Derived Intransitivity: A Contrastive Analysis of Certain Reflexive Verbs in German, Russian and English 9783111341026, 9783484102538

Citation preview

Linguistische Arbeiten

38

Herausgegeben von Herbert E. Brekle, Hans Jürgen Heringer, Christian Rohrer, Heinz Vater und Otmar Werner

David'J. Granmer

Derived Intransitivity: a Contrastive Analysis of Certain Reflexive Verbs in German, Russian and English

Max Niemeyer Verlag Tübingen 1976

CIP-Kurztitelaufnahme der Deutschen Bibliothek

Cranmer , David J. Derived intransitivity, a contrastive analysis of certain reflexive verbs in German, Russian and English. - 1. Aufl. - Tübingen : Niemeyer, 1976. (Linguistische Arbeiten ; 38) ISBN 3-484-10253-5

ISBN 3-484-10253-5 © Max Niemeyer Verlag Tübingen 1976 Alle Rechte vorbehalten. Ohne ausdrückliche Genehmigung des, Verlages ist es auch nicht gestattet, dieses Buch oder Teile daraus auf photomechanischem Wege zu vervielfältigen. Printed in Germany

TABLE OF CONTENTS Acknowledgements

.

.

.

.

.

.

.

.

.

.

1. Preliminaries . . . . . . . . . . 1.0 Introduction . . . . . . . . . 1.1 Purpose of this study . . . . . . . 1.2 Subject matter of this study . . . . . . 1.2.1 Reflexive pronoun in English, German and Russian 1.2.2 SICH/SJA-verbs to be studied 1.2.3 Lexical meaning . . . . . . . 1.2.4 Derived Intransitivity . . . . . . 1.2.5 Sumxary paradigm of SICH/SJA-verbs . . 1.3 Methodological considerations . . . . . 1.3.1 Contrast!ve analysis and linguistic investigation 1.3.2 Choice of theoretical framework . . . 2. Background . . . . . . . . . 2.1 Previous classifications of SICH/SJA-verbs . . 2.2 Transitivity 2.3 Voice 2.4 Transitivity and voice . . . . . . . 2.5 Comparison of some recent treatments of SICH/SJA-verbs 2.5.1 Bierwisch 2.5.2 Schaarschmidt . . . . . . . 2.5.3 Ruziöka

.

VII

. . . . .

1 1 1 2 . 2 5 6 9 .10 .11 . 11 .12

. . . . . . . . . .

.13 .13 18 21 28 . 28 28 30 32

3. The contrastive analysis . . . . . . . . . 38 3.1 Explanation of the theoretical framework . . . . 38 3.2 Derived Intransitivity in Russian . . . . . . 39 3.2.1 Evidence from selectional restrictions . . . 39 3.2.2 Ctoject-Preposing . . . . . . . . 43 3.2.3 Subcategorization features . . . . . .45 3.2.4 Object deletion . . . . . . . . 47 3.2.5 Unfilled accusative-NP's 50 3.2.6 SJA-verbs with more than one interpretation . . 54 3.2.7 Sunmary of Derived Intransitivity in Russian . . 54 3.3 Derived Intransitivity in German . . . . . . 55 3.3.1 Evidence from selectional restrictions . . . 55 3.3.2 Gernen phrase structure . . . . . . 58 3.3.3 Object-Preposing, object deletion, and unfilled accusative-NP1s . . . . . . . . 62 3.3.4 SICH-Introduction 71 3.3.5 Exceptions to SICH-Introduction . . . . . 73 3.4 Derived Intransitivity in English . . . . . . 75 3.4.1 Evidence from selectional restrictions . . .75 3.4.2 Discussion of open . . . . . . . 78 3.4.3 Discussion of object-NP's . . . . . . 81 3.4.4 SELF-verbs 82 3.5 Contrastive analysis statements . . . . . . 86 4. Further remarks on SICH/-SJA 4.1 Subjectless sentences in German . . 4.2 SICH-verbs formed from intransitive verbs 4.3 SICH and valence 4.4 English subject and valence . . . 4.5 Lack of subject and no valence reduction

.

. . . . .

.

. .

.

. .

. . . .

88 90 90 92 94 95

VI

4.6 4.7 4.8 4.9 5.

-SJA, intransitive verbs and valence . . . . . Comparison of Derived Intransitive marker and verb subcategorization for subject . . . . . . . Syntactic and lexical derivation . . . . . Conclusions: restatement of SICH-Introduction and the contrast!ve analysis in chapter 3 . . . . .

Summary

101 .102 .105 .

Literatur

108 111

List of verbs Register

99

.114 .

.

.

.

.

.

.

.

.

.

.

.116

ACKNOWLEDGMENTS Many of my teachers and fellow students discussed with me the material presented in this work.

I am particularly grateful to my teachers Frederick

Agard, Leonard Babby, Joseph Grimes, Richard Leed and Frans van Coetsem.

To

Renate Born, Sylvia Skorge and Alia Novosilzov, who acted as native speaker informants for German and Russian, I give a special thanks.

This work was

made possible by a grant from the Humanities and Social Sciences fund at Cornell University, to which this work was originally presented as part of my work for a doctor of philosophy degree in general linguistics.

1. 1.0

PRELIMINARIES Introduction.

IsaCenko (1968:456) writes:

"We have seen how in the Russian reflexive gram-

matical and lexical functions criss-cross."

[Wir haben gesehen, wie sich im

russischen Ref lexivum grarmiatische und lexikalische Funktionen überschneiden. ] This present work shows one way of separating these granmatical und lexical functions which are associated with what have been traditionally called "reflexive verbs" in Russian and German.

The so-called reflexive verbs in

English do not form such a grammatically significant verb class.

In fact, it

is the class of English verbs which can be used both transitively and intransitively that is of more relevance to a contrastive study of German and Russian "reflexive verbs".

(See section 1.3 where I explain how my use of contrastive

analysis differs frcm the more usual pedagogical one.) verb analyses are not mine.

The Russian and English

I have used Babby's (forthcoming - Lingua) analysis

of Russian SJA-verbs and Bowers's (ms.) analysis of transitive-intransitive verb pairs in English. 1.1

Purpose of this study.

The purpose of this work is to extend our knewledge of the syntax of English, German and Russian and, additionally to explore the possibilities of using contrastive analysis as a tool for linguistic investigation. main reasons why I chose to work on the "reflexive verbs":

There are two 1) to clarify sane

of the problems involved in analyzing them; and 2) to uncover certain syntactic differences among these three languages. A recent article by Jucquois (1973) entitled "The triple function of the reflexive in seme languages" brings out the basis of the problems which have arisen in the analysis of "reflexive verbs".

He notes that "reflexive verbs"

occur in sentences which have three different interpretations: middle, passive.

reflexive,

Because of this fact grartmarians have been unable to satis-

factorily analyze "reflexive verbs". Leonard Babby's work on SJA-verbs has done much to clarify the analysis of Russian "reflexive verbs".

He drops the term "reflexive verb" and uses the

term SJA-verb since there is nothing reflexive about many SJA-verbs.

I have

applied his syntactic theory to G e m e n "reflexive verbs" and have also followed him by dropping the "reflexive" label.

I call these SICH-verbs which occur ι with seme form of the setj-pronoun I call SELF-verbs. By using the same theory

2 for German, Russian and English, I have been able to highlight some of the syntactic differences among these three languages. 1.2

Subject matter of this study.

The verbs which this study covers form a subset of the SICH/SJA-verbs in German and Russian and the English transitive-intransitive verb equivalents. verbs in this subset have one defining characteristic:

The

all their members are

the result of syntactic Derived Intransitivity, i.e. lexically transitive verbs which are intransitive in the surface structure (see my section 1.2.4 and Bafcby forthcoming - Lingua).

Since SICH/SELF-verbs employ forms which are harophonous

to the reflexive pronouns, I present a brief description of the reflexive pronouns of English and Gerxran in order to help the reader follow the example sentences.

The Russian reflexive pronoun is included in order to show up sane

of the differences in morphology between Russian, German and English. 1.2.1

The reflexive pronoun in English, German and Russian.

The English reflexive pronoun has been described by Heike (1970:30) as consisting of the noun self preceded by a possessive pronoun in the determiner node. Although this analysis is not the usual one, it has the advantage of being able to relate within one phrase structure the following two sentences: (l)a b

I am not feeling myself

today.

I am n o t feeling m y usual self today.

The third person reflexive pronouns, however, do not occur with a possessive determiner but with the object pronoun forms: I we you

my our your

house house house

myself ourselves yourself/yourselves

he she it they

his her its their

house house house house

himself (*hisself) herself itself (*itsself) themselves (*theirselves)

-

Heike posits a rule to change the underlying possessive forms, i.e. his, its, their} one's into object forms, i.e. him3 it3 them3 one. vacuously on her.

Heike's rule operates

In English, then, the third person is treated differently

1. F u r t h e r m o t i v a t i o n for this step is p r o v i d e d by the fact t h a t it is g e n erally r e c o g n i z e d that the s o - c a l l e d "reflexive v e r b s " do n o t have a r e flexive m e a n i n g (see P o u t s m a 1926:143-144; IsaSenko 1968:457-458; Stötzel 1970:164).

3

fron the first and second persons. The diagram at the end of this section lists the English reflexive pronouns and the parameters by which they are characterized.

2 The German reflexive pronoun inflects for case, nuntoer and gender.

In

German the reflexive pronoun for the first persons and the familiar second persons declines for case, number and person. non-reflexive reflexive accusative/dative ich Ί ' wir 'we' du 'you' ihr 'you' (plur.)

mich/mir uns/uns dich/dir euch/euch

mich/mir uns/uns dich/dir euch/euch

Because the accusative and dative reflexives are the same as the non-reflexives, seme grammarians (e.g. Bierwisch 1965:95) treat these pronoun forms as the sane as the non-reflexive forms.

In the formal second persons and the third persons,

hcv/ever, there is a special reflexive form sich which retains invariant, i.e. it does not inflect for case, number or gender. Sie er sie es sie

' you1 'he' •she' •if •they'

Sie/Ihnen ihn/ihm sie/ihr es/ihm sie/ihnen

sich sich sich sich sich

Historically, the foriral second person is a special usage of the third person plural. Considering the formal person,· then, as a third person form, we can also say that in German, as in English, the third person is treated differently than the first and second persons.

I use SICH to refer to all members of the

reflexive paradigm collectively. The diagram on page 4 gives the reflexive pronoun forms and the parameters which characterize them. The reflexive pronoun in Russian inflects for case only and does not reflect 3 person, number, or gender. The diagram on page 4 gives the forms of this 2. There are two numbers in German (singular and plural), three grammatical genders (masculine, feminine and neuter) and four cases (nominative, accusative, dative and genitive). Within the pronoun system there is a fourth person category: the second person is divided into familiar and polite/ formal forms. 3. There are two numbers in Russian (singular and plural), three grammatical genders (masculine, feminine and neuter) and six cases (nominative, accusative, dative, instrumental, genitive and prepositional). Jacobson (1958 [1971:173f]) posits two other cases, which are listed as genitive II and prepositional II. Their use is marginal and in textbooks they are not

4 pronoun and its parameters. The verbal suffix -SJA has been called a reflexive (cf. Lyons 1967:362; Schaarschmidt 1968:24-26) because often a verb with -SJA can be substituted for a verb and the reflexive pronoun, e.g.

can replace

myt'sja

myt'

sebja

'wash

oneself'. This suffix is attached to the end of the verb, after the other verbal endings of tense, person, gender, number, infinitive, etc. It has two forms, which are phonologically conditioned, except when used with the active participle: -sja occurs after consonants and -s ' after vowels. The participle occurs only with -sja. Hereafter, I refer to this suffix as -SJA, and it is intended that both phonological variants be included. Babby (forthcoming - Lingua) has demonstrated, however, that -SJA cannot be considered a reflexive because it arises fron sources other than accusative objects which are coreferential with the subject. In his article Babby discusses the fact that although is not the case that

myt'

ubit'

and

sebja sebja

myt'sja

can both mean 'wash oneself', it

'kill oneself' has the same meaning as

'get killed'. There is no reflexive "meaning" at all associated with

ubit*sja

ubit'sja

and with many other SJA-verbs. Therefore, -SJA cannot be said to be a reflexive. I present Babby's analysis in chapter 3. Reflexive

Pronouns

English: pers.

first

second

third

sing.

myself

yourself

him-/her-/it-/oneself

ourselves

yourselves

themselves

first

second (formal)

jplur.

pers. | case

dat. acc.

second (familiar) datacc.

Ising.

mir

dir

plur.: Russian: case

mich : uns

dich

third

sich

euch

gen./acc.

dat./prep.

sebja

sebe

, instrumental soboj

A comparison of the above three diagrams reveals certain points of contrast and of similarity among the three languages. These three languages form a

usually listed as separate cases. The accusative and genitive case forms are the same in certain pronouns — masculine and feminine singular and the plural, which does not inflect for gender. Within the noun system, these two case forms are the same for animate and masculine singular nouns and for animate masculine and feminine plural nouns.

5 progression along the continuum of "overt expression of referent information" in the reflexive pronoun:

English gives the greatest amount of information,

since it is the only one to reflect gender and number in the third person, Russian the least amount, since it does not distinguish person, and German 4 stands between the tνιο.

There are also certain relations among these three

languages in types of marking.

Gerrran in the formal second person and the

third person is similar to Russian in that there is no indication of gender and number, but German is similar to English in the first persons and the familiar second persons in that number is expressed. 1.2.2

SICH-verbs and SJA-verbs to be studied.

Not all SICH/SJA-verbs are relevant to a study of syntactic Derived Intransitivity.

Certain verbs are lexically, and not syntactically, related to the cor-

responding verbs without SICH/-SJA (e.g. r e c i p r o c a l s ) M u c h of the confusion in previous treatments of SICH/SJA-verbs arises through an insistence on treating all these verbs in one section (e.g. see Vinogradov 1947:629-639). The first criterion which a SICH/SJA-verb must have is that there is a corresponding transitive verb without SICH/—SJA (see section 1.2.3). automatically excludes certain SICH/SJA-verbs:

This

1) those which are never found

without SICH/-SJA such as sich erholen*recover, get well', sich schämen 'be ashamed', bojat 's.ja 'be afraid', smejat'sja 'laugh';

2) those which have

intransitive equivalents without SICH/-SJA such as (sich) irren 'err, be mistaken', (sich) leben 'live' (see section 4.2 for a discussion of this type of verb), belet' (sja) 'shew white';

3) those which simultaneously add a prefix

with SICH/-SJA such as schreiben - sich verschreiben 'write - nrake a mistake in writing', meStat' - zameZtat'sja 'dream - lose oneself in dreams'. There are also certain other SICH-verbs which I have excluded because they involve other syntactic phenecnena simultaneous with the addition of SICH.

One

of these phenomena is the addition of an adjective such as in essen - sich satt essen 'eat - eat one's fill, be satisfied'. exclusively associated with SICH.

The addition of adjectives is not

Erben (1959:111) gives seme examples; one of

4. An observation from Isacenko (1968:465) is relevant here. He lists several German verbs which have a dative reflexive pronoun with an accusative noun. Their Russian equivalents, however, do not have this dative reflexive. Here there is another instance of more person marking in German than in Russian. 5. Reciprocal verbs are complex enough to form the topic of a separate dissertation. I touch on them briefly in section 4.8.

6 than is weinen - die Augen rot weinen 'cry - make one's eyes red with crying'. Another phenomenon involves the relation between two usages of a given verb, one with a dative reflexive pronoun and accusative object, the other with just the accusative reflexive pronoun, as in (2). (2)a

Ich wasche Ί

b

mir (dat.) die Hände.

am-washing my

Ich wasche 'I

mich

am-washing

hands.'

(acc.).

myself.'

This too is a phenomenon which is not exclusively associated with reflexive pronouns.

The example fron Erben (1970:121) shows that the same alternation

between dative and accusative of the person occurs with nouns: (3)a

Man bindet dem Verbrecher (dat.,I die Hände (acc.) auf den Rücken. 'They bind the criminal's hands behind his back.'

b

(acc.).

Man bindet den Verbrecher 'They bind the criminal.'

Additionally, since I am considering only syntactic functions or SICH/-SJA, there must be no change in lexical meaning between the verb without SICH/-SJA and the verb with SICH/-SJA.

Since there is no agreement as to what lexical

meaning encompasses, I treat it at length in a separate section. 1.2.3

Lexical meaning.

Not all meaning differences are differences in lexical meaning and linguists do not agree as to which meaning differences are lexical and which are not.

In

this section I give sane meaning differences which I consider not to be differences in lexical meaning.

In general, I adhere to a view of lexical meaning

which posits as few lexical items as possible. Many meaning changes associated with different uses of a verb I regard as the result of different interpretations of that verb in different syntactic configurations. Fillmore (1968:27-30) makes a similar distinction.

He observes that meaning

differences which are lexical are differences in semantic characterization of the verb and non-lexical meaning differences are differences in semantic interpretation of the verb in given contexts.

Brinkmann (1971:198-199) also speaks

to this issue in that he mentions the difficulty of distinguishing between Inhaltswert 'content-value1 of a verb and its Satzwert 'sentence-value'. wert is relevant to a verb's semantic interpretation. lists the following things:

Satz-

Under Satzwert Brinkmann

1) finite vs. non-finite, 2) relationship between

subject and predicate, and 3) valence.

Both subject-predicate and valence

7 relations are manifested in syntactic configurations.

(As we shall see in the

other chapters, these are relevant for a syntactic analysis of SICH/SJA-verbs.) Fillmore illustrates the distinction between semantic characterization and semantic interpretation with the verb cook.

He identifies certain semantic

cases such as Agent, Object, Instrument, Dative and a few others.

A verb's

lexical entry will include the case frames into which the verb can be inserted. In the case of cook, Fillmore gives three sentences to shew different case frames into which cook can be inserted: (4)a

Mother is cooking the potatoes.

[

Ο A]

b

The potatoes are cooking.

[

0]

c

Mother is cooking.

[

A]

Sentence (4c) demonstrates that cook is one of those verbs in English which allcw the object to be deleted Wien it is a "typical NP for the verb" (Fillmore 1968:29; see my section 3.4.3). (5)

cook:

[

Fillmore's case frame for cook is given in (5).

Ο (A) ]

Fillmore then writes: The semantic description of the verb will do no more than identify a particular activity having a result of a particular kind on the object identified by the Ο element. The same semantic entry, in other words, will account for the use of cook in all of the sentences [in (5)]. Instead of saying that the verb has three different meanings, we can be satisfied to say that there is a certain variety in the case frames which accept it, and that it is one of the 'deletable object' verbs ... The example with cook shows that the lexicon need not contain as many semantic entry tokens under the present proposal as it would in a subject/object grammar. It will now be shown that this same flexibility makes it possible to reduce the number of semantic entry types, for now it is feasible to show that some syntactically different words are in fact semantically identical (with respect to that aspect of their meanings which is independent of the contribution of the associated cases). (Fillmore 1968:29-30) Lyons (1968:383) agrees with Fillmore's analysis when he writes that the semantic feature CAUSATIVE is somehow added when a verb occurs in a certain configuration.

Bcwers (1973:198-200) points out the difficulties incurred

when this feature (which he labels CAUSE) is included as part of the lexical entry of verbs, such as burn.

Therefore, he also analyzes CAUSE as a feature

which is not part, of a verb's lexical entry. Recently, Dieter Kastovsky (1973) has taken issue with Fillmore's analysis and questions Fillmore's ability to account for different semantic interpretation of basically transitive verbs, e.g. read, and of causative-transitive

8 verbs, e.g. cook, on the basis of syntactic structure alone.

Kastovsky claims

that Fillmore is not correct because Kastovsky believes that there can be no feature addition to a verb without a corresponding phonological change.

Thus,

read and cool·.: cannot sometimes carry the feature [+CAUSATIVE] and sometimes not carry this feature.

I reject Kastovsky1s assumption.

Another linguist who disagrees with the position taken by Fillmore and Lyons is Korolev.

Korolev (1969:212) specifically states that the feature difference

CAUSE between otkryt' 'open (transitive)1 and otkryt's,ja 'open (intransitive)' constitutes a difference in lexical meaning.

I note here that Korolev's line of

reasoning can be further pursued in regard to various other verb features.

The

sentence pair in (6) (from Lyons 1968:365) gives two usages of the verb fly. (6) a b

The bird flew through the air. The stone flew through the air.

It could be argued, ä la Korolev, that there are two different lexical items fly in English and that they differ only in that one of them has the feature Agentive which the other does not have.

However, Jespersen (1927:332ff) lists a number of

verbs just like fly and classifies verbs of this type as the "move and change class" verbs.

He notes that each verb paix of this class has two meanings:

to produce a movement or change in something; undergo that change.

1)

2) to perform that movement or

This is a productive class of English verbs, and to list

them separately in the lexicon is to miss an important generalization about these verbs.

Presumably, fly and also otkryt '/otkryt 'sja should be entered as only one

item each in the lexicon; the meaning differences will then be a matter of semantic interpretation of the lexical item in conjunction with the different contextual features. It is interesting to note that Korolev and IsaCenko cannot agree as to which SJA-verbs are separate lexical items and which are not.

Korolev (1969:212) lists

under verbs which do not shew lexical meaning difference the verb uVit'/u&it'sja ' teach/leam'.

IsaCenko (1968:453) states emphatically that these are two

separate lexical items. I wish to nake clear that I do believe there is sane meaning difference in the sentences in (4) as well as in those in (7). (7)a

b

Dver' otkrylas'. Die-Tür öffnete-sich. 'The-door opened. 1 Anton otkryl dver'. Anton öffnete die-Tür. 'Anton opened the-door.'

9 However, I agree with Lyons, Fillmore and Bowers, that the meaning difference is the result of different semantic interpretations.

In other words, according to

this view there is no difference in the semantic characterization of open, otkrytöffnen

in (7a-b), only a difference in semantic interpretation.®

Whether or not the reader agrees with my judgments, it should new be clear that part of the subject matter of this study will be those SICH/SJA-verbs for which it might be claimed that they differ from their transitive correspondents by the feature CAUSE.

Therefore, I have not used the criterion of same lexical

meaning as the sole definition of SICH/SJA-verbs to be covered here but have based my selection on the syntactic criterion of Derived Intransitivity. 1.2.4

Derived Intransitivity.

This term, coined by Babby (forthcoming - Lingua) to discuss SJA-verbs, is the label which I use for SICH/SJA-verbs which are derived from transitive verbs syntactically, that is, they involve no change in lexical meaning (see sections 3.2.1, 3.3.1, 3.4.1 for examples).

The intransitivity of SJA-verbs has been

recognized as the only feature which these verbs have in ccmmon (see the Academy Gramtar [AG] 1960:413; the names of several Soviet grammarians who have made this observation can be found in Janko-Trinickaja 1962:57-59).

The intransitiv-

ity of SJA-verbs is therefore not in question. There is a different situation in regard to SICH-verbs. and not an affix as is -SJA. pronoun objects.

SICH is a clitic

This means that SICH is syntactically similar to

Stötzel (1970:21-28) discusses the question as to whether SICH

is always a full direct object, full in the sense that it has the same syntax as other pronominal objects.

He quotes sentences frcm various German graititBrians

to show that they are divided on this issue.

For example, Erben (1959:165)

analyzes SICH as a proper syntactic object: (8)a b

Fritz entrüstet-sich. 'Fritz became-angry.' Katzen fangen Mäuse. 'Cats catch mice.'

He states that both (8a) and (8b) are sentences which involve a verb and two

6. This view underlies Babby (forthcoming - Lingua) and Babby and Brecht (1975) In Russian the SJA-form is used for imperfective middle and passive verbs. Babby and Brecht 1 s analysis uses semantic interpretation of the elements in the phrase marker to determine whether the verb is middle or passive. Grossly stated: if there is an animate instrumental NP, then the sentence is passive; otherwise it is middle.

10 NP's.

Duden (1966:70,471), however, states that sentences such as (8a) contain

intransitive verbs despite the presence of the accusative reflexive pronoun. Duden gives the following examples of intransitive verbs: (9)a

Ich schäme-mich. Ί am-ashamed.'

b

Ich beeile-mich. Ί am-hurrying.'

c

Karl freut-sich. 'Karl is-glad.1

d

Der Hund bellt. 'The dog is-barking.'

and claims that (9a-c) are intransitive just as is (9d) . 1.2.5

Summary paradigm of SICH/SJA-verbs to be studied.

Belcw are listed some examples of the types of verbs whose syntax I propose to explain in this work. tions.

They are classed according to subject and object rela-

Not all the verbs listed here are Derived Intransitives (see section

3.4.3 for conments concerning beißen, bite, waschen, wash), but in each set the verbs from at least one language are. I.

verbs in which the subject of the intransitive verb is identical with the object of the transitive verb:

a)

verbs which have unrestricted usage (Russian and German mark Derived Intransitivity): English - open Russian - otkryt'/otkryt'sja German - öffnen/sich öffnen

b) verbs in which the subject of the intransitive verb is identical with the subject of the transitive verb: English - melt Russian - plavit'/plavit'sja German - schmelzen c)

verbs which have restricted usage (require an adverb): English - read Russian - citat'/citat'sja German - lesen/sich lesen

II. a)

verbs in which the subject of the intransitive verb is identical with the subject of the transitive verb: verbs which mark Derived Intransitivity in all languages: English - express/express oneself Russian - vyrazit'/vyrazit'sja German - ausdrücken/sich ausdrücken

11

b)

verbs which mark Derived Intransitivity in Russian and German: English Russian German -

c)

verbs which are Derived Intransitive in Russian: English Russian German -

III.

move dvinut'/dvinut'sja bewegen/sich bewegen

bite kusat'/kusat'sja beißen

verbs derived from deep structure reflexive constructions: English Russian German -

wash (oneself) myt'/myt'sja waschen/(sich)

waschen

1.3 Methodological considerations. 1.3.1 Contrast!ve analysis and linguistic investigation. Contrastive analysis, as it has developed, is a tool to aid in language pedagogy. It is a means by which the language teacher can predict student errors and determine which areas to devote more teaching time to. In this use of contrastive analysis two languages are compared: the student's own language and the language to be learned. Such a comparison reveals where the two languages differ and where they are the same. The teacher devotes time to where the languages differ and does not worry so much about where they are the same. No more than two languages are ever compared at any one time since in the usual language-learning situation only two languages are involved. In this work I use contrastive analysis differently. Like the language teacher, I am interested in where languages differ and where they are the same. But unlike the language teacher, I am not chiefly concerned with differences and I do not have to be restricted to only two languages at a time. I use contrastive analysis to supplement the information obtained fron analysis of the individual languages. There is a parallel between transfornational studies and those of historical linguistics.· Both analyses are concerned with selecting base forms from which other forms can be derived. This fact became particularly evident during the 1960's when objections to generative phonological studies arose because they seemed to be synchronic recapitulations of diachronic processes. I am not concerned with the validity of these objections. I only wish to point out that a certain parallelism between the two methodologies does exist. It is this parallelism which led me to recognize that contrastive analysis could have a role in synchronic transformational investigations similar to that of comparative

12 reconstruction in diachronic studies. Historical linguistics uses a two-part methodology to determine its base forms.

One part is internal reconstruction.

language.

This draws on data from only one

The other part is comparative reconstruction.

from more than one language.

This draws on data

Internal reconstruction alone cannot provide all

the answers to questions of base forms (earlier stages of the language).

His-

torical linguistics uses comparative reconstruction to bring in more information to give answers to sate of these questions.

Comparative reconstruction supple-

ments internal reconstruction. I believe that synchronic linguistic investigations can also be conducted with a two-part methodology:

one-language analysis and contrastive analysis.

And just as the assumption of genetic relationship legitimizes the use of information from one language in another, so the assumption of linguistic universals legitimizes information fron one language being applied to help analyze another. An excellent example of transforirational grammar as a synchronic parallel to internal reconstruction is found in Bowers (1973:129-136).

In this section on

Indirect Objects, Bowers examines occurrences of annoy (i.e. as verb, as adjectives annoyed and annoying, and as noun annoyance) to determine its base subcategorization.

I present this work as an example of contrastive analysis used

to supplement one-language transformational analysis. 1.3.2

Choice of theoretical framework.

I wish to draw the reader's attention to the list of correspondence sets given in section 1.2.5.

Sets I and II are distinguished by NP-verb relations.

The

verbs in set I have their intransitive subjects identical with their transitive objects.

However, the verbs in set II have their intransitive subjects identical

with their transitive subjects.

Whatever theory is used to analyze the problem

of Derived Intransitivity, it must be able to formalize both of these NP-verb relations simultaneously. Lyons (1968:380-388) notes that most transformational work (up until 1967) uses syntactic formulations which equate all subjects of intransitive verbs with the subjects of transitive verbs.

The Standard Theory, as given in Chomsky

(1965), is most representative of the theories which Lyons is refering to. can characterize properly only the verbs in set II.

It

Chomsky himself notes his

inability to harrlle verbs of set I and specifically mentions break (Chomsky 1965:214). Studies by post-Fillmorian generative-senanticists dealing with NP-verb

13 relations of both sets above, are not rigorously enough formalized yet and do not have a satisfactory treatment of actual lexical items in the languages. They have concentrated on stating relations between semantics and surface structure in as general a way as possible-

This has led them to posit pre-

lexical syntax (McCawley 1971) and to hypothesize that languages differ not so much in syntax as in lexicon.

The result has been a loss of appreciation for

the fact that each language has its cwn syntactic structuring. Case studies (on the model of Fillmore 1968) are also not satisfactory. There is general agreement about certain semantic cases, but there is no one set of cases which has been accepted by most linguists. And there is no formalism for relating these semantic cases to syntactic structures. There does exist a theoretical framework which combines a rigorous formalism of the caliber of the Standard Theory with certain insights of the generativesemantic ist approach and is thus able to treat all Derived Intransitive verbs with the same formlism.

This theory has been developed by John Bowers (1973

and ms.) and is a combination of Chomsky's work (1965, 1970) and that of Fillmore (1968), both fitted together within the Structure-Preserving Hypothesis proposed by Emonds (1970).

I have chosen this fairly rigorous syntactic frame-

work and will explain it in chapter 3, after I have reviewed the literature in order to better acquaint the reader with the background of the problem of SlCH/SJA-verbs. 2. 2.1

BACKGROUND Previous classifications of SICH/SJA/SELF-verbs.

In this section I wish to review certain previous treatments of these verbs.

The

SELF-verbs have not been as extensively treated by English granmarians as have the SICH/SJA-verbs.

In fact, the status of -SELF as a marker of Derived Intran-

sitivity is questionable (see section 3.4.4), and the comparison of English with German and Russian indicates that Derived Intransitivity is normally not morphologically marked in English. Russian grammarians have devoted much attention to discussions of hew to analyze SJA-verbs.

In the last century the discussions centered around SJA-verbs

and voice; this culminated in the article on Russian voice by Fortunatov (1899). During this century discussion has involved "meanings" of these verbs; this tradition was established by Saxnatov (1925, 1941) . German granmarians, while concerned with the syntax of SICH-verbs, have not devoted attention to questions of these verbs and voice or of "meanings" of these verbs.

14

My review of previous treatments of SICH/SJA/(SELF)-verbs proceeds fron a consideration of a narrow range of phenomena.

In this work I am concerned with

Derived Intransitivity, that is, with what happens when the object-NP of a verb is empty, when it is either not filled by lexical insertion or by transformation or its filler is deleted.

I wish to stress, therefore, that my assessment of

earlier analyses of SICH/SJA-verbs is not a once for all statement.

These

analyses may be valid when these verbs are studied from other points of view. In giving a background against which chapter 3 can be better understood, I am not interested in presenting a detailed picture of how Russian and German gramtiarians have differed among themselves on the analysis of SICH/SJA-verbs. Treatments of previous SJA-verb analyses can be found in Vinogradov (1947:607629) and in Korolev (1969), and SICH-verb analyses in Stötzel (1970, chapter 5). For a classification of SJA-verbs by a non-Russian see Margulies (1924). What is important for a contrastive analysis analysis of SICH/SJA-verbs is an awareness of the gross differences between the general Russian approach to SJA-verbs and the general German approach to SICH-verbs.

It seems to me that

the different approaches used by these two groups is the result of two things: 1) the morphological difference between German SICH-verbs, in which SICH is detached fron the verb and acts like other clitic pronouns, and Russian SJAverbs, in which -SJA is attached to the end of the verb and acts like an affix; 2) the fact that in Russian, SJA-verbs are used to form passive constructions for imperfective verbs (see (8) at the beginning of section 2.3), whereas in German SICH-verbs are not found in passive constructions, although they are said to be "passivnah" 'near-passive' (see Brinker 1969:1). The major result of the morphological difference is that SJA-verbs have been given special voice labels, restricted to only those verbs (e.g. "middle", or "reflexive" (AG 1960:413)) whereas SICH-verbs either are considered active voice because of their morphological sameness to active verb forms (Duden 1966:474) or are excluded from having voice because they are syntactically different from active and passive verbs (Brinkmann 1971:205; see section 2.3 below).

Another

result of the morphological difference is that SJA-verbs are said to be intransitive (AG 1960:413; Isaienko 1968:453) while there is disagreement about the transitivity status of SICH-verbs among German grammarians (Erben 1959:30 and Duden 1966:70 consider them intransitive, but Erben 1970:45 considers them transitive). The major emphasis of classifications of SJA-verbs has been on their "meanings" (see especially Vinogradov 1947:629-637) . Although many Russian gramnarians have noted the fact that some SJA-verbs are related to transitive verbs

15 without -SJA, others related to intransitive verbs without -SJA, and others are not related to any verb without -SJA, these relationships have not been treated syntactically but have been integrated into a classification by "meanings". A striking exception to this general Russian trend is found in IsaCenko (1968:453-456).

IsaCenko seeks a more syntactic classification of SJA-verbs.

He distinguishes between "Reflexiwerben" and "Reflexivformen" of verbs. "Reflexiwerben" are SJA-verbs which do not occur in the same meaning without -SJA.

He gives the example of uSit'sja 'learn' (versus uSit' 'teach') and

radovat'sja 'be happy' (versus rado vat' 'make h a p p y " R e f l e x i v f o r m e n " are SJA-verbs which are either imperfective passives (as in 1) or irrpersonal constructions (as in 2). (1)

Dom stroitsja. '(The)-house is-being-built.'

(2)

Mne ne rabotaetsja. to-me not working Ί don't feel like working.'

It is interesting to note that the term "reflexive form of a verb" is used to mean different things by IsaCenko and by the Academy Grammar.

In the Academy

Grairrar "reflexive form of a verb" is used to designate verbs which are not found in the language without -SJA; one such verb is bojat'sja 'be afraid*. IsaCenko lists bojat'sja as a sub-type of his Reflexiwerben. IsaCenko's analysis of SJA-verbs is much closer to the fornat of the majority of Gerrtan analyses of SICH-verbs. into twD groups.

Most Gernan analyses divide SICH-verbs

One group is characterized by the fact that its verbs cannot

occur without SICH.

The verbs in the second group may occur without SICH.

Different analyses, however, do not agree on what is meant by being able to occur without SICH. Curme (1952:330) speaks of verbs that are "usually" reflexive (such as siah schämen 'be ashamed', and sieh freuen 'be happy') versus verbs which are "regularly" transitive or intransitive (such as (siah) baden 'bathe (oneself)'). Duden's distinction between "echte reflexive Verben" 'true reflexive verbs' and "unechte reflexive Verben" 'not genuine reflexive verbs' and Stötzel's distinction between "reflexive Verben" and "partimreflexive Verben" 'partly nonreflexive verbs' parallel Curme's distinction. When I speak of being able to occur without SICH/-SJA (see section 1.2.5),

1. I do not necessarily agree with Isacenko as to what constitutes different meaning.

16

I mean something quite different than IsaCenko, Stötzel or Duden. Stötzel probably has the clearest test of what he means by his distinction of reflexive and partly non-reflexive verbs.

His test does involve occurrence of a verb

without SICH. However, when performing this test, he requires that the subject remain constant. Thus, within the reflexive verb category he not only includes verbs such as sich schämen 'be ashamed' (see 3), for which there is no other use of * schämen, but also verbs such as sich öffnen 'open (intrans.)' as well (see 4). (3)a

sich. Das Kind schämte { *ihn. •seinen Freund. 'The child was-ashamed.'

b

*Der Fehler schämte das Kind. the mistake was-ashamed the child

(4)a

b

sich. Die Tür öffnete { *sie. *das Fenster. •The door opened.' Fritz öffnete die Tür. 'Fritz opened the door.'

The Russian verb-type correlate to sich schämen is represented by bojat's,ja 'be afraid1, for which there is no use of *bojat'. And the verb-type correlate to sich öffnen is otkryvat'sja/otkryt'sja 'open1. Although Stötzel employs a very strict, easily verifiable criterion to distinguish his reflexive verbs from his partly non-reflexive verbs, he rasks the distinction between sich schämen and sich öffnen since he does not concern himself with the difference between these two verbs as shown in the ^-sentences above (see my section 3.3.1). Stötzel, Isaienko and other gramnarians support their observations based on examples like (3a) and (4a) with a much narrower definition of "same meaning" than I use.

I discussed this topic earlier in section 1.2.3.

Earlier in this section I pointed out the fact that many analyses of SJAverbs are based on different "meanings" of these verbs. The question arises as to what is responsible for these different meanings. Vinogradov (1947), and to a great extent Margulies (1924), locates these different meanings in -SJA itself. The following quote is from Vinogradov (1947:648): With reflexive verbs, which are produced from transitive verbs, different meanings of -sja very often are combined in the structure of one and the same word. [U vozvratnyx glagolov, kotorye proizvedeny ot perexodnyx glagolov, raznye znaCenija -sja o5en' iasto sovme5£ajutsja ν struktur odnogo

17 i togo ze slova.] Leonard Babby (forthcoming - Lingua, section 8.3.1) has pointed out that to locate all meaning differences in -SJA is counter-productive.

It leads to an

open-ended classification of SJA-verbs, since a different class of verbs will have to be posited for each different NP which -SJA is related to.

IsaCenko

(1968:456) states a similar opinion: After all, it is highly improbable a priori that a single formal element, namely the reflexive ending -sja, could/should have so many different grammatical functions. [Schon a priori ist es höchst unwahrscheinlich, daß ein einziges formelles Element, nämlich die Reflexivendung -sja., so viele verschiedene grammatische Funktionen haben könnte.] He goes on to conclude (same page):

2

that the so-called "meanings" of reflexive verbs are not connected with the reflexive element -sja, but with the concrete lexical meaning of the reflexive verb itself. [daß die sogenannten "Bedeutungen" der Reflexivverben nicht mit dem reflexiven Element -sja, sondern mit der konkreten lexikalischen Bedeutung der Reflexiwerben selbst zusammenhängen.] Vinogradov (1947) has a statement in his grainier ( - it could be seen to contradict his statement which I quoted above), which is a slightly different claim than IsaCenko makes.

On page 630 Vinogradov states:

The fact of the matter is that meanings and nuances of the affix -sja depend on lexical meanings of those verb bases to which this affix is joined. [Delo ν torn, £to znaCenija i ottenki affikxa -sja zavisjat ot leksi£eskix znaCenij tex glagol'nyx osnov, k kotorym ötot affiks prisoedinjaetsja.] He does not seem to be aware of any contradiction between his two statements, later, he opts for a lexico-sanantic analysis rather than a syntactic one (1947:648): It is necessary to outline methods and norms of combination and matching of various reflexive meanings in the semantic system of one and the same word. [Neobxodimo oiertit1 priemy i normy soCetanija i sovmesSenija raznyx vozvratnyx znaCenij ν semantiSeskoj sisteme odnogo i togo ze slova.] One of the fundamental premises of the theory I am using in this work is that much of what has been termed semantic in the past can be shown to be

2. I wish to remind the reader that I am claiming that -SJA is not a reflexive element at all and that SJA-verbs are not reflexive verbs.

18

syntactic (see, far example, Fillmore 1968). Babby (forthccming - Lingua) has already given a demonstration of this for SJA-verbs. I am following his general conclusions and with him I claim that irany of these different "meanings" of -SJA are not due to different -SJA1 s or to the semantics of the transitive verbs involved or to the semantics of the verbs which are formed once -SJA has been attached to a verb. The different meanings of SJA-verbs result from different syntactic processes which accanpany the introduction of -SJA into the phrase marker (see chapter 3). The following question seems reasonable to ask: why have German grammarians not made a similarly elaborate SICH-verb classification by meanings? It seems to me that the answer lies in the fact that SICH-verbs do not form a special morphological category as do SJA-verbs. Since SICH is a clitic, it is not attached to the verb and its analysis has been kept separate frcm the analysis of verb morphology traits. English SELF-verbs are similar to SICH-verbs in this respect. This similarity coupled with the small number of occurrences of SELF-verbs is a good explanation of why SELF-verbs have received even less treatment than have SICH-verbs. 2.2 Transitivity. Traditional grarrmarians classify verbs according to whether or not they do take a direct object (i.e. accusative NP in German and Russian) . The German and Russian gramnarians have differed somewhat as to the status accorded to SICHverbs and SJA-verbs, respectively. In general, scholars of Russian classify SJA-verbs as intransitive. The Academy Grammar (1960:413) states that the reflexive-middle voice (which does not include all of the SJA-verbs, however) renders verbs intransitive. Vinogradov (1947:630) states that there are two general meanings of -SJA: 1) to show loss of transitivity, 2) to show strengthening of intransitivity. IsaCenko (1968:446) writes that all SJA-verbs are intransitive. However, he also states (pp. 444-445) that transitivity/intransitivity is not an inherent property of the verb, that it is a capability, a potential property, which is given by the context.3 3. A few pages later (p. 453) he seems to make a contradictory claim in his argument against analyzing uSit' 'teach' and unit's ja 'learn/study' as the same lexical item. "Somit kann ein Verb wie ucit'sja schon deshalb nicht als n u£it' sebja" gedeutet werden, weil u5it' transitiv, uSit'sja intransitiv ist." This argument makes it clear that no SJA-verb is to be syntactically related to the corresponding verb with sebja. Later in the same

19 One exception to the general analysis that SJA-verbs are intransitive is found in Schaarschmidt1s work (1968, 1970), which is based on the hypothesis that -SJA is the weak form of sebja, the reflexive pronoun (1968:26) . Accordingly, he then describes all non-passive SJA-verbs as transitive.

On pages

21-23 he discusses seme of the differences between pisat' 'write* and zapisyvat' 'copy' - zapisyvat'sja 'enroll'.

Because (5) (below) is grammatical but not

(6a), he concludes that pisat' but not zapisyvat' is intransitive. (5)

On xoroso piset. he well writes 'He writes well.'

(6)a

*0n zapisyvaet na letnyj kurs. he is-enrolling for winter semester

b

On zapisyvaetsja na letnyj kurs. 'He is-enrolling for (the)-winter semester.'

He analyzes sentences with zapisyvat'sja as containing objects in order to account for their grammaticality.

Later in his dissertation he postulates an

underlying structure with subject-object identity for (6b). The passive SJA-verbs are not included in Schaarschmidt1s analysis, since he cannot relate them to the non-passive SJA-verbs within his theory. Within the German tradition there has not been as much unanimity that SICHverbs are intransitive as there has been about SJA-verbs. Curme (1952:331) talks about SICH-verbs having intransitive force, but his wording leads the reader to believe that SICH-verbs are in fact transitive after all.^

The main

thesis of Stötzel's work (1970) is that many SICH-verbs are syntactically transitive but semantically intransitive (p. 216).

I will return to a discussion

of whether transitivity is a semantic or syntactic notion shortly. Stötzel is aware of the confusion about SICH-verbs and devotes a chapter to sorting out the reasons for this confusion. about these verbs.

He reaches a definite conclusion

There are other granmarians, however, who have taken differ-

ent stands at different times.

One of these is Erben.

In one of his recent

works (Erben 1970:45), he classifies SICH-verbs as a subcategory of transitive

chapter there is a two-page section devoted to comparing SJA-verbs versus verbs plus sebja (pp. 463-464), in which IsaSenko maintains that it is only exceptionally the case that an instance of these two forms can have the same meaning. 4. Curme (1952:331): "In another group [i.e. different from verbs such as sich kleiden 'dress oneself'] reflexive form has a force closely related to intransitive use, so that the subject is not thought of as acting on himself." He gives the following examples: sich erholen 'recover/get well', sich bessern 'get better/improve', sich zeigen 'appear/become evident'.

20 verbs. Earlier (Erben 1959:30, 139) he stated that using SICH was one way to make a verb intransitive. There are other grammarians who analyze SICH-verbs as intransitive. Duden (1966:70) states that "echte reflexive Verben" are intransitive. These are verbs such as sich schämen 'be ashamed' and sich sorgen 'be vrorried'. The "unechte reflexive Verben" are considered to be transitive. An example of an "unechtes reflexives Verb" is sich waschen 'wash oneself'. My own analysis claims that sich sorgen is derived fron a transitive verb and is therefore not intransitive in the same way as sich schämen. Within the English tradition Poutsma (1926) gives a rather thorough treatment of SELF-verbs. Concerning verbs which are sometimes used with and seretimes without the reflexive pronoun, he states (p. 14, and again p. 54) that these verbs are intransitive when the reflexive pronoun is emitted. He too, as Stötzel, speaks of syntactic versus seirantic transitivity in describing SELFverbs and other types of verbs which may emit the direct object (e.g. read, eat) . In a later section (p. 143) he speaks of transitive and intransitive reflexives. There are gramtarians who do not class verbs according to the transitiveintransitive division. Otto Jespersen, for example, preferred to speak about transitive arid intransitive uses of a verb in English (Jespersen 1927:319). Kastcrvsky (1973:303) makes a similar claim. For him the transitive and intransitive uses of jump are part of word-formation and are not syntactic.-' In analyzing the verbs of German, Rupp (1965:183-185) uses a five-way verb-class division which he says makes the transitive-intransitive analysis untenable. This division differentiates among different types of transitive verbs but has nothing new to contribute to the question of the place of SICHverbs, although they would presumably be fitted into several of Rupp's classes as follows: sich schämen 'be ashamed' into the same class as einen Tod sterben 'die a death' because both verbs can take only one specific object; sich waschen 'wash oneself' into the same class as essen 'eat' because the object does not have to expressed in the surface structure; sich kleiden 'dress oneself' into the sane class as geben 'give' because the object must be stated in the surface structure. The confusion about SICH/SJA-verbs and the question of transitivity leads to the question of the nature of transitivity. Lyons (1968, chapter 8:350-371)

5. This is opposite to Bowers base of my analysis here.

(1973:213-217), whose syntactic theory is the

21 correctly points out that transitivity is not a semantic notion after all (above, I noted that Stötzel and Poutsma do refer to semantic transitivity). He states that the question "What does X do?" can be answered with either a transitive verb plus object or an intransitive verb (Lyons 1968:341).

Later he

quotes fron Robins (1964:266) to shew that the "notional" view of transitivity (which I understand to be similar to semantic intransitivity) just does not square with the facts of language (Lyons 1968:350-351). In this work, therefore, I am following Bowers (1973) and Babby (forthcoming - Lingua) in my use of the term transitivity.

It is a syntactic term.

A verb

is transitive if there is an object-NP in its subcategorization feature.

It can

be the case that this object-NP is vacated in the derivation of an utterance frctn its underlying structure.

When this happens the verb becomes intransitive.

For example, the verb open is a transitive verb because it requires an object in its base form. (7)a b c

John opened the door, The door opened. *John opened.

The sentences in (7a) and (7b) differ in their underlying structure only in that (7b) has no Agent but (7a) does. object of the verb.

Both of them have the door as the underlying

It is only in the derivation of (7b) that:

vacates the object-NP, being moved into the subject-NP, and

1) the door

2) the vacated

object-NP deletes. The relationship posited in English between (7a) and (7b) holds for the corresponding German and Russian sentence pairs, as will be made clear in chapter 3.

English differs fron German and Russian in what happens to the

vacated object-NP.

German and Russian mark a vacated NP by SICH/-SJA, respec-

tively; English does not possess a morpheme with this function.

And this fact

is the major reason I have not devoted more time to SELF-verbs. 2.3

Voice.

Gernan, Russian and English differ in the number of voices which are generally ascribed to than. sive.

Both English and German have two voices —

active and pas-

The passive in both languages is formed with the use of an auxiliary {be

in English, werden 'become' in German) and the past participle. the other hand, has three voices —

active, passive, middle.

Russian, on

(Not all grantna-

rians agree to this, and I present seme of their opinions below.)

The middle

voice is what I refer to as Derived Intransitivity; it includes certain SJAverbs.

Certain English intransitive verbs and certain SICH-verbs can also be

22 considered middle voice. The Russian passive voice is rrost often formed in tvro ways, depending on the aspect of the verb.

The verbs of perfective aspect form the passive similarly

as English and German, i.e. with the use of an auxiliary (byt' 'be') and the past participle. belcw.

The imperfective passive is formed with SJA-verbs, as in (8a)

This passive and the middle both use the same verb forms.

In Russian, then, SJA-verbs are said to have passive voice

(8a), middle

voice (8b-c), or not voice (8d) (AG 1960:413). (8)a

Dver1 otkryvalas' Antonom. '(The)-door was-opened by-Anton.

b c

Dver' otkryvalas'. Dver' otkrylas'. '(The)-door opened.'

(imperfective) (perfective)

d

Devu5ka boitsja sobaki (gen.), '(The)-girl is-afraid of-(the)-dog.'

Middle SJA-verbs differ from passives in that there is no Agent; hence the ungranmaticality of (9).^ (9)

*Dver' otkrylas' Antonom. *Tür öffnete-sich von-Anton *door opened by-Anton

The category of voice is a confused area of grammar.

Helbig (1968:132)

notes that different G e m e n authors have used different criteria (and have even mixed criteria) in determining and defining voice.

Lyons (1968:372) writes:

The Greek grammarians failed to appreciate the true nature of the distinctions marked by the verbal inflexions which they referred to as 'active', 'passive' (i.e. signifying the state of 'being acted upon' or 'suffering the effects of the action') and 'middle'; and this has left us with a legacy of contradictory statements about the role of voice, not only in the classical languages, but also in English and other modern languages, the description of which has been strongly influenced by traditional grammar.7 There have been three main sets of criteria which have been used in defining voice:

morphology (e.g. Aksakov 1875), NP-verb relations (see Korolev 1969 for

6. There is at least one usage of these verbs with an Instrumental pronoun: (i) Dver' otkrylas' sama-soboj. Die-Tür öffnete-sich von-(sich)-selbst. The-door opened by-itself. However, these sentences do not contain Agents. In fact, they are used to convince the hearer that there is no Agent. 7. McKerrow (1922:163): "If we were now starting for the first time to construct a grammar of modern English, without knowledge of or reference to the classics, it might never occur to us to postulate a passive voice at all."

23

a list), functions/meanings/uses (see Halliday 1967-68 and Grimes 1975, chapter 17). A review of Vinogradov (1947:606-629) and Korolev (1969:199-211) presents convincing evidence to support the above statements by Helbig and Lyons. Brinkmann's definitions of active and passive voices as "Wirkung nach außen" 'effect outside/away from the subject' and "Wirkung von außen" 'effect frcm outside the subject' (Brinkmann 1971:205) make for a verb classification which in general coincides with Weisgerber's (1963) definitions of active as "täterbezogen" 'agent-oriented' and passive as "täterabgezogen" 'agent-withdrawn' . These two granrarians are representative of the German tradition, which uses "notional" definitions in granmar. (See Lyons 1966, 1968 for discussions of notional definitions in syntax.) Concerning the Russian tradition, Vinogradov's section on SJA-verbs gives a good review of how various Russian granmarians have viewed voice (Vinogradov 1947:606-651). Vinogradov himself states that the category of voice "lies on the very boundaries between graitmar, lexicology and phraseology, and in the area of grarrmar — closer to the syntax of the sentence than to the morphology of the word." [... leZit u2e na samoj pograniönoj öerte meSdu grartmatikoj, leksikologiej i frazeologiej, a ν oblast gramiatiki — bliZe k sintaksisu predloSenija, öem k morfologii slova. ] (Vinogradov 1947:606) Later he writes that in certain cases "the category of voice goes beyond the limits of graitmar." [ kategorija zaloga vyxodit za predely grairmatiki. ] (Vinogradov 1947:638) In other vrords, voice is not entirely grairmatical. Once he claims this, he does not have to account for voice in a rigorous manner. Frcm a study of the various treatments of voice one is led to question what this term even captures within the grannars of modem languages (see footnote 7) Part of this confusion stares frcm the fact that whereas Greek and Latin had verb paradigms very closely associated with certain syntactic NP configurations, English, German and Russian do not. In Latin, for example, there is a verb paradigm which is used when an accusative NP is moved into the nominative and the original nominative is moved into an ablative (which can then be optionally deleted). The modern languages do not have a verb paradigm exclusively associated with this type of syntactic derivation. All three languages make use of an analytic construction involving a verb auxiliary and a participle. In English and Russian, however, this analytic construction is also used as a stative adjective, as in the following exarrples: (10)a b

Mary was annoyed by John, Mary was annoyed at John.

24

(10)c

*Mary was annoyed at John by Bill.

(11)a

Okna byli otkryty Antonom. ' (The)-windows were opened by-Anton.1

b

Okna byli otkryty ves' den'. '(The)-windows were open all day.'

c

*0kna byli otkryty ves1 den1 Antonom. windows were open all day by-Anton

Russian uses SJA-verbs to form the imperfective passive. However, when the instrumental Agent-NP is deleted in the passive, the remainder is then subject to the "middle" interpretation, as in (12). (12)

Dver' otkryvalas'. (The)-door was-opening.'

1

Thus, although Russian has tvro verb forms which are used in the same syntactic configuration associated with the Latin passive paradigm, the Russian forms are not exclusively associated with this configuration. It can be claimed, therefore, that passive in Russian does not have the same syntactic status as it doe§ in Latin. German has been said (IsaCenko 1968:442) to have a special set of forms for the passive. These are an analytical construction consisting of the auxiliary verb werden 'beccme' and the past participle. The difference that arises between German on the one hand and English and Russian on the other is that since the "passive" auxiliary is werden and not sein 'be', there is no ambiguity between passive constructions and Stative adjectives. (13)a

b

Die Stadt wird vom Feind zerstört, the city becomes by-the enemy destroyed 'The city is-being-destroyed by the enemy.' Die Stadt ist (*vom Feind) zerstört. 'The city is destroyed (*by the enemy).

The construction involving sein + ge-participle (ge

0, if the verb has an

inseparable prefix as is the case with zerstören or has the suffix -ier-) has been called the Zustandspassiv 'statal passive'. But (13b) does not render a passive at all since no Agent is allowed; it contains an adjective which has g been lexically derived from the corresponding verb. Therefore, we can say that

8. The rules of German are changing since sentences such as the following are found in everyday speech: (i) Das Essen ist von mir gekocht. 'The food was cooked by me.' (ii) Das Haus ist von dem-und-dem gebaut. 'The house was built by so-and-so.'

25 German does possess only one verb construction whenever the ncminative-NP has been vacated and the contents of the accusative-NP have moved into the nominative-NP.

However, werden + ge-participle is not exclusively associated with

the construction above; it is also used in sentences with intransitive verbs, as in (14). (14)

Es wird hier sonntags getanzt, it becomes here Sundays danced 'There's dancing here on Sundays.'

In an.intransitive sentence, such as (14), although there is lexical raterial moved out of the naninative-NP, there is no object moved into that NP.

Thus,

German like English and Russian is also a language which has no verb paradigm corresponding exactly to the Latin passive paradigm. In spite of the differences between modern languages and ancient languages, as Lyons noted in the quote above, the majority of the graimers of modern languages have been greatly influenced by Latin and Greek granmars.

One result

of this influence has been the inclusion of the category of voice within the verb classifications for modern languages.

In the preceding paragraphs I

illustrated hew German, Russian and English verbs do not correspond to Latin as far as voice is concerned.

It is no wonder, then, that there has been no

unanimity among grairmarians as to how rrany voices there are in each language or even as to whether all verbs do have voice. Within the Russian tradition, opinions range fron that of Nekrasov (1865), who denies that voice exists in Russian, to those of early nineteenth-century grairmarians, who list six voices for Russian —

active (=transitive verbs),

passive (=opposite of active), reflexive, reciprocal, middle (=intransitive), and general (e.g. bojat'sja 'be afraid').

In general, however, most Russian

grammarians have concluded that Russian has either two or three voices.

Fortu-

natov (1899), Jakobson (1957 [1971:131-147]) and IsaCenko (1968) are three grairmarians who analyze Russian as having two voices.

Fortunatov and Jakobson

speak of reflexive and non-reflexive voices; IsaCenko speaks of active and passive voices (SJA-verbs are distributed between the two voices). son and IsaCenko speak of marked and unmarked voices. urmarked in both analyses.

Both Jakob-

The active voice is

Fortunatov states that not all verbs have voice,

only those which shew an alternation between transitive verb without -SJA and verb with -SJA (Vinogradov 1947:614) . IsaCenko, hewever, states that all verbs have voice (IsaCenko 1968:272). Fortunatov1s claim that not all verbs have voice, only those which show an opposition between verb without -SJA and verb with -SJA, is an application of

26

the structural linguistic concept that language analysis is based on systems of 9

oppositions. He divorced the term active voice fron a particular set of verb endings and in essence claimed that there is no voice when there is only one arrangement of NP's in a clause. This carries with it the corollary that any voice besides active voice involves a replacement of the original active voice configuration with a new configuration.^ Within the German tradition, there has rot been such a great diversity of opinion as to numbers of voices; only two voices — active and passive — have in general been posited. However, as in Russian, the question has arisen as to whether all verbs do or do not have voice. For instance, Brinkmann (1971:205) claims that seme verbs do not have voice because they do not fit notional definitions of active and passive voice. Brinkmann's definitions af active and passive as "effect away fron the subject" and "effect frcm outside onto the subject" exclude SICH-verbs from falling into either voice category. Duden, however, reaches a different conclusion about SICH-verbs. They have active voice, since the definition of this voice is not syntactic but notional. His definition of passive voice consists of two parts: one is notional (see belcw) and the other is a syntactic correlation between active and passive constructions. The syntactic correlation is that the subject of the passive verb is the object of the active verb. Once this definition has been given, there fellow two sections dealing with the exceptions to this correlation (i.e. passive of intransitive verbs and lack of passive with SICH-verbs; Duden 1966: 107-108). The notional part of the passive definition is that the passive is a "Sehweise" 'manner of viewing' which takes the carrier of the happening out of the limelight, so to speak (Duden 1966:106). Although it is not stated in Duden, it is implicit that within his analysis anything which is not passive is active. Glinz (1971:19, 24) has a similar conclusion and further caiments that the passive is marked and the active unmarked. The concept of structural oppositions, with the exception of Brinkmann (1971), plays no role in the German treatment of voice. All verbs have voica, and some verbs have two voices. The definition of voice which I adhere to recognizes that voice is a property of all verbs. It follcws from the work of 9. According to Korolev (1969), a recent criticism of IsaSenko is precisely this — that his classification is based on verb form and not on oppositions within a given verb paradigm (Guxman 1964). 10. Such a view is also found within transformational grammar, i.e. the activepassive relationship as posited by Chomsky (1957).

27 adherents to sqne of the ideas expressed by Fillmore (1968) and Tesniere (1959). Mel'Cufc 3nd Xplqdoviö (1970) discuss voice in terms of mappings from semantic unite (which are yery similar to Fillmore's deep structure cases) into syntactic units (which are very similar to surface structure cases).

They discuss

especially fcoth that a given set of semantic units nay be mapped into more than ope configuration qf

syntactic units and that these different mappings can be

registered in the verb in various ways. diathesis (Tesniere 1959).

To the mappings they apply the term

And to the registering of these different mappings

in the verb they apply the term voice. Two things in Mel'Cuk and XolodoviC's work are important.

First, since

there is always a mapping between syntactic and semantic units, there will always be diathesis. 3II verts ft^ve voice.)

(Translated into traditional terminology, this means that Second, passive is not derived from active.

lated, this means that passive does not replace active.)

(Trans-

And by making a

distinctiqn fcjetween the mappings and the registering of those mappings in the verb fgrm, M/el'Spk and XolodoviC have been able to separate questions of form from g|ue§tiqRs of syntax. The definition of voice which I use is given in Babby and Brecht (1975:342). They def^rjß yoipe ''as the relationship between a verb's subcategorization fe^tijpe and the surface form of the sentence in which it occurs. tionship pay be optionally expressed by a voice morpheme."

This rela-

The superficial

difference between Babby and Brecht and Mel'Cuk and Xolodovic is that the formier substitute the term voice in place of the latter's term diathesis and the tern) voice morpheme in place of the latter*s term voice.

However, there is a

more pnpfqund difference between these two approaches to voice. Py defining voice in terms of a verb's subcategorization feature, Babby and BreGht brip,g into the picture the possibility of mis-matchings between the number of semantic units actually present in a given sentence and the nunrber of units required by a verb's subcategorization feature. when this occurs and thus account for -SJA.

They explain what happens

Mel'Cuk and XolodoviC's work does

not 4©al ^ith this possibility and thus does not account for -SJA.

The ramifi-

cations of jE^atjby and Brecht's approach are that reductions in the number of NP's actually filled in a verb's subcategorization feature are just as much a part of voice as ape the different syntactic arrangements of a given set of semantic NP's.

Mei'Cuk and XolodoviC (1970:123) even argue against this when they list

as a non-ypice