Selecciona una palabra y presiona la tecla d para obtener su definición.

––––––––   144   ––––––––


- 3 -

––––––––   145   ––––––––

––––––––   146   ––––––––




Theoretical Linguistics

Jorge Guitart87

A Functional Approach to the Accusative A

Maureen Weissenrieder

Ohio University

1. Introduction

This paper is a preliminary study of the use of the Spanish preposition a with inanimate direct object nouns (DOs) such as those used in sentences 1 through 5 below88:

1. A primera vista pudiera aparecer que los gerundios modifican al sujeto...


2. Llamaremos oración sólo a las oraciones independientes89.


3. Algunos gramáticos no consideran oración a la secuencia con verbo.

(LB: 12)                

4. Propone considerar al adverbio... como un intermedio.

(NR: 133)                

5. Predicado abarca el predicado y al atributo de los franceses.

(NR: 15)                

My goal is to show that one principle based on the notion of functional noun saliency explains the distinct circumstances under which the preposition a is used obligatorily in sentences 1 and 2, optionally in 3 and 4, and optionally and rather uniquely in 5. I will proceed by describing the properties of such constructions at the lexical, sentence, and discourse levels, and will end by discussing the general principles that condition the preposition’s appearance.

2. Background

It is a well established fact that Spanish marks some, but not all direct object nouns. Marking is highly correlated to human, and to a lesser extent, animate noun phrases (NPs) and to the degree of referentiality of those NPs. Thus, proper nouns referring to people are marked almost without exception because of their unique reference to a single individual, while DOs referring to inanimates are generally unmarked. The correlation of the preposition’s appearance with the [+human] feature has been prominent enough for the preposition to be characterized pedagogically as the ‘personal a’. Contrastively, the less frequent appearance of the preposition with inanimates has been considered exceptional and unpredictable. Thus, Ramsey (45) describes its use with inanimate NPs in conjunction with the comparative como (see example 4 above) as «according to taste». In a like vein, Bello (283) explains the preposition’s use with the verbs preceder and seguir as «caprichos de la lengua».

Despite these claims for exceptionality, two principles arose which attempted to explain the preposition’s use as systematic. One, the ambiguity principle [AP] (Ramsey, 1956; Solé/Solé, 1977; Butt, 1988), describes the preposition as marking DOs in the event that they cannot be distinguished from their subjects. Ambiguity of this type normally occurs where inherent NP characteristics (i.e., number, animacy, etc.) are identical in subject and DO nouns. In languages such as Spanish where freedom of word order putatively assigns no fixed position to subject and object, distinction of syntactic roles, it is claimed, is impossible to determine90. Thus in the phrase El más simple adjetivo que acompaña al sustantivo..., one could argue that without the preposition, one would not know which noun phrase accompanied which. In fact, native informants offer such explanations as objections when one tries to remove the preposition from sentences where it is required.

It has been argued elsewhere (Gili y Gaya, 1964; Isenberg, 1968; Lujdn, 1978; Weissenrieder, 1985) that the AP has several weak points. It is appealing to believe that rules of language are motivated by principles, such as the AP, that dictate that ambiguity should be avoided. At the same time, however, it is equally clear that natural

––––––––   147   ––––––––

languages have a tolerance for ambiguity and the AP does little to further an explanation of why certain types of ambiguity should not be acceptable. More specifically for Spanish, the AP falls short for those constructions where the preposition is necessary but where no ambiguity would result from its removal (as in Carlos vio a sus amigos). In addition, since it can be shown that Spanish does not have complete freedom of word order for subject and DO noun phrases (Weissenfeder, 1985), the foundation on which the AP rests is greatly undermined.

The second principle offered to explain the preposition’s use is elaborated in Fish (67) where marked nouns are defined as «conspicuous» or «singled out»91. Conspicuousness occurs in the event that the noun is an «individual» which in turn, is defined by the relative value assigned the noun’s referent in the natural world, or by its equality or superiority to a subject/actor noun in the grammatical world. For inanimate direct objects, Fish subsumes several categories of verbs and constructions whose DOs are conspicuous and thus marked. Many of these categories overlap with constructions treated here and thus deserve mentioning. They include:

l. Verbs of naming such as llamar which result in the DO becoming individualized.

2. Verbs of relative order and sequence such as preceder and seguir.

3. Verbs of «close association». Fish includes in this category verbs of inclusion, exclusion, union and separation as well as disparate items such as caracterizar, acompañar, and sustituir.

4. Unusual, laudatory, or derogatory comparisons. In this category are included sentences such as En otros tiempos yo le hubiera tomado a esta casucha por un palacio or Un día verás al valle verde de alfalfa [emphasis mine].

Fish offers no particular explanation of why many of these constructions should make their DO noun phrases more conspicuous than others, nor why an individual noun should somehow be more «conspicuous» than other noun phrases. In other words, Fish lacked an encompassing, communication-based theory which would enable him to explain the interrelationship of noun phrases within the sentence. Ultimately, Fish resorts to taxonomy to explain «conspicuousness». As with most taxonomies, categories often turn out to appear «somewhat arbitrary», to use Fish’s own words. Despite these weaknesses, I believe that Fish’s approach was basically correct.

In order to reconfirm Fish’s findings while refining my understanding of the preposition’s use with inanimates, I collected cases of marked DOs from three corpora of expository prose. These corpora are listed below with the initials with which they will be identified in this study:

LB:   Lope Blanch, Juan. 1979. El concepto de oración en la lingüística española. México: UNAM.

LM:   Luján, Marta. 1980. Sintaxis y semántica del adjetivo. Madrid: Cátedra.

NR:   Navas Ruiz, R. 1963. Ser y estar: estudio sobre el sistema atributivo en español. Salamanca: Universidad de Salamanca.

I realized that I would need corpora that contained not only a high number of inanimate DOs but also a high number of inanimate subjects acting upon them. I hypothesized that scientific discourse would provide a greater number of examples than other text types92. Indeed, the three corpora -550 pages in length- produced 271 cases of marked DOs, compared to literary corpora of a previous study (Weissenrieder, 1990), which although similar in size, produced markings of inanimates in a few isolated cases. I collected samples on two separate readings of the corpora. On the first reading, I gathered all examples of marked DO nouns. On the second reading, I collected sentences which contained verbs identical to those gathered in the first reading, but whose DO nouns were not marked. From these data, patterns at the lexical, sentential, and discoursal levels were extracted.

3. The Lexical Level

The direct object of 57 distinct lexical verbs occurred with the preposition, as listed below, in an order reflecting a descending frequency of occurrence with regard to marking. Items in bold indicate that the verb did not appear without the preposition.

Marked more than once with a: modificar; llamar; considerar; designar; definir; acompañar; seguir; caracterizar; concretar; preceder; calificar; distinguir; excluir; especificar; implicar; contener; introducir; situar; afectar; colocar; clasificar; diferenciar; explicar; encuadrar; escapar; integrar; separar; ver.

Marked only once with a: analizar; abarcar; agrupar; asumir; ayudar; asimilar; concordar; circunstancear; confundir; concebir; desbancar; encasillar; encerrar, enumerar; gobernar; hallar; identificar; juzgar; mantener; marcar; poner; mencionar; realizar; regir; reemplazar; señalar; tratar; tildar; transformar.

It should be noted that these occurrences are

––––––––   148   ––––––––

not as categorical as they may seem. The verb llamar, for example, appears in italics because it occurs categorically with a in the event that it is construed as a verb of naming. Other instances of llamar, such as in the phrase llamar la atención do not carry the marker. It is possible to explain marking by separating llamar, ‘to name’ from llamar ‘to call [attention]’. Consonant with this approach, Ramsey and Fish recognize the category of verbs of ‘naming’ as part of their explanation of the preposition’s use. An alternate approach would be to construe the lexicon in such a way that verbs such as llamar would be considered single dictionary entries whose diverse meanings are derived inferentially from the contexts in which they appear. In other words, sub-meanings of llamar and other verbs like it, would be derived from the context, which would include the presence or absence of marking.

The majority of verbs that repeat in the corpora do not appear categorically with the preposition. Moreover, the first eight most frequently marked verbs demonstrate that frequency of marking varies with different items. Note Table I below, where the instances of marking and nonmarking for particular verbs shows that modificar favors the appearance of the preposition (50 marked instances; 30 unmarked instances) while designar occurs more frequently without it (14 marked instances; 33 unmarked instances)93.

Number of Cases --Marking to Non-Marking
ML LB NR Totals
modificar 31/29 0/0 19/1 50/30
considerar 7/3 7/1 4/14 18/18
designar 4/11 5/10 5/12 14/33
definir 2/9 3/7 7/5 12/21
caracterizar 3/1 3/1 1/0 7/2

Given the limited number of occurrences for each verb, it is impossible to make statements of significance. Nonetheless, it is still interesting to note the extent to which there is consistency across corpora, not only in terms of which lexical items appear with the preposition, but with which verbs the preposition’s use is variable. In the case of the verb designar, even the rate of occurrence is consistent. If indeed the preposition’s optional appearance occurs «according to taste» as Ramsey has de scribed it, it is remarkable to what extent these three authors have essentially the same taste.

As in the case of Fish’s earlier study, the items of these corpora seem to form lexical classes readily. In fact, the notion of placement seems to characterize a great number of them. There are verbs of direct placement such as colocar, poner; verbs of separation, distinction, or placement apart such as distinguir, diferenciar, encasillar, encuadrar; verbs of sequencing or placement before or after such as seguir, preceder, and the list could go on. Also notable are the verbs of naming and singling out such as llamar, identificar, especificar, etc. Despite the lexical patterning, the use of the preposition with these verbs is not always categorical, thus demonstrating that the appearance of a is not a part of the lexical entry for the verb. It is, therefore, necessary to investigate properties at other levels of analysis to discover why the preposition is used.

Finally, there are many verbs, most occurring only once, that do not fit into the categories of placement and naming mentioned above. The great semantic diversity in these verbs vitiates any attempt to explain the preposition’s use by way of lexical categories alone.

4. The Sentence Level

The majority of marked DO noun phrases within the corpora occur within two construction types. Marking occurs when:

1. In one-complement constructions, DO noun phrases demonstrate activity levels equal to or greater than activity levels of their corresponding subjects.

2. In two-complement constructions, DO noun phrases occur in opaque syntactic constructions where DOs have subject-like qualities.

In what follows, I will explain these two patterns in some detail, give examples of each type, and finally show how one principle underlies marking in both construction types.

In order to understand the single principle underlying marking, it is important to consider the roles that NPs play in overall language use. From a communicative point of view, nouns can be defined as linguistic referents for real-world participants, who, in turn, are involved in real world activities, processes, or states94. These activities, processes or states are represented linguistically through verbs. Some participants are more active than others. The level of activity assigned to a particular individual may depend on the real world. Human beings, for example, by their nature are more active than inanimates. Even within categories of individuals, levels of activity can vary. Thus, in general, humans who

––––––––   149   ––––––––

are «named» may be considered more capable of individual control and responsibility than nameless groups of humans.

Activity level can also be assigned linguistically through case role. In a given sentence, one human referent may be more active than another because it is assigned subject-as-agent status. The willful control of an action exercised by the «doer» represents a higher level of activity than that of a direct object noun functioning as patient, even if the DO is another human being.

It is logical to assume that in the variations of sentence texture that are created by the interplay of relationships between NP participants, one NP should demand the most attention of the language processor. Generally, this NP is the subject. Restricted to one and only one constituent of the clause, the subject has properties that define it as unique. Its propensity towards highly individualized nouns, defined as more human and referential, and with agentness, correlate with the tendency for human beings to pay more attention to members of their own species95. As a focalized participant, the subject tends to be what the sentence «is about». In this pragmatic role as sentence topic, the subject again demands a higher level of attention from the listener.

Since the subject demands greater focus from the language processor, we may say that it displays a general saliency in comparison to other NP participants. Saliency, however, can be usurped by other grammatical functions. Noun phrases in other syntactic roles can display the same characteristics (animateness, for example) that are associated with subject NPs, They thereby compete with the subject for saliency. In fact, when such competition between subject and DO nouns occurs, the latter often carry the so called «personal a».

Independent noun phrase characteristics, however, are not the only element involved in giving status to DO noun phrases. Lexical verbs themselves may endow DO noun phrases with features that make them vie for saliency with regard to their subjects. Thus, verbs can assign their DO’s unexpected characteristics. This is common with cases of «personification» where inanimate direct objects are elevated to animateness through the verb. The verb imitar in the sentence El niño imitaba al tren requires animacy features to be assigned to the inanimate tren. Not fortuitously, it also occasions marking.

Case assignments can also be made through the lexical verb. Consider the clear example used by Fish. If one were to see a man and a dog walking down the street and wished to use the English verb ‘accompany’ to describe the situation, one would be more likely to state that the dog was accompanying its master and not that the master was accompanying his dog. The lexical verb accompany assigns to its DO a level of importance at least equal, if not superior to that of its subject. Thus, the grammatical DO is usurping the pragmatic role generally assigned to the subject. This is unusual and in Spanish it would be marked accordingly: El perro acompañaba al dueño/El adjetivo acompaña al sustantivo.

4.1 Pattern I: Saliency and One-complement Constructions

In a previous study (Weissenrieder, 1985), I dealt with verbs of placement such as acompañar, preceder, seguir, and reemplazar, which are generally described as occurring obligatorily with the preposition, either as due to the AP or to their member ship in an exceptional lexical class. My findings demonstrated that some of these verbs were marked, not due to ambiguity nor as a reflex of the lexical verb. Rather, as predicted, they were marked when the activity level of the DO noun was equal to or greater than the activity level of the subject noun. Thus, as indicated in sentence 6 below, the object is obligatorily marked with regard to its subject, which is interpreted as non-agentive. The object of sentence 7, on the other hand, cannot be marked in the case that its subject is given agentive interpretation.

6. El profesor reemplaza al libro. ‘The professor takes the place of the book’.

7. El profesor reemplaza el libro. ‘The professor replaces the hook (with something else)’.

These findings are reminiscent of Fish’s statement that an object noun is marked when «an inanimate object acts conspicuously upon an object of its own or a similar kind». Similarity of subject and object noun, according to my findings, would be defined more specifically not only with regard to noun class or inherent NP characteristics (e.g., animacy) but with regard to the activity level in the case role assigned to the grammatical functions at the sentence level. Since higher activity correlates to saliency, marked noun phrases call for greater attention than unmarked nouns.

Within this construction type, verbs such as preceder are categorically marked since they always assign the same case roles to their NP participants. On the other hand, verbs such as

––––––––   150   ––––––––

reemplazar can assign various sets of case markings to NP participants. Duplicity of case assignments explains some of the variability of marking that is found in the case of such verbs.

Within our corpora, some of the most frequently marked verbs fit within Pattern 1: modificar, acompañar, preceder, etc.

4.2 Pattern II: Saliency in Two-complement Constructions

Although the activity level of subject and DO nouns, along with their related properties, helps to explain the use of the preposition within pattern I, they do not seem to have immediate applications to other verbs found in the corpora. This is particularly obvious in the case of verbs that have human, active subjects and typically inanimate DOs such as is exemplified in sentences 3 and 4, repeated here for reader ease.

3. Algunos gramáticos... no consideran oración... a la secuencia con verbos.

4. (Él) propone considerar al adverbio como un intermedio.

Two questions immediately occur with regard to these constructions: Why are they marked more than other DO noun phrases and to what extent can an explanation of marking be related to previous findings? Insight into some of these questions can be found in a study by Davison (1984) which deals with the relationship of certain syntactic constructions to the definition of sentence topic. Although the subject is generally the topic of the sentence, certain types of derived sentences have DO nouns which, because of their saliency, may cancel out or upstage saliency on part of their grammatical subjects. It is Davison’s purpose to prove that sentences such as 9 below are indeed marked in English, that marking leads to noun phrase saliency for the DO in terms of processing, and that such DO nouns are, therefore, often more pragmatically interpreted as sentence topic. Since topic is a role usually given to the subject, the DO noun phrases of the se sentences usurp the subject’s spotlight. The argument that these constructions are unusual and thus marked follows from the fact that there are two constructions, one normal and underived (as in sentence 8) and the other marked and derived (as in sentence 9).

8. The chief inspector believed that Smith’s murderer was insane.

9. The chief inspector believed Smith’s murderer to be insane.

As can be seen, what Davison is referring to is a case of what was once known as «raising». In sentence 9, the object of the verb believe is the subject of the underlying predication Smith’s murderer is insane. Such raised nouns tend to evince subject-like properties, not only by the semantic role they play in underlying interpretations, but in that nouns that are highly nonreferential are somehow less felicitous in these constructions. Thus, non-referential NPs receive shaky readings in sentences such as 10.

10. The chief inspector believes a murderer to be insane.96

11. The chief inspector believes that a murderer is insane.

It is also the case that subject NPs in general tend to be referential (Givón 1979, 1984).

Based on these and similar constructions, Davison creates a hierarchy of syntactic constructions whose nouns are marked, salient, and thus most likely to usurp the subject’s default role as sentence topic. Raising-to-object is highly ranked in terms of its association with markedness. The similarity between the constructions Davison treats and the Spanish sentences 12 and 13 below are obvious:

12. Algunos gramáticos a quienes nosotros seguiremos, no consideran oración a la secuencia con verbo...

13. Algunos gramáticos no consideran que la secuencia con verbo sea oración...

The element functioning as the subject of the embedded clause in sentence 13 is the DO in sentence 12. The resulting grammatical opacity of the DO role in 12 confers upon it a saliency normally reserved for the subject. By virtue of its saliency, the DO in sentence 12 is more likely to be processed as sentence topic-a role generally associated with the subject.

Indeterminacy regarding roles only occurs when there are two complements. Without the second complement, the considerar construction is a prototypically transitive one, where the active subject dominates its inanimate DO. Such constructions as Consideren estos datos ‘Consider these data’ occur seven times in the data. As might be expected in these cases, marking is categorically absent.

Many of the marked constructions found in the corpora are similar, but not identical to, the considerar pattern. Llamar, when accompanied by a, is found with what is traditionally known as an object complement, but unlike considerar, has no corresponding transform with a sentential complement. More striking are the number of

––––––––   151   ––––––––

verbs with a como complement as in Propone considerar al adverbio... como un intermedio. Again, this sentence cannot be derived by some relation-changing transformation but the amorphous duality of the marked NP as DO of the verb considerar and as the underlying subject of the attributive predicate intermedio is clear. Interestingly, over a third of all verbs marked with a in the corpora are of this construction type.

4.3 Pattern III: The Case of Caracterizar

The item caracterizar seems to be a mid-verb whose profile spans the two verb groups described above. As can be seen in sentences 14-17 below, caracterizar can occur with both an animate and inanimate subject, as well as carry a como complement.

14. Por un lado, [él] caracteriza la oración por su contenido semántico.


15. Un análisis adecuado debe caracterizar esta relación de hiponimia.


16. La autonomía semántica sirve solo para caracterizar a la cláusula.


17. Resulta sorprendente que trate aquí Amado Alonso de caracterizar a la proposición como «miembro de oración con forma de oración».

(LB: 32)                

Sentences 14 and 15 are similar in that they contain active, dominant subjects and their DO’s remain unmarked. Sentence 16, on the other hand, contains an inanimate subject which is not dominant. The inanimate subject of sentence 16, la autonomía semántica is a characteristic or part of the DO, la cláusula. Thus, this relationship might even be rephrased by saying that La cláusula tiene autonomía semántica. Compare this with sentence 15, which also has an inanimate subject. The subject, un análisis adecuado, however, is not a characteristic of the DO, esta relación. In fact, análisis adecuado can be replaced by an animate subject such as Ese investigador. In other words, the inanimate subject of 15 is much more animate-like, and thus subject-like, than the inanimate subject of sentence 16. The resulting relationship between subjects and their DO is thus different despite the same ness in verb and the NP characteristic of animacy. The saliency of the DO over subject in sentence 16 occasions the marking. In sentence 17, an active agent dominates a less dominant DO, comparable to 14, but the como complement confers subject-like saliency to the DO in a fashion similar to the case of considerar discussed above.

From these data, it can be hypothesized that the DO of caracterizar will be most marked when non-agentive subjects are involved, variably marked when the agentive subject and a como complement are present, and least marked when an agent subject acts on a single complement. Although data within the corpora confirm this hypothesis, insufficient numbers keep us from making any statement of significance concerning it.

5. The Discourse Level

We have seen that NPs take on saliency through role assignment, which is conditioned by lexical verbs and their corresponding semantic cases. Actually, this is somewhat misleading. It treats language as a series of isolated constructions at sentence level. Inasmuch as most communication is multi-propositional, nominals may be said to derive primary saliency, not from sentence-level phenomena alone, but from their overall role within discourse. While subject and DO appear at the sentence level to be syntactic categories with varying case assignments (subject-as-agent, direct object-as-patient, etc.), at the level of discourse, they are defined as pragmatic categories whose primary role is that of topic assignment. Subjects are the most likely candidates for topicality, DOs follow as second most likely topics. Not coincidentally, subjects are most likely to be agents and, in turn, agents are most likely to be human, definite, active individuals. Thus, the implications of case assignments, discussed above with regard to marking at the sentence level, actually work within discourse, rather than independently of it. This insight explains why no description based solely on NP characteristics, lexical verbs, or case assignment, will account for all uses of the preposition. Although these semantic elements are discourse sensitive, there are other, discours emotivated and sometimes discourse-exclusive elements which condition the preposition’s appearance. I will turn to some of these now.

5.1 Referentiality

It is widely known that the preposition a is used to mark DO human beings, provided that the speaker has a specific individual in mind. When no such individual is envisioned by the speaker, the DO is unmarked, despite the [+human] feature. Thus, in the sentences below, un hombre is non-referential; a un hombre, referential:

18. Busco un hombre que hable inglés. ‘I am looking for [any] man who speaks English’.

––––––––   152   ––––––––

19. Busco a un hombre que habla inglés. ‘I am looking for a [specific] man who speaks English’.

Since referentiality is not necessarily a characteristic inherent to the noun, but rather, a perspective given to it within discourse, it can be concluded that discourse-related issues such as the importance of the noun’s overall role in discourse are pertinent to the understanding of the preposition’s use.

The corpora under study contain patterns similar to sentences 18 and 19, but for inanimate DO nouns. Consider the case of modificar, the most frequent Pattern I verb in the corpora. Out of 80 total cases, there are 50 marked and 30 unmarked instances. Two questions arise. Why should DOs with modificar be marked at all, and given that they do carry the preposition, why should 38% of the cases remain unmarked? Using the analysis given above for sentence-level phenomena, we find that modificar is a prime candidate for such inconsistency. The item modificar can assign several case markings to its subject. Thus, one would not normally expect marking when the subject is agent as in the sentence El empleado modificó el horario del tren. Even with the over whelmingly inanimate subjects of our corpora, marking is less likely when the subject is considered «relatively more active». Higher levels of activity are assigned to the subject when modificar is interpreted as «to change», as is shown below:

20. De ahí que estar modifique en cierta medida la naturaleza del sustantivo hasta convertirlo en adjetivo con frecuencia.

(NR: 162)                

‘Thus, [the verb] estar somehow changes the nature of the noun converting it frequently into an adjective.’

21. Este resultado se podría corregir postulando una transformación que modificara el orden de estos elementos...


‘This result could be corrected by postulating a transformation that would change the order of these elements...’.

Since two of the corpora deal with a linguistic description of adjectives, however, the most common use of modificar is with the meaning ‘to match grammatically’. As we might expect, there is a strong tendency towards marking when modificar is interpreted with this meaning. Both subject and DO nouns are equally inanimate and inactive, as in sentence 1, repeated here for reader ease:

1. A primera vista pudiera aparecer que los gerundios modifican al sujeto...

Nonetheless, 30 cases of modificar ‘to match grammatically’ remain unmarked. With the majority of these cases, there are linguistic correlates that indicate that the noun phrase’s saliency has been «down-played» in the discourse through a lack of referentiality. As might be expected, most marked nouns with modificar are definite, and thus referential (only 16% or 8 cases out of 50 are indefinite) while a higher proportion of non-marked cases are indefinites (70% or 21 instances of the 30 non-marked cases). As predicted, wherever non-referentiality is clear, marking is absent, as in shown in sentences 22 and 23 below. The relative clause containing the subjunctive in sentence 22 signals the nonreferentiality of the head NP, as does the indefinite pronoun algo in sentence 23.

22. Este análisis conduce a una explicación natural del hecho relacionado de que las restrictivas puedan modificar sustantivos propios que estén construidos con determinantes.


‘This analysis leads to a natural explanation of the related fact that restrictivas [restrictive clauses] can modify [any] proper nouns that are construed with determiners’.

23. Si los adjetivos sustantivados derivan de cláusulas relativas restrictivas de la estructura subyacente, entonces hay que suponer que estas cláusulas modifican algo...

(ML: 124)                

‘If the nominalized adjectives derive from relative restrictive clauses in the deep structure, then it must be supposed that these clauses modify something’.

For the most frequent Pattern II verb, considerar, different referentiality patterns appear. With the exception of one unmarked case, all the DOs of the item considerar are definite. Marked items, however, are more «individualized» than unmarked items. Often, they refer to the names of specific items rather than to abstractions. Thus, marked items include (el pronombre) lo, el adverbio bien, la interjección, el verbo atributo, los adjetivos adverbiales. Unmarked items include abstractions such as la concordancia, la medida, la valoración, la única razón. These findings suggest that referentiality depends upon inherent NP characteristics, as well as on the perspective given to the noun in discourse.

5.2 Topicality

The fact that a should correlate highly with referential nouns is not an isolated, semantic fact about its usage. Research on other languages, (See, e.g., Givón, 1984) has demonstrated that referentiality correlates highly with greater topicality within discourse. In other words, referential nouns may correlate more with what the

––––––––   153   ––––––––

discourse «is about». Since topical referents are considered to be important to the message, referential nouns, if found to be more topical, may also be considered to be more important.

In order to determine whether marked nouns were more topical than unmarked nouns for the verbs in Patterns I and II, I studied the 109 paragraphs in which the verbs modificar/considerar occurred. I worked under the assumption that a referent that is mentioned repeatedly within the same paragraph is more likely to be topical than unrepeated referents. As a rough estimate of topicality, I identified NP referents in clauses immediately preceding or following the DO that was being studied. If the DO referent was repeated in a previous or subsequent clause, that NP was considered topical. If it was not repeated, it was considered to be an independent referent97. Thus, in paragraph 24, the noun interjección, mentioned seven times in full and pronominal form, is considered topical.

24. De la complejidad del problema que supone clasificar las interjecciones son buena prueba las vacilaciones que se advierten en algunos autores. García de Diego, por ejemplo, después de haber considerado a la interjección como una simple parte de la oración (...), cambió de parecer, aunque sin llegar a una solución precisa y definitiva. Por un lado opina que hay interjecciones de función verdaderamente oracional, a las cuales «lo mismo da llamarlas oraciones que equivalentes» de oración, en tanto que hay otras «que no pueden llamarse lenguaje más que en un amplio sentido, como llamamos lenguaje a algún gesto que intencionalmente es expresivo»; y además, «interjecciones conversacionales hay que, gramaticalmente, no hay medio de explicar, porque no se da con el tipo de la frase completa reconstruida».

(LB: 67)                

On the other hand, the noun un núcleo mentioned only once in paragraph 25, is considered less topical.

25. En cambio, las frases introducidas por sino en los ejemplos (109) y (110) presentan una de las cláusulas relativas, pero no llevan implícitas la otra cláusula restrictiva. Es válido, entonces, concluir que el orden de las cláusulas restrictivas en una secuencia que modifica un núcleo sustantivo conlleva una interpretación semántica determinada.

(ML: 75)                

Findings regarding topicality vary for the two verb patterns. As is demonstrated in Table II, the DOs of modificar, marked or unmarked, tended lobe independent referents. Moreover, topicality did not seem to correlate significantly with marking.

Topicality Count for DO’s of the verb Modificar
Marked Unmarked
Independent reference 30 (60%) 19 (63%)
Topical reference 20 (40%) 11 (37%)

This lack of correlation between topicality and marking is exemplified in passages 26 and 27. In 26, the less topical noun sintagma, which is mentioned only once, is marked. (See first sentence):

26. Finalmente pasemos al hecho de que las cláusulas restrictivas y no las apositivas, puedan modificar libremente a los sintagmas nominales con los determinantes cualquier-, todo, y cad-. La explicación de este hecho proporciona el apoyo más fuerte para mi análisis, pero predice ciertas observaciones que, de ser correctas, revelan que ciertas generalizaciones que se aceptan sobre las cláusulas son falsas.

(ML: 63)                

In contrast, the more topical noun proposición, mentioned five times in either full or pronominal form, remains unmarked. (See first sentence):

27. Una explicación similar a la que precede puede ofrecerse para el hecho de que las cláusulas apositivas puedan modificar proposiciones completas, no así las restrictivas. Una proposición parece singular; por lo menos no hay manera de pluralizarlas. Las proposiciones no designan a un conjunto. Considérese, por ejemplo, la siguiente conjunción de proposiciones.

(ML: 62)                

For considerar, on the other hand, DOs were more topical overall. (See Table III). More importantly, marked nouns tended towards topical reference98.

Topicality Count for DO’s of the verb Considerar
Marked Unmarked
Independent reference 4 (22%) 8 (73%)
Topical reference 14 (78%) 3 (27%)

This difference between the marked and unmarked cases of considerar is demonstrated in the following two passages. In 28, the author is discussing the use of the copula ser with the adverb bien. The marked noun bien, because it is the name of an individual adverb, is considered highly specific (referential). Moreover, it is used repeatedly in the paragraph. (The sentence containing considerar is the last in the paragraph.)

28. Más interés tiene otro artículo de Bolingeren que se refiere al empleo de ser con el adverbio bien. Alude a la solución de Cuervo que afirma que ser bien equivale a ser cosa

––––––––   154   ––––––––

oportuna y que, por lo tanto, bien tiene valor de good en inglés: well y bien significan una satisfacción social; good y bueno, una satisfacción personal. Propone considerar al adverbio bien en este caso como intermedio de adjetivo y adverbio.

(NR: 133)                

In 29, the author is discussing how to analyze nominalized adjectives. The DO in question, la única razón (see last sentence of paragraph), is abstract and refers to no specific referent in the paragraph:

29. En la sección primera recitaron tres argumentos convincentes que demuestran que los adjetivos sustantivados no funcionan como sustantivos, sino que en estas frases cumplen la misma función modificadora que tienen cuando aparecen después de un sustantivo. Por consiguiente, considero anulada la única razón en la que se basaba el análisis de lo como el artículo neutro.

(ML: 129)                

The data suggest a significant difference for Pattern I and Pattern II verbs. Marking with the item modificar correlates highly with referentiality but not with topicality; marking with considerar, on the other hand, seems tied to both referentiality and topicality. These differences suggest that the preposition’s appearance with modificar is motivated more by case assignment at the sentence level. In this case, the DOs high activity level (relative to its subject) causes the DO to be marked. The DOs lack of referentiality, which is a discourse-based factor, cancels the importance of the DO and thus, the need for marking. The preposition’s appearance with considerar, on the other hand, although conditioned by sentence-level features, seems to be based on discourse factors (referentiality and topicality). In other words, marked nouns with modificar are «down-played» in discourse through lack of referentiality. Unmarked nouns with considerar, on the other hand, seem «upstaged» to marking when they are more topical. If this interpretation turns out to be the case, it would explain why native informant judgements on isolated sentences involving modificar seem much more categorical than such judgements with regard to considerar. Direct objects with modificar seem more obligatorily marked than similar objects with considerar.

5.3 Thematic importance

There are infrequent instances of the preposition’s appearance that fit into none of the patterns described above and seem to defy analysis. These «stylistic» variations again seem to depend on the importance given the NP within discourse. When the preposition appears exceptionally, informants will explain that such marked nouns indicate that the speaker found the noun somehow unusual, surprising, or otherwise worthy of being singled out. In other words, these nouns demand attention and the saliency normally assigned the topic of discourse may be displaced to the DO noun. This is the case in the following passage where the verb abarcar has conjoined DOs, one unmarked, the other marked:

30. En la terminología germánica, por to tanto, el término atributo designa globalmente lo que en español diferenciamos mediante términos como aposición, calificativo, determinativo, etc. Predicado abarca el predicado y al atributo de los franceses.


‘In Germanic terminology, of course, the term attributive globally designates what in Spanish, we would distinguish through terms such as apposition, qualificative, determiner, etc. Predicate includes the predicate, AND, the French attributive’.

An interpretation to this unusual usage is easily found. Out of context, the phrase Predicado abarca el predicado might seem tautological. Within context, the author contrasts the unexpected definition of predicado as el atributo de los franceses. The surprise with which the author wishes to introduce the DO shows up in the marking. Other instances of abarcar in the corpora are unmarked.

Although informants will confirm this interpretation as consistent with their intuitions, such interpretations are dissatisfying as linguistic descriptions. In such cases, there are no measurable «correlates» that support the notion that the NP is indeed more important in discourse than other unmarked nouns. It is clear why such optional uses would lead one to believe that the use of a is purely «according to taste». Yet, the fact that language users recognize that the preposition singles out the noun phrase remains in accord with the general principle of noun phrase saliency.

6. Concluding Remarks

In this study, I have used the principle of NP importance within context to explain the appearance of the so-called «personal a». Importance is defined from the functional perspective of the listener. Listeners must know which noun to single out in order to process more efficiently what they hear. Nouns that are somehow unusual call attention to themselves and thus, are foregrounded in the discourse. These nouns are considered salient.

Saliency is conditioned by phenomena at

––––––––   155   ––––––––

lexical, sentence, and discourse levels of analysis, but cannot be explained exclusively by any one of these levels. At the lexical level, some NPs are more important than others by virtue of their semantic characteristics alone. Thus, human nouns are more important than inanimate ones. Lexical verbs also play a role in assigning importance to the noun. Through personification, for example, inanimate DO nouns can be elevated to human status, and require marking accordingly.

At the sentence level, the saliency of the DO is determined relative to the status of its subject. Subjects are salient by virtue of their pragmatic role as topic, but DOs may usurp that status when they are made conspicuous. Conspicuousness occurs when DOs, by virtue of their inherent NP characteristics or by their case assignments, are equal to or greater than their subjects in activity level.

Select verbs have the potential of assigning several case roles to their subjects. Inanimate subjects are interpreted as having «lower» activity level when the corresponding DO noun is marked. This explains why certain lexical verbs appear with and without the preposition, often with distinct lexical interpretations (for example modificar ‘to match grammatically’/modificar ‘to change’).

Where two complements appear in the same sentence, one stands out more than the other. In this study, those DO noun phrases that evidenced subject-like qualities relative to the secondary complement were marked. Fish explained the preposition’s appearance in the sentence Un día verás al valle verde de alfalfa as being an «unusual» comparison. Indeed, underscoring the surprising nature of the comparison may have been the communicative purpose of the preposition in the discourse in which the sentence was found. Nonetheless, the syntax also favors Marking by semantically equating valle with verde de alfalfa.

The association of marking with a particular verb, syntactic structure, or case marking is only part of the picture, however. Correlates that normally favor or disfavor marking at lower levels of analysis seem to take a back seat to the nouns’ overall role in the larger context. Animate and inanimate nouns, which one would otherwise expect to be marked, remain unmarked when non-referential. At the other extreme, highly unpredictable, «stylistic» occurrences of a follow from general patterns of discourse-related phenomena such as topicality. Thus, it is impossible to state that certain lexical items such as modificar ‘to modify grammatically’ appear categorically with the preposition, or that others such as considerar occur randomly with the preposition «according to taste».

Every student of Spanish has at some time or another sensed the frustration of finding a plethora of exceptions to every categorical rule. I can offer no real remedy for that frustration. Quite the contrary, I have exacerbated it, by showing that within real language data, categorical rules seem to have less than categorical application. If there is aconsolation, it lies in knowing that variablility is a predictable part of language, and assuming that we uncover correct principles, we should be able to explain where and how seemingly random «exceptions» are likely to occur.

––––––––   156   ––––––––

Works Cited

Bello, Andrés and Rufino J. Cuervo. 1945. Gramática de la lengua castellana. Buenos Aires: Sopena.

Butt, John and Carmen Benjamin. 1988. A New Reference Grammar of Modern Spanish. London: Edward Arnold.

Bentivolgio, P. «Topic Continuity and Discontinuity in Discourse: A Study of Spoken Latin-American Spanish», in T. Givón, (ed.) Typological Studies in Language, Vol 3.1983. Philadelphia: John Benjamins Publishing. 255-311.

Chafe, William L. 1980. The Pear Stories: Cognitive, Cultural and Linguistic Aspects of Narrative Production. Norwood, NJ: Ablex.

Davison, Alice. 1984. «Syntactic Markedness and the Definition of Sentence Topic». Language: 60: 797-846.

Fish, Gordon T. 1967. «‘A’ with Spanish Direct Object» Hispania 50: 80-85.

Gili y Gaya, Samuel. 1964. Curso Superior de Sintaxis Española. Barcelona: SPES y VOX.

Givón, Talmy. 1984. Syntax: A Functional-Typological lntroduction. Philadelphia: John Benjamins Publishing.

_____. (ed.) 1983. Typological Studies in Language. Philadelphia: John Benjamins Publishing.

_____. 1979. On Understanding Grammar. London: Academic Press.

Isenberg, Horst. 1968. Das Direkte Object im Spanischen. Berlin: Akademie Verlag.

Lope Blanch, Juan. 1979. El concepto de oración en la lingüística española. México: UNAM.

Luján, Marta. 1978. «Direct Object Nouns and the Preposition Win Spanish». Texas Linguistic Forum 10: 30-52.

_____. 1980. Sintaxis y semántica del adjetivo. Madrid: Cátedra.

Moravcsik, Edith. 1978. «On the limits of subject-object ambiguity tolerance». Papers in Linguistics 11: 255-59.

Navas Ruiz, R. 1963. Ser y estar: estudio sobre el sistema atributivo en español. Salamanca: Universidad de Salamanca.

Ramsey, M. 1956. A Textbook of Modern Spanish. New York: Holt.

Soló, Yolanda R. and Carlos A. Solé. 1977. Modern Spanish Syntax. Lexington: D. C. Heath.

Suñer, Margarita. 1982. Syntax and Semantics of Spanish Presentational Sentences-Types. Washington, D.C.: Georgetown University Press.

Weissenrieder, Maureen. 1985. «Exceptional Uses of the Accusative A». Hispania 68: 393-98.

_____. 1990. «Variable Uses of the Direct-Object Marker A». Hispania 73: 223-31.