227 114 5MB
English Pages 82 [84] Year 1973
THE ENGLISH VERB SYSTEM
JANUA LINGUARUM S T U D I A MEMORIAE N I C O L A I VAN WIJK D E D I C A T A edenda curai
C. H. VAN S C H O O N E V E L D INDIANA
UNIVERSITY
SERIES M I N O R NR. XXIV
1973 MOUTON THE HAGUE • PARIS
THE ENGLISH VERB SYSTEM by ALPHONSE
JUILLAND
STANFORD U N I V E R S I T Y and JAMES
MACRIS
HUNTER COLLEGE
Second
printing
1973 MOUTON THE HAGUE • PARIS
© Copyright 1962 by Mouton & Co., Publishers, The Hague, The Netherlands. No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.
First printing
1962
Printed in The Netherlands
FOREWORD
The need for a new analysis of the English verb system may seem questionable, considering that every manual or textbook on English grammar devotes at least one chapter to the subject. Appropriate sections dealing extensively with English verb morphology can be found in Sweet's A New English Grammar (1892), Palmer's A Grammar of Spoken English (1930), Curme's A Grammar of the English Language (1935), Fries's American English Grammar (1942), and Jespersen's A Modern English Grammar (1942), to quote only a few of the more authoritative sources. Deliberately structural interpretations, exploiting the concepts and the techniques devised in recent decades by modern linguistics, essentially derive from Bernard Bloch's "English verb inflection" (1947, 1957); following Charles Hockett's study of the same title (1942), this study underlies a number of versions included in Englishoriented treatises on general linguistics such as H. A. Gleason, Jr.'s An Introduction to Descriptive Linguistics (1955, 1961) and Archibald Hill's Introduction to Linguistic Structures (1958), or in structurally oriented textbooks on English such as those in Gleason's selected bibliography. This refers to verbal forms rather than verbal phrases, to simple rather than compound forms, which constitute the main object of the present analysis. For the latter, especially for the intricate problem of the auxiliaries, see Twaddell's English Verb Auxiliaries (1960); and for a binary interpretation of the whole system, see Jakobson's study in the American Anthropologist (1960). It is hoped that, despite such an abundant treatment, the results brought forward in this monograph will justify an attempt prompted by our conviction that the resources of structural linguistics have not been fully exploited, and that more could be said on
6
THE ENGLISH VERB SYSTEM
one of the most thoroughly explored topics of English morphology. Essentially, the present analysis differs from its predecessors in that it distinguishes much more sharply between the various aspects of the system, the inflectional from the thematic and, for the latter, the vocalic or syllabic part of the contrastive segment from the consonantal or non-syllabic. First, each of these features is examined independently, with the ensuing classes established separately; then only, criteria are hierarchized and mutually subordinated, with the particular groupings combined in a general system of classes, subclasses, and sub-subclasses. Advantages of the hierarchical approach over those which prefer to coordinate thematic and inflectional characteristics (rather than subordinate the former to the latter) and treat the contrastive segments of stems as indivisible wholes (rather than distinguish their syllabic section from the non-syllabic) are both theoretical and practical, and affect both grammar and lexicon. Theoretically, it is a more logical procedure, which hierarchizes criteria to proceed from the more general features to the particular instead of placing the two on the same level; practically, the more discriminate approach reveals certain previously overlooked relations that hold between the thematic and the inflectional properties of verbal bases (see Section 5.0). As a result, a considerable simplification is achieved in the grammar, where the English verb system can be reduced to a fewer number of classes, and also to a fewer number of rules that govern the occurrence of their members; and in the lexicon, where economy is obtained in the morpholexical transcripts, i.e., in the number of symbols of formal class-membership necessary and sufficient to generate or construct any and all desired forms. In outlining an analysis intended mainly for specialists, a detailed account of previous interpretations was felt unnecessary, especially since a closer examination reveals the more recent ones to be versions of Bernard Bloch's "English verb inflection", a study which, for economy's sake, was taken to represent the thoroughest and soundest alternative to the analysis advanced in this essay.
FOREWORD
7
The analysis is divided into inflectional (see Section 2.0) and thematic (see Sections 3.0 and ff.), the latter further distinguishing the consonantal aspect (see Section 4.0) from the vocalic (see Section 5.0). At each step, relevant criteria are examined, classes established accordingly, memberships tabulated, and their inventories listed. Consequences of each analytic decision, both in the grammar and in the lexicon, are examined in the conclusions of each section. Finally, the two main classifications, inflectional and thematic, are combined in the general conclusions in which the English verb system is determined by subordinating the consonantal to the vocalic properties of the stems, and the thematic characteristics to the inflectional. Section 6.0 outlines the verall system of English verb classes and exhaustively lists their iaegular membership, which it follows by an alphabetical inventory of the morpholexical transcripts of irregular English verb bases. We wish to thank Mrs. Lillian Shargel for devoted help in the preparation of the manuscript, and Professor Cornelis van Schooneveld for precious editorial advice.
TABLE OF CONTENTS
FOREWORD
5
1.0 INTRODUCTION 13 1.1 Data 13 1.11 Thematically Regular and Irregular Verbs. . 13 1.12 Inflectionally Regular and Irregular Verbs. . 14 1.13 Irregular Verbs 14 1.131 Thematically Irregular verbs . . .14 1.132 Inflectionally Irregular verbs . . 14 1.133 Thematically-Infiectionally Irregular verbs 15 1.14 Anomalous, Auxiliary, and Defective Verbs . 16 1.2 Categories and Concepts 17 1.21 Suffixless Categories 17 1.22 Morphemic Environment 19 1.23 Complementary Distribution 20 2.0 THE INFLECTIONAL SYSTEM 2.1 Inflectional Morphs 2.11 Infinitive 2.12 Present 3 2.13 Preterit 2.14 Participle 2.15 Gerund 2.2 Inflectional Morphemes 2.21 Infinitive 2.22 Present 3 2.23 Preterit 2.24 Participle
22 22 22 22 22 22 23 23 23 23 24 24
10
THE ENGLISH VERB SYSTEM
2.25 Gerund 2.26 Inventory 2.3 Inflectional Classes 2.31 Inflectional Sets 2.32 Infinitive 2.33 Present 3 and Gerund 2.34 Preterit and Participle 2.35 Inflectional Classes 2.36 Inventory 2.4 The Inflectional System 2.41 Classes and Types 2.42 Strong, Weak, and Mixed Verbs . . . . 2.43 Regular, Mixed, and Irregular Verbs . . . 2.5 Inflectional Conclusions 2.51 Grammar 2.52 Lexicon
25 25 26 26 26 27 27 28 30 31 31 32 33 35 35 36
3.0 THE THEMATIC SYSTEM 3.1 The Thematic Criteria 3.11 Number of Alternants 3.12 Shape of Alternants 3.13 Selection of Alternants 3.2 Thematic Analysis 3.21 Inflectional and Thematic Classes . . . . 3.22 Syllabic, Non-Syllabic, and Syllabic-NonSyllabic Alternances 3.23 Vocalic and Consonantal Alternances . . .
37 37 38 38 39 40 40
4.0 THE THEMATIC SYSTEM: CONSONANTAL ASPECT 4.1 Consonantal Analysis 4.11 Number of Alternants 4.12 Shape of Alternants 4.13 Selection of Alternants 4.131 Syntagmatic relations 4.132 Exceptions 4.133 Selection rules
41 44 46 46 46 46 48 49 49 50
TABLE OF CONTENTS
4.2 Consonantal Classification 4.3 Consonantal Conclusions 4.31 Grammar 4.32 Lexicon
11
50 51 51 51
5.0 THE THEMATIC SYSTEM: VOCALIC ASPECT . 5.1 Vocalic Analysis 5.11 Number of Alternants 5.12 Shape of Alternants 5.121 First alternant included 5.122 First alternant excluded 5.123 Classes and subclasses 5.13 Selection of Alternants 5.131 The seven base groups 5.132 Base groups 1, 4, 5, 6, and 7. . . . 5.133 Base groups 2 and 3 5.134 Exceptions 5.135 Selection rules 5.136 Classes and subclasses 5.2 Vocalic Classification 5.3 Vocalic Conclusions 5.31 Grammar 5.32 Lexicon
53 53 53 54 54 55 57 58 59 60 62 64 65 65 66 68 68 69
6.0 GENERAL CONCLUSIONS 6.1 Grammar 6.2 Lexicon
70 70 75
1
INTRODUCTION
This introduction justifies the data underlying our investigation and discusses some of the analytic principles which have guided their analysis and interpretation; it also defines some of the general concepts exploited, especially insofar as they differ from American usage. 1.1 D A T A
Assuming that all but certain anomalous verb forms consist of a stem and an ending, English verbs can be classified into thematically Regular and thematically Irregular on the one hand, into inflectionally Regular and inflectionally Irregular on the other. Consequently, English verbs may be considered Irregular either on account of their stems, or on account of their endings, or on account of both. 1.11 Thematically Regular and Irregular Verbs Thematically Regular verbs are those which exhibit only one stem alternant in the various forms, i.e., whose stems display no variation in shape in the various categories. This is the case for a majority of English verbs, e.g., praise, pass, show, put, etc., which are not included in our corpus on account of their stems. Thematically Irregular verbs are those which exhibit more than one stem alternant in the various forms, i.e., whose stems exhibit variations in shape in the various categories. There are 149 such verbs, e.g., come, take, break, sing, etc., which are included on account of their stems.1 1
See 1.131 and 1.133.
14
THE ENGLISH VERB SYSTEM
1.12 Inflectionally Regular and Irregular Verbs Inflectionally Regular verbs are those which exhibit the inflectional set -0, -S, -D, -D, -ing in the five representative categories,2 or, more simply, the set -D, -D in the Preterit-Participle.3 This is the case for a majority of English verbs, e.g., wait, live, praise, say, etc., which are not included in our corpus on account of their endings. Inflectionally Irregular verbs are those which exhibit an inflectional set other than -0, -S, -D, -D, -ing in the five representative categories or, more simply, a set other than -D, -D in the PreteritParticiple. There are 182 such verbs, e.g., pass, fall, put, show, etc., all of which are included on account of their endings.4 1.13 Irregular Verbs English Irregular verbs can be divided into three categories, as follows: 1.131 Thematically Irregular verbs. Verbs incorporated on account of their stems, i.e., which exhibit more than one shape in the various forms, but are inflectionally Regular, i.e., which exhibit the -D, -D set in the Preterit-Participle, are: bid (bid), bid (bade), flee, have 'possess, etc.', hear, make, say, sell, shoe, stand, tell. 1.132 Inflectionally Irregular verbs. Verbs incorporated on account of their endings, i.e., which exhibit a set other than -D, -D in the Preterit-Participle, but are thematically Regular, i.e., which do not exhibit stem alternants, are: beat (beat, beat), beat (beat, beaten), bet, bid, (bid, bidden), burn, burst, cast, cost, cut, dwell, fit, hew, hit, hurt, knit, learn, let, outbid, overbid, pen, prove, put, quit, rid, set, sew, shed, show, shut, slit, smell, sow, spell, spill, spit, split, spoil, spread, strew, sweat, thrust, underbid, wed, wet. 1 3 4
See 1.2. See 2.34. See 1.132 and 1.133.
INTRODUCTION
15
1.133 Thematically-Inflectionally Irregular verbs. Verbs included both on account of their stems and on account of their endings, i.e., which exhibit more than one alternant as well as an inflectional set other than -D, -D in the Preterit-Participle, are: bare, beget, begin, bend, bereave, besiege, bid (bade, bidden), bide, bind, bite (bit bit), bite (bit, bitten), bleed, blow, break, breed, bring, build, buy, catch, choose, cleave (cleft, cleft), cleave (clove, cloven), cling, come, creep, crow, deal, dig, dive, do, draw, dream, drink, drive, eat, fall, feed, feel, fight, find, fling, fly, forbid, forget, forsake, freeze, get (got, got), get (got, gotten), gird, give, grind, grow, hang, heave, hide (hid, hid), hide (hid, hidden), hold, keep, kneel, know, lead, lean, leap, leave, lend, lie, light, lose, mean, meet, plead, read, rend, ride, ring, rise, run, see, seek, send, shake, shear, shine, shoot, shrink (shrunk, shrunk), shrink (shrunk, shrunken), shrink (shrank, shrunk), shrink (shrank, shrunken), shrive, sing, sink, sit, slay, sleep, slide, sling, slink, smite, speak, speed, spend, spin, spit, spring (sprung, sprung), spring (sprang, sprung), stave, steal, stick, sting, stink (stunk, stunk), stink (stank, stunk), stride (strode, strode), stride (strode, stridden), strike, string, strive, swear, sweep, swell, swim, take, teach, tear, think, thrive, throw, tread (trod, trod), tread (trod, trodden), wake (woke, woke), wake (woke, woken), wear, weave, weep, win, wind, wreak, wring, write. There are 11 thematically Irregular verbs, 44 inflectionally Irregular, and 138 thematically-inflectionally Irregular, giving a total of 193 Irregular verbs. All others, exhibiting only one stem in the various forms and the set -D, -D in the Preterit-Participle, are considered Regulars. The total of 193 Irregular verbs is obtained by adding 6 "prefixed" forms, i.e., beget, forbid, forget, outbid, overbid, and underbid, to the 201 entries listed by Bernard Bloch, 5 and by subtracting 14 anomalous, auxiliary, and defective entries, be, better, can, dare, 5
Article originally published in Language, XXIII (1947), pp. 399-418, and reprinted in Readings in Linguistics, ed. by Martin Joos (Washington, D.C., 1957), pp. 243-54.
16
THE ENGLISH VERB SYSTEM
go, have 'be obliged', may, might, must, need, ought, shall, used, and will, to be treated separately. 1.14 Anomalous, Auxiliary, and Defective Verbs Traditionally designated as "anomalous", "auxiliary", and "defective", members of Bloch's G, H, I, and J classes6 are such that to treat them on the same footing as the Irregulars proper would be uneconomical and misleading: as many as four classes are necessary to accommodate as few as 13 verbs, which would blur the main outlines of the English verb system. The structural properties which justify separate treatment of anomalous, auxiliary, and defective verbs are: (1) Reduced membership: class J, one member, be', class G, three members, dare, have 'be obliged', need; class H, three members, can, shall, will; and class I, six members, better, may, might, must, ought, used, in contrast to the average membership of Irregular classes, which is considerably higher; (2) Much higher than average frequency of occurrence for such anomalous verbs as can, shall, will, may, must, and be, in contrast to the considerably lower frequency of the non-anomalous, including irregulars; (3) Dystrophic paradigms, e.g., only one suffix (H) or two (G), or hypertrophic, e.g., six suffixes (J), in contrast to the regular four or five; (4) Special types of stem alternance, e.g., can, shall, will, and be, in contrast to the Irregulars, whose alternants differ according to certain established patterns; (5) Special syntactic function, generally auxiliary, in contrast to the non-anomalous verbs, including Irregulars; (6) Special accentual status of auxiliaries, generally atonic, in contrast to the non-anomalous, usually stressed.
6 In addition to the 13 members of these classes, we include the verb go in the Irregular category, on account of its stem as well as its ending. Cf. footnote 14 in Section 2.
INTRODUCTION
17
1.2 CATEGORIES A N D CONCEPTS
The verbal categories of English are basically those recognized by Bloch, according to whom "A verbal base, in English, is used without any suffix in several different functions: as an infinitive (I can't wait; I don't want to wait), as an imperative (Wait a minute), as a finite present with a subject in the 1st or 2d person singular or in the plural [or as A. A. Hill puts it, 'with a genderless subject'] (I wait here every day; If you wait for him; They wait in line for it), and in some other ways. In other functions, the verbal base appears with a following inflectional suffix: as a finite present with a subject in the 3d singular ['a gender subject'] (He waits here every day), as a finite preterit (I waited for him), as a participle (I've waited long enough), and as a gerund (I'm still waiting; perhaps also Waiting is tiresome)."7 1.21 Sujfixless Categories A minor difference must be registered, whose consequences are nevertheless significant. While Bloch's analysis is based on forms exhibited in the Present 3, the Preterit, the Participle, and the Gerund, we reckon at the outset with a fifth category, which includes the three "bare" or "sufBxless" forms, the Infinitive (to come), the Imperative (come!), and the "genderless" Presents (/, you, we, they come, as well as all compound forms, e.g., have come, will come, should come), inflectionally characterized by the morpheme /0/, referred to hereafter as the Infinitive. We must briefly state our reasons for doing so. The decision to discard "suffixless" forms is related to Bloch's third assumption, that "One of the alternants of a given morpheme may be zero; but no morpheme has zero as its only alternant" : 8 to take "suffixless" forms into account would enforce recognition of an inflectional morpheme having / 0 / as its only alternant, in the Infinitive, the Imperative, and the "genderless" Presents. 7
•
Cf. Readings, p. 244.
Ibid., p. 245; italics ours.
18
THE ENGLISH VERB SYSTEM
Since linguistic entities or invariants are best conceived of as pure abstractions, i.e., as ways of conceptualizing linguistic contrast, and "emes" essentially as ways of symbolizing relations between or among "alios", it is only in a highly metaphorical sense that phonemes, morphemes, or logemes can be said to "occur" in this or that shape, in this or that position, in this or that order. Considering that it is basically contrast which forms the object of linguistic symbolizations, hence that appropriate morphemic representations must account for every morphological contrast, the assumption that "no morpheme has zero as its only alternant" must be rejected, for there is as much contrast between, say, live and lives /liv-0~liv-z/, as between, say, lives and lived /liv-z~ liv-d/. The fact that the contrast between live and lives is expressed through the presence versus the absence of a shape, that between lives and lived through the presence of a shape versus the presence of another, is only a fact of substance. Combined with the assumption that verbal forms consist of a stem and an ending, the more fundamental postulate according to which every contrast must be actualized in an overt symbol requires that transcripts of English verbs consist of two symbols or groups of symbols, one to account for the stem, the other for the ending. An overt symbol is therefore needed to represent such inflectional contrasts, and insofar as absence of shape is involved for the "suffixless" categories, established conventions require that it be /0/. This avoids a dangerous distinction between degrees of absence or zero-ness, such as between absence symbolized by zero, e.g., Preterit put interpreted as bi-morphematic /put-0/, and absence symbolized by no space, e.g., Infinitive put, interpreted as mono-morphematic /put/. Furthermore, it avoids an awkward distinction between verbal forms consisting exclusively of a stem (those covered here by the label "Infinitive") and verbal forms consisting of a stem and an ending (all others). It is true that to reckon with a / 0 / inflection in the suffixless forms might create difficulties, since another inflectional morpheme of similar "shape" (an allomorph, in Bloch's interpretation) has been recognized in the Preterit and in the Participle, e.g., put /put-0,
INTRODUCTION
19
put-0, put-0/, come /kAm-0, keym-0, kAm-0/, and do /duw-0, did-0, dAn-0/. However, contrasts such as between the Preterit /put-0/ and the Participle /put-0/ show there is nothing unusual about different morphemes having allomorphs of identical shape; this is a simple case of homonymy or "syncretism" at the morphological level. The desire to avoid homonymy cannot override the more basic postulate that every contrast must be accounted for in terms of an overt symbol. Since other cases of homonymy must be recognized in the English verb system, cf., Preterit - 0 , -D, -t and Participle - 0 , -D, -t, paradigmatic discreteness can be secured by specifying the category of the morpheme occurring in homonymous cases. As we must in any event specify whether /D/ is the Preterit morpheme, i.e., /DP*/, or the Participle morpheme, i.e., /DPa/, or whether /t/ is the Preterit morpheme, i.e., /tP r /, or the Participle morpheme, i.e., /tPa/, we may similarly distinguish between the suffixless /0/, the Preterit /0P r /, and the Participle /0P»/.
1.22 Morphemic Environment The method followed here in reducing allomorphs to morphemes differs from American procedure, insofar as it uses the morpheme concept in a restricted sense, closer to the European tradition. Considering that the word is the natural matrix which restricts the morpheme's extension as well as its domain of occurrence, the environment relevant to morphemic reduction is necessarily intra-word environment. Extra-word environment, we believe, cannot be invoked in reducing allomorphs to morphemes, but only in reducing allologs to logemes. This means that segments such as am, are, is, etc., cannot be identified as allomorphs of the same morpheme with reference to, say, I, You, He, etc., since environments of this extra-word order are relevant only at the higher level at which am, are, and is are normally reduced as allologs of the same logeme. In terms of intra-word environment, which is the same, i.e., zero on either side, the segments in question must be interpreted as occurrences of three different morphemes, am, are, and is; but in terms of extra-word environment, which is
20
THE ENGLISH VERB SYSTEM
different, e.g., /, You, He, etc., they must be interpreted as occurrences of one and the same logeme, be.9 1.23 Complementary Distribution The concept of morpheme we use is further restricted in that it considers only morphs in phonological complementary distribution as allomorphs of one and the same morpheme. Thus, -s in bat-s, -z in boy-s, and -ez in ros-es (or puts, comes, and prais-es) are allomorphs of an inflectional morpheme -S, whose occurrence can be inferred from certain phonological properties of the stem. However, since the occurrence of -ren in child-ren, -en in ox-en, -i in gladiol-i, -a in dat-a, - 0 in men-0, etc., cannot be similarly inferred, we choose to reckon with more than one morpheme in this category and to consider that -S, -ren, -en, -i, -a, etc. (if not interpreted lexically) are different pluralizing morphemes rather than allomorphs of the same plural morpheme. There are several reasons for doing so. The first is economy, which may seem a paradoxical claim on behalf of a method which tends to increase the number of morphemes per category. Indeed, to reduce morphs in non-phonological complementary distribution is to recognize by definition only one morpheme per category. This narrow understanding of economy is self-defeating if we consider that grammar consists not only of inventories of entities ("emes"), but also of inventories of rules. While the more inclusive concept slightly lowers the number of invariants, it also increases the number of rules, which must be redundantly cross-referenced in the lexicon, at least for members of one-member classes; on the other hand, the more restricted concept somewhat increases the number of morphemes, for which it compensates by substantially reducing the rules of their occurrence. After all, the formal difference between grammar and lexicon is one of degree: rules which apply to many or all items belong to the grammar; rules which apply to one or a few items belong to * Cf. A. Juilland, Word and Grammar: Towards a General Theory (= Janua Linguarum, Series Minor, Nr. XVII) (The Hague, 1962), Section 7.2.
INTRODUCTION
21
the lexicon.10 Hence, rules governing the occurrence or selection of "allomorphs" such as -ren in child-ren and i- in i-s, which apply only to one or a few items, are much more lexical than grammatical in character. Therefore, it is preferable to convey this information in the lexicon rather than in the grammar, the latter referring to the former, not vice versa. Another reason is redundancy. The method which identifies morphemes on the basis of lexical complementarity inevitably reduces all morphs in a category to one single morpheme. Hence, there shall be by definition one grammeme (grammatical or nonroot morpheme) per category and one category per grammeme, which renders one of the two concepts superfluous. To preserve the useful distinction between morphological categories and morphological entities or morphemes — a category being a class of morphemes characterized by occurrence in the same general environment — one must be able to reckon with more than one morpheme per category. In Bloch's analysis, the distinction between the morphological concepts of category and entity is rescued by using the symbol ~ to express relations between "allomorphs" in phonological complementary distribution; and the symbol ; to express relations between "allomorphs" in lexical or non-phonological complementary distribution, where the choice of the allomorph "depends on the fact that the base ... is need and not some other base, such as lead".11
10
Cf. A. Juilland and P. M. H. Edwards, The Rumanian Verb System (in preparation), Section 1.2 11 Cf. Readings, p. 244. E.g., "Suffix 1 (3d singular): /ez/ after sibilant — /s/ after voiceless non-sibilant ~ /z/ elsewhere; /O/ (zero)". [Ibid., p. 245.]
2
T H E I N F L E C T I O N A L SYSTEM
The inflectional analysis consists of the four operations which characterize any exhaustive morphological investigation: segmentation of the chain into morphs, reduction of allomorphs to morphemes, grouping of morphemes into classes, and further grouping of classes into a system of categories.1
2.1 INFLECTIONAL MORPHS
Inflectional morphs exhibited by English verb forms in the various categories are as follows: 2.11 Infinitive The only morph in this category is - 0 , e.g., live-0, fall-0, put-0, show-0, dive-0.2
pass-0,
2.12 Present 3 The morphs in this category are -ez, e.g., pass-es, -z, e.g., live-s, fall-s, shows, dives, -s, e.g., puts, and - 0 , e.g., must. 2.13 Preterit The morphs in this category are -ed, e.g., wait-ed, -d, e.g., live-d, show-ed, dive-d, -t, e.g., pass-ed, and -0, e.g., put-0. 2.14 Participle The morphs in this category are -ed, e.g., wait-ed, -d, e.g., live-d, 1
Cf. A. Juilland, Outline of a General Theory of Structural Relations ( = Jantia Linguarum, Series Minor, Nr. XVI) (The Hague, 1961), Section 3.2. 2 See 1.21.
THE INFLECTIONAL SYSTEM
23
dive-d, -t, e.g., pass-ed, -0, e.g., put-0, -n, e.g., show-n, and -an, e.g., bitt-en. 2.15 Gerund The only morph in this category is -ing, e.g., liv-ing, pass-ing, fall-ing, putt-ing, show-ing, div-ing.
2.2 I N F L E C T I O N A L
MORPHEMES
Inflectional morphs in complementary distribution are reduced to one another and interpreted as allomorphs of the same inflectional morpheme. In operating these reductions, lexical complementarity is excluded: only morphs in phonological complementary distribution, whose occurrence can be inferred from certain properties of the stem, are subject to morphemic reduction.3 2.21 Infinitive Inflectionally characterized by absence of endings, Infinitive forms such as give, prove, or start are interpreted as exhibiting a / 0 / inflection, i.e., /giv-0/, /pruwv-0/, and /start-0/. To account for the inflectional contrast with forms in other categories, we posit an inflectional morpheme /0/, which "occurs" in the Infinitive, the Imperative, the Present 1, and the Present 2 forms. We register in this category one inflectional morpheme, /0/. 2.22 Present 3 The first three morphs, -ez, -z, and -s, are in complementary distribution with classes of stems established with reference to the final consonant: -ez is selected by stems ending in sibilants, e.g., praises /preyz-ez/; -s by stems ending in voiceless non-sibilants, e.g., parts /part-s/; and -z by all other stems, e.g., gives jgiv-z/. Hence, they are interpreted as allomorphs of one inflectional morpheme, symbolized as /S/.4 3 4
See 1.21. In Bloch's analysis, the symbol is /z/.
24
THE ENGLISH VERB SYSTEM
The remaining verbs, e.g., dare, need, and have 'possess', interpreted as /dehr-0/, /niyd-0/, and /hav-0/, can be said to exhibit a - 0 ending, in non-complementary distribution with other endings in this category, to postulate a / 0 / invariant. We register two inflectional morphemes in this category: /S/, allomorphs -s, -z, -ez; and / 0 / . 2.23 Preterit The morphs -ed and -d are in complementary distribution: -ed occurs with stems ending in an apical stop, e.g., parted /part-ed/, -d with stems ending in a voiced stop other than an apical, e.g., praised /preyz-d/. Interpreted as occurrences of the same morpheme, they are symbolized as /D/. 6 The other two endings, -t and - 0 , are not in complementary distribution with the allomorphs of /D/ (cf. /dwel-t/ and /pul-d/, /keym-0/ and /kleym-d/) or with one another (cp. /geyv-0/ and /reyv-d/). Hence, two more inflectional morphemes are added, transcribed respectively as /t/ and / 0 / . We register in this category three inflectional morphemes, /D/, allomorphs -d, -ed; /t/; and / 0 / . 2.24 Participle The morphs -ed and -d exhibit the same relations as in the Preterit, -ed occurring after stems ending in an apical stop, e.g., (have) waited /weyt-ed/, and -d after stems ending in a voiced sound other than an apical stop, e.g., (have) lived /liv-d/. Hence, we posit an inflectional morpheme /D/, distinguished from the homonymous inflection of the Preterit by transcribing respectively /Dp»/ and /DP'/. The two morphs, -t and - 0 , which are not in complementary distribution with the above, cp. /paehs-t/ and /part-ed/, /put-0/ and /smel-d/, must be interpreted as occurrences of two different inflectional morphemes, symbolized respectively as /t/ and / 0 / . In the Participle, there are two additional morphs in comple5
In Bloch's analysis, the symbol is /d/.
THE INFLECTIONAL SYSTEM
25
mentary distribution, -n, and -an; -n is selected by stems ending in a syllabic,6 e.g., shown /sow-n/, -an by all other stems, e.g., bitten /bit-an/. These are not in complementary distribution with any of the above. Since no relations hold with /D/, cp. /teyk-an/ and /feyk-t/, or with -t, cp. /powk-t/ and /browk-an/, or with - 0 , cf. /bid-0/ and /bid-an/, -n and -an are interpreted as occurrences of an inflectional morpheme symbolized /N/. We register in this category four inflectional morphemes, /D/, allomorphs -d, -ed; /N/, 7 allomorphs -n, -an; /t/; and /0/. 2.25 Gerund
The only inflectional morph in the gerund is /-ing/, e.g., giv-ing, prov-ing,
start-ing.
We register one inflectional morpheme in this category, /ing/. 8 2.26 Inventory
The inventory of inflectional morphemes is as follows: (1) Infinitive: / 0 / (2) Present 3: /S/ (3) Preterit: /D/, /t/, / 0 / (4) Participle: /D/, /t/, /0/, /N/ (5) Gerund: /ing/ Besides differences in notation,® two departures from Bloch's treatment must be registered: since we reckon with an additional category, the Infinitive, we also recognize an additional morpheme, / 0 / ; and since we treat anomalous, auxiliary, and defective verbs separately, we eliminate the Present 3 morpheme /0/, to reckon with only one inflectional morpheme in this category, /S/. • "A syllabic is a vowel alone or a vowel plus /y, w, h, r, hr/;..." {Readings, p. 245, fn. 11). ' In Bloch's analysis the symbol is /n/. 8 Suggested by a number of scholars, the bi-phonematic interpretation of [q] as /ng/ seems to us preferable to the mono-phonematic / g / o n distributional, economical, and systematic grounds. * See footnotes 4, 5, and 7.
26
THE ENGLISH VERB SYSTEM 2.3 I N F L E C T I O N A L
CLASSES
Given the inflectional morphemes in the five categories, the relations in which they stand to verbal bases must be examined, and the paradigmatic sets which inflectionally characterize English verbs determined. 2.31 Inflectional Sets A system of sets similar to the one presupposed by the Strong, Weak, and Mixed division of traditional grammar has been established by Bloch.10 Parallel sets, differing only in that they reckon with the Infinitive while they eliminate the / 0 / in Present 3, are given below: TABLE l
Class
Infin.
Pr. 3
Prêt.
Part.
Ger.
Ex.
A
0
S
D
D
ing
live
B
0
S
t
t
ing
pass
C
0
S
0
N
ing
fall
D
0
S
0
0
ing
put
E
0
S
d
N
ing
show
F
0
S
0
D
ing
dive
2.32 Infinitive The addition of the Infinitive does not affect the classification, which consists of the same number of classes with the same membership, thus showing that inflectionally, the Infinitive is an irrelevant category. As a matter of fact, this is true for all one-morpheme categories, which do not allow for inflectional differentiation: there being only one ending in the Infinitive, all English verbs are inflectionally alike in regard to this category. Incapable of inflection10
Cf. Readings, p. 245.
THE INFLECTIONAL SYSTEM
27
al contrast, the Infinitive can be discarded, a decision confirmed by the fact that its elimination does not affect the ensuing classification : whether included or excluded, the same classes of English verbs are determined, with the same membership. 2.33 Present 3 and Gerund This calls attention to other categories which are similarly incapable of securing morphological contrast: characterized by only one ending, they must also be considered inflectionally irrelevant.11 This is the case for the Present 3 and the Gerund, whose unique ending offers no possibility of inflectional contrast and which may therefore be discarded, together with the Infinitive. The inflectional irrelevancy of these categories is confirmed by the fact that their exclusion affects neither the inflectional classes nor their membership. 2.34 Preterit and Participle Inflectional sets established with reference to the Preterit and the Participle yield the same classes with the same membership as inflectional sets including also the Infinitive, the Present 3, and the Gerund. As the latter are inflectionally irrelevant, only the Preterit and the Participle are necessary to establish the inflectional classes of the English verb system. This is expressed by the equation (1X 1 x ) 3 x 4 ( x 1) = 12, whose digits stand, in order, for the number of inflectional morphemes per category. The total of twelve represents all possible combinations of ten inflectional morphemes distributed into five categories, Infinitive, Present 3, Preterit, Participle, and Gerund ( 0 ; S; 0 , D, t; 0 , D, t, N; ing), i.e., 1 x 1 x 3 x 4 x 1 = 1 2 ; or of seven inflectional morphemes distributed into two, Preterit and Participle (0, D, t; 0 , D, t, N), subdivided into four combinations (0, D, t, N), i.e., 3 x 4 = 12. It is significant that only six out of twelve possible combinations are realized, the inflectional aspect of the English verb system being characterized by 11
/ 0 / in need, etc., has been eliminated.
28
THE ENGLISH VERB SYSTEM
the six "holes in the pattern" as much as it is by the six realized combinations.12 Nevertheless, it would be premature to consider these categories irrelevant and to discard them for being unable morphologically to differentiate English verbs, since we have shown only that the forms they exhibit are inflectionally irrelevant; to eliminate them altogether, it must be shown that they are also thematically irrelevant, i.e., that English verbs cannot differ from one another in terms of the stem alternances they exhibit in those categories.13 2.35 Inflectional Classes With the Infinitive, the Present 3, and the Gerund discarded, there should be as many inflectional classes as there are sets of endings displayed in the Preterit-Participle. Class-constitutive sets are combinations of two inflectional morphemes, each set consisting of one of three Preterit endings, /D/, /t/, or / 0 / ; and one of four Participle endings, /D/, /t/, /0/, or /N/, as indicated below:
liv-d, liv-d paehs-t, paehs-t put-0, put-0
Pr.
Pa.
D *
D
dowv-0, dayv-d
N
§ow-d, sow-n
t 0 ^
\
fel-0, fohl-an By exploiting the phonological similarity of the inflectional morphemes in the Preterit and the Participle, /D/ morphemes, either Preterit or Participle, may be represented by digit 1; /t/ morphemes, Preterit or Participle, by digit 2; / 0 / morphemes, 12 13
See 2.36. See 4.13 and 5.13.
29
THE INFLECTIONAL SYSTEM
Preterit or Participle, by digit 3 ; and the /N/ morpheme, Participle only, by digit 4. Numerically, the paradigmatic relations that constitute the inflectional sets may be symbolized as follows : Pr. ©
Pa. >• CD
The six inflectional classes can thus be identified by two-digit symbols, the first standing for the inflectional morpheme selected in the Preterit, the second for the inflectional morpheme selected in the Participle: 11 (live), 14 (show); 22 (pass); 34 (fall).
31 (dive), 33 (put),
The numerical symbolization has the advantage of expressing more clearly the substantial relations which obtain between the constitutive morphemes and also to show the position of the actual or realized sets, 11, 14, 22, 31, 33, and 34, relative to the potential or unrealized ones, 12 (/D t/), 13 (/D 0/), 21 (/t d/), 23 (/t 0/), 24 (It N/), 14 and 32 ( / 0 t/), as follows:
e 0 © 0 TABLE 2
Dt
tD
0t
14
D 0
TABLE 3
©
t0
tN
21
(00)
(0N)
©
12
13
©
©
23
24
32
©©
Inflectionally, the verb go (wen-t, go-ne) could be considered a member of the otherwise "unrealized" Class 24. However, instead of setting up a onemember class, we prefer to treat this verb lexically.
30
THE ENGLISH VERB SYSTEM
Listed in alphabetical and numerical order, the inflectional classes of the English verb system are: TABLE 4
Traditional Proposed Class Class
Set
Example
A
11
D D
live
B
14
D N
show
C
22
11
pass
D
31
0 D
dive
E
33
0
0
put
F
34
0 N
fall
2.36 Inventory
There are 9 verbs in Class 11, 7 in Class 14, 39 in Class 22, 2 in Class 31, 82 in Class 33, and 54 in Class 34, as follows: (1) Class 11: flee, have, hear, make, say, sell, shoe, stand, tell. (2) Class 14: do, hew, prove, sew, show, sow, strew. (3) Class 22: bend, bereave, beseech, bring, build, burn, buy, catch, cleave, creep, deal, dream, dwell, feel, gird, keep, kneel, lean, leap, learn, leave, lend, lose, mean, pen, rend, seek, send, sleep, smell, spell, spend, spill, spoil, sweep, teach, think, weep, wreak. (4) Class 31: crow, dive. (5) Class 33: beat, begin, bet, bid (bid, bid), bid (bade, bade), bide, bind, bite, bleed, breed, burst, cast, cling, come, cost, cut, dig, drink, feed, fight, find, fit, fling, get, grind, hang, heave, hide, hit, hold, hurt, knit, lead, let, light, meet, plead, put, quit, read, rid, ring, run, set, shed, shine, shrink (shrunk, shrunk), shrink (shrank, shrunk), shut, sing, sink, sit, slide, sling, slink, slit, speed, spin, spit (spit,
THE INFLECTIONAL SYSTEM
31
spit), spit, (spat, spat), split, spread, spring (sprung sprung), spring (sprang, sprung), stave, stick, sting, stink (stunk, stunk), stink (stank, stunk), stride, strike, string, sweat, swim, thrust, tread, wake, wed, wet, win, wind, wring. (6) Class 34: bear, beat, beget, bid (bid, bidden), bid (bade, bidden), bite, blow, break, choose, cleave, draw, drive, eat, fall, fly, forbid, forget, forsake, freeze, get, give, grow, hide, know, lie, outbid, overbid, ride, rise, see, shake, shear, shoot, shrink (shrunk, shrunken), shrink (shrank, shrunken), shrive, slay, smite, speak, steal, stride, swear, swell, take, tear, thrive, throw, tread, underbid, wake, wear, weave, write.
2.4 T H E I N F L E C T I O N A L S Y S T E M
We may now consider the grouping of inflectional classes into more general and inclusive types based on the relations that hold among their members. 2.41 Classes and Types The preceding classification parallels Bloch's, which in turn parallels the conventional, but without grouping alphabetical classes into the larger categories of traditional grammar. Such larger groupings are normally based on the relations which hold among classes, and no such relations seem to justify a tripartite division into Weak, Strong, and Mixed. Bloch's alphabetical classes correlate with the traditional grouping in that A and B correspond to the Weak, C and D to the Strong, and E and F to the Mixed.15 Numerical and traditional classes, however, do not match in orderly fashion : 11 and 22 correspond to the Weak, 33 and 34 to the Strong, and 14 and 31 to the Mixed. This lack of parallelism 14
Bloch's G, H, I, and J classes correspond to the traditional Irregulars, i.e., anomalous, auxiliary, and defective verbs.
32
THE ENGLISH VERB SYSTEM
can be dismissed as long as the traditional classification cannot be justified on structural grounds; the fact that Bloch's classes match the traditional ones is inconclusive, since the alphabetical order, A, B, C, etc., is chosen precisely for this purpose. 2.42 Strong, Weak, and Mixed Verbs
Whatever the value of the traditional classification into Weak, Strong, and Mixed, the question is whether it can be formally justified. Weak verbs are characterized by the inflectional sets ID D/ and /t t/, i.e., 11 and 22; Strong verbs by the inflectional sets / 0 0 / and / 0 N/, i.e., 33 and 34; and Mixed verbs by the inflectional sets /D N/ and ¡0 D/, i.e., 14 and 31. Synchronically, it is difficult to see why /D D/ and /t t/ should be "Weak" while / 0 0/ is "Strong", and even more difficult to understand why /D NI should be "Weak" while / 0 N/ is "Mixed". On descriptive grounds, the tripartite classification could be justified by dividing the four inflections, /D/, /t/, /0/, and /N/, into two groups: /D/ and /t/ would be considered intrinsically "Weak", /01 and /N/ intrinsically "Strong". It goes without saying that "Weak" and "Strong" here, are only labels, hence arbitrary. Inflectional sets can be labelled in accordance with the division and labelling of inflections, as follows: (1) sets combining Weak and Weak inflections, i.e., /D D/ and /t t/, would be grouped in a category labelled "Weak"; (2) sets combining Strong and Strong inflections, i.e., / 0 0 / and / 0 N/, would be grouped in a category labelled "Strong"; (3) sets combining Strong and Weak, i.e., / 0 D/, or Weak and Strong inflections, i.e., /D N/, would be grouped in a category labelled Mixed. Thus, the "substantial" characterization of inflections and inflectional sets can be said to underlie the traditional classification of English verbs into Weak, Strong, and Mixed. A brief comment on the traditional classification seems in order. Descriptively, it is based on inflectional sets that include the morphologically irrelevant suffixless forms, here covered by the label "Infinitive". If sets including Infinitive forms are considered - a
THE INFLECTIONAL SYSTEM
33
normal procedure in traditional systems which take the Infinitive or the Present 1 as entry forms representing all verb variants (allologs) - paradigmatic or intra-logemic contrast between Regular verbs such as live /liv-0, liv-d, liv-d/ and meet /miyt-0, met-0, met-0/ or put /put-0, put-0, put-0/ inevitably leads to two different classes, Weak /D D/, e.g., live, and Strong / 0 0/, e.g., meet or put. If. however, only those forms through which English verbs may differ inflectionally are considered, the morphologically irrelevant difference between live and meet or put disappears. But even if morphologically irrelevant forms are taken into account, facts appear in a different light if the Present is represented by Present 3 instead of Present 1: the contrast between the Weak liv-0 ~ liv-d ~ liv-d and the Strong put-0 ~ put-0 ~ put-0 disappears if liv-S~ liv-d ~ liv-d on the one hand, and put-S ~ put-0 ~ put-0 on the other, are considered. The difference between the two, namely, that Weak verbs take the inflectional set /D D/, e.g. live, while Strong verbs take / 0 0/, e.g., put, is no more and no less significant than the difference between, say, the Weak /liv-S~ liv-d ~ liv-d/ and the Weak /pshs-S ~ pxhs-t ~ paehs-t/, which also exhibits a different set in the Preterit-Participle, namely, /t t/. Clearly, the difference between /t t/ and / 0 0 / is one of substance, not of form. 2.43 Regular, Mixed, and Irregular Verbs A more fundamental objection can be raised to this interpretation: by relying upon the arbitrary distinction between the intrinsically "weak" and "strong" nature of inflectional morphemes, the main question, whether the morphemes which constitute the relevant sets are the same or different, is bypassed. If the basic question is asked first, the inflectional sets fall naturally into two rather than three groups: (1) those which display the same ending, i.e., homonymous or syncretic sets in the relevant categories, namely, /D D/, /1 t/, and ¡0 0 / ; these may be called Regular. (2) those which display different endings, i.e., non-homonymous or discrete sets in the relevant categories, namely, /D N/, / 0 D/, and / 0 N/; these may be called Mixed.
34
THE ENGLISH VERB SYSTEM
In this systematization, the traditionally Weak category, which includes / 0 0 / and / 0 N/ verbs, is eliminated by assigning its / 0 0 / members to the traditionally Strong, together with which they form the structurally Regular class; and its / 0 N/ members to the traditionally Mixed, together with which they form the structurally Mixed class. The division into Weak, Strong, and Mixed, each consisting in turn of two classes, may be replaced by a division into Regular and Mixed, each containing three classes obtained by dividing the traditionally Weak into Regulars ( / 0 0 / 33) and Mixed (/ 0 N/ 34). Of course, a class of Irregulars including the anomalous, auxiliary, and defective members of Bloch's G, H, I, and J classes should be added. The classifications discussed can be compared in Table 5 below. TABLE 5
Traditional
Endings
Proposed 11
/DD/
A
22
It t /
B
33
10 01
C
34
\0 N /
D
14
/DN/
E
31
10 Til
F
Strong Regular
Mixed
Weak
Mixed
Irregulars
The breakdown by inflectional categories appears in Table 6. There are 130 verbs in the Regular category (9 in Class 11, 39 in Class 22, and 82 in Class 33),16 63 in the Mixed (7 in Class 14, 2 in 16
The members of these classes are listeii .n 2.36.
35
THE INFLECTIONAL SYSTEM TABLE 6
Regular Weak
Irregular
Mixed Strong
Irregular
Mixed m
DD
11
0 0
1-1
2-2
3-3
(9)
(39)
(82)
D N
(7)
b
3-4
30 a
*