. The result is an ego-network of hrw, that is the network consisting of all nodes connected directly to hrw (its neighbourhood) and the edges between these nodes (the alteri). A straightforward procedure would be to collect the edges between all words in the corpus, and then select only the ego network of hrw. However, this approach would also include those edges between alteri that do not occur in the context of hrw, but elsewhere in the corpus. The resulting network would then reflect, to a large extent, the general semantic structure of the corpus, but not necessarily that of hrw. Considering this, we take only those edges between alteri into account that can be identified in the context of hrw. Our first step for network extraction is to generate a sub-corpus that contains all sentences9 in which the lexeme hrw10 occurs. Experiments show that using sentences as the delimiting text unit for alter-alter relations produces better results than using larger units. Additionally, sentences are taken into account when determining word co-occurrences. Words are only linked by an edge if the window does not cross a sentence boundary. Our tests showed that sentences seem to approximately express semantic contexts, while word relations across sentences are often less interpretable. Linguistic analysis of co-occurrences usually takes the frequency of co-occurrences into account. More frequent co-occurrences will express a stronger semantic relation. Instead of using raw frequency counts (Bullinaria and Levy 2007: 513), statistical measures such as likelihood ratios (Manning and Schütze 1999: 172) or pointwise mutual information (ibid, 178; Bullinaria and Levy 2007: 514) are used. Another
7 Nouns include the names of gods and goddesses. The definition of the grammatical terms is based on TLA annotation schema. 8 The term co-occurrence denotes two words appearing in a defined context, e.g. a sentence or a certain number of subsequent words, Bubenhofer 2009: 113. 9 The Egyptian writing systems under study usually do not mark word or sentence boundaries. Our use of word and sentence follows the editorial decisions expressed in the TLA encoding. 10 TLA Lemma No. 99060.
Tracing Concepts – Semantic Network Analysis
319
approach is to use word proximity to define edge weights. Not all words within the selected window size are treated as equally connected; instead, words appearing closer to the target term are linked with a higher weight. An innovative solution in the domain of text network analysis has been described by Paranyushkin (2011). This procedure uses a rolling window of five, i.e. selecting the first five words and linking them, then selecting words two to six and linking them, and so on. If a link already exists, its weight is incremented by one. Then, the procedure is repeated with a window of two, i.e. linking directly neighbouring words again. As a result, words closer to each other are linked through edges with a higher degree compared to words with a higher distance in the text. Our implementation of this approach differs from the original one in that we stop at sentence boundaries and not only at paragraph boundaries. Fig. 1 shows this principle using only the nouns of the sentence: Some years ago – never mind how long precisely – having little or no money in my purse, and nothing particular to interest me on shore, I thought I would sail about a little and see the watery part of the world. (from Herman Melville’s Moby Dick [1851: 1]).
Our experiments comparing this approach to a statistical significance measure (loglikelihood ratio) show that this kind of weighted window produces results that are more interpretable in the context of network analysis.
Fig. 1: Word network of nouns in a single sentence
In the corpus, there are 678 different11 references to hrw, with the earliest originating from the Old Kingdom12 and the latest from the Greco-Roman Period13. Every unique sentence that contained the lemma hrw was collected into a subcorpus. In a second step, two additional subcorpora were extracted for a diachronic perspective. The networks of the Old Kingdom (2700–2200 BC) and New Kingdom (1550–1070 BC) have been selected because they contained enough references to allow for a more detailed analysis. The assignment of texts to periods follows the metadata contained in the TLA database.
11 Text passages that occur more than once were removed manually. 12 Tomb of Rawer in Giza, from the 5th dynasty BC, PM III, 269. 13 pFlorenz PSI inv. I 72 from the 1st century CE.
320
Frederik Elwert / Simone Gerhards
The co-occurrence networks were created from the corpus using the TCFnetworks package (Elwert 2014). The networks were then imported into Gephi (Bastian, Heymann, and Jacomy 2009) for analysis and visualization.
5.2 Network analysis: Detecting semantic clusters Following the idea that the context of a word is indicative for its meaning, we assume that usage, and consequently meaning, of a word are not uniform. Meaning can change not only diachronically as semantic shift, but also synchronically, as different meanings of a word (or nuances of a word’s meaning) can be highlighted, depending on the situation. In computational linguistics, this assumption informs a wide range of research in the domain of word sense disambiguation. For text processing, it is often necessary to identify the meaning of an ambiguous word. The basic approach tries to assign an occurrence of a word to one of its senses as listed in a lexicon (Knopp, Völker, and Ponzetto 2013: 97). Since the corpus of the TLA already contains disambiguated lemma information for each word, this is not relevant to our research. Instead, we are interested in more subtle shifts of meaning that are not discriminated in a dictionary. To this end, techniques of word sense induction can be applied, since they aim at discovering senses of a word from a given corpus without using additional resources like a lexicon (ibid). Semantic network analysis can be applied for the task of word sense induction (Dorow and Widdows 2003). A semantic network allows the identification of context clusters, and consequently of distinct meanings of a word. Using techniques of community detection (see above), the semantic network around a given term can be partitioned into sub-networks. Since the network edges represent co-occurrences of words in the context of the term under study, in this case hrw, these communities indicate relatively distinct contexts in which hrw is used.
5.3 Network representation: opportunities and challenges In the following section we will discuss some exemplary networks in order to show the method with its opportunities as well as some challenges it poses. It should not be seen as an approach to generate new Egyptological knowledge, but as a heuristic instrument that is able to indicate the places in the original textual sources, which can be consulted to reconstruct ancient Egyptian concepts and/or classification schemes.
Tracing Concepts – Semantic Network Analysis
321
Firstly we will discuss the entire network made of all references for the lemma hrw. It is suitable in order to get a general overview of the words, their relational structure, and the context information, because it spans all time periods and text categories. But considered only by itself, it is not possible to draw conclusive statements from it, precisely because it spans different time periods and text categories. Thus, in a second step, the entire network will be compared with the networks of the Old and New Kingdom. This will provide a diachronic perspective and allow for more specific conclusions. In comparison to the full network, it is interesting to observe how communities change their size and content, and how new communities arise. This may indicate time dependent conceptualizations of hrw. The entire network, constructed on the basis of all 678 references, consists of 1368 nodes and 8194 edges. In order to clarify the structure of the network, we selected only nodes directly connected to hrw, but removed the node hrw itself, because it has a relation to all other words in the network by definition. In this process, single nodes and small groups that have no connection to the majority of the nodes – the so-called giant component – are also removed. This yields a smaller graph consisting of merely 963 nodes and 5267 edges. However, this is still too big to perceive all relevant information at a glance (for an excerpt of the network see fig. 2). Therefore, we will discuss the community structure that was detected by the above-mentioned algorithm by Bondel et al. (2008). We can identify 15 conceptual communities within the network, the largest of which consists of 10.49% of all nodes and the smallest of 3.32%. None of the communities is clearly dominant. In the network every community is visualized using a different colour (see fig. 2). The central nodes of the network are: grḥ (TLA no. 167920), jwi (TLA no. 21930), nṯr (TLA no. 90260), jb (TLA no. 23290), RꜤw (TLA no. 400015), ḥm (TLA no. 104690), and pri (TLA no. 60920). The interpretation of the results will be carried out in relation to two different levels. On the one hand it is visually supported by the two-dimensional network and on the other hand by different centrality measurements and algorithmic calculations. In the following tables, four exemplary communities and their first members are listed. For easy reference the unique TLA number is added in brackets. The translation corresponds to the one given in the TLA. In addition, you will find both the count of a certain word in the extracted text corpus and its degree in the particular subnetwork. Although count and degree correlate with each other, the degree can be seen as a relational measure of the importance of a word with regard to hrw.
322
Frederik Elwert / Simone Gerhards
Fig. 2: Excerpt of the entire network constructed on the basis of all 678 references of hrw
Nodes
Count
Degree
Translation
jri (hrw) (29030)
35
22
to spend (the day)
pr (60220)
16
14
house
ṯꜢi (174260)
6
10
to take
rmṯ (94530)
12
10
human being, man
ḥmsi (105780)
8
10
to sit
pꜢ-hrw (58940)
3
9
this day
bꜢk.w (53820)
4
8
work
spr (132830)
11
8
to reach
mjni (70060)
12
7
to arrive at, to reach
Table 1: Community structure around jri (hrw)
Tracing Concepts – Semantic Network Analysis
323
Tab. 1 shows an excerpt of the second largest community, which consists of 8.31% of the total nodes. The dominating word is jri (hrw) with a meaning of ‘to spend the day with sth.’ (WB I, 109.24 f.). Most of the following terms are verbs, which directly refer to a certain action, for instance, to take, to sit or to reach. Together with the terms work and man/human being this would suggest that this community shows the concept of hrw as moment or time of activity. Other words with a lower degree (< 8) are, for example, sšr with the meaning of ‘corn’ (TLA no. 851818; WB IV, 541.11–542.1) or wrš, another word translated with ‘to spend the day with sth./to be awake’ (TLA no. 48130; WB I, 335.10–18). Fig. 3 shows the community and its structure in detail. jri (hrw) is in a central position of the sub-network while all other community members are grouped around it. The words with the highest degree (see tab. 1) are the ones which hold the community together. In this way, this subgraph is rather coherent in its structure.
Fig. 3: Visualization of the community structure around jri (hrw)
324
Frederik Elwert / Simone Gerhards
Nodes
Count
Degree
Translation
nṯr (90260)
35
21
god
mri (72470)
16
16
to love
ꜤḥꜢ (39930)
15
16
battle
sꜤḥꜢ (129190)
6
11
to erect; to set up
nḥm (86430)
12
11
to take away, to rescue
mḥ (73290)
10
11
to fill
mšꜤ (76300)
7
11
army, workforce
sbtj (132260)
3
10
wall, rampart
mnfꜢ.t (70670)
2
10
infantry, soldiers
Table 2: Community structure around nṯr
The community listed in tab. 2 consists of 7.58% of all nodes. It is obvious that the words mentioned here (except for the first two) belong to a topic of fighting and battle. That means day in this community (and context) is characterized by war. At first sight, it seems that god and love do not fit into a war community, but by looking into the text sources (here it is inevitable to go back to close reading), one finds examples in which divine intervention is mentioned in a context of a battle.14 It is interesting that some words with a lower degree (< 8) refer to a very different context. psḏ ‘to shine’ (TLA no. 62420; WB I, 556.14–558.3), Ꜣḫ.t ‘horizon’ (TLA no. 227; WB I, 17.12–23), wbn ‘to shine/to rise’ (TLA no. 45050; WB I, 292.9–294.3) and Jtn ‘Aten’ (TLA no. 850317; WB I, 145.4–7) belong to the context of hrw as part of the solar circle. But this special connection of the words just occurs in one sentence from the ‘great hymn to the Aten’15 dating back to the time of king Akhenaton. They are linked to the community by their common connection with terms and entities like nṯr (TLA no. 90260; WB II, 358.1–360.14), RꜤw-Ḥr.w-Ꜣḫ.tj ‘Re-Herakhty’ (TLA no. 70002) and MꜢꜤ.t ‘Maat’ (TLA no. 66630; WB II, 20.10–13). Close reading reveals that this particular sub-group can be regarded as an artefact and should be excluded for interpretation. Otherwise, the community shows a thematically coherent conceptualization of day in the context of war/battle (see fig. 4).
14 For example pNew York MMA 35.9.21, col. 27.14–27.15. 15 The reference is from Tell el-Amarna, tomb 25 (Eje), entrance, west wall, PM IV, 228.
Tracing Concepts – Semantic Network Analysis
Fig. 4: Visualization of the community structure around nṯr
Nodes
Count
Degree
Translation
grḥ (167920)
60
34
night
RꜤw (400015)
38
26
Re
wnw.t (46420)
21
21
hour
ḥtp (111230)
14
14
to be pleased, to set (of sun)
ḥr (107510)
19
11
face
ꜤḥꜤ.yt (40380)
4
9
midday
tp-Ꜣbd (852477)
4
8
beginning of month
sḫm (142180)
2
7
might, power
psḏn.tjw (850569)
5
7
new-moon day
Table 3: Community structure around grḥ
325
326
Frederik Elwert / Simone Gerhards
The community consisting of 7.17% of all nodes is characterized by the word night (tab. 3). In addition, some group members like hour, midday, beginning of the month, etc. refer to time and date designations. The second predominating term is the divine entity Re. This gives a focus on the religious field of the connection between day and night. If you look at the network visualization (fig. 5), the direct link between grḥ and RꜤw is very thin. They are indirectly connected by the links with ḥtp ‘to be pleased’ (TLA no. 111230; WB III, 188.2–192.10), Ꜥnḫ.t ‘the west (lit. place of life)’ (TLA no. 38780; WB I, 205.16), and mkw.t ‘protection’ (TLA no. 76880; WB II, 160.22–161.4). As a result, the community can be interpreted as consisting of two aspects of hrw, its relation to other time specifications and its religious connotation. However, they are not strictly separable, but interrelated through shared references.
Fig. 5: Visualization of the community structure around grḥ
Nodes
Tracing Concepts – Semantic Network Analysis
Count
Degree
Translation
ꜤꜢ (450158)
24
18
great, large
sn (136230)
10
17
brother
wꜢḥ (43010)
12
14
to put
wnm (46710)
8
12
to eat
ḥm.t (104730)
4
12
wife
Ꜥš (40940)
6
11
fir tree
ꜤšꜢ (41011)
4
10
numerous
BꜢtꜢ (850439)
4
10
Bata
sḫ.t (141480)
10
9
field
4
9
valley
jn.t (26780)
327
Table 4: Community structure around ꜤꜢ
The next community (tab. 4) is one of the smallest and consists of only 6.44% of all nodes. Their relation can be ascribed to a specific context they have in common: they occur in sentences after expressions like ḫr jr m-ḫt hrw.w qn.w – After many days (sth. happens). In this context day has no extensive meaning, because it occurs not in its function of a literal day, but as a stylistic expression to mark that a new text section begins (see Erman 1933: 286; Junge 1999: 276–277). One can find this expression, for example, in The tale of the two brothers16 or The tale of the doomed prince17. This finding highlights the fact that especially smaller communities might not only stem from conceptual complexes, but also can express stylistic functions. Depending on the research question, they might or might not be relevant for further interpretation. In comparison to the general network constructed on the basis of all references, the one generated from the Old Kingdom texts will be presented below. Here just 84 unique references for hrw are to be found. Most of them stem from pyramid text spells and funerary tomb inscriptions, only a very few are from letters or administrative documents. This suggests that in the network religious and funeral aspects will be more visible than in the aforementioned network from all references. It consists of 305 nodes and 1225 edges. After applying the procedure of dropping unconnected nodes (see above), 209 nodes and 649 edges remain. Because it is considerably smaller than the one constructed from all references, the network is already legible as a whole (see fig. 6). The nodes with the highest degree are grḥ, pri, nṯr, Wsjr, Ḥrw, tp (TLA no. 170880) and RꜤw. Comparison with the full network shows that Wsjr
16 pBM EA 10183 (pD’Orbiney), recto, 2,7. 17 pHarris 500, verso, 7,5.
328
Frederik Elwert / Simone Gerhards
and Ḥrw did not play such a central role there. This supports the assumption that the communities will be characterized mostly by religious concepts due to the pyramid text spells. In the following, we will discuss one new community and the one around grḥ that was already present.
Fig. 6: Excerpt of the network constructed on the basis of 84 references of hrw from the Old Kingdom.
Node
Tracing Concepts – Semantic Network Analysis
Count
Degree
329
Translation
Wsjr (49460)
8
11
Ꜥ (34320)
4
9
arm
jp (24070)
2
8
to count, to asses
qs (162200)
3
8
bone
ḥꜢb (850109)
2
6
to fish
tm (172000)
1
6
to be complete
mw (69000)
1
6
water
sn (136230)
2
5
brother
twt (170470)
2
5
statue, image
smnḫ (135360)
2
4
to make distinguished
Osiris
Table 5: Old Kingdom community structure around Wsjr
This community around the divine entity Wsjr – the Egyptian god usually identified as the god of the underworld, afterlife and dead, but also as the god of regeneration and resurrection – was not visible in the full network. Wsjr connects all words and is the only link between two sub-structures (see fig. 7). This could explain why the community as a whole is hardly interpretable. The sub-community in the right upper corner (fig. 7) goes back to a spell of the pyramid texts (PT 536, §1297a–§1297c) in which Osiris and the words jp, qs, ḥꜢb, tm, mw, and sn occur (cf. tab. 5). Here, a day is characterized by the concepts of fishing, counting of bones, and making endure the soles (of the foot). Some of the links of the lower left corner of fig. 7 can be traced back to PT 505 §1089a–§190f in which the day of the landing of Osiris is mentioned. This refers to the day of Osiris’ death. This community highlights the importance of going back to the text in order to understand the origin of certain connections. Additionally, it shows how references to gods tend to influence the community structure. Generally speaking, a day in this community is conceptualized by (the death/dying of) Osiris. This is a result of the time and genre of the pyramid texts.
330
Frederik Elwert / Simone Gerhards
Fig. 7: Visualization of the Old Kingdom community structure around Wsjr
Nodes
Count
Degree
Translation
grḥ (167920)
11
16
night
ḥtp (111230)
4
11
to be pleased, to rest
t (168810)
2
7
bread
jꜢḫi (20870)
1
5
to be flooded
mjn.t (68370)
1
5
stagnant water (?)
Ꜣwi (50)
1
4
to be extended
šꜢ (151110)
1
4
marsh, meadow
Ꜣḫ.w (253)
1
4
(magical) power, mastery
hꜢi (97350)
2
4
to descend, to fall, to strike
zšp (144830)
1
4
to polish, to smooth
Table 6: Old Kingdom community structure around grḥ
Tracing Concepts – Semantic Network Analysis
331
This community (tab. 6.) is the largest one with 13.88% of all nodes. The term night is dominant, but in comparison to the community of night from the full network, only ḥtp is left. In the references from the Old Kingdom the term is not used with the meaning of to set (of the sun), but rather represents concepts from the pyramid text spells, in which is said that a night should be merciful (PT 44, §34b) or a night is satisfying (PT 406, §708a–708c). Looking at the counts of the words in the community, one can see that most nodes are just mentioned once, making it difficult to interpret the remaining words from the community (see fig. 8). This example demonstrates clearly that a certain number of references is urgently needed to obtain an interpretable result.
Fig. 8: Visualization of the Old Kingdom community structure around grḥ
332
Frederik Elwert / Simone Gerhards
In the following, the network constructed from the New Kingdom references (1550– 1070 BC) will be presented. There are 349 references in total – this actually represents half of all unique references of hrw in the TLA database. The network consists of 833 nodes and 4328 edges; after removing unconnected nodes, 581 nodes and 2574 edges remain. If you look at the network of the New Kingdom (fig. 9 shows an excerpt), one can see in comparison to the one from the Old Kingdom that the influence of Ḥrw and nṯr fades into the background, while new context fields like the ones around rnp.t and msi arise. Only grḥ remains one of the most central nodes in all networks. In the following, we will discuss the community around rnp.t and again the one around grḥ in the New Kingdom in order to highlight significant differences.
Fig. 9: Excerpt of the network constructed on the basis of 349 references of hrw from the Old Kingdom
Nodes
Tracing Concepts – Semantic Network Analysis
Count
Degree
333
Translation
rnp.t (94920)
20
24
year
ḫꜢb (103300)
9
22
festival
tꜢ (400096)
10
16
earth
msw.t (75070)
10
15
birth
Ꜣbd (93)
14
13
month
ḥtp.w (111260)
7
9
peace, contentment
pꜢq.t (59330)
3
9
fine linen
rꜢ (92560)
2
8
mouth
ḫꜢs.t (114300)
8
7
foreign land, desert
Ꜣ.t (5)
3
7
moment, instant
Table 7: New Kingdom community structure around rnp.t
Tab. 7 shows the largest community, consisting of 13.79% of all nodes. It extends around the words rnp.t ‘year’ and ḥꜢb ‘festival’ and characterizes hrw as a feast day, when earth is at peace and festivals take place. This community, too, exists in the network constructed on the basis of all references (where it consists of 7.68% of all nodes). Since the New Kingdom references account for a large share of all references, it is plausible that basic structures that can be found in this sub-network also show in the full network. However, it is absent in the Old Kingdom network.
Fig. 10: Visualization of the New Kingdom community structure around rnp.t
334
Frederik Elwert / Simone Gerhards
Nodes
Count
Degree
Translation
grḥ (167920)
23
24
night
jwi (21930)
23
14
to come
RꜤw (400015)
13
14
Re
ḥr (107510)
10
13
face
nw (80840)
6
7
moment
Psḏ.t (62500)
4
7
divine ennead
Ḏḥw.tj (185290)
2
7
Thot
Table 8: New Kingdom community structure around grḥ
Tab. 8 shows the community that is built around night. In comparison to the Old Kingdom network, the link to the god Re becomes visible and at the same time central. Therefore, it is present in the full network as well. Very interesting is that the relation to ḥtp does not exist any more. Nevertheless, a new node jwi ‘to come’ (TLA no. 21930, WB I, 44.6–45.6) appears (cf. fig. 11). Although jwi has the same count as grḥ its degree is much lower. Because hrw and grḥ are very close to each other in the texts, the weight between the nodes becomes higher. By comparing all three sub-networks (fig. 5, fig. 8 and fig. 11) in which grḥ is the central node, the different composition of each becomes visible. This result indicates shifts in the conceptualization of the relation between day and night and can serve as a starting point for further analysis.
Fig. 11: Visualization of the New Kingdom community structure around grḥ
Tracing Concepts – Semantic Network Analysis
335
6 Conclusions: why network analysis for tracing concepts and categorizations? The exemplary analyses of hrw in the TLA corpus show how semantic network analysis enables new approaches in the study of textual corpora. The representation of word co-occurrences as a network and computational analysis of this network allows the detection of semantic structures. In this way, it is possible to see at a glance indicators for different conceptualizations, like the example above showed: day connected to feasts, war and certain activities; day connected to entities like Osiris or Re; day as a stylistic expression to mark the beginning of a new text section, and the most frequent: day as part of a categorization pattern with other time specifications like night, midday, month or year. From an emic perspective, an ancient Egyptian day was closely connected to the solar cycle and the resulting change between light and dark (Hornung 1956: 26 f.). Thus, the expression hrw means, roughly speaking, either a whole day (corresponds approximately to 24 hours) or a time unit for a literal day (corresponds to the time when sun is shining), also referred to as daytime as the direct opposite of night (see Hornung 1956: 26 f.; Spalinger 1992: 147; Hoffmann 1996: 184; Westendorf 2003: 89–91). Through diachronic comparison, the relation to night becomes visible over the time. While this in itself is neither new nor surprising, semantic network analysis allows changes to be traced in the connotations of this relation through different time periods. The networks of themselves should not be seen as a final result or as an answer to research questions, but as a tool to get from distant to close reading. Analysis does not stop on the level of the network. Interesting relationships between words, visible by distant reading, can be reconstructed through close reading. Because each edge in the graph can be traced back to specific source passages in the text, the network can be used as an alternative index. Instead of just listing words, it also allows word cooccurrences to be looked up. This kind of relationality can be seen as the main advantage. It shows words in their context, allowing relations to be studied, instead of isolated lexical items. In addition, the visualization of these relations allows relevant structures to be spotted more easily than traditional methods of corpus analysis do. At the same time, there are also limitations. The method unfolds its full potential only if a sufficient number of references are available in the corpus. Only then do structural patterns begin to emerge that go beyond individual texts. Moreover, computational methods tend to make their results seem objective and unbiased. It requires a certain level of familiarity with the underlying methods to understand the implicit assumptions hidden in the algorithms and potential sources of bias. We thus see semantic network analysis as a heuristic device for tracing concepts. However, it should not be used without going back to the original source to understand the origin of a certain word connection in the corpus.
336
Frederik Elwert / Simone Gerhards
References Atteveldt, Wouter van. 2008. Semantic Network Analysis: Techniques for Extracting, Representing and Querying Media Content. Amsterdam: Vrije Universiteit. http://vanatteveldt.com/p/ vanatteveldt_semanticnetworkanalysis.pdf. [Accessed 1/2017] Bastian, Mathieu, Sebastien Heymann, and Mathieu Jacomy. 2009. “Gephi: An Open Source Software for Exploring and Manipulating Networks.” In International AAAI Conference on Weblogs and Social Media, 361 f. ICWSM. The AAAI Press. http://aaai.org/ocs/index.php/ ICWSM/09/paper/view/154. [Accessed 1/2017] Biemann, Chris. 2006. “Chinese Whispers: An Efficient Graph Clustering Algorithm and Its Application to Natural Language Processing Problems.” In Proceedings of the First Workshop on Graph Based Methods for Natural Language Processing, 73–80. http://dl.acm.org/citation. cfm?id=1654774. [Accessed 1/2017] Blondel, Vincent D., Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre. 2008. “Fast Unfolding of Communities in Large Networks.” Journal of Statistical Mechanics: Theory and Experiment 2008 (10). doi:10.1088/1742-5468/2008/10/P10008. Borgatti, Stephen P., and Martin G. Everett. 2006. “A Graph-Theoretic Perspective on Centrality.” Social Networks 28 (4): 466–84. doi:10.1016/j.socnet.2005.11.005. Bubenhofer, Noah. 2009. Sprachgebrauchsmuster: Korpuslinguistik als Methode der Diskurs- und Kulturanalyse. Berlin: de Gruyter. Bullinaria, John A., and Joseph P. Levy. 2007. “Extracting Semantic Representations from Word Co-Occurrence Statistics: A Computational Study.” Behavior Research Methods 39 (3): 510–26. doi:10.3758/BF03193020. Dorow, Beate, and Dominic Widdows. 2003. “Discovering Corpus-Specific Word Senses.” In Proceedings of the Tenth Conference on European Chapter of the Association for Computational Linguistics - Volume 2, 79–82. EACL ’03. Stroudsburg, PA, USA: Association for Computational Linguistics. doi:10.3115/1067737.1067753. Elwert, Frederik. 2014. TCFnetworks. https://github.com/SeNeReKo/TCFnetworks (23.01.2017). Elwert, Frederik. 2016. “Network Analysis between Distant Reading and Close Reading.” In DHLU 2013. Proceedings of the Third Conference on Digital Humanities in Luxembourg with a Special Focus on Reading Historical Sources in the Digital Age, edited by Lars Wieneke, Catherine Jones, Marten Düring, Florentina Armaselu, and René Leboutte. CEUR Workshop Proceedings 1681. Luxembourg: University of Luxembourg. http://ceur-ws.org/Vol-1681/Elwert_Network_ analysis.pdf. [Accessed 1/2017] Erman, Adolf. 21933. Neuägyptische Grammatik. Leipzig: Wilhelm von Engelmann. Freeman, Linton C. 1978. “Centrality in Social Networks Conceptual Clarification.” Social Networks 1 (3): 215–39. doi:10.1016/0378-8733(78)90021-7. Ganesan, Kavita. 2014. “All About Stop Words for Text Mining and Information Retrieval.” Text Mining, Analytics & More. http://www.text-analytics101.com/2014/10/all-about-stop-wordsfor-text-mining.html. [Accessed 1/2017] Hoffmann, Friedhelm. 1996. “Der Literarische Demotische Papyrus Wien D6920–22.” Studien zur Altägyptischen Kultur 23: 167–200. Hofmann, Beate, and Frederik Elwert. 2014. “Heka und Maat. Netzwerkanalyse als Instrument ägyptologischer Bedeutungsanalyse.” In ‘Vom Leben Umfangen.’ Ägypten, das Alte Testament und das Gespräch der Religionen. Gedenkschrift für Manfred Görg, edited by Georg Gafus and Stefan Wimmer. (Ägypten und Altes Testament 80, 235–245). Wiesbaden: Harrassowitz. Hornung, Erik. 1956. Nacht und Finsternis im Weltbild der Alten Ägypter. Tübingen: Universität Tübingen. Junge, Friedrich. 21999. Neuägyptisch. Einführung in die Grammatik. Wiesbaden: Harrassowitz.
Tracing Concepts – Semantic Network Analysis
337
Knopp, Johannes, Johanna Völker, and Simone Paolo Ponzetto. 2013. “Topic Modeling for Word Sense Induction.” In Language Processing and Knowledge in the Web, edited by Iryna Gurevych, Chris Biemann, and Torsten Zesch, 97–103. (Lecture Notes in Computer Science 8105). Berlin and Heidelberg: Springer. http://link.springer.com/chapter/10.1007/978-3-64240722-2_10. [Accessed 1/2017] Lietz, Haiko. 2007. “Mit neuen Methoden zu neuen Aussagen: Semantische Netzwerkanalyse am Beispiel der Europäischen Verfassung.” http://www.haikolietz.de/docs/verfassung.pdf. [Accessed 1/2017] Manning, Christopher D., and Hinrich Schütze. 1999. Foundations of Statistical Natural Language Processing. Cambridge (Mass.)/London: MIT press. Melville, Herman. 1851. Moby-Dick, Or, the Whale. New York: Harper & Brothers; London: Richard Bentley. Moretti, Franco. 2013. Distant Reading. London: Verso. Paranyushkin, Dmitry. 2011. Identifying the Pathways for Meaning Circulation Using Text Network Analysis. Berlin: Nodus Labs. http://noduslabs.com/research/pathways-meaning-circulationtext-network-analysis. [Accessed 1/2017] PM = Porter, Bertha, Moss, Rosalind L. B. 1960–1981. Topographical Bibliography of Ancient Egyptian Hieroglyphic Texts, Reliefs and Paintings, 7 vols. Oxford: Oxford University Press. Rieger, Burghard B. 1988. “Bedeutungsanalyse und Dispositionsstrukturen. Zum Problem einer empirischen Komponente der Situationssemantik.” In Angewandte Linguistik und Computer. Kongreßbeiträge zur 18. Jahrestagung der Gesellschaft für Angewandte Linguistik (GAL), 44–47. (Forum Angewandte Linguistik, Band 16). Tübingen: Narr. Rieger, Burghard B. 1989. Unscharfe Semantik: Die Empirische Analyse, Quantitative Beschreibung, Formale Repräsentation Und Prozedurale Modellierung Vager Wortbedeutungen in Texten, edited by Bernd Spillner. Frankfurt am Main u. a.: Lang. Snijders, Chris, Uwe Matzat, and Ulf-Dietrich Reips. 2012. “Big Data: Big Gaps of Knowledge in the Field of Internet Science.” International Journal of Internet Science 7 (1): 1–5. http://iscience. deusto.es/wp-content/uploads/2012/08/ijis7_1_editorial.pdf. [Accessed 1/2017] Spalinger, Anthony J. 1992. “Night into Day.” Zeitschrift für Ägyptische Sprache und Altertumskunde 119 (2): 144–157. Strauss, Anselm L., and Juliet M. Corbin. 1996. Grounded Theory: Grundlagen qualitativer Sozialforschung. Weinheim: Psychologie Verlags Union. Teubert, Wolfgang. 2005. “My Version of Corpus Linguistics.” International Journal of Corpus Linguistics 10 (1): 1–13. doi:10.1075/ijcl.10.1.01teu. WB = Erman, Adolf, and Hermann Grapow. 1926–1963. Wörterbuch der Ägyptischen Sprache, 5 vols. Berlin: Akademie-Verlag. Westendorf, Wolfhart. 2003. “‘Der kleine oder vollendete Tag’ im Ägyptischen Weltreich.” Göttinger Miszellen 193: 87–94. Wittgenstein, Ludwig. (1953) 2008. Philosophische Untersuchungen. Edited by Joachim Schulte. Frankfurt am Main: Suhrkamp.
Figures Fig. 1–11: Frederik Elwert / Simone Gerhards
Index Page numbers in italics refer to notes on that page.
I. Subject1 acculturation lexical ~ 12 f., 45–65 linguistic ~ 57 f. action 21, 81, 103 f., 108–110, 125, 129, 143, 239 f. action-orientation 133 actant 297, 299 action-chain 79, 296 f. action-frame 301–303 actor 11, 79, 103 acts 40, 162, 244, 283–286, 313 human bodily ~ 246 inter~, s. interaction limitation of the ~ 131 modal ~ pattern 240 modes of ~ 145 therapeutic(al) ~ (operationes) 136–138, 152, 155 type of ~ 168 age, s. criteria of categorization/classification agency 14, 239–260 agent 26, 53, 58, 210, 239–241, 256, 258, 264, 289 f., 296, algorithm 233, 311–335 alphabet, s. criteria of categorization/classification analogy 15, 26, 184, 251, 253, 266, 293 f., 293–294, 298–301 animacy 206 f., 217, 221, 239–260, 269 animals antelope 61 f., 80–81 bat 56, 81 f., 96, 98 beast 65, 79, 86, 95, 99, 261 bird 4, 9, 41, 60, 62, 69–71, 74 f., 78–86, 88, 90 f., 93, 95 f., 171, 174, 176, 250, 252, 261, 266, 291 bovine 48 f., 51, 54, 60 bull 178, 180 butterfly 78 1
1 Compiled by Peggy Zogbaum. DOI 10.1515/9783110538779–013
camel 56 carnivore 61 f. cat 57, 61 f., 77, 178, 252 catfish 69, 77 chicken 57, 265 chimpanzee 276 crocodile 62, 69, 78, 83, 175, 188 dog 41, 59, 61 f., 69, 73 f., 223, 239–241, 243, 253, 259 f., 290, 296 donkey 48 f., 55–56, 59 f., 62, 69, 84 f., 177 dove 69, 83 feline 80, 86 fish 41, 60, 62, 69–71, 75, 77–81, 83–84, 86, 88, 90, 93, 174–176, 250, 266 gazelle 69, 80 f., 84–85 giraffe 64, 69, 96, 194 goose 69, 71, 182 hedgehog 78 f., 81, 97 hippo 62 horse 12, 45–60, 62, 84–85 insect 41, 69, 79, 83, 98, 250, 256 lion 61 f., 69, 81, 178, 273 mammal 9, 41, 61, 82 monkey 61, 69 f. mosquito 189 ox 72, 84 panther 61 perch 71 pig 62, 69, 242 porcupine 79 seaworms 250 sheep 56, 60, 69, 84 f., 240 snake 9, 37, 61 f., 69, 72 f., 83–85, 91, 189, 191, 206 f., 217 animate 200 f., 213, 221, 239–260 anthropology 202, 229, 239–241, 247, 256, 263–275 arrangement (s. also criteria of categorization/ classification) 86, 112, 122, 184, 200 f., 231 f., 251, 293, 317 ~ of categories 273 ~ of drugs 13, 169, 192
340
Index
~ of frames 268, 289 ~ of knowledge 290 ~ of prescriptions/formulae 13, 167–169, 184, 190, 192 attribute 36, 39 f., 69, 72, 101, 153, 177–180, 192, 195, 239–244, 247, 251–254, 256–259, 264–266 ~ of frames (slots) 288–292, 294 f., 297, 299, 301 biology/ethnobiology (s. also criteria of categorization/classification) 1 f., 6, 12, 14, 21, 223, 231–236, 239–260, 263–275 body parts 103, 110, 115, 123, 125, 131, 137, 153, 171, 175 f., 243 botany, s. criteria of categorization/classification canonization 49, 74, 67, 86, 94 f., 215, 312 f. categorization/classification (s. also criteria of categorization/classification) anthropological ~14, 239–241, 258 basic level/top-level ~170, 176, 268, 298 ~ in context of interaction 78 f., 81, 144, 184, 201, 221–223, 247, 256–259, 264, 313 culture (ancient/past) 1 f., 6, 8–12, 45–63, 75, 104, 114, 120, 136, 156, 158, 181 culture (modern/contemporary) 5–7, 14, 22, 199–227, 248–255, 263–275 flexibility of ~ 1, 5 f., 171 motivation of ~ 5, 13, 167, 171, 178, 181, 191, 216, 220, 226 subordinate level ~/subcategory 8, 116, 167, 170, 200–202, 289 f. superordinate level ~ 170, 179, 176, 225 classification, s. categorization classifier (s. also determinative) 9 f., 13, 45–63, 69 f., 72, 83, 167–193, 232, 265, types of ~ 199–227 relational ~ 199–227 cognition 1–5, 15, 47, 57, 60 f., 168, 199 f., 216, 218, 220 f., 239–241, 247, 252–256, 258, 263–267, 269, 272 f., 281, 283–289, 293 f., 297–300, 305 cognitive experiments 6 cognitive foundation 200, 221, 252 cognitive organization 218 cognitive-semantic properties 220 communicate/communication 2, 68, 86, 95, 242, 258, 263, 283–286
concept(s) 24, 31, 58 f., 61–63, 86, 94, 105, 216–219, 268 f., 281–306, 311–335 agreement of ~ 192, 199 f., 203 f., 215–216, 219–221, 224, 226 fuzzy ~ 168, 171, 239, 241, 247, 258, 265, 269, 274 ~ of animals 67–95 ~ of body 171, 184 ~ of concept 281–283 meta-~ 167 f., 172, 190 ~ of drugs 13, 167–169, 174, 177–181, 183 f., 192 ~ of healing/medical ~ 171, 192, 202 ~ of nature 39, 247, 329 ~ of plants 3, 224 conception/conceptualization 48, 168, 171, 199 f., 223–227, 252, 263 f., 281–306, 311–335 ~of therapeutic qualities 116, 124, 133 taxonomic ~, s. taxonomy cosmogony/cosmology 23 f., 27, 29, 34, 40, 270 criteria of categorization/classification (s. also hierarchy) 2–14 age 33–35, 69, 178, 222 alphabet 13, 87, 101–158 arrangement 172, 184, 190, 200 f., 231 f., 251, 317 behavioural properties biological/ethnobiological ~ 6, 202, 223, 249, 251–253, 263–275 botanical/ethnobotancial ~ 40, 175, 231, 235, 263–275 chronological 87, 89, 93, 128, 303 colour 26, 36, 70 f., 73 f., 125, 130, 143, 148, 151–155, 170, 173, 178, 181, 202, 231, 255, 269, 292, 299 consistency 170, 172 f., 200, 211, 221 dimension 170, 172, 201, 206 f., 217, 221, 224 effect 105, 110, 131, 134, 178 Galenic ~ 102 gender 55, 86, 169, 178, 199 f., 202 f., 205, 214–216, 218–222, 226 genetics 6, 12, 14, 231–236, 255, 258, 268–271 habitus 9, 77, 174 location 26, 34, 38, 40, 111, 116, 174, 189, 200 f., 206 f., 221, 269 material 22, 24, 151, 170, 173, 200 f., 223, 225, 267, 286, 292, 298 f.
Index
morphology 13, 49, 61, 172–174, 203, 228 f., 215,234–236, 254–258 name 105 naturalistic ~ 102, 134 origin 10, 116 f., 123, 126, 132 f., 138, 175, 178, 181, 191 f. processing 178, 192, 216 quantum 200 quality, 25 f., 38, 61, 103, 105, 123, 129, 133 f., 141, 144, 146, 149–154, 156, 178, 255, 257 relevance 101 shape 9, 78, 108, 170, 172–174, 200 f., 206 f., 211, 217, 221, 255 similarity 4, 6, 10–14, 25–27, 32–34, 38–41, 51, 56, 59 f., 104, 109, 117, 119, 128, 138, 152, 155, 171, 182–184, 214 f., 231–235, 239, 250, 252 f., 270, 296, 302, 304 size 10, 28, 62, 69, 71, 78, 118, 200 f., 204, 206 f., 221, 292, 299 smell 125, 130, 143, 148, 150 f., 173, 178, 188 f., 254, 305 source 103, 112–114, 167 f., 174, 192, 246, 303 taste 125, 130, 148, 150 f., 178 definiteness 171, 199, 205, 213, 216, 218–220, 226, 285 dependency (s. also independency; interdependency) 168, 170 f., 304 ~of context 1, 4, 8–14, 36, 86, 95, 295 ~ of concept 4 ~ of time 321 determinative (s. also classifier) 3, 6, 8 f., 13 f., 67–70, 77, 86, 91 f., 95, 101, 110–117, 124, 146, 149 f., 152, 155, 167–193, 203, 211, 218, 221, 224, 231, 264, 283, 285, 290 f., 299 dimension, s. criteria of categorization/classification disease symptoms 117, 171, 174, 175 f., 191 distinctiveness 249, 255, 266, 288 drug terminology 174, 177 episteme 127, 252, 285–290, 292–295, 298–301, 306 ethnobiology, s. biology (s. also criteria of categorization/classification) ethnobotany, s. botany (s. also criteria of categorization/classification) etymology 29, 41, 115
341
frame ~ -semantic, s. methods ~-to-frame, s. methods genera 40 f., 95, 231–234, 242, 251, 267–269 generic category 60, 62 generic name 193 generic term 62, 72, 85, 317 group (s. also categorization) 1, 4–6, 21, 32–38, 41 f., 67–95, 101–158, 170 f., 181 f., 184–192, 201, 231–236, 248, 250, 255, 265, 267–269, 288, 292, 301 f., 315, 321, 323, 326 sub-group (s. also subordinate level categorization) 32, 42, 234, 324 hermeneutics 21, 67, 83, 118, 179, 205, 266, 282, 286, 311–335 hierarchy 6, 10, 32, 61, 67, 86–89, 93–95, 135, 140, 150, 175, 184, 201, 235–236, 252, 264, 266 f., 295 f., 298 f., 302 hyperonym 202, 296, 298 icon 47, 54, 58, 61, 270 iconography 13, 68–84 independency 37, 301, 117, 121, 127, 181, 224, 246, 268, 301 ~ of agent 240–242, 256 ~ of context 4, 6, 12, 14 ~ of language 283 f. individuation 200, 216 f., 226, 229 interaction 15, 53, 78 f., 81, 144, 184, 199–227, 240 f., 247, 256–259, 264, 292 functional/physical ~ 201, 221 social ~ 200 f., 221 f., 227, 264, 313 interdependency 8, 314 inventory 74, 200, 220 knowledge 3, 5 f., 8, 11, 13, 68, 75, 79, 95, 102, 114, 122, 146, 148, 150–153, 156, 179, 181–184, 189 f., 192, 253 f., 264 f., 268, 270, 272 f., 275, 281–306 acquisition of ~ 101 f., 150, 152, 156, 282 elements of ~ 15, 168, 289 f., 299, 302 concept(s) of ~ 15, 281, 304 organisation of ~ 60, 80, 265, 270, 273, 282, 285, 288, 292 f., 301 theoretical ~ 136, 138, 190 languages Akkadian 53 f. Ancient Egyptian 45–63, 67–195, 167–193, 311–335 Aztec 55, 57 Bantu 203–205
342
Index
Chinese 56, 206–208, 216–218 elements from ~ 4, 304 Fiji 209 f. French 47, 203 German 7, 11, 22, 47, 203, 283, 294, 304, 312, 316 Hmong 213 Jacaltec 205 f., 221–223 Miraña 224 Nahuatl 12, 45–63 Palikur 214 Panare 211 f. Spanish 57 f. Sumerian 12 f., 45–63 Swahili 203–205, 215, 220 Tariana 214 f. Thai 206 f., 224 f. Vietnamese 225 f. lexicography 90, 249, 266 list 29–31, 36 f., 47, 55, 67–95, 101–158, 167–193, 208, 264, 311–335 medical texts 171, 175 f., 179, 192, 273 f., 316 Egyptian ~ Papyrus Ebers 169 f., 182, 185, 190 f. Papyrus Berlin 3038 182, 190, 192 medieval ~ De materia medica (DMM) 107–112 De medicinis simplicibus (DMS) 127, 132–135, 162 De probanda medicina (DPM) 121, 122, 124, 126–128, 130–135, 143, 147, 148 De simplici medicina (DSM) 127, 132–135, 162 De simplicibus ad Paternianum (SAP) 116–118 De simplicium medicamentorum temperamentis et facultatibus (DSF) 07, 108, 112 Herbarius attributed to Apuleius Platonicus (PAH) 113–116 Liber canonis (LC) 124, 125, 135–136, 138–149, 156 Liber de gradibus (LDG) 121–129, 133–135 Lorscher Arzneibuch 107, 160 Macer floridus (MF) 114–116 Pantegni Practica (PP) 122, 124, 127, 130, 135, 148, 156 Tabulae Salerni (TS) 132, 133, 135
medicine (s. also arrangement) 67, 101–158, 167–193 memory 4, 218, 220, 235, 254, 270–273, 287, 289, 300, 305 cultural ~ 11 metaphor 23, 48, 75, 183, 208, 256, 286, 288, 293, 301, 305, 314 methods 1, 6 f., 11, 15, 248, 298, 311, 313, 316, 335 frame–semantic 1, 281–306 (automated) network analysis 1, 12, 15, 311–335 modularity 253, 269 naming (patterns) 14, 37, 19, 88, 169, 218, 231–236, 255, 298 onomastica, 68–72, 74 f., 80 f. ontology 15, 241, 246 f., 250, 252, 258 f., 281, 293 f., 296, 300 pair 2, 26 f., 31, 34, 48, 81 f., 93, 117, 241, 258 paradigm 57, 117, 135, 142, 284 plants 8–10, 16, 39, 102 f., 115, 123, 128, 137 f., 155, 174 f., 178, 192, 201, 222 f., 225 f., 231, 235, 237, 249, 253, 255, 265, 269 f., 273, 276 bamboo 242 f. banana 224 barley 177 basil 247, 260, 308 bean 69 figs 9, 10 flora 248, 250 flower 9, 78, 126, 133, 201, 207, 224, 231–236, 253 fruit 9 f., 31, 81, 126, 133 f., 172, 174 f., 178, 182, 212, 224 f., 233–234, 244 garden 201, 212, 231, 264 grain 126, 172–176 herbs 102, 126, 133 f., 172, 176 mushroom 255 oak 3, 201 palm 9, 245, 268 papaver/poppy 233–237 pine 201 rose 138, 224 f. seaweed 248, 255 tobacco 231, 244, 264 tree 3 f., 6, 9 f., 29, 31, 89, 91, 170, 172, 174–176, 201, 204, 207, 225 f., 233, 236, 250, 255, 268, 274 f., 327 wheat 223
Index prototype (s. also prototype semantics) 1, 4, 8, 10, 13, 49, 56, 59, 61, 77, 168, 171 f., 176, 179, 191, 203, 215, 243, 255, 266 f., 291, 301, 303 qualifier 255, 266 quantifier 227 relation 12, 21–42, 50, 62, 113, 115, 149, 156 f., 172–174, 183, 199–227, 231–236, 250 f., 263–275, 281–306, 311–335 ~al classifier, s. classifier epistemic relation 289 meronymic ~ 170, 172–174 reciprocal ~ 8 schematic ~ 170, 172 f. semantic ~ 7, 223, 318 taxonomic ~ (s. also taxonomy) 170, 172–174, 176, 223 thematic ~ 170, 174 repeater 54, 61 f., 211 f. scheme/schematization 5, 27, 71, 88 f., 111, 116, 125, 170, 172–174, 182, 248, 250, 258, 264, 267 f., 275, 285–287, 291, 300, 311, 316–318, 320 semantics 8, 13, 48, 52, 93, 168, 170, 199–227, 247 f., 254–256, 264 f., 269, 281–306, 311–335 frame ~, s. methods prototype ~ 1, 8, 10, 171, 191 ~ network analysis, s. methods
343
sign (s. also symbol) 3, 7 f., 11, 47, 52, 70–72, 74, 81, 83–85, 89, 91 f., 150 f., 170, 172–174, 281, 283–287, 292, 298 signifier 54 similarity, s. criteria of categorization/classification structuralism 270 symbols/symbolism (s. also sign) 2–5, 7 f., 51, 170, 183, 252, 256, 258, 270, 274, 286 synonym 25, 31, 34 f., 40 f., 90, 105, 115, 283 f. syntactic unit 220 f. taxonomy (s. also taxonomic relation) 5 f., 14 f., 61, 67, 71, 82, 85, 170–176, 181, 193, 200–202, 223–225, 231 f., 239, 241, 247–260, 263–268, 293–302, 313 theories 16, 179, 287–289, 293 f. cultural interference theory 53 frame-theory 281–306 prototype theory, s. prototype semantics schema theory 286, 294 semiotics 1–4, 7, 10 f., 47, 54, 313 token 282, 289–291, 293, 298 toponym 316 universals/universality 1, 10, 14 f., 145, 152–155, 199–201, 221, 227, 248, 251 f., 255, 257, 263, 267–269, 282 visualization 300, 311–335 writing system 13, 56, 169–171, 175, 264, 318, hieratic 47–50, 52, 61 f., 171, 175 hieroglyphic 50, 53 f., 62, 67, 69–71, 74, 81–83, 92, 171, 175, 316
II. Author Abelson, Robert P. 287, 303, 304, 308 Adamson, Peter 136, 160 Adanson, Michel 231, 237 Ahrens, Kathleen 207, 217, 227 Aikhenvald, Alexandra 199, 202, 205 f., 208–211, 212, 213 f., 227 Al-Ayedi, Abdul Rahman 84, 95 Alderotti, Taddeo 144, 164 Allaby, Michael 9, 16 Allan, Keith 200, 205, 221, 227 Allen, James P. 169, 194, 200, 227 Althoff, Jochen 12, 21, 342 Apuleius Platonicus (s. also Herbarius attributed to Apuleius Platonicus) 113–116
Aristotle (Aristoteles) 2, 4, 12, 21, 22, 24, 41, 42, 43, 68, 102, 136, 144, 251 Arnaldus de Villanova 16, 141 f., 158 Arnold, Dorothea 75, 95 Assmann, Aleida 10 f., 16 Assmann, Jan 82 f., 85–91, 93, 95 Astuti, Rita 253, 260 Atran, Scott 239 f., 251 f., 254, 260, 268, 270, 275 Atteveldt, Wouter van 311 f., 336 Aufrère, Sydney 78, 95 Ausécache, Mireille 132, 160 Auslasser, Vitus 115, 159
344
Index
Avicenna (s. also Liber canonis) 102–104, 111, 113, 120, 124 f., 135–156, 159, 160, 162, 164, 165. Ayari-Lassueur, Sylvie 136, 160 Bakir, Abd el Mohsen 87, 89, 91 f., 95 Baldwin, Bruce G. 235, 237 Bally, Charles 209, 227 Barsalou, Lawrence 5, 16, 266, 275, 287 f., 290–295, 297 f., 303, 305 f., 307, 309 Bartlett, Lawrence W. 286 f., 294 f., 307 Bartholomaeus Salernitanus 101, 130–132, 166 Bastian, Mathieu 320, 336 Bechtel, Friedrich 37, 42 Beekes, Robert 41, 42 Bénézet, Jean-Pierre 150, 160 Bentrovato S. 218, 227 Berlin, Brent 1, 16, 200 f., 227, 247, 249, 251, 258, 260, 265, 267 f., 275, 276 Bernardus Provincialis 132, 159 Biemann, Chris 314, 336 f. Bietak, Manfred 45, 63, 65 Bisang, Walter 1, 4, 14, 16, 199 f., 206, 213, 215–220, 224, 226, 227, 229, 342 Black, Mary 249, 250, 260 Bloch, Maurice 247, 257, 260 Blondel, Vincent 315, 336 Blümer, Wilhelm 21 f., 42 Boessneck, Joachim 45, 63 Borgatti, Stephen 315, 336 Bottiglieri, Corinna 104, 160 Boyer, Pascal 252, 260 Boys-Stones, George R. 22, 42 Bracken, Louise J. 254, 260 Breedlove, Denise A. 200 f., 227, 275, 276. Bremer, Dieter 22, 42 Bresnan, Joan 204, 227 Briggs, Davis 231, 237 Brovarski, Edward 71 f., 80, 95, 98 Brown, Cecil H. 13, 16, 45, 47, 51, 56 f., 63, 64, 251, 260, 262 Brunner-Traut, Emma 81, 95 Bubenhofer, Noah 313 f., 318, 336 Budge, Ernest A. Wallis 87, 96 Bullinaria, John A. 318, 336 Burkard, Günther 68, 96 Burnett, Charles 121, 136, 160, 161, 162, 163, 166 Busse, Dietrich 4 f., 15, 281, 283, 285–288, 294, 306, 307, 342 Cantino, Philip D. 236 f.
Carey, Susan 252, 260 Carlson, Robert 211, 228 Carolan, James C. 234, 237 Carrithers, Michael 254, 256, 260 Casson, Ronald W. 53, 63 Chandelier, Joël 136, 143, 160, 161 Chappell, Hilary 209, 228 Chen, Lai-Shen Lisa 217, 228 Chierchia, Gennaro 217, 226, 228 Conklin, Harold C. 265 f., 276 Constantinus Africanus (s. also De probanda medicina; Liber de gradibus; Pantegni Practica) 101, 111–113, 120–125, 127–135, 143, 147 f., 156, 159, 160, 162, 163, 164, 165, 166 Cooper, Alix 250, 260 Coppinger, Raymond 239, 260 Corbett, Greville G. 199, 203, 215, 228 Corbin, Juliet M. 313, 337 Craig, Colette 205, 221–223, 228 Croft, William A. 201 f., 228 Cronier, Marie 112, 161 Crum, Walter E. 83, 96 Cruse, Alan 2 f., 16, 201 f., 228 Dahan, D. 218, 228 Darwin, Charles 12, 21, 42, 232, 237 Davies, W. Vivian 54, 63, 65, 81 f., 96 Davis, Peter H. 231 f., 237 de Queiroz, Kevin 236 f. De Renzi 130, 132, 159, 160 de Vaan, Michiel 41, 42 DeVos, Paula 105, 161 Debru, Armelle 107, 125, 149, 161, 162, 164 Decker, Wolfgang 79, 81, 96 Deines, Hildegard von (GdM VI) 172, 175, 180, 194 Demuth, Katherine 204, 228 Denny, John P. 60, 63, 200 f., 221, 227, 228 Descola, Philippe 241, 259 f. Diller, Hans 22, 42 Dioscorides Pedanius (s. also De materia medica) 13, 101 f., 107–118, 159, 161, 164, 175 Dixon, R. M. W. 209, 228 Dorow, Beate 314, 320, 336 Douglas, Mary 248, 265 Droste zu Hülshoff, Vera von 78 f., 96 Eco, Umberto 16, 47, 57, 63 Edel, Elmar 49, 63, 71, 80, 96 Egmond, Florike 128, 161
Index Ellen, Roy 5, 11, 14, 16, 202, 239, 242–244, 246, 252–256, 258, 260, 261, 262, 268, 274, 276, 342 Elwert, Frederik 15, 311 f., 314, 320, 336, 337, 342 Emery, Stephen 254, 260 Erbar, Claudia 237 f. Erichsen, Wolja 83, 96 Erman, Adolf 81, 83, 85, 90–92, 96, 98, 195, 327, 336, 337 Erroux-Morfin, Marguerite 78, 95 Even-Zohar, Itamar 53, 63 Everett, Martin G. 315, 336 Everett, Nicolas 116–118, 158 Faulkner, Raymond 48, 51, 62, 63 Fedden, Sebastian 215, 228 Felsenstein, Joseph 233, 237 Ferraces Rodriguez, Arsenio 113, 161 Fillmore, Charles John 5, 16, 286–288, 291 f., 294–297, 300 f., 303, 305, 306, 307, 308, 309 Fischer-Elfert, Hans-Werner 68, 74, 96 Fischer, Klaus-Dietrich 113, 116, 161, 276 Flashar, Hellmuth 22, 42 Forth, Gregory 271, 276 Fortuna, Stephania 130, 162 Frake, Charles 251, 261 Frati, Lodovico 138, 162 Freeman, Linton C. 237, 315, 336 French, Roger 143, 162 Friederici, Angela D. 218, 228 Friedrichsen, George W. 256, 261 Frippiat, Hélène 96, 132, 162 Galenus (s. also De simplicium medicamentorum temperamentis et facultatibus) 13, 101–103, 106–118, 123, 125, 128 f., 133–135, 139, 142, 144, 149, 155, 157, 158, 159, 160, 161, 162, 164, 166, 175 Gamer-Wallert, Ingrid 77, 96 Ganesan, Kavita 317, 336 Garbo, Dinus de (Dino del Garbo) 143 f., 159 Gardiner, Alan Henderson 47, 52, 63, 64, 68 f., 71 f., 74, 84, 87, 92, 96, 99, 194 Garnham A. 218, 228 Gell, Alfred 224, 247, 256, 261 Gelman, Susan 252, 261 Gentile da Foligno 143 f., 162 Geraghty, Paul 209, 228 Gerard of Cremona 104, 112 f., 136, 162, 229 Gerhards, Simone 15, 311, 337, 342
345
Gerke, Sonja 10, 13, 67, 79, 81, 86, 96, 181, 194, 342 Goedicke, Hans 48, 63 Goldstone, Robert 3, 16 Goldwasser, Orly 10, 12 f., 45, 47, 49 f., 54–56, 60–63, 64, 65, 69 f., 77, 92, 96, 170 f., 176, 194, 264 f., 276, 342 Goody, Jack 267, 273, 276 Grandi, Nicola 205, 228 Grapow, Hermann 81, 83, 85, 90–92, 96, 184, 194, 195, 337 Green, Diana 214, 227 Green, Monica 120 f., 127, 162 Greenberg, Joseph 216, 226, 228 Greimas, Algirdas Julien 304 f., 308 Grethlein, Jonas 39, 42 Griffith, Francis Ll. 63, 65, 74, 96 Grinevald, Colette 47, 49, 55 f., 60–62, 64, 65, 228, 264 f., 276 Grosjean, F. 218, 228 Grossman, Eitan 64, 276 Guichard, Helène 75, 77, 96 Gutas, Dimitri 136, 162 Guthrie, Malcolm 203, 229 Habachi, Labib 52, 64, 65 Haller, Rudolf 283, 308 Hallowell, A. Irving 250, 261 Hardy, Frederick G. 105, 162 Harris, Marvin 8, 16 Hawkins, John A. 220, 229 Heinemann, Gottfried 39, 42 Heinimann, Felix 39, 42 Heitsch, Ernst 42, 43 Helck, Wolfgang 45, 64, 95, 97, 98 Hennig, Willi 235, 237 Herb, Michael, 79, 81, 96, 97 Herodotus 29, 41, 257 Hesiod 12, 21–41, 42, 43 Heuken, Ulrike 105, 162 Heymann, Sebastien 320, 336 Heywood, Vernon H. 231 f., 237 Hillis, David 233, 237 Hippocrates 108, 100 f., 116 f., 163, 166 Hoch, James 50, 55, 64 Hoffmann, Friedhelm 82, 95, 97, 112, 159, 335, 336 Hofmann, Beate 312, 336 Holttum, Richard 232, 237 Homer 21, 29, 30, 39, 42, 43 Hornung, Erik 69, 82, 84, 95, 97, 335 f.
346
Index
Hoskins, Janet 254, 261 Houlihan, Patrick F. 77 f., 97 Humboldt, Wilhelm von 285 f., 308 Hundius, Harald 207, 216, 229 Hunn, Eugene S. 60, 64, 249, 261 Husserl, Edmund 286 f., 308 Ibn al-Abbas al-Majusi / al-Magusi (s. also De probanda medicina; Pantegni Practica, ) 101, 111, 113, 121 f., 124–127, 143, 147 f., 156, 162, 163, 165, 166 Ibn al-Jazzar 101, 113, 121 Imhausen, Annette 8, 16, 17, 195 Ingold, Tim 257–259, 261, 262 Iohannes de Sancto Amando, s. also John of Saint-Amand 103, 121, 132, 159 Iohannes de Sancto Paulo, s. also John of Saint-Paul‘s 132, 159 Isaac Iudaeus 106 f., 121, 159, 165 Jacobsen, Thomas 218, 228, 229 Jacomy, Mathieu 320, 336 Jacquart, Danielle 120, 136, 139, 142–144, 160, 162, 163, 165, 166 Jacques, Jean-Marie 99, 111, 161, 162 Jansen, Ulrike 114, 160, 162 John of Saint-Amand (s. also Iohannes de Sancto Amando) 102–104, 139, 142, 164 John of Saint-Paul’s (s. also De medicinis simplicibus; Iohannes de Sancto Paulo) 132–135 Johnson, Kaitlin R. 64, 229 Jones, Dilwyn 84, 97, 336 Junge, Friedrich 327, 336 Kadereit, Joachim W. 6, 14, 231, 234 f., 237 f., 262, 342 Kahl, Jochem 170, 194 Kammerzell, Frank 47, 49, 64, 170 f., 195 Kanawati, Naguib 77 f., 97 Kaplony, Peter 90, 97 Katamba, Francis 203, 229 Keil, Frank 252, 261 Keil, Gundolf 102, 103, 163 Kersten, Alan 3, 16 Kitchen, Kenneth A. 83, 87, 97 Klebs, Luise 77 f., 97 Kleinsgütl, Dagmar 80, 97 Klimaschewski-Bock, Ingrid 105, 163 Knigge Salis, Carsten 65, 86, 97 Knopp, Johannes 320, 337 Köhler, Ines 183, 195 Kölver, Ulrike 207, 216, 229
Konerding, Klaus-Peter 296, 298, 308 Koning, Hugo H. 22, 43 Körtvélyessy, Lívia 228, 229 Koselleck, Reinhart 306, 308 Krauss, Rolf 69, 97 Kristeller, Paul-Oskar 130, 163 Kroemer, Georg-Heinrich 132, 159, 163 Kullmann, Wolfgang 41, 43 Lakoff, George 16, 77, 97, 168, 195, 264, 276 Lambert, W. E. 203, 229 Lambrecht, Knut 219, 229 Lancy, David F. 267, 276 Langermann, Y. Tzivi 142, 163 Leach, Edmund 248 f., 253, 261, 262 Lehman, Frederic K. 53, 57, 58, 64 Lehmann, Christian 219, 229 Leitz, Christian 72 f., 97 Leonard, Albert Jr. 78, 97 Leonti, Marco 273, 276 Lesky, Albin 29, 36, 43 Lesur-Gebremariam, Joséphine 61, 64 Lévi-Strauss, Claude 248, 251, 253, 261 Lévy-Bruhl, Lucien 209, 229 Levy, Joseph P. 318, 336 Lichtenberk, Frantisek 209, 229 Lieberknecht, Sieglinde 105, 163 Lietz, Haiko 312, 337 Lincke, Eliese-Sophia 47, 49, 64, 170, 195 Linnaeus, Carl (Carolus; Carl von Linné) 21, 40 f., 42, 43, 67, 231, 235–237, 250, 261, 264, 266–268 Liszka, Kate 71, 97 Little, William 256, 261 Lloyd, Geoffrey Ernest Richard 26, 43 Löbel, Elisabeth 225, 229 Lönneker, Birte 298, 308 Loprieno, Antonio 7, 16 Loret, Victor 84, 97 Luiselli, Maria Michela 86–87, 91, 97, 98 Lyons, Christopher 218, 229 MacKinney, Loren 116, 163 Maire, Brigitte 113, 116, 161, 163 Makowski, Maciej 45, 64 Manning, Christopher D. 317 f., 337 Martín Ferreira, Ana Isabel 132, 163 Mattéi-Muller, Marie-Claude 211, 229 Matzat, Uwe 316, 337 Mauss, Marcel 253, 258, 261 Mazzini, Innocenzo 116, 163 McGregor, William 209, 228
Index McVaugh, Michael R. 139, 142, 158, 163 Medin, Douglas L. 270, 275 Meeks, Dimitri 83, 98 Melville, Herman 319, 337 Meyer, Fernand 274, 276 Micheau, Françoise 136, 143, 162 Miller, Daniel 247, 261 Minsky, Marvin 286–288, 294 f., 297, 300–302, 304 f., 308 Mithen, Steven 254, 261, 272, 276 Mittelstrass, Jürgen 283, 308 Möller, Georg 88, 98 Montanari, Franco 22, 43 Moret, Alexandre 87, 98 Moretti, Franco 312 f., 337 Most, Glenn W. 23, 30, 31, 35, 43 Mottin, Jean 213, 229 Müller, Matthias 61 f., 64, 307 Murphy, Gregory L. 16, 168, 171 f., 179, 195 Newberry, Percy E. 79, 81, 98 Newton, Francis 121, 163 f. Nibbi, Alessandra 90, 98 Nicoud, Marilyn 132, 164 Nosofsky, Robert 4, 16 Odo Magdunensis (Odo of Meung, s. also Macer floridus) 101, 103, 113, 160 Ogden, C. K. 3, 7, 16 Ogilvie, Brian W. 128, 164 Onions, Charles T. 256, 261 Orozco, Inti Aedo 205, 229 Osing, Jürgen 69–72, 74, 81, 98 Oswalt, Robert L. 57, 64 Oswalt, Wendell H. 241, 261 Paavilainen, Helena 136 f., 164, 165 Pagel, Julius Leopold 103, 142, 159, 160 Pantalacci, Laure 61, 64 Paranyushkin, Dmitry 319, 337 Paravicini Bagliani, Agostino 132, 160, 164, 165, 166 Parfionovitch, Yuri 274, 276 Passer, Matthias Benjamin 215, 221, 229 Paulsen, Thomas 21, 43 Payne, Anick 56, 64 Payne, Doris 211, 228. Peck, Arthur 42, 43 Peirce, Charles Sander 3, 7, 16 Petit, Caroline 107, 161, 164 Petrie, William M. Flinders 74, 96 Petrus de Sancto Floro 142, 159, 160 Philippson, Paula 22, 27, 43
347
Platearius 101, 129, 131, 160, 165 Plato 24, 41 f., 283 Pliny the Elder 68, 250 Pommerening, Tanja 1, 8, 10, 13, 16, 17, 69, 98, 167, 174, 176–178, 183–185, 195, 342 Ponzetto, Simone Paolo 320, 337 Pormann, Peter E. 120, 136, 164 Porphyry 274, 275 Port, Robert 3–5, 17 Praet, Istvan 259, 261 Pries, Andreas Henning 72, 98 Pseudo-Apuleius 101, 160 Pseudo-Galen (s. also De simplicibus ad Paternianum) 116–118, 161 Pseudo-Mesue 105, 142, 163 Quack, Joachim Friedrich 74, 98 Quirke, Stephen 84, 98 Ramscar, Michael 3–5, 17 Ray, John 231, 237 Rechenauer, Georg 22, 42 Reed, Edward S. 256, 262 Reichel, Michael 21, 43 Reichelt, Gregor 1, 17 Reips, Ulf-Dietrich 316, 337 Reisner, George 71, 98 Rengakos, Antonios 22, 43 Richards, I. A. 3, 7, 16 Richards, Paul 271, 276 Ridley, Mark 237 f. Rieger, Burghard B. 312 f., 337 Rigault, A. A. 203, 229 Rips, Lance 3, 17 Rival, Laura 242, 258 f. 260, 262, 276 Romney, Kimball 267, 277 Rosch, Eleonor 4, 17, 266, 268 f., 277 Rumphius, Georgius E. 250, 262 Sahrhage, Dietrich 77, 98, 99 Salernitanus, Bartholomaeus 101, 130–132, 166 Salernus (magister) (s. also Tabulae Salerni) 101, 132 f., 160 Sandman, Maj 85, 87, 98 Santamaría Hernández, María Teresa 113, 164 Sauneron, Serge 72 f., 88, 93, 98 Savage-Smith, Emilie 136, 164 Scarborough, John 105, 164 Schalick, Walton O. 139, 164 Schank, Roger C. 287, 303, 308 Schenkel, Wolfgang 177, 195 Schmitt, Bernd H. 218, 230 Schmitz, Rudolf 105, 164
348
Index
Schneider, Richard 239, 260 Schneider, Thomas 49 f., 59, 65 Schulman, Alan R. 47, 51, 64 Schulz, Raimund 29, 43 Schulze, Christian 105, 164 Schütz, Albert J. 209 f., 229 Schütze, Hinrich 317 f., 337 Schwarz, Monika 168, 195 Schwarzbach, Andrea 234, 237 Seifart, Frank 224, 229 Seiler, Hansjakob 177, 195, 200, 209, 229 Sera, Maria D. 218, 219 Serrano , Jose M.90, 98 Shaw, Ian 45, 65 Shepardson, Kenneth N. 205, 229 Siraisi, Nancy G. 143, 164 Smith, Justin 251, 262 Sneath, Peter H. A. 231, 237 Snell, Bruno 22, 43 Snijders, Chris 316, 337 Sokal, Robert R. 231, 237 Sowa, John F. 275, 277 Spalinger, Anthony J. 335, 337 Spradley, James P. 251, 262 Srinivasan, Mahesh 218, 229 Stearn, William T. 231, 237 Stephen of Antiochia 121 f., 126 f. Störk, Lothar 82, 98 Strathern, Andrew J. 267, 276 Strauss, Anselm L. 313, 337 Stuessy, Tod F. 236 f. Sybesma, Rint 217, 228 Taylor, Paul M. 246–248, 262 Tesnière, Lucien 289, 308 Teubert, Wolfgang 306, 307, 313, 337 Theophrastus 10, 17, 107, 250 Theuerkauf, Gerhard 7, 17 Thissen, Heinz-Josef 68, 96 Toepfer, Georg 41, 42, 43 Totelin, Laurence M. V. 105, 162 Touwaide, Alain 108, 112, 116, 119, 164, 166 Trigger, Bruce 8, 17, 284 Tsagalis, Christos 22, 43 Tucker, G. R. 203, 229 Tyler, Stephen A. 260, 266, 277 Ullmann, Manfred 127, 136, 165 Urso, Anna Maria 130, 162 van Steenis, Cornelius 235, 237 Van’t Land, Karine 136, 165 Vandier, Jacques 73, 99
Veit, Raphaela 136, 165 Ventura, Iolanda 10, 13, 101, 112, 122, 131, 165, 166, 175, 342 Vernus, Pascal 45, 47, 51, 53 f., 65, 75, 78 f., 82 f., 86, 95, 99 Viveiros de Castro, Eduardo 254, 262 Vogt, Sabine 107, 166 Voigts, Linda 113, 166 Völker, Johanna 320, 337 von den Driesch, Angela 45, 63 von Lieven, Alexandra 72–75, 97 von Recklinghausen, Daniel 72, 98 Wack, Mary Frances 121, 126, 127, 166 Wahrig, Gerhard 283, 308 Wallis, Faith 96, 120 f., 123, 127, 130, 166 Walters, S. Max 231, 237 Warburton, Dvid A. 69, 97 Weill-Parot, Nicolas 144, 166 Welmers, William A. 203, 229 Wenig, Steffen 71, 80, 96, 99 West, Martin L. 21, 23–30, 32–38, 43 Westendorf, Wolfhart 9, 17, 167, 171, 182, 185, 194, 195, 335, 337 Widdows, Dominic 314, 320, 336 Wiedenhof, Jereon 208, 230 Wild, Henri 78, 99 Wittgenstein, Ludwig 4, 17, 283 f., 308, 314, 337 Wölfel, Hans 131, 160 Wolters, Gereon 2, 17 Yoyotte, Jean 45, 47, 65, 75, 78 f., 82–83, 86, 88, 93, 95, 98, 99 Zandee, Jan 87, 99 Zeidler, Jürgen 49, 65 Zhang, Hong 208, 230 Zhang, Shi 218, 230 Ziem, Alexander 287, 295, 309 Zierlein, Stephan 41, 43
List of contributors
349
List of Contributors Prof. Dr. Jochen Althoff Johannes Gutenberg-Universität Mainz FB 07, Institut für Altertumswissenschaften, Klassische Philologie Philosophicum, Jakob-Welder-Weg 18 D-55128 Mainz
Dr. des. Sonja Gerke Johannes Gutenberg-Universität Mainz FB 07, Institut für Altertumswissenschaften, Ägyptologie Hegelstraße 59 D-55122 Mainz
Prof. Dr. Walter Bisang Johannes Gutenberg-Universität Mainz FB 05, Department of English and Linguistics General Linguistics/Language Typology Jakob-Welder-Weg 18 D-55128 Mainz
Prof. Dr. Orly Goldwasser Hebrew University Jerusalem Dept. of Ancient Near Eastern Languages & Civilizations, Faculty of Humanities 91905 Jerusalem Israel
Prof. Dr. Dietrich Busse Heinrich-Heine-Universität Düsseldorf Lehrstuhl für Germanistische Sprachwissenschaft D-40204 Düsseldorf
Prof. Dr. Joachim W. Kadereit Johannes Gutenberg-Universität Mainz Fachbereich Biologie Direktor Botanischer Garten D-55099 Mainz
Prof. Dr. Roy Ellen University of Kent School of Anthropology and Conservation Prof. of Anthropology & Human Ecology (em.) Canterbury, Kent, CT2 7NR United Kingdom
Prof. Dr. Tanja Pommerening Johannes Gutenberg-Universität Mainz FB 07, Institut für Altertumswissenschaften, Ägyptologie Hegelstraße 59 D-55122 Mainz
Dr. Frederik Elwert Ruhr-Universität Bochum Digital Humanities, Religionssoziologie, Religion und Migration, Netzwerkforschung Universitätsstraße 90a D-44789 Bochum
Prof. Dr. Iolanda Ventura Università di Bologna Dipartimento di Filologia Classica e Italianistica Via Zamboni 32 40126 Bologna Italien
Simone Gerhards, M.A. Johannes Gutenberg-Universität Mainz FB 07, Institut für Altertumswissenschaften, Ägyptologie Hegelstraße 59 D-55122 Mainz