On The Relation Between Structural Case, Determiners, And Verbs In Ism A Study Of Hebrew And Dutch

APHASIOLOGY, 2008, 22 (9), 948–969

On the relation between structural case, determiners, and verbs in agrammatism: A study of Hebrew and Dutch Esther Ruigendijk Carl von Ossietzky University, Oldenburg, Germany

Naama Friedmann Tel Aviv University, Israel

Background: This study explored the relation between the production of determiners and case markers and the production of verbs and verb inflections in agrammatism. Determiners and case markers require case and therefore depend on the existence of case-assigning constituents. Aims: Since verbs and verb inflections are case assigners, and are impaired in agrammatism, we tested whether the presence of verbs and verb inflection affects the production of determiners and case markers in Dutch and Hebrew agrammatism. Methods & Procedures: A total of 11 Hebrew-speaking and 8 Dutch-speaking individuals with agrammatism participated in picture description and sentence elicitation tasks, and their spontaneous speech was analysed. Outcomes & Results: The production of case-related morphemes was closely connected to the presence of a case assigner in the sentence. In Hebrew, object case was produced correctly 98% of the time, and always when a transitive verb was present in the sentence. In Dutch the production of determiners on the subject was related to the presence of a finite verb. The production of complete object noun phrases related to the presence of a transitive verb. Conclusions: The results indicate that case itself, as well as determiners and case markers, which depend on case, are not impaired in agrammatic production. The apparent deficit is rather tightly related to the deficit in verbs and verb inflection. This suggests that the production of determiners and pronouns should be treated within sentence context, in which a special emphasis should be given to the production of correctly inflected verbs.

Individuals with agrammatic aphasia encounter difficulties in the production of grammatical morphemes such as determiners, case markers, and verb inflection, and often their sentences lack verbs. Recent studies show that not all grammatical morphemes are equally susceptible to impairment and that the pattern of omission Address correspondence to: Esther Ruigendijk, Carl von Ossietzky Universita¨t, Fak. III, Institut fu¨r Fremdsprachenphilologien, Ammerla¨ nder Heerstr. 114-118, 26111 Oldenburg, Germany. E-mail: [email protected] The project on case assignment in Dutch has been carried out under auspices of the Graduate School of Behavioral and Cognitive Neurosciences in Groningen (BCN) and the Center for Language and Cognition in Groningen (CLCG). Naama Friedmann was supported by a university grant for the encouragement of research. We are grateful to Roel Jonkers for providing the Dutch data, and to Roelien Bastiaanse and Aviah Gvion for their comments on a previous version of this paper. # 2008 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business http://www.psypress.com/aphasiology DOI: 10.1080/02687030701831482

CASE IN AGRAMMATIC PRODUCTION

949

and substitution is determined by linguistic constraints (De Bleser & Luzzatti, 1994; Friedmann, 1994, 2001, 2006; Grodzinsky, 1990; Hagiwara, 1995; Ruigendijk, van Zonneveld, & Bastiaanse, 1999). In this study we explore the relations among the impaired morphemes, specifically the relation between the production of determiners, pronouns, and case markers on the one hand, and verbs and verb inflections on the other. Agrammatic speakers omit and substitute determiners, and produce only a small number of pronouns in their free speech (see e.g., Menn & Obler, 1990; Nespoulous et al., 1988; Ruigendijk & Bastiaanse, 2002; Saffran, Berndt, & Schwartz, 1989). Linguistically, pronouns and noun phrases with determiners have something in common: both need case, a syntactic mechanism that marks syntactic roles such as subject and object in the sentence (for a detailed explanation of case see the next section, ‘‘What is case?’’). The main question we asked in this study was whether grammatical case is impaired in agrammatism, or whether what seems to be impairment in case, manifesting in omissions of case markers and determiners for example, should actually be ascribed to a deficit in another component of syntactic ability that influences case. Specifically, we examine a hypothesis that case in itself is not impaired in agrammatism. The impaired production of morphemes related to case, such as case markers, determiners, and pronouns in agrammatism is related to a deficit in the production of verb and verb inflection, which assign case. In order to describe the syntactic requirements for case assignment, the next sections present a brief linguistic background regarding case in general, case in agrammatism, and case in Hebrew and Dutch in particular, after which we describe the experimental investigation and the results.

LINGUISTIC BACKGROUND What is case? Case is a mechanism that specifies the syntactic relationship between, for example, a verb and the subject and object. It marks the function of each noun phrase in the sentence. The subject receives nominative case, and the object usually receives accusative case.1 Since the Government and Binding Theory (Chomsky, 1981, 1986), it is assumed that every pronounced noun phrase must have (exactly one) case. This requirement is called the case filter. A sentence with a noun phrase that has not been assigned case is thus ungrammatical. For example, a sentence like ‘‘*I am proud my students’’ (Chomsky, 1995, p. 113) is ungrammatical because my students does not receive case. According to Chomsky (1981, 1986), all languages have case. In some languages, like Russian and Hungarian, case is overtly realised on nouns and pronouns. In some other languages, such as Chinese, case is invisible. In other languages, it is sometimes realised morphologically while at other times it remains invisible, as in the languages under discussion in this study: Hebrew and Dutch. Even when case is invisible, it is 1 Note that we present a somewhat simplified overview of the theory on case covering only the most basic cases that are relevant for our study. A more detailed discussion would be beyond the scope of this paper.

950

RUIGENDIJK AND FRIEDMANN

assumed to be there on an abstract level. When we use the term case in this study, we refer to this syntactic notion of abstract case that is present in all languages. The assignment of case to the subject and the object, which is the topic of the current study, is dependent on the structural position of these noun phrases in relation to the verb and the inflection and is therefore called structural (or syntactic) case (Chomsky, 1981).2,3 Noun phrases get their case from a case assigner. Nominative case is assigned to a noun phrase in subject position by verb inflection, accusative is assigned by the verb to its object. This study thus explored these case assigners—verbs and verb inflection. Verbs and verb inflection play a major part in the agrammatic deficit, as illustrated in the introduction, and we surmised that the deficit in determiners and pronouns might be related to the deficit in verb production. In a simple subject-verb-object sentence, such as (1a), the modal ‘‘will’’ (or, in other sentences, the inflection of the verb) assigns nominative case to the subject noun phrase ‘‘the man’’. The transitive verb ‘‘meet’’ assigns accusative case to the object ‘‘the boy’’. In this example in English, abstract case is assigned to the noun phrases, but it is not visible. Case becomes visible in English when pronouns are used. As seen in (1b), the subject ‘‘he’’ has nominative case, whereas the object ‘‘him’’ has accusative case. (1) a.

b.

The man will meet the boy. NOM ACC He will meet him.

Thus, subjects depend on the presence of the finite inflection of a verb, whereas objects depend on the presence of a transitive verb. Subjects will not receive case if there is no finite verb, and objects will not receive case if there is no transitive verb (for objects the verb does not need to be finite). When the subject or the object do not receive case, the case filter will be violated. One constraint on the case filter was suggested by Ouhalla (1993). According to Ouhalla, the case filter applies only to complete noun phrases such as nouns with a determiner and pronouns. Importantly for the current study, noun phrases without a determiner can be caseless.

Case in agrammatism Several empirical investigations of the production of case-related morphemes4 in agrammatism have yielded an unclear pattern of results. Some indicate the preservation 2 Case can also be lexically specified, and then it is called inherent (or lexical) case. For the study of inherent case assignment, languages that show a clear distinction between inherent and structural case assignment, like German or Russian, are more interesting (see Ruigendijk & Bastiaanse, 2002, and Ruigendijk, 2002, for a study of these languages). When we speak about case here, we always mean structural case. 3 But see, for instance, Landau (2006) for an alternative analysis of case. 4 Case-related morphemes can be bound or free. In Russian and Standard Arabic, for example, case is morphologically realised as a suffix on the noun, in German it is realised on the determiner. Note that although the presence of a determiner always requires case assignment to the noun phrase, determiners are not case-marked in all languages. English and Dutch determiners, for example, are not specified for case.

CASE IN AGRAMMATIC PRODUCTION

951

of the production of case-related morphemes (De Bleser, Bayer, & Luzzatti, 1996; Jarema & Kadzielawa, 1990; Ruigendijk & Bastiaanse, 2002), while others report a deficit in case-related morphemes, manifested in the overuse of the default form (which is nominative in the languages that were examined, i.e., Russian and Serbo-Croatian, see Luria, 1976, and Zei & Sˇikic´, 1990). The aim of the current study is to assess the conditions in which case is impaired in agrammatism. One suggestion for the description of the syntactic deficit in agrammatism is the tree-pruning hypothesis (TPH, Friedmann, 2001, 2002, 2006; Friedmann & Grodzinsky, 1997). According to the TPH, the deficit of individuals with agrammatism is related to the projection of the syntactic tree up to its highest nodes. This results in impaired production of structures and grammatical morphemes that involve high nodes, whereas structures that involve only low nodes remain intact. Crucially for the current study, tense inflection of the verb, which is associated with the high part of the tree, is impaired in the speech production of many agrammatic speakers,5 when the verb cannot move to high nodes to get tense inflection. As a result, the verb is often produced either in a non-finite form rather than a finite form—as is the case in Germanic languages— and in a low node, namely in sentence-final position (Bastiaanse & Jonkers, 1998; Kolk & Heeschen, 1992; Ruigendijk & Bastiaanse, 2002), or it is produced in the wrong tense inflection, as in other languages such as Hebrew and Arabic (Friedmann, 2000, 2001, 2006).6 Frequent verb omissions are also explained in this framework. Individuals with agrammatism produce fewer verbs than nonbrain-damaged speakers (Luzzatti et al., 2002; Saffran et al., 1989). This was found to be closely related to the position of the verb on the syntactic tree, as more verbs are omitted when the verb should have been produced in a high node (Bastiaanse & van Zonneveld, 1998; Friedmann, 2000, 2006; Friedmann & Gil, 2001; Friedmann, Gvion, Biran, & Novogrodsky, 2006). So when verbs have to move to pruned nodes on the syntactic tree, they either do not move and then appear with the wrong tense inflection and in a different sentential position, or they get omitted. In the current study we explore the possibility that the deficit in tense inflection and the omission of verbs cause a deficit in syntactic case, because tense inflection and verbs are necessary to assign case. Specifically, we will examine the realisation of nominative case, the case of the subject, which is assigned by the verb inflection, and of accusative case, the case usually assigned by the verb to its object. Importantly, there is an additional side to this generalisation. Given that according to the TPH only structures that involve the high nodes are impaired in agrammatism, case that is assigned (or checked) in low nodes should be intact. Because object case is assigned in low nodes, it is not expected to be impaired under the TPH assumptions, that is, when the case-assigning verb has been realised. 5 There are different degrees of severity in agrammatism. The individuals who are impaired at the tense phrase (TP) level have tense impairment and no impairment in agreement. Those who are impaired above TP are not impaired in either tense or agreement (for a description of degrees of severity see Friedmann, 2001, 2005). 6 Under a checking account for tense inflection, the verb enters the tree randomly inflected and its inflection is checked in T. If the tense is correct, the derivation converges, but when the tense is incorrect, the derivation crashes. If checking in T is impossible, the verb can be produced with its random tense.

952

RUIGENDIJK AND FRIEDMANN

Given these considerations, we suggest the preserved case hypothesis, which we will examine in the present study.

The Preserved Case Hypothesis Morphemes that depend on case and case assignment are not directly impaired in agrammatism. Impaired production of case and case-related elements in a sentence is a by-product of an impairment in related syntactic domains. Recent results from Dutch and German agrammatism support this hypothesis. Ruigendijk et al. (1999) demonstrated that the production of determiners and pronouns in Dutch and German was related to the production of a case assigner, such as a (finite) verb or a preposition. Individuals with agrammatism could produce determiners and pronouns in spontaneous speech when a case assigner was realised; when no case assigner was present, they tended to omit determiners or produced determiners and pronouns in the default nominative case. Similar results were found for German speakers with agrammatism in spontaneous speech as well as in several production tasks (Ruigendijk, 2002; Ruigendijk & Bastiaanse, 2000, 2002).

CASE IN HEBREW AND DUTCH, AND SPECIFIC PREDICTIONS Hebrew In Hebrew case is visible on definite object noun phrases and on pronouns. Nominative case is not marked overtly. Accusative case on objects is marked with the free morpheme et, which appears before the object. Only definite noun phrases can occur with the accusative marker (Berman, 1978; Danon, 2001, 2006; Shlonsky, 1997), as shown in the examples in (2) and (3). Definite noun phrases are either marked with the definite article ha- (2a), with bound possession marking (2b), as a part of a construct state nominal in which the complement of the head noun is definite (2c), or as a proper name (2d), and also before the demonstrative pronoun ze, ‘‘this’’ (2e) (examples 2a–e are grammatical and are taken from the speech of participants in this study). (2) a.

b.

c.

d.

e.

ha-yeled xipes et ha-kadur the-child searched ACC the-ball ‘‘The child looked for the ball’’ hikarti et kol-ex recognised-1st.sg.past ACC voice-your ‘‘I recognised your voice’’ Yakov shama et ne’um rosh ha-memshala Jacob heard ACC speech-head-the-government ‘‘Jacob heard the prime minister’s speech’’ gvina cehuba mazkira li et holand cheese yellow reminds to-me ACC Netherlands ‘‘Yellow cheese reminds me of the Netherlands’’ eifo macat et ze? where found-2nd.sg.fem.past ACC this? ‘‘Where did you find it?’’

CASE IN AGRAMMATIC PRODUCTION

953

With indefinite objects, the accusative marker is not allowed in Hebrew, and according to Danon (2006) indefinite objects in Hebrew lack case altogether. Therefore a sentence that contains an indefinite object is grammatical without a case marker (3a) and is ungrammatical with an accusative marker (3b). (3) a. Ha-yeled xipes kadur the-child searched ball b. * Ha-yeled xipes et kadur the-child searched ACC ball

Dutch In Dutch as in English, case is visible on pronouns only (e.g., ik vs mij, ‘‘I’’ vs ‘‘me’’ or hij vs hem, ‘‘he’’ vs ‘‘him’’). Determiners are not marked for case, only for number and gender. All singular count nouns obligatorily take a determiner (and therefore 4a is grammatical but 4b is not), except for mass nouns and plural count nouns, which do not require a determiner (see 4c, 4d), and incorporate nouns, which must occur without a determiner (4e). (4)

a.

b.

c.

d.

e.

Ik kocht een broodje kaas I bought a roll cheese ‘‘I bought a cheese roll’’ * Ik drink graag glaasje wijn I drink gladly glass wine ‘‘I like to drink glass of wine’’ Ik vind (deze) kaas erg lekker I think (this) cheese very nice ‘‘I like (this) cheese very much’’ Ik vind (deze) broodjes kaas lekker I think (these) rolls cheese nice ‘‘I like (these) cheese rolls’’ De jongen houdt van auto rijden The boy likes car driving ‘‘The boy likes car driving’’

Although Dutch determiners are not marked for case, it is assumed that their realisation depends on having case, and thus on the presence of a case assigner (following Ouhalla, 1993). According to the Preserved Case Hypothesis, it is expected that agrammatic speakers will be able to produce case-dependent morphemes such as case markers and determiners as long as they have the proper syntactic preconditions. Given these properties of case assignment in Hebrew and Dutch, for Hebrew this means that if a transitive verb is produced, and the object is definite, an accusative case marker should appear. For Dutch this means that if a finite verb is present, complete subject noun phrases can be realised, and if a transitive verb is present, complete object noun phrases can appear. Notice that the tree-pruning hypothesis does not mean that individuals with agrammatism can never access high nodes; it does, however, predict frequent failure in verb movement. The exact expectation is thus that when there is

954

RUIGENDIJK AND FRIEDMANN

no case assigner there will be more omissions than when a case assigner (a verb or verb inflection) exists. Testing two structurally different languages like Hebrew and Dutch thus allowed us to test different aspects of our hypothesis. In Hebrew, where definite articles are not obligatory but where an overt accusative marker for definite objects exists, we tested the relationship between the presence of a transitive verb for definite objects and the accusative marker. We expected that, if the presence of a case assigner is a critical factor, definite objects should appear more often with than without accusative marker in the presence of a transitive verb. In Dutch we tested the production of complete noun phrases. We investigated the relationship between the presence of a nominative case-assigning finite verb and completeness of the subject noun phrase and between an accusative case assigning transitive verb and completeness of the object noun phrase. If the presence of a case assigner is indeed a critical factor in the production of a complete noun phrase, we expect a higher rate of complete-to-incomplete noun phrases when a case assigner is present than when a case assigner is not present.

EXPERIMENTAL INVESTIGATION Case in Hebrew Participants. A total of 11 Hebrew-speaking individuals with agrammatic aphasia participated in the Hebrew part of the study. They all had non-fluent aphasia, and were diagnosed with Broca’s aphasia with agrammatism by the neuropsychological batteries used in Israeli rehabilitation centres—the Hebrew versions of the WAB (Kertesz, 1982; Hebrew version by Soroker, 1997), the PALPA (Kay, Lesser, & Coltheart, 1992; Hebrew version by Gil & Edelstein, 2001), and the BAFLA battery for assessment of syntactic abilities (Friedmann, 1998), and by clinical workup. All participants had a lesion in the left cerebral hemisphere and were right-handed. Their mean age was 39 years 6 months (SD 5 17.1), and mean years of education 12 years 5 months. All patients had characteristic agrammatic speech: non-fluent and short incomplete utterances, reduction of sentence structure, and tense inflection errors. They produced very few, if any, well-formed Wh-questions, relative clauses, or sentential complements, and they could not repeat sentences with verb movement to second position, omitting the verb or leaving it in a position after the subject (see Friedmann, 2005, for a detailed description of their syntactic abilities). Eight of the participants also had severe impairment in tense inflection. Participants MA, ML, and IE had relatively spared TP. Crucially all of them had unimpaired production of agreement inflection, indicating that at least the lower part of the syntactic tree was available for them. Only patients who had at least two-word utterances were included in the study. Method. To assess the use of accusative markers with verbs and definite and indefinite object noun phrases, we used analysis of spontaneous speech as well as elicitation of sentences. Spontaneous speech was collected and analysed for six participants; the rest of the participants did not produce enough spontaneous speech or produced only very short utterances without objects. Two more structured methods were also used to elicit transitive verbs in a sentence. Seven individuals were asked to describe in one sentence 40 pictures that depicted a transitive verb with one

CASE IN AGRAMMATIC PRODUCTION

955

Figure 1. An example of a picture used in the Hebrew sentence elicitation task ‘‘The girl combs the chicken’’.

figure that performs an action on another (an example is given in Figure 1). Four of these seven individuals also participated in an additional task, in which they were asked to produce a sentence with a given inflected verb (e.g., ‘‘Say a sentence with the word ‘fixed’.’’). This elicitation task included 100 verbs, 20 of which were transitive verbs (and the rest were verbs that take sentential complements and intransitive verbs: unaccusatives, reflexives, and unergatives, which were not analysed for the current study except for four cases in which the participants produced a sentence with accusative case as a response). Only sentences that included an object noun phrase were included in the analysis, and responses that did not include an object were excluded (for example, for the picture given in Figure 1, one of the participants said Ha-tarnegolet mistareket ‘‘The chicken combs-self’’ instead of ‘‘The girl combs the chicken’’, using the reflexive instead of the target transitive verb, so this response was not included in the analysis). Two individuals participated both in the spontaneous speech analysis and in the elicitation tests. The elicited speech and the spontaneous speech were tape-recorded and transcribed. If the patients corrected themselves, only the last attempt utterance was analysed. The different syntactic properties of Hebrew compared to Dutch allowed us to run a different type of analysis for Hebrew—recall that Hebrew includes an overt accusative case marker, et, which appears before definite objects. This allowed us to directly test the appearance of an accusative case marker in the context of definite objects, and whether they appeared only when a verb was present. Recall also that Hebrew does not have an indefinite article, and indefinite objects appear bare, without a determiner; thus Hebrew sentences in which both the determiner and the accusative marker are absent are perfectly grammatical, and do not necessarily indicate omission of the accusative marker or of definiteness markers. Therefore we only tested the appearance of accusative case markers with respect to definite objects, and were not interested in sentences in which the object was indefinite and the case marker was absent. Such sentences were not included in the analysis. For the same reason, bare subjects are interpreted as indefinite subjects, and are also

956

RUIGENDIJK AND FRIEDMANN

perfectly grammatical, and therefore could not be used in the analysis to indicate a case problem as they do in Dutch. Thus, the main question for Hebrew was whether each time a definite object occurs in the sentence it is preceded by the accusative marker, and whether each time an accusative case marker occurs, it occurs before a definite object. Then the question was whether these objects appeared when a verb was realised in the sentence. For these aims, sentences with definite objects were collected from both spontaneous speech and the elicitation tests. For each definite object it was determined whether it appeared after the obligatory the accusative case marker or not. In addition, all sentences with an accusative case marker were analysed to test whether accusative case marker occurred only before definite objects. Finally we examined whether object noun phrases with accusative case marking were produced in the presence of a case-assigning transitive verb. In the sentence-toverb construction task the verb was given to the participant so, naturally, the two other tasks—the elicitation with the pictures and the spontaneous speech analysis— were more informative with respect to the production of the verb. Hebrew results. The Hebrew-speaking participants presented excellent ability in their use of the accusative case marker ‘‘et’’. A summary of the results of the Hebrew experiment is given in Table 1 (see Appendix A for individual data). The data for the spontaneous speech and for the elicitation tasks were similar (the rate of correct and incorrect responses in both tasks did not differ significantly, using Mann Whitney, z 5 1.14, p 5 .25), and the statistical analysis therefore collapsed the data together for the two individuals who participated in both spontaneous speech and elicitation tasks. The participants produced a total of 319 definite objects. Most of these definite objects (98%) were produced correctly with an accusative marker. The accusative marker was omitted only before six definite objects – one proper name, one construct state nominal, and four nouns with a definite article, one of them following a long pause. We used the non-parametric Wilcoxon Signed-Ranks Test for comparisons with an alpha level of .05. The difference between definite objects with accusative case marker and definite objects without accusative case marker was significant, T 5 0, p , .001. In only five sentences was the accusative marker erroneously used before an indefinite object noun phrase (this happened significantly less than using the accusative marker correctly before a definite object noun; T 5 0, p , .001). In one sentence the accusative marker was substituted by a preposition. Importantly, all 313 definite object noun phrases with an accusative marker were produced in sentences with a case assigning transitive verb.

TABLE 1 Hebrew: Number of definite and indefinite object noun phrases with and without accusative case marker in spontaneous speech and in elicited sentence production task

Spontaneous speech (n56) Elicited sentence production (n57) Total

Accusative + definite NP

No accusative + definite NP

Accusative + indefinite NP

94 219 313

2 4 6

1 4 5

CASE IN AGRAMMATIC PRODUCTION

957

Case in Dutch Participants. Eight Dutch-speaking individuals with agrammatic aphasia (mean age 60 years 1 month) participated in the study. They were right-handed and aphasic due to a single stroke in the left hemisphere. All patients were at least a year postonset and had been diagnosed with agrammatic Broca’s aphasia using a standard assessment battery (Dutch version of the Aachen Aphasia Test; Graetz, De Bleser, & Willmes, 1992). The type of aphasia was confirmed by two aphasiologists. The speech production of all patients was agrammatic, and their speech was characterised by problems with finiteness of the verbs and/or a low number of verbs, relatively few pronouns, and omission of determiners. Their spontaneous speech included no Wh-questions or embedded sentences. Procedure. A picture description task was devised to elicit sentences (developed by Jonkers, 1998). The picture descriptions were taken from Jonkers (1998). This task consisted of 30 pictures depicting an action representing a transitive verb (See Figure 2 for an example for a picture used for the verb aaien, to pet). The patients were asked to describe in one sentence what was happening in the picture. The elicited speech was tape-recorded and transcribed. If the patients corrected themselves, only the last attempt utterance was analysed. For each item it was established whether a verb was produced and which syntactic roles (subject and/or object) were realised, and whether they were realised as complete noun phrases. The complete noun phrases in our analysis included nouns with a determiner, mass

Figure 2. An example of the Dutch sentence production task ‘‘The man pets the dog’’.

958

RUIGENDIJK AND FRIEDMANN

nouns, bare plural count nouns7 and pronouns, or as incomplete noun phrases— bare nouns that should have a determiner but were produced without one. The subject noun phrases were divided into three groups: subjects that occurred in a sentence with a case assigning finite verb (5a), subjects with a non-finite verb (5b), and subjects without a verb (5c). (5)

(6)

a. Subject with a finite verb: Target: De man maait het gras The man mows the grass Response: De man maait The man mows b. Subject with a non-finite verb: Target: De man maait het gras The man mows the grass. Response: Die kerel… dat gras aan het maaien. That fellow… that grass on the mow (5 mowing). c. Subject without a verb: Target: De vrouw veegt de straat The woman sweeps the street Response: vrouw…straat woman…street a. Object with a verb: Target: De vrouw veegt de stoep The woman sweeps the pavement. Response: De straat vegen The street sweep-infinitive b.Object without a verb: Target: De jongen aait de hond. The boy pets the dog. Response: Jongen hond, lieve hond Boy dog, sweet dog

For the objects we analysed whether a verb was produced in the sentence or not (6a and b). Subsequently, all objects with a case-assigning verb were divided into two groups: objects with a finite verb and objects with a non-finite verb. This was done to evaluate whether verb presence or verb finiteness was the important factor for the production of complete noun phrases. Finally, we also counted how many subjects and objects were not realised and how many finite or non-finite verbs were produced in isolation, i.e., without any arguments. Apart from the elicited sentence production data, we analysed spontaneous speech production of each patient with respect to the production of complete and incomplete subject and object noun phrases. The spontaneous speech production came from the interviews that were part of the AAT and included questions like ‘‘Could you tell me how your speech problems started?’’, and ‘‘Could you tell me something about your job/ family/ hobbies?’’. These samples consisted of 175–480 7 Mass nouns and plural count nouns do not need a determiner in languages like English and Dutch. According to Longobardi (1994) these noun phrases should still be analysed as DPs (cf. Abney, 1987); that is, as complete noun phrases. De Roo (1999) suggested the same for Dutch mass nouns and plural count nouns. We follow Longobardi and de Roo in our analysis and refer to their work for a technical discussion of this issue.

CASE IN AGRAMMATIC PRODUCTION

959

TABLE 2 Dutch: Number of complete and incomplete subject noun phrases in relation to the presence of a case assigning finite verb

Finite verb Non-finite verb No verb

Complete

Incomplete

128 (97%) 14 (73%) 13 (65%)

4 (3%) 5 (27%) 7 (35%)

words per participant. To be able to determine whether a noun phrase or a pronoun was used as an object or a subject, only spontaneous utterances with a verb (finite or non-finite) were analysed. Fixed expressions (e.g., weet ik niet, ‘‘I don’t know’’) were excluded from the analysis. For each utterance, it was established whether a verb was produced and which syntactic roles (subject and/or object) were realised, and whether they were realised as complete noun phrases or as incomplete noun phrases. We used the non-parametric Wilcoxon Signed-Rank Test for all comparisons, with an alpha level of .05 for all statistical tests. Dutch results. The results of the Dutch elicitation study are presented in Tables 2 and 3. Table 2 presents the number of subject noun phrases in the various conditions, and Table 3 presents the distribution of object noun phrases (see Appendix B for individual data). We analysed 201 of the responses to the pictures (83.8%); of these 39 (16.2%) could not be analysed with respect to subject and object production due to zero reactions (‘‘I don’t know’’), perseverations, paraphasias, or circumlocutions. None of these included a verb that described the action on the picture even roughly. In total, only 57.9% of the 201 analysable utterances contained a finite verb, which was always realised in the second position as is obligatory in Dutch matrix clauses, 17.1% included a non-finite verb, and in 8.8% no verb was realised. In total, 171 subject noun phrases and 119 object noun phrases were produced. The patients produced more subjects than objects due to the fact that some of the verbs could also be used without an object. About half of the subjects were pronouns; all subject pronouns appeared in the nominative case as required. No case errors were made on the pronouns. No object pronouns were produced. Of all nouns with a determiner (n 5 146), only two nouns appeared with an incorrect determiner, both due to a gender error. Significantly more complete subject noun phrases than subject noun phrases without a determiner were produced when the relevant case assigner, a finite verb, was present, T 5 0, p 5 .008. When the verb in the sentence was non-finite, there was no difference between the number of complete subject noun phrases and subject noun phrases without a determiner, T 5 4, p 5 .22. No significant difference was found between the number of complete subject noun phrases and incomplete subject noun phrases also when there was no verb at all, T 5 2, p 5 .18 (see Table 2). A total of 69 of the subjects were realised as a pronoun. The majority of these pronouns were produced in the presence of a finite verb (65 out of 69). Only three of these pronouns were produced with a non-finite verb and only one was produced without a verb. The completeness of objects was also found to depend on the verbs, but this time on the existence rather than on the finiteness of the verbs (see Tables 3 and 4). When a case-assigning transitive verb was used, significantly more complete than incomplete object noun phrases were produced, T 5 0, p 5 .004. However, when there was no case-assigning verb present, more incomplete than complete objects appeared,

960

RUIGENDIJK AND FRIEDMANN

TABLE 3 Dutch: The total number of complete and incomplete objects in relation to the presence of their case-assigning verb

Verb No verb

Complete

Incomplete

81 (79%) 6 (35%)

21 (21%) 11 (65%)

TABLE 4 Dutch: The number of complete and incomplete objects in relation to the presence of finite and non-finite verbs

Finite verb Non-finite verb

Complete

Incomplete

56 (80%) 25 (78%)

14 (20%) 7 (22%)

but these cases were too few to reach significance (Because of the number of ties, the Wilcoxon test could not be used. A chi-square test yielded x2 5 2.94, p 5 .08.) Table 3 shows that, unlike for subjects, and in line with the predictions the finiteness of the verb did not play a role in the realisation of case on objects (as manifested by determiner production). When the verb was finite, 80% of the objects were complete noun phrases; when the verb was non-finite, 78% of the objects were complete. Thus, for both the finite and non-finite verbs, the majority of the object nouns were complete, with no significant difference between finite and non-finite verbs with respect to the rate of complete noun phrases (x2 5 0.05, p 5 .83: chi-square for the group was run instead of Wilcoxon here because four participants did not produce any object in one of the conditions). This means that for the production of complete object noun phrases, the existence of a case-assigning transitive verb, rather than verb finiteness, is needed. The analysis of the spontaneous speech data shows exactly the same pattern as the data from the elicitation task. Whenever there was a finite verb, subjects were realised as a complete noun phrase and not as an incomplete noun phrase (90 vs 0, which is a significant difference, T 5 0, p 5 .02). In the corpus, subjects usually appeared with a finite verb and therefore there were not enough instances of subjects with a non-finite verb to allow for a comparison between complete and incomplete noun phrases (there were only three such instances) or for a comparison between complete noun phrase subjects with and without verb finiteness. When there was a verb in the sentence, objects were realised as a complete noun phrase (n 5 35) significantly more times than as an incomplete noun phrase (n 5 9), T 5 0, p 5 .02. As in the elicitation task, finiteness did not play a role for the objects, and no significant difference was found between the number of complete object noun phrases with finite and non-finite verbs, T 5 5, p 5 .31. Furthermore, the rate of complete object noun phrases was not significantly different between finite and nonfinite verbs, 85% and 72% respectively, x2 5 1, p 5 .32.

DISCUSSION The results from both Hebrew and Dutch indicate that the production of case itself is not impaired in agrammatism, and that it is tightly related to syntactic preconditions and specifically to the presence of a proper case assigner in the sentence. Moreover, the

CASE IN AGRAMMATIC PRODUCTION

961

results show that agrammatic speakers respect the syntactic principles of case (case filter). The main findings of the study are that in Hebrew the accusative case marker is unimpaired and is produced correctly for 98% of the definite object nouns. In Dutch pronouns are never produced in a wrong case and determiners are produced correctly whenever a case assigner is present. In most utterances, subjects are produced with a determiner when a finite verb is present in the sentence, and objects are produced with a determiner when a verb (irrespective of its finiteness) is present. These findings have several implications. First, in line with the findings of de Bleser et al. (1996) and Ruigendijk and Bastiaanse (2002) as well as with our preserved case hypothesis, they indicate that case is unimpaired in agrammatic production, and they add support for the general claim that not all grammatical morphemes are impaired in agrammatism. These results also indicate that case, determiners, inflection, and verbs are interrelated. The results from Hebrew indicate a tight relation between case realisation and the production of the determiner, as in 98% of the sentences in which a determiner appeared on the object noun the accusative case marker was produced, and in 98% of the sentences in which the case marker was produced the object was definite. Furthermore, the results from both Dutch and Hebrew demonstrate a close connection between the production of caseassigning verbs and the production of determiners and case markers: In Hebrew, object case markers appeared only in sentences that included a verb. In Dutch, the large majority of the subjects in sentences that included a case-assigning finite verb, and of objects in sentences that included a verb, was produced with a determiner. So the most important finding here is that there was a significant difference between sentences with a case assigner, in which much more noun phrases were complete than incomplete, and sentences without a case assigner, in which this was not true. When the conditions for case were met—i.e., when the proper case assigner was present—case was realised on the noun phrases, and they appeared as complete noun phrases significantly more often. The results of the current study are readily explained by the combination of current linguistic theory and theories of agrammatic production. According to current syntactic theory, within the framework of transformational grammar and the minimalist programme (Chomsky, 1995, 2000, 2001), the subject and the object check their case with a case assigner. Subjects check their case against the tense of the verb that is in the tense phrase (TP). Objects check their case against the verb on a low node of the syntactic tree.8 Thus finiteness, or the tense inflection of the verb, as well as the movement of the subject (and the verb) to TP are crucial for successful case assignment to subjects, whereas for objects the verb itself, rather than its tense inflection, is the crucial factor, and therefore movement to higher nodes is unnecessary. 8 Specifically: According to the minimalist programme (Chomsky, 1995), the case of the subject and the object is checked in spec-head configuration of the noun phrase (DP) and the case assigner. That is to say, the noun phrase is in the specifier position of the phrase, and the case assigner is at the head of the phrase. Subject DPs raise to spec-TP to check their case against the verb and its tense, which are in T0 (following the movement of the verb from VP to T0). Objects check their case with the verb at AgroP according to Chomsky (1995), or at the light v layer according to Chomsky (2000, 2001). Note that although several different frameworks have been suggested for structural case, such as assignment of case with and without AgroP, the aspects that are relevant to our study remain the same: structural case is dependent on the relation between subject DP and finite V and the relation between object DP and a transitive V, and case for the object NP is assigned in a lower node of the tree.

962

RUIGENDIJK AND FRIEDMANN

When agrammatic speakers fail to move the verb and the subject to TP, the checking of the subject case against the verb and its tense cannot take place, and therefore when verbs do not move high up and are uninflected for tense (when they are non-finite or omitted) case assignment to the subject fails and the determiner of the subject is omitted. Objects, on the other hand, do not require movement to high nodes or tense inflection, only the existence of the verb is necessary for them. Thus, when the verb is present in the sentence, even if it is non-finite and has not moved to a high node, it can assign case to the object. Verbs that cannot move to TP to check their tense inflection, or to CP, to which the matrix verb moves in Germanic languages like Dutch (in order to be in the second sentential position), are either omitted or left in a non-finite form in a low node (which in Dutch and German is a sentence-final position, and in Hebrew the position within VP after the subject) (Bastiaanse & van Zonneveld, 1998; Friedmann, 2000; Kolk & Heeschen, 1992). However, when the verb is omitted, the assignment of object case is deficient, which in turn might lead to the production of incomplete object noun phrases. The relation between determiners and case or, more specifically, the reason for determiner omission when case is not assigned, is related to the distinction between complete noun phrases, such as noun phrases with a definite article or pronouns, which are called determiner phrases (DPs) in Abney’s (1987) terminology,9 and incomplete noun phrases (NPs), noun phrases without a determiner. According to Ouhalla (1993), the case filter applies to DPs rather than to NPs, and case is actually a property of complete noun phrases, and not of incomplete noun phrases. Thus, DPs in utterances with no suitable case assigner receive no case, and therefore violate the case filter and are ungrammatical, but incomplete NPs without a case assigner do not violate the case filter. This distinction between NPs and DPs explains the omissions of determiners in our study, which occurred when the subject or the object lacked case. When there is no case, a determiner cannot appear because a caseless DP is ungrammatical. Therefore, an NP without a determiner, which is not subjected to the case filter, is produced instead. These results are in line with earlier studies that showed that the production of determiners in German depends on the realisation of a case assigning verb (Ruigendijk, 2002; Ruigendijk & Bastiaanse, 2002). Ruigendijk and Bastiaanse (2002) show that German agrammatic speakers produce more complete than incomplete noun phrases when a case assigner is realised, both in spontaneous speech and in sentence-elicitation tasks. When the German individuals with agrammatism do not realise a case assigner, they omit the determiner much more often than they produce it. Interestingly, these data also emphasise why testing Dutch is important. Whereas these findings on German have already shown the strong relationship between determiner realisation and the presence of a case assigner, they could also be argued to be related to morphological case, which is shown on German determiners. The argumentation could then be that if no case assigner is present, no morphological case can be determined, and therefore the determiner’s morphological form remains unspecified, which might lead to 9 In the government and binding framework (e.g., Chomsky, 1986), the noun was assumed to be the head of a noun phrase (NP), with the determiner in the specifier position. Abney (1987) presented an alternative analysis, the DP analysis: D (the determiner) is a functional head that takes a noun phrase as its complement, forming a DP, determiner phrase.

CASE IN AGRAMMATIC PRODUCTION

963

determiner omission. The results we have presented here from Dutch show that it is the syntactic relationship between case assigners and noun phrases that is important, rather than morphology. Since Dutch does not have morphological case on determiners, there is no morphological reason for article omission here.10 This does not mean, however, that morphology does not play a role at all: first results from a close comparison of Dutch and German show that German speakers omit determiners more often than Dutch speakers (Ruigendijk, 2007). Recently, the results on Dutch have been replicated in a study in which the spontaneous speech of eight Dutch-speaking agrammatic aphasic speakers has been analysed with regard to the production and omission rates of determiners and pronouns and, among other things, their relationship with the presence of a case assigner (Ruigendijk & Baauw, 2007). Ruigendijk and Baauw (2007) also show that many more complete noun phrases and pronouns are realised if a case assigner is present than absent. They furthermore demonstrate that it is mainly this syntactic factor of case assignment that affects the production and omission of determiners in agrammatic speech, whereas pragmatic factors (realising a definite or an indefinite determiner) and lexical and semantic factors (i.e., gender of the determiner and pronoun respectively) do not play a role. The results of the current study thus strongly suggest that when there is a relevant case assigner, T or V, the subject and object noun phrases (respectively) are complete. A question remains regarding the other direction of the implication: the finding that sometimes when no case assigner was present, the Dutch participants still realised some complete noun phrases. How did these noun phrases receive case? One possibility is that—at least for the subject noun phrases—patients can adopt a strategy; so-called default case assignment. Default case assignment is an option that has been proposed for normal elliptical utterances where structural case assignment fails: if the subject does not check/receive its case from I, it gets nominative by default (van Zonneveld, 1994). As was also suggested in Ruigendijk et al. (1999), agrammatic speakers may be able to use this default option as a strategy when normal case assignment fails. Notice, however, that this default strategy cannot be the whole story, because there was a significant difference in the production of complete subjects when the verb was finite compared to when it was not finite. Thus, the presence of an appropriate case assigner in the sentence clearly made a difference, over and above the default case. This default strategy cannot explain the six object noun phrases that appeared with a determiner without a case-assigning verb. Another problem for this default explanation is that it is not immediately clear at what level default case is applied. According to van Zonneveld (1994) it is indeed an alternative abstract case that is assigned if normal case assignment fails, and this could explain the fact that some complete subject noun phrases occurred without a proper case assigner without violating the case filter in our study. However, this characterisation would render the case filter vacuous. Schu¨tze (2001, p. 206) therefore characterises default case as ‘‘… forms of a language […] that are used to spell out nominal expressions (e.g., DPs) that are no associated with any case feature assigned or otherwise determined by syntactic means’’. As such it is morphological case that is ‘‘neither necessary nor sufficient for satisfying the case filter’’ (Schu¨tze, 2001, p.208). The case filter is, according to Schu¨tze, not 10

We thank the reviewer for pointing out this important point to us.

964

RUIGENDIJK AND FRIEDMANN

morphologically motivated, but a purely configurational requirement. When default case is morphological case, it is unclear how it can be applied to Dutch determiners that are not specified (any more) for morphological case. Another possible explanation comes from the nature of the task that was used. The patients were examined with a picture description task. And although they were asked to describe the pictures in one sentence, they sometimes completely failed to produce a sentence or even a fragment. Probably some of the patients still wanted to describe the picture as well as possible within the limits of their impairment, and simply started naming the objects and figures they saw in the picture. When naming items in a picture in Dutch, it is possible to use a determiner in a deictic way (outside a case assigning context; that is, these noun phrases do not have abstract case), especially if both the patient and the experimenter are looking at the same picture, which was the case in our study. Results from former studies in which spontaneous speech was analysed support this explanation. The data in Ruigendijk et al. (1999) and Ruigendijk and Baauw (2007) show that the number of incomplete noun phrases in the speech of Dutch agrammatic speakers is (much) higher than the number of complete noun phrases when no case assigner is present. In other words, when the task does not allow for a naming strategy, the relationship with no case assigner is clearer: when there is no case assigner there is no complete noun phrase. However, the more important finding is that as soon as a case assigner is present many more complete than incomplete noun phrases are realised. This close relation between a case assigner and the determiner has also been reported for another type of case assigner: prepositions. De Roo (1995) showed that Dutch-speaking agrammatic patients almost never omit determiners from within a prepositional phrase (they do not omit ‘‘the’’ from the PP ‘‘in-the-garden’’), although they omit determiners that do not appear in a PP approximately 20% of the time (de Roo did not analyse these omitted determiners with respect to whether or not a verb existed in the sentence). Ruigendijk (2002) showed that German- and Dutch-speaking agrammatic patients produced virtually no incomplete noun phrases on a noun phrase insertion task in which the preposition was provided. Given that prepositions are case assigners, these findings constitute further support for the claim that determiners are not omitted when a case assigner is present. To summarise, the causal chain that leads to determiner omissions, even though case and determiners themselves are unimpaired, unfolds in the following way. An impairment in syntactic structure building, causes difficulties in the movement of verbs to TP and CP, and therefore in many sentences the verb is either omitted, left uninflected, or appears in a wrong inflection at a low node. When a verb is omitted case cannot be assigned to either the subject or the object; a verb that has not moved to TP cannot assign case to the subject. When the subject or the object are caseless, they cannot be complete noun phrases because complete noun phrases require case, and therefore they appear only as incomplete noun phrases; that is, noun phrases without a determiner. This leads to determiner omissions. The results have interesting implications for the treatment of individuals with agrammatism. They indicate that training the production of isolated noun phrases to improve determiner and/or case marker production will not be enough, since the determiners and case markers are related to case assigners—verbs. Instead, the results of the present study suggest that determiners and case markers should be treated in the context of a sentence, and should be accompanied by treatment of verb production. Treatment that will improve the production of verbs will also improve

CASE IN AGRAMMATIC PRODUCTION

965

the production of complete noun phrases—namely the production of determiners and case. Results of a study performed by Springer, Huber, Schlenck, and Schlenk (2000) support this clinical direction. Springer et al. found that in some of their severely agrammatic patients the production of complete noun phrases increased after treating these patients with a programme that aimed at the production of (infinite) verbs combined with noun phrases (note, however, that treatment that ignores verb inflection will be inefficient with respect to the case of subjects). Furthermore, treatment programmes that are aimed at improving the accessibility of high syntactic nodes (such as TP and CP, see Friedmann, Wenkert-Olenik, & Gil, 2000; Shapiro & Thompson, 2006) should also affect the production of determiners and case markers by increasing the rate of verb production and inflection production by allowing movement to high nodes. Manuscript received 31 August 2007 Manuscript accepted 28 November 2007 First published online 27 May 2008

REFERENCES Abney, S. (1987). The English noun phrase in its sentential aspect. Doctoral dissertation. Cambridge, MA: MIT Press. Bastiaanse, R., & Jonkers, R. (1998). Verb retrieval in action naming and spontaneous speech in agrammatic and anomic aphasia. Aphasiology, 12, 951–969. Bastiaanse, R., & van Zonneveld, R. (1998). On the relation between verb inflection and verb position in Dutch agrammatic aphasics. Brain and Language, 64, 165–181. Berman, R. A. (1978). Modern Hebrew structure. Tel-Aviv: University Publishing Projects. Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris. Chomsky, N. (1986). Knowledge of language, its nature, origin and use. New York: Praeger. Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press. Chomsky, N. (2000). Minimalist inquiries: The framework. In R. Martin, D. Michaels, & J. Uriagereka (Eds.), Step by step: Essays on minimalist syntax in honor of Howard Lasnik (pp. 89–155). Cambridge, MA: MIT Press. Chomsky, N. (2001). Derivation by phase. In M. Kenstowicz (Ed.), Ken Hale: A life in language. Cambridge, MA: MIT Press. Danon, G. (2001). Syntactic definiteness in the grammar of Modern Hebrew. Linguistics, 39, 1071–1116. Danon, G. (2006). Caseless nominals and the projection of DP. Natural Language and Linguistic Theory, 24, 977–1008. De Bleser, R., Bayer, J., & Luzzatti, C. (1996). Linguistic theory and morphosyntactic impairments in German and Italian aphasics. Journal of Neurolinguistics, 9, 175–185. De Bleser, R., & Luzzatti, C. (1994). Morphological processing in Italian agrammatic speakers syntactic implementation of inflectional morphology. Brain and Language, 46, 21–40. de Roo, E. (1995). Articles and finite verb inflections in Dutch agrammatism. Poster presented at the RuGSAN-VKL conference on aphasiology, Groningen, The Netherlands. de Roo, E. (1999). Agrammatic grammar: Functional categories in agrammatic speech. Den Haag: Holland Academic Graphics. Friedmann, N. (1994). Morphology in agrammatism: A dissociation between tense and agreement. Unpublished Master’s thesis, Tel Aviv University, Israel. Friedmann, N. (1998). BAFLA: Friedmann Battery for Agrammatism. Tel Aviv University, Israel. Friedmann, N. (2000). Moving verbs in agrammatic production. In R. Bastiaanse & Y. Grodzinsky (Eds.), Grammatical disorders in aphasia: A neurolinguistic perspective (pp. 152–170). London: Whurr. Friedmann, N. (2001). Agrammatism and the psychological reality of the syntactic tree. Journal of Psycholinguistic Research, 30, 71–90.

966

RUIGENDIJK AND FRIEDMANN

Friedmann, N. (2002). Question production in agrammatism: The tree pruning hypothesis. Brain and Language, 80, 160–187. Friedmann, N. (2005). Degrees of severity and recovery in agrammatism: Climbing up the syntactic tree. Aphasiology, 19, 1037–1051. Friedmann, N. (2006). Speech production in Broca’s agrammatic aphasia: Syntactic tree pruning. In Y. Grodzinsky & K. Amunts (Eds.), Broca’s region (pp. 63–82). New York: Oxford University Press. Friedmann, N., & Gil, M. (2001). Verb movement in agrammatic comprehension and production. Paper presented at the 2nd ‘‘Science of Aphasia’’ conference, Giens, France. Friedmann, N., & Grodzinsky, Y. (1997). Tense and agreement in agrammatic production: Pruning the syntactic tree. Brain and Language, 56, 397–425. Friedmann, N., Gvion, A., Biran, M., & Novogrodsky, R. (2006). Do people with agrammatic aphasia understand verb movement? Aphasiology, 20, 136–153. Friedmann, N., Wenkert-Olenik, D., & Gil, M. (2000). From theory to practice: Treatment of agrammatic production in Hebrew based on the tree pruning hypothesis. Journal of Neurolinguistics, 13, 250–254. Gil, M., & Edelstein, C. (2001). PALPA – girsa ivrit [Hebrew version of the PALPA]. Ra’anana, Israel: Loewenstein Hospital Rehabilitation Center. Graetz, P. R., de Bleser, R., & Willmes, K. (1992). Akense Afasie Test [Aachen aphasia test]. Lisse: Swets Zeitlinger. Grodzinsky, Y. (1990). Theoretical perspectives on language deficits. Cambridge, MA: MIT Press. Hagiwara, H. (1995). The breakdown of functional categories and the economy of derivation. Brain and Language, 50, 92–116. Jarema, G. E., & Ka˛dzielawa, D. (1990). Agrammatism in Polish: A case study. In L. Menn & L. K. Obler (Eds.), Agrammatic aphasia: A cross-language narrative sourcebook (pp. 1695–1715). Amsterdam: John Benjamins. Jonkers, R. (1998). Comprehension and production of verbs in aphasic speakers. Doctoral Dissertation, Groningen University, Groningen, The Netherlands. Kay, J., Lesser, R., & Coltheart, M. (1992). PALPA: Psycholinguistic Assessments of Language Processing in Aphasia. Hove, UK: Lawrence Erlbaum Associates Ltd. Kertesz, A. (1982). Western Aphasia Battery. Orlando, FL: Grune & Stratton. Kolk, H. H. J., & Heeschen, C. (1992). Agrammatism, paragrammatism, and the management of language. Language and Cognitive Processes, 7, 89–129. Landau, I. (2006). Severing the distribution of PRO from case. Syntax, 9, 153–170. Longobardi, G. (1994). Reference and proper names: A theory of N-movement in syntax and logical form. Linguistic Inquiry, 16, 609–665. Luria, A. (1976). Basic problems of neurolinguistics. The Hague: Mouton. Luzzatti, C., Raggi, R., Zonca, G., Pistarini, C., Contardi, A., & Pinna, G-D. (2002). Verb–noun double dissociation in aphasic lexical impairments: The role of word frequency and imageability. Brain and Language, 81, 432–444. Menn, L., & Obler, L. (Eds.). (1990). Agrammatic aphasia: A cross-language narrative sourcebook. Philadelphia: John Benjamins. Nespoulous, J-L., Dordain, M., Perron, C., Ska, B., Bub, D., & Caplan, D. et al. (1988). Agrammatism in sentence production without comprehension deficits: Reduced availability of syntactic structures or of grammatical morphemes? A case study. Brain and Language, 33, 273–295. Ouhalla, J. (1993). Functional categories, agrammatism, and language acquisition. Linguistische Berichte, 143, 3–36. Ruigendijk, E. (2002). Case assignment in agrammatism: A cross-linguistic study. Doctoral dissertation, Groningen University, The Netherlands. Ruigendijk, E. (2007). Pronomina en determinatoren: Een taalvergelijkend onderzoek naar de problemen van Nederlands – en Duitstalige mensen met afasie. Neerlandica Extra Muros, 45(2), 33–50. Ruigendijk, E., & Baauw, S. (2007). Syntactic and pragmatic aspects of determiner and pronoun production in Dutch agrammatic Broca’s aphasia. Aphasiology, 21, 535–547. Ruigendijk, E., & Bastiaanse, R. (2000). Determiners in agrammatic speech production. Journal of Neurolinguistics, 13, 247–250. Ruigendijk, E., & Bastiaanse, R. (2002). Two characteristics of agrammatic speech: Omission of verbs and omission of determiners, is there a relation? Aphasiology, 16, 383–395. Ruigendijk, E., van Zonneveld, R., & Bastiaanse, R. (1999). Case assignment in agrammatism. Journal of Speech Language and Hearing Research, 42, 962–971.

CASE IN AGRAMMATIC PRODUCTION

967

Saffran, E. M., Berndt, R. S., & Schwartz, M. F. (1989). The quantitative analysis of agrammatic production: Procedure and data. Brain and Language, 37, 440–479. Schu¨tze, C. T. (2001). On the nature of default case. Syntax, 4, 205–238. Shapiro, L. P., & Thompson, C. K. (2006). Treating language deficits in Broca’s aphasia. In Y. Grodzinsky & K. Amunts (Eds.), Broca’s region (pp. 119–134). New York: Oxford University Press. Shlonsky, U. (1997). Clause structure and word order in Hebrew and Arabic. New York: Oxford University Press. Soroker, N. (1997). Western Aphasia Battery – girsa ivrit [Hebrew Western Aphasia Battery]. Ra’anana, Israel: Loewenstein Hospital Rehabilitation Center. Springer, L., Huber, W., Schlenck, K. J., & Schlenck, C. (2000). Agrammatism: Deficit or compensation? Consequences for aphasia therapy. Neuropsychological Rehabilitation, 10, 279–309. van Zonneveld, R. (1994). Kleine Syntaxis van het Nederlands. Dordrecht: ICG Publications. Zei, B., & Sˇikic´, N. (1990). Agrammatism in Serbo-Croatian: Two case studies. In L. Menn & L. K. Obler (Eds.), Agrammatic aphasia: A cross-language narrative source book. Amsterdam: Benjamins.

968

RUIGENDIJK AND FRIEDMANN

APPENDIX A: INDIVIDUAL DATA – HEBREW

TABLE A1 Object noun phrases with and without an accusative marker: Spontaneous speech Accusative + definite NP

No accusative + definite NP

Accusative + indefinite NP

12 8 11 5 34 24 94

0 0 0 0 1 1 2

0 0 0 0 1 0 1

AL RA RN IE RS GR Total

TABLE A2 Object noun phrases with and without the accusative marker: Sentence elicitation tasks Accusative + definite NP

No accusative + definite NP

Accusative + indefinite NP

43 48 15 31 60 5 17 219

2 0 0 0 1 0 1 4

1 0 1 1 1 0 0 4

AL RA HY ML SB MA AE Total

APPENDIX B: INDIVIDUAL DATA – DUTCH

TABLE B1 Subject DPs (complete noun phrases) and NPs (incomplete noun phrases) with a finite verb, a non-finite verb and without a verb Subjects With finite verb

With non-finite verb

Without a verb

Participant

DP

NP

DP

NP

DP

NP

1 2 3 4 5 6 7 8 Total

15 22 6 0 27 26 28 4 128

0 0 1 0 0 0 1 2 4

1 4 0 2 0 0 0 7 14

0 0 0 3 0 0 1 1 5

2 3 1 4 1 0 0 2 13

0 0 0 4 0 0 0 3 7

Total subjects

Omitted subjects

18 29 8 13 28 26 30 19 171

4 1 11 11 0 1 0 4 32

CASE IN AGRAMMATIC PRODUCTION

969

TABLE B2 Object DPs (complete noun phrases) and NPs (incomplete noun phrases) with a finite verb, a non-finite verb and without a verb. Object With a verb

Without a verb

Participant

DP

NP

DP

NP

1 2 3 4 5 6 7 8 Total

14 12 8 9 1 8 23 6 81

2 2 2 4 0 6 2 3 21

1 2 0 2 0 0 0 1 6

0 2 0 7 1 0 0 1 11

Total objects

Omitted objects

17 18 10 22 2 14 25 11 119

5 12 9 2 26 13 5 10 82