Journal of Pragmatics North-Holland
145
7 (1983) 145-168
SOURCES OF VARIATION IN SPANISH VERB CONSTRUCTION USAGE: THE ACTIVE, THE DATIVE, AND THE REFLEXIVE PASSIVE
Susan
*
BERK-SELIGSON
A study of grammatical case usage in Costa Rican Spanish confirms the hypothesis that non-active constructions, specifically the dative and the reflexive passive, are used for the purpose of avoiding the attribution of blame to persons involved in adverse situations. In addition to choice of case, mentioning of agent, choice of lexical item. and discourse strategies such as qualifying expressions are shown to accomplish the attribution of responsibility. Linguistic universals in blame avoidance are suggested.
The present study constitutes an effort to mesh discourse and variationist approaches to sociolinguistic analyses, focussing on three verb constructions in modern-day Spanish. Whereas many scholars have analyzed what will be called in this study “the active”, “the dative of interest”, and “the reflexive passive”, these analyses have been entirely linguistic [l]. None has tried to analyze these grammatical forms from the point of view of language use. The present analysis attempts to do just that, using an experimental design for its data-gathering methodology. The results of this study, in conjunction with those of a similar one (Berk-Seligson 1971) [2] which had used informants from a wide range of Latin American Spanish backgrounds, indicate that the patterns discussed below are not unique to Costa Rican Spanish, but are common to many of the regional dialects of New World Spanish. The findings of this study confirm the existence of a culturally prominent mode of discourse in Costa Rican Spanish, namely, one which avoids both the mention of agents and the use of active voice in adverse situations. The * The author wishes to thank Keith Basso, Adrienne Lehrer, Susan U. Philips. and Mitchell A. Seligson for their helpful comments on earlier drafts of this article. Author’s address: Susan Berk-Seligson, Dept. of Foreign Languages and Literatures, Purdue University, West Lafayette, IN 47907, USA. [I] These grammatical constructions have been analyzed by Babcock (1970). Bolinger (1969). Carrasco (1973). Cartagena (1971). Cary (1972). Alonso Cortts (1939). Fazio (1971). Goldin (1968), Jordan (1973). Langacker (1970). Otero (1972), Perlmutter (1970) and Schroten (1972). [2] In 1971 a similar study was first carried out. Informants consisted of Puerto Rican bilingual school children living in New York City and college students coming from a number of Latin American countries and studying in the U.S.
0378-2166/83/0000-0000/$03.00
0 1983 North-Holland
146
S. Berk
-Seligson / Variafron in Spanish verb usage
findings also demonstrate a not unconscious awareness by Costa Ricans of the social implications of using one linguistic alternative rather than another for a given adverse situation. Furthermore, while it was hypothesized at the outset of the study that use of the grammatical alternatives would vary with the sex of the person describing the situation, this hypothesis was not borne out. Specifically, it was thought that Costa Rican female speakers, being in a socially subservient position relative to men, would tend to use grammatically passive constructions more often t,han male speakers. However, no such difference in usage was found, disconfirming the role of speaker sex as a source of variation for at least the present discourse object of study. Finally, the study discovers additional sources of linguistic options that were found to be available to speakers. These sources of variation were found to exist in both the internal components of the grammar (in the lexicon and the morpho-syntax) as well as in the sphere of discourse.
Indirect speech and the avoidance of blaming A prominent theme of Costa Rican culture is an unwillingness to attribute responsibility to human agents for the occurrence of some unfortunate or damaging circumstance. Essentially, this is a desire to avoid laying blame on persons for untoward events, especially if the person is oneself. Consequently, a common way of “covering” for small mishaps is to personify inanimate objects. For example, upon accidentally knocking over a container, a Costa Rican will refer to it with a phrase such as, Quiere uolar (‘It wants to fly’), which is semantically anomalous since ‘wanting’ can be said only of something animate. Hence, an expression such as this can be interpreted only in a metaphorical sense. Similarly, a reaction to seeing a guava fruit fall down, nearly knocking over a bottle of milk, might be the following utterance: Ese yuayabo, casi se me come la /e&e (‘That guava almost eats up the milk (on me)‘). Since there is no literal way in which a piece of fruit can ‘eat up’ some milk, this sentence, too, must be interpreted metaphorically if it is to be accepted as meaningful. Just as susceptible to personification is hot lard or cooking oil: in its capacity to produce injury by splattering, it may be portrayed as a culprit, or scoundrel. In fact, any inanimate object which is giving trouble in any way - even a jar that will not open - will be referred to as Ese bandido! (‘That bandit’) Another linguistic tactic for removing an actor from his/her action in the context of adverse events is obfuscation of the action to the point where nothing in the surface meaning of the sentence denotes what it is that actually happened. For example, when a women complains that the clothes on her clothesline frequently “get lost”, what she may really be saying is that her laundry is often the target of theft. The only way in which a non-native can
S. Berk -Sehgson
/
Variation
rn Spunish
cerb usage
147
know that the speaker’s sentence is not to be taken literally is by a combination of additional, clarifying remarks (e.g., “There’s a lot of theft around here, you know”) and non-verbal communication (e.g., conspiratorial eye winks accompanied by head nods pointing in the direction of a neighbor’s house). Thus, Costa Ricans make great use of indirect, roundabout ways of saying things to avoid openly having to accuse others of improper conduct. But perhaps the most widespread method for removing from an actor any responsibility for the event that he/she is somehow involved in is avoiding the mention of an agent in the statement of an action (including inexplicitly stated agents, such as verb endings which represent lst, 2nd, and 3rd person singular and plural, and which may or may not be accompanied by a preceding personal pronoun). Thus, in making a statement about the fact that John lost a letter, at least three common options are open to a speaker [3]: (1) Juan perdih lu curtu (‘John lost the letter’). (Or simply: lost the letter’).) (2) Se perdib lu curtu (‘The letter got lost’). (3) Se le perdib /a carta (‘The letter got lost (on him)‘).
Perdib la curtu (‘He
The first sentence employs an active subject, which is to say that the action of the verb centers on the subject. The second sentence uses what Woehr et al. (1974) call a sujeto comprometido (an involved subject), defined by them as a construction in which “the speaker is involved in the action or is emotionally affected by the action which the verb expresses; it introduces the hearer as a spectator to an event which unfolds before him” (Woehr et al. 1974: 156). This construction, a reflexive one which uses the reflexive particle se, has been traditionally labelled the “dative of participation” or “dative of interest”, but has been categorized as one of the several uses of the reflexive as a substitute for the true passive voice (i.e., the verb ser + past participle). Ramsey (I 894: 290-291), for example, classifies the dative of interest as a reflexive applied Halliday (1967) distinguishes between ‘agent-oriented’ and ‘process-oriented’ sentences in the analysis of a similar, but not isomorphic, set of English sentences. Of the following three sentences, (1) is considered to be ‘process-oriented’. whereas (2). and even more so (3). is thought of as ‘agent-oriented’.
[3]
(I) The books sold quickly. (2) The books were sold quickly. (3) They sold the books quickly. Although Halliday’s distinction is a good one, one would need lo verify empirically the claim that native speakers view (2) as an agent-oriented, rather than a process-oriented sentence, despite the fact that an unspecified agent may be responsible for the rapidity with which the books were sold. Similarly, of the three Spanish sentences that refer to the loss of a letter, (2) ‘Se perdio la carta’ would seem to be semantically closer in meaning to (3) ‘Se le perdio la carta’, than it would be to (1) ‘Juan perdio la carta’, despite the presence of an implicit agent in (3).
14R
S. Berk - Selipm
/ Variation in Spanrsh cerb usage
personally as a substitute for the passive. In its operation, “The reflexive force disappears, and the meaning conveyed is purely passive” (Ramsey 1894: 291). He cites examples such as the following ones (1894: 29 1): (4) Eso se hard fucilmente (‘That will be easily done’). (5) La cena se despuchb pronto (‘The supper was quickly dispatched’). (6) Por fortuna sus temores no se realiznron (‘Fortunately his fears were not realized’). (7) Su crbdito se ha ugotudo (‘His credit has been exhausted’). Ramsey (1894: 290) unwittingly makes an insightful observation about the relative use of the true passive vis a vis the reflexive, but is incorrect in his explanation for their disproportionate use when he notes that, “In modern Spanish the formal passive is but little used, as it is considered too lengthy and roundabout; its place is usually taken by the much over-worked reflexive form of the verb”. It is not, I believe, the lengthy, roundabout nature of the formal passive which limits its use. Rather, it is the fact that the passive construction incorporates an agent of the action, albeit one which can be optionally deleted. It is precisely this agent which these Spanish speakers apparently wish to do away with. An example of the formal passive construction is the following: La pelotu fue tiradu (por Juan) (‘The ball was thrown (by John)‘). Even when the agent is deleted, its semantic force is somewhat retained, in that the very use of the passive construction implies the existence of a known agent. A perfectly logical response to La pelota fue tirade would be: i Por quibn? (‘By who(m)?‘). Whereas such a question could not follow the sentence Se tit-b la pelota (‘The ball was thrown’) whose meaning is synonymous with that of the example just cited. The passive construction, then, poses a psychologically real agent, even when he/she is not overtly referred to. As for sentence (3), above, it is often referred to as the “reflexive passive” (Lado and Blansitt 1967: 63; Soto 1969: 252) for accidental, or unintentional or unexpected events. Recently it has been referred to as the reflexive for a non-responsible subject (Woehr et al. 1974); however, this definition, too, entails the notion of accidentality and non-deliberateness as being an essential characteristic of the action in question. The non-responsible subject conception simply goes further, in that it asserts that “The one who is speaking makes the action more impersonal and makes himself less responsible” (Woehr et al. 1974: 156). In fact, the doer of the action is portrayed as its unsuspecting victim. Typical constructions are the following: se me cuyd (‘I dropped it’ or ‘It fell on me’) [4], se te rompib (‘you tore it’ or, ‘it got torn on you’), se le oluidi [4] Note that the ‘on’ of ‘on me’ is not a preposition dative” (e.g., Menn 1972).
here. but what has been called an “ethical
S. Berk
-Seligson
/
Variation
in Spanish
verb tcvoge
149
(‘he forgot it’ or, ‘it slipped his mind’), and se nosperdih (‘we lost it’ or, ‘it got lost on us’). What the dative of interest and reflexive passive for accidental occurrences have in common is the lack of any manifested, that is to say, overt, agent. And in this respect they are similar to each other and to the true passive with an unstated agent in that rather than to connect a given action with a particular doer, they all leave that doer, or agent, unspecified. Furthermore, the relationship between an active construction, a dative of interest and a reflexive passive may be viewed as a continuum, where at one extreme one sees responsibility for an action manifestly attributed, at a point not midway but closer to the opposite extreme an occurrence is overtly dissociated from an agent, and at the extreme. not only is there no agent, but the individual most directly associated with the circumstance is not merely not its cause, but is viewed as its unsuspecting victim. Graphically, the relationship between the three sentence types can be portrayed in the following way: Attribution
(+I a (a)
(8) (9) (10) (b) (11) (12) (13) (c) (14) (15) (16)
of responsibility
to an agent
c--j
_________---------------b
C
DejP caer el dinero (‘I dropped the money’). Perdi el dinero (‘I lost the money’). Rompi el case (‘I broke the glass’). Se cay6 et dinero (‘The money fell down’). Se perdid el dinero (‘The money got (was) lost’). Se rompib el cuso (‘The glass broke’). Se me cay6 e/ dinero (‘The money fell down (on me)‘). Se me perdid et dinero (‘The money got lost (on me)‘). Se me rompib el DUSO(‘The glass broke (got broken) (on me)‘).
The significance of the existence of these three construction types is precisely that all three exist, and that they can be used alternatively to describe one and the same situation. It is the fact that a speaker of Spanish has these options open to him/her whenever he/she describes an adverse situation that gives each construction its own connotative meaning, or implication, in regard to the attribution of responsibility. Each of the three constructions then colors a given situation in particular way. Traditional grammar books (in the Jespersen genre), rather than modern-day structuralist or transformational generative ones, and teaching grammars in particular, turn out to be closest to the ideal grammar envisioned by sociolinguists. for they approach the goal of not merely analyzing a grammatical entity as such. but of stating the rules for its use. Thus, for example, labels which in
effect define usage (e.g., “dative of participation”. “involved subject”. “reflexive passive for accidental or unintended events”, and so on) were so eschewed by post-Bloomfieldian structural linguists that some scholars (e.g. Fries 1952), in an effort to achieve greater objectivity. reached the point of labelling form classes by such completely semantically neutral, and. therefore. meaningless terms as ‘A’, ‘B’, ‘C’, and ‘l’, ‘2’. and ‘3’. Linguists who carry out their analyses within a deep/surface structure model of language are much less apt to use such “objective” levels. However. they, too, often miss the point precisely because they do not take into account the notion of language use. Thus. for example, in the much debated question of how to best analyze the numerous constructions using the particle se (for citations see footnote I), Langacker (1970) disagrees with Goldin’s (1968) analysis of a sentence such as Se me olvidb truer el dinero (‘1 forgot to bring the money’). incorrectly characterizing that sentence just because he overlooks its semantic implications. Langacker criticizes Goldin’s contention that se can be inserted into a sentence like the one above only if the verb potentially takes an Agent. According to Langacker (1970: 177). “. . . there is no particular reason to believe that . . . [the Spanish sentence referred to above] is an impersonal sentence at all. One could perfectly well treat oluidur as one of the class of verbs (including gustar (‘please’), encantar (‘enchant’). and others) that allow an Objective to become the subject despite the presence of a Dative. imposing the requirement that olvidur in this case takes an automatic reflexive”. What Langacker overlooks. however, is that olvidar does belong to the class of verbs which may take an agent, as Goldin points out by his example. 1.0 olvidk traer el dinero (or, alternatively, Yo me olcidk de traer el dinero). And this fact is not happenstance, but lies at the very core of the issue that is of interest here. The fact that certain verbs may. either take an agent or not. so that a speaker may elect to produce a sentence like (9) versus a sentence like (15). or one like (10) as opposed to one like (16), has important semantic and behavioral implications. It is the fact that one may describe a given event in alternative grammatical ways - which is to say, the fact that an option exists for the speaker - that constitutes the meaningfulness of the grammatical alternatives. And whether one chooses the notion of a “potential” agent (Goldin 1968), that is, one which is potentially but not actually present. or the notion of an “unspecified”, or “dummy” agent (Langacker 1970). one is acknowledging the psychological reality of a non-surface structure agent. When that agent is kept in the deep structure, away from overt presence. the utterance conveys a particular implied meaning, one which contrasts with the implied meaning that the utterance has when the agent is present at the surface. Thus, Langacker’s (1970: 177) already cited claim that, “there is no particular reason to believe that (se me olvidb truer et dinero) is an impersonal sentence at all” misses the mark, since there is a good reason for believing so.
S. Berk
-Seligson
/
Variution
in Spanish
verb usage
151
namely that (JW) (me) olvidk (de) traer el dinero is an alternative to se me oluidd truer el dinero, an alternative whose connotation is quite distinct from that of its counterpart. And to suggest that verbs such as olvidar when used as in the sentence last cited be considered comparable to verbs such as gustar and encantar is to be guilty of exactly that which transformational generative grammarians have accused structuralist grammarians of, namely, disallowing considerations of meaning from grammatical analyses. Meaning in this case, however, must be thought of as existing in the realm of discourse, rather than in the lexical component of a grammar. Thus, to classify oluidur together with gustar and encantar on purely formalistic grounds is to ignore both a grammatical fact (i.e., that verbs such as the latter two may take only objectives as a subject, whereas verbs such as olvidar, caer, perder, romper, and so on, have the option to do so or not), and a fact on the level of ethnolinguistics (i.e., that by using olvidur with an Objective for a subject is to imply something about the human participants involved which is quite different from what one implies upon using an overt agent as subject). Thus, criteria of an ethnolinguistic sort are equally as necessary in making grammatical analyses as are criteria of the strictly linguistic type - which is to say that evidence beyond that which is purely syntactic is needed in a complete explanation of Spanish active and non-active (i.e., se) constructions.
Speaker/hearer
awareness of the social meaning of grammatical alternatives
An experimental design was conceived, in order to test the hypothesis that Costa Rican speaker/hearers behave with some degree of intentionality, that is. awareness of the implications of what they are saying, when they select one grammatical construction rather than another. In particular, it was set out to discover if people are in any way conscious of what they are implying when they choose to describe a situation in terms of an active construction rather than a se construction, be it “dative of interest” or “reflexive passive”. Thus, the study was designed to find out if non-active constructions are intended by speakers and interpreted by hearers as a way of avoiding attributing responsibility to an agent. be it themselves or non-co-present parties.
This investigation, which grew out of a broader scale study of Costa Rican Spanish (Berk-Seligson 1978). took place in the Spring of 1976. Eighty high school students [S]. the overwhelming majority of them 9th graders and some [S] The
Alfaro.
students were drawn from two non-co-cd public high schools, the Colegio an all-girls school, and the Lice0 de Costa Rica, an all-boys school.
Anastasio
10th graders, 40 female and 40 male, participated in a paper-and-pencil type of test, the test consisting of performing three written tasks for each of two sets of drawings. The first drawing depicted a boy reaching for his back pants pocket. something resembling a handkerchief dangling from the pocket. and a large coin falling from the pocket to the ground. The second drawing consists of two frames: a girl dressed in pyjamas, stretching out her arms, her mouth open in a wide, circular fashion, and a vase full of flowers tipping over near her outstretched fingers. The second frame shows only the vase. broken. with flowers strewn all over the floor [6]. The boy and girl seem to be somewhere between 8 and 12 years old, that is, no longer young children but not yet full-fledged teenagers. The tasks asked of the respondents were that they (1) answer, in a single, short sentence the question “What happened?“: (2) explain more fully what happened; (3) write a dialog. an imagined conversation. between the actors in the drawings and their respective mothers. (See fig. 1.) Findings
There is a great deal of evidence present in the students’ responses which would indicate a clear (although probably somewhat unconscious) notion of the differences implied in the use of an active construction as opposed to a non-active one. This understanding becomes particularly manifest in the hypothetical dialogs [7]. Thus, interchanges such as the following ones strikingly demonstrate that tacit knowledge: Dialog I [a]
Al entrar, la mama diria: (Upon entering the M: iQuC hiciste! (What did you do!) D: Se cuyd el florero. (The vase fell down.) M: iSe cayo! Lo botuste, que es diferente! (It which is something different!) D: Esta bien, pero fue sin culpa. (Okay, but (You’re a careless M; iEres una descuidada! D: Fue sin querer. (It was unintentional.)
mother
would
say:)
fell down! You knocked
it down,
it wasn’t my fault.) thing!)
Dialog 2 M:
~Quk paso? (What
[6] Future
happened?)
studies of this sort should consider varying the sex of the actors in the pictures. so as IO bias which this variable might introduce. [7] In all the dialogs and individual sentences that are cited here, orthographic corrections have been made. in order to avoid the numerous “[sic]” enwies that otherwise would have been necessary. [8] For all the dialogs, “mother” will be abbreviated as “M”. “son” as “s”. and “daughter” as LSD”.
control for any possible
S. Berk - Sehgson /
Fig.
Variorion
in Spanish
verb usage
153
I.
D: Se CC& el Morero. (The vase fell down.) M: iSe cayo! (It fell down!) D: Bueno, lo borP. (All right, I knocked it down). M: De ahi, mi hijita. esta comida de yegua. (My, you certainly young lady.) D: Si lo hice sin querer. (But I did it without meaning to.)
are ill-natured,
Dialog 3 D: Mami, se rompib el florero de la sala. (Mommy, the living room vase broke.) M: Pero no se rompio solo iverdad? (But it didn’t break all by itself, did it?) D: No, fue que yo lo botk con el brazo. (No, what happened was that I knocked it over with my arm) What is evident in each of these three dialogs is that the mother interprets the girl’s using the dative of interest (se CC@ in the first two instances and se rompid in the third) as an evasion of responsibility for the unfortunate event. In fact, in each case the mother is implying that her daughter is lying to her, or,
S. Berk
154
-Seligson
/
Vmurion
in Spmish
wrb
urge
at the very least, is not being completely honest with her. Evidence of this is the daughter’s “changing her story”, by way of changing both the lexical item (from ‘fall’ to ‘knock over’) and the case of the verb (from non-active to active). Perhaps the best illustration of the awareness of speakers of the contrast between active and non-active constructions is the following interchange: D: M:
Se quebrb el florero. (The vase broke.) $e quebro o lo quebraste, malcriada?
you ill-mannered
(Did it break or did you break it.
thing?)
Here the mother’s reply directly challenges the veracity of the girl’s statement. and the challenge consists completely of a change in case. Not even was a change in lexical item necessary to make the accusation. A somewhat less obvious indication that the respondents were making a semantic distinction between the use of the active and that of the non-active is the tendency on the part of the imagined mother, upon initiating the conversation, to make her accusations in the active form. This pattern. too. as exemplified below, implies the association of responsibility with an active form, and avoidance of responsibility with the non-active. Of the following dialogs, Dialog 6 brings to full consciousness the link between linguistic structure and a socio-cultural notion such as responsibility. Dialog 4 M: D:
iDe ahi! iQue quebruste! (Well, for heaven’s (The vase broke.)
sake! What
did you break!)
Se quebr6 el florero.
Dialog 5 M: D:
LQuien quebrb el florero? (Who broke the vase?) Se me ccryd; no tuve la culpa. Lo siento. (It fell down (on me): it’s not my fault. I’m sorry.)
Dialog 6 S:
Mama, se me perdid una moneda. (Mom, a coin got lost on me.) M: iQue lastima! (What a pity!) S: iMe podrias dar otra? (Could you give me another one?) M: No, ya te di. Si la perdiste es cosa tuya. (No. I already gave you one. If you lost it that’s your affair.) Diulog 7 S:
Mama, se me perdid la moneda cuando venia para la casa. (Mom, the coin got lost on me when I was coming home.) M: Cuando no siempre vos de atarantado, por andar pensando en otra cosa la
S. Berk -Seligson
/
Variation in Spanrsh
perdisre. (When aren’t you always dazed: about something else that you lost it.)
verb usage
(It was) on account
155
of thinking
Dialog 8 S: Mami, eche la moneda en mi balsa de atras y se me perdib. (Mommy, I stuck the coin in my back pocket and it got lost on me.) M: ;Adonde la perdiste? Puiia, ya estas muy grande para que pierdas la plata. (Where did you lose it? Boy, you’re too old already to be losing money.) S: Lo siento mama. No lo volvere a hater. (I’m sorry, Mom. I won’t do it again.) M: iVamos a ver! (We’ll see!) Dialog 9 (Beginning of dialog): The mother asks her son for the change. The son looks for it in his pocket and gives it back. The mother asks him why there is so little change, to which he replies that he had been given another coin, and acts upset about not being able to find it. M: &La perdisre o la gastaste? (Did you lose it or did you spend it?) S: No, no, creame. Quiza se me cay6 (No, no, believe me. Maybe it fell down (on me)). M: La hayas perdido o no, siempre ten&s que ser responsable; ademas, nunca te eches la plats en la bolsa trasera de1 pantalon. (Whether you have lost it or not, you still have to be responsible. Besides don’t ever put money in your back pocket.) Dialog 9 makes clear the fact that the reflexive passive for accidental circumstances (se me cayb) is being used by the speaker as a protective mechanism against the attribution of full blame, and - what is more important - the listener acknowledges its use as such, and rejects it as a real excuse. The mother’s insistence on the active and on the lexical choice of ‘lose’ rather than the son’s choice of ‘fall down’ in effect says to the boy, “You may claim all you want that it ‘fell down’ but 1 consider you to have ‘lost’ it, am consequently holding you accountable, and am reprimanding you for your negligence”.
Usage patterns: overall frequencies versus sex-based frequencies That speakers are aware of the social meaning behind the use of active versus passive constructions is made clear by the examples presented above [9]. An [9] ‘The term “social meaning” used in conjunction alternatives” is intended to reflect the relationship
with linguistic entities such as “grammatical between language and social use, or function.
S. Berk - Seligson / Variation m Spamsh cerh usagr
156
analysis of their content has illuminated that meaning. What such an analysis fails to do, however, is demonstrate empirically that non-active forms are in fact used more often then passive ones when actors are perceived as not being responsible for adverse events. The frequencies in table 1 provide quantitative evidence in support of this contention [lo]. The figures corresponding to the picture of the boy and his,coin revcal an overwhelming use of non-active verb constructions: 83.6%. taking the dative of interest together with the reflexive passive. Of the two non-active forms. the more passive one, the reflexive passive, was unequivocally the preferred form for such a situation. The boy apparently was viewed as not being responsible for what happened to the coin. As most respondents interpreted the scene. the boy was considered an innocent victim of the event: not only was it totally accidental, and therefore not his fault, that the coin fell out of his pocket, but he was not even aware of its happening at the moment when it did. and onI) later found out about it. The situation of the girl and the vase. on the other hand. was interpreted [IO] Although it would have been preferable for rhe respondents 10 have reacted to a larger number of situations (i.e., more drawings), the sample size and strength of relationship. as sho\\n in table I, provide such strong support for the hypothesis. that it is highly unlikely that radical11 different results would have been obtained by the addition of a greater number of siruatlons. Nevertheless. increasing the number of situations might have provided a more discriminating. and hence more precise. measure. --. ---Ftxmore
9 (wnr’d)
That is to say that members of a given speech community unconsciously associate certain linguistic forms with specific social norms, so that the linguistic forms come 10 symbolize those social norms. The correlation between linguistic usage and social meaning is perhaps most vividly illuswatcd b! pronominal address patterns. It has been noted in many societies Hherc the second-person pronoun used for addressing persons has two or more variant5 (commonly referred to as TU/VOUS usage). ihe choice of one variant rather than another can often symbolize intimaq and/or solidarity, or. conversely. social distance and/or power. In this sense. each variant becomes a symbol of a social relationship. And whereas the social situarion encapsulated in the rules of pronominal usage is an objective social fact (i.e.. that speaker A. a factory worker. is not intimate with, and is in a subservient position relative lo his employer). the fact that a given linguistic altcrnant is used for a specific social situation (that the worker addresses the employer with VOUS, while the employer addresses him with TU) means that the linguistic al~emants embody that social relationship, and thus have a definite social meaning. Similarly. in the case of the three Spanish constructions being analyzed in this article, the fact rhat a mother and child can use different verb forms lo describe one and the same event means that they are focussing on different aspects of that event, and that the choice of one ahernant rather than another is a reflection of the interpretation which the observer gives to that event. In particular. it is an interpretation of his or her involvement in the event. By dissociating him- or herself from the event. the observer interprets it as a process (*‘It happened”); by associating him- or herself with it. the event becomes an action (“I did it”). Thus, social meaning becomes embodied in linguistic form through the mediating process of interpretation in the face of real-life events. This notion is not original with this writer, but has its roots in the phenomenological theory of Schiitz (1970) and in the more recent work of Taylor (1973) and Ricoeur (1970).
S. Herk Seligwn Table Overall
/
Varimion in Spanish cerh wage
1 usage of active/non-active
vserb constructions.
” Boy and Coin” drawing Active Dative of interest Rcflcxive passive Total
157
16.5% (38) 10.0 (23) 73.6 (170) 100.0 (Gamma
(231)
“Girl and Vase” drawing 69.4% 22.7 7.9 100.0
(159) (52) (18) (229)
= - 0.68, sig. = < 0.001)
Noret The frequencies (numbers in parentheses) represent the sum total of verbs used in the three sections of the questionnaire. Percentages are based on non-missing data. Missing data were usually the result of respondents not following directions, writing in narrative rather than dialog form in section 3 of the questionnaire. In such instances no verb was coded.
rather differently than that of the boy and the coin. The girl was seen as being somewhat responsible for the breaking of the vase. Respondents attributed the unfortunate event to her carelessness: she should have paid more attention to her physical surroundings when she stretched her arms out, should have been more in control of her body movements. The interpretation of negligence and, consequently, blameworthiness, is correlated with a greater use of active verb forms. The second column in table 1 shows that nearly 70% of the verbs used to describe the situation were used in a grammatically active form. This is a sharp contrast to the low (16.5%) active usage for the boy and coin situation. And of the two non-active verb constructions used for describing the boy, whereas the more passive one was overwhelmingly preferred (73.6% of the verbs were in the reflexive passive form), in referring to the girl, respondents clearly did not see her as the “innocent victim”, judging by the 7.9% use of the reflexive passive as compared to the 22.7% use of the dative of interest. While it might be believed that societal sexism is the explanatory factor behind the finding that the boy was thought of as less blameworthy than the girl, it is important to point out that male and female respondents in the study were in agreement on these interpretations. If one looks at tables 2 and 3, which break down verb usage by respondents’ sex, one finds that with one exception, there is no difference between male and female usage of the grammatical constructions under investigation. Thus, on the whole there are no statistically significant differences between males and females in the use of active and non-active verb forms, the exception being in the lengthier responses to “ What is going on. 7” in the pictures of the boy and the coin. Only in this instance (see the column labelled ‘Verb in Section 2’, in table 2) is there a
100%
Total
(40)
(4) (31)
(5)
1
100%
7.5% 10.0 82.5
Female
(40)
(4) (33)
(3)
(40)
(3) (23)
(14)
100%
5.0% 17.5 77.5
Female
(sig. = < 0.001. Gamma
100%
35.0% 7.5 57.5
Male
Verb in section 2
“Boy and Coin” drawing
= -0.49)
(40)
(7) (31)
(2)
(40)
(0) (27)
(7)
(sig. = NS)
100%
20.6% 0.0 79.4
Male
Verb in section 3
100%
18.9% 13.5 67.6
Female
(40)
(5) (25)
(7)
Nor,: gee explanatory
100%
Total
(39)
(9) (0)
(30)
note to table 2.
(sig. = NS)
76.9% 32. I 0.0
Male
Verb in section
use of verb constructions:
Active Dative of interest Reflexive passive
Table 3 Male/female 1
(26) (IO) (3) (39)
100%
(40)
(6) (0)
(34)
(sig. = NS)
100%
85.0% 15.0 0.0
Male
Verb in section 2
and Vase” drawing.
66.7% 25.6 7.7
Female
“Girl
100%
82.5% 15.0 2.5
Female
(40)
(6) (I)
(33)
(34)
(7)
(17) (IO)
(sip. = NS)
100%
50.0% 29.4 20.6
Male
Verb in section
3
(37)
100%
(19) (II) (7)
51.4% 29.7 18.9
Female
Nore; Numbers in parentheses are frequencies. Percentages are based on non-missing data. Missing data were usually the result of respondents not following directions, writing a narrative rather than a dialog in section 3 of the questionnaire. In such instances no verb was coded. Sections 1. 2, and 3 refer to the three sections of the questionnaire.
(sig. = NS)
12.5% 10.0 77.5
Male
Verb in section
use of verb constructions:
Active Dative of interest Reflexive passive
Table 2 Male/female
S. Berk
-Selig.wn
/
Variation
in Spanish
verb usage
159
statistically significant difference. What the data show is greater female use of non-active forms than male use of such forms. Thus, if figures under the headings “dative of interest” and “reflexive passive” are collapsed, to form one, more general, non-active category, female respondents are found to use such verb constructions to a significantly greater extent than male respondents. When the two non-active forms are not collapsed into one category, and the three verb constructions are compared with each other separately, there is still a statistically significant greater tendency of females to use non-active forms (significant at 0.10, gamma = - 0.49). The overall lack of differentiation between male and female active/non-active verb form usage comes as a surprise. It was expected that women would use non-active forms more frequently than men. This expectation arose from personal observation of male/female behavior in Costa Rica, and from the findings of research on sex-based variation in American English grammatical case. Specifically, Barron (197 1) has found that the choice of case in English is sex-typed, and that this sex-typing can be explained by men’s speech being marked by action, such that men project themselves as actors upon their environment, whereas women are more concerned with internal psychological states and the functions of objects for interpersonal use. Similarly, Shuster (1973) has found that the use of active versus passive voice and transitive versus intransitive verbs is different for American men and women: passive and intransitive forms occur more often when women are speaking or are being referred to, whereas active and transitive forms are more common when men are speaking or are being referred to. Shuster interprets this finding as non-reciprocal grammatical forms being a reflection of an asymmetrical power relationship of men over women. Furthermore, Crosby and Nyquist (1977: 320). in testing out Lakoffs (1975) hypotheses regarding women’s speech have concluded that, at least for American English, what most basically characterizes the female register is non-assertiveness. Finally, Berko-Gleason (1975). reporting on her own research as well as that of Kriedberg, brings evidence to support the hypothesis that female speech is less forceful or dominant than male speech. While it is surprising to find that the speech of Costa Rican young men and women is not asymmetrical in assertiveness or forcefulness, at least insofar as choice of verb is concerned, since the culture is in fact characterized by an asymmetrical power relationship between men and women (i.e., Latin American machismo is a cultural phenomenon present in Costa Rica), the fact that men and women cannot be distinguished linguistically in one regard fits in with another fact of Costa Rican speech variation, namely, that men’s and women’s subjective reactions to phonological variation are for the most part indistinguishable, this despite the fact the women’s speech is significantly closer to the phonologically prestige norm than is men’s (Berk-Seligson 1978). Specifically, it was found that in reacting to the taped speech of three speakers,
160
S. Berk - Seligson / Vurintion m Spanish verb usage
each of which varied only in the use of prestige and stigmatized phonological variables, women reacted to the voices in a manner that was virtually the same as men. In only a few instances did statistically significant differences appear between the subjective reactions of the 440 men and women in the study. However, when the productive phonology of Costa Rican men and women was compared, it was found that within each of five contextual levels of formality, women used prestigeful phonological variants with a higher frequency than did men (sample N = 88). The negative findings with respect to sex differentiation in Costa Rican Spanish active/non-active verb usage are particularly significant in that they both contrast and compare favorably with other sociolinguistic facts about this dialect of Spanish. More importantly, they lead one to the conclusion that to talk of a “female register” in any given society. it is imperative that male/female differences be investigated in many realms of speech. Only such an empirical approach is capable of discounting those linguistic traits which merely appear intuitively to vary along gender lines, from those which do so in reality. The comparisons presented above regarding sex-linked linguistic variation in Costa Rican Spanish lead one also to a tentative generalization. one which would need further testing, but which seems to stand up well in the face of research findings to date, namely, that where speech variants fall under the category of “correctness” (i.e., approximation to a prestige norm, or standard), correlations with speaker sex usually appear (Anshen 1969; Cedergren 1973; Fasold 1968; Fischer 1964; Labov 1972; Phillips 1972; Poplack n.d.; Sankoff 1974; Sankoff and Cedergren 1971). Such variants, then, derive from the phonological, morphosyntactic, or semantic components of a grammar. Where sex-differentiation in speech is less regularly found, apparently. is in the realm of language outside the immediate, inner core: what is broadly referred to as “communicative behavior”. This encompasses discourse, stylistics. in short, pragmatics. Thus, just as the choice of case has been shown not to vary with speaker/hearer sex in Costa Rica, so, too, have such discourse variables as overall speech quantity, diminutives, civilities, requests for information, and hesitations been shown not to vary with speaker sex in non-Spanish settings (Brouwer et al. 1979). The distinction between internal linguistic rules and discourse phenomena, then, might turn out to be a significant one insofar as gender differentiation is concerned, one in need of investigation [ 111. [I I] It has been pointed out that discourse phenomena
have in fact been shown to vary along the lines of speaker sex. A case in point is the work of Keenan (1974) on Malagasy. which demonstrates that men’s speech is clearly distinct from women’s along the dimension of directness/indirectness. Other large-scale stylistic cleavages associated with speaker sex apparently have been found in traditional cultures in Africa and South America. What this leads one to is a qualification of the hypothesis proposed above, namely. that whereas in Western. technologically developed societies, verbal differences are correlated with speaker sex only in the area of inner-core linguistic features; in tribal or village types of societies larger. more broadly scoped aspects of verbal behavior may be involved in socially differentiating men and women.
S. Berk - Sdigson
Additional
/ Variation in Spanish verb usage
sources of variation related to active/non-active
161
verb usage
The preceding analysis has shown that in Costa Rican Spanish active voice is used for attributing responsibility to actors for their actions, whereas non-active grammatical voice is employed to imply that actors arc not responsible, or blameworthy, when it comes to adverse events that they are somehow involved in. Of the linguistic options available to speaker in describing the relationship of actors to adverse situations, choice of case is but one, however. A look at the responses of students who had participated in this study reveals that choice of lexical item is yet another source of variation. The following list of verbs comprises the entire set of lexical items that was used to describe the drawings. For the drawing of the boy and the coin, the following verbs were used: perder (‘lose’), extruviar (‘misplace’), caer (‘fall’), salir (‘leave, come out’), romper (‘tear’), poner (‘put’), echar, meter (‘put in’), tit-at-, hotar (‘hurl, throw’), zufur (‘slip away’). For the situation of the girl and the vase the following verbs were employed: quebrar, romper (‘break’), golpear (‘strike, knock’), chocur (‘collide’), dur con (‘hit up against’), empujar (‘push’), mover (‘move’), tocnr (‘touch’), botar, volcar (‘overturn’), derribar and lumbar (both ‘knock down’), caer (‘fall’) estirur (‘stretch’), extender (‘extend’), and levuntur (‘raise’). The very range of meanings reflected in these two sets of verbs tells us that more than just one type of perception, or interpretation, of the drawings was going on in the minds of the respondents. At a minimum, a distinction was being made regarding cause and effect, and this distinction was made not in the one-sentence description, but in the lengthier explanation of the drawings as well as in some of the dialogs. Thus, responses such as the following demonstrate that the respondents were aware that the vase’s breaking was a result of some movement by the girl’s arm/hand/fingers, and so on. (17) Al levantarse se acercb donde siempre ha estado el florero, y resulta ser que sin quererlo rompe, es decir quiebra, el florero, lo bota, lo hate caerse tocandolo con la mano izquierda. Sin querer. (Upon getting up she approached the place where the vase has always been, and it turns out that without meaning to she breaks, that is to say, she breaks, the vase. She knocks it over, makes it fall down, by touching it with her left hand. Without meaning to.) (18) Esta nifia, coma se ve en el dibujo, se va a dormir puesto que esta con una pijama, y ella al bostezar extendio demasiado las manos y con la mano izquierda golpe un florero que estaba sobre una mesa y el florero por el golpe cay6 y se rompio. (This little girl, as you can see in the drawing, is going to bed for she has pyjamas on, and upon yawning she stretched her hands out too much,
S. Rerk -Selrgson / Variarion m Spanish verb wcrge
162
and with her left hand hits up against a vase which was on the table, and the vase, because of the blow, fell and broke.) (19) La nifia se venia levantando para ir a desayunar. Entonccs al hater sus ejercicios estaba detras un florero. Al estirar las manos se le quebro el florero, se le hizo pedazos, se le cayeron las flores. Tambien se ensucio el suelo con el agua de1 florero. (The little girl was getting up in order to go have breakfast. So then, upon doing her exercises, a vase was behind her. Upon stretching out her hands, the vase broke (on her), it broke into pieces (on her) the flowers fell down (on her). Also the floor got messed up with the water from the vase. Dialog 10 D:
M: D: M: D:
Mami, quebre el florero de1 cuarto mio. (Mommy, I broke the vase that’s in my room.) LPor que? (Why?) Lo hice sin querer. (I did it without meaning to.) LQue paso? (What happened?) Me estiraba y de pronto mi brazo choco con el florero y lo quebre. (I was stretching and suddenly my arm collided with the vase and I broke it.)
There is, as evidenced by the preceding examples, an overwhelming consensus among the respondents concerning “what is going on” in the diagrams. The perceptions of the sequence of events, of the accidental nature of the events, are very much agreed upon. Nevertheless, there is great variation in the to each drawing set, insofar as some responses to the first question pertaining refer to the cause while others refer to the effect. Paralleling the relation cause ‘1 did it, I am responsible for this versus effect, is the concomitant implication, event’ versus ‘It happened (to the coin/vase, which in turn affected me)‘. Thus, the brief, one-sentence perception is the most revealing one, for it reflects the initial basic, unanalyzed, sort of ‘gut reaction’ of the respondents to the situations presented to them. And it is in this response that one most clearly sees the extent of variation. Thus, some people choose to talk of the coin (‘it fell’, ‘it was lost’) and the vase (‘it fell’, ‘it broke’), whereas others talk of the characters involved and what they did (‘he dropped it’, ‘he lost it’, ‘she knocked it over’, ‘she broke it’). So, it appears that lexical choice in itself will greatly affect the reading given to a text [12], in this case. the drawings. And the decision to use one lexical item rather than another says something about how the reader wishes to place the textual characters (in this case a young boy and girl) in relation to the events they are involved in. [ 121 The notion of applying textual analysis to other types of meaningful work of Ricoeur (1970). Taylor (1973). and Geertz. (1973).
behavior
is based on the
S. Berk
-Seligson
/
Variarion
m Spanish
cerb usage
163
A second sphere of options is that of transitivity, an important area of choice since it leads one into the rather tricky question of whether or not Spanish speakers’ options are more restricted than one would normally be led to believe. In other words, is it not true that only certain verbs may take a human agent, that only certain other ones may enter into a dative of interest construction, and that a third set may participate in a reflexive passive for accidental occurrences? The answer, based on the empirical evidence derived from the questionnaires. is that the three are by no means mutually exclusive, and that there are a large variety of verbs which can be used for all the constructions that are under investigation. Consider, for example, the following actually occurring forms: (20) (a) (b) (c) (21) (a) (b) (c) (22) (a) (b) (c)
quebre el florero (I broke the vase) se quebro el florero (the vase broke) se me quebro el florero (the vase broke (on me)) cayo la moneda (the coin fell) se cay6 la moneda (the coin fell down) se le cay6 la moneda (the coin fell down (on him)) perdio la moneda (he lost the coin) se perdio la moneda (the coin got lost) se le perdio la moneda (the coin got lost (on him))
However,
some verbs seem to be used only in active forms (e.g., hotar, turnbar, whereas others (e.g., quebrar, romper) are used usually in the active form, but can and do appear in either of the two non-active forms; and still others are found usually in the non-active constructions, and much less so in the active one (e.g., caer, perder). Apparently, any verb that could enter into a dative of interest construction could just as easily enter into a reflexive passive construction, and vice-versa. And, contrary to expectation, there were no verbs which were never used in active constructions. This grammatical fact can be understood better in light of the factor to be discussed next. A third source of options, alluded to at the beginning of this article, is the present or absence of expressions which in some way serve to qualify, or reinforce the meaning which results from the combined impact of lexical meaning and grammatical form. The expressions 1 am referring to are explicit statements of unintentionality, many of which have already appeared in the examples presented throughout this paper. The most common among them is sin querer (literally, ‘without wanting to’, but best glossed as ‘without meaning to’), but also frequent are sin intention (‘unintentionally’), sin culpa (‘not at fault’, ‘not to be blamed’) and sin durse cuenta (‘without realizing’, ‘without being aware’). Any of these expressions may be added on to any of the lexical items listed above, and to any of the grammatical constructions discussed above. What is significant about these expressions is not their use with
derribur,
golpear);
164
S. Rerk - Seligson /
Variation in Spamsh Gerb usage
non-active forms (in which case they serve to reinforce the unintentionality that is already expressed by the verbal construction), but their use with active ones, for it is in this capacity that they make viable for a speaker the option of coupling the notion of accidentality or unintentionality with the admission of a certain degree of responsibility for an adverse circumstance. In other words, the fact that such expressions can be used with active forms allows the speaker a greater degree of honesty (or, less evasion) in talking about an adverse event which he/she in some respect is responsible for. A child, thus, in relating to his/her mother an adversative_incident. will be considered by her to be more honest if he/she uses an active form of a verb, accompanied by a “But it wasn’t my fault” type of expression, than if he/she uses such an expression with a non-active form. In fact, the use of such a latter combination smacks more of intentional deceit than does the use of a non-active form without such an expression. The combination “protests too much”, so to speak.
Conclusion This article has tried to demonstrate three things: (1) that speakers of Costa Rican Spanish, when they use constructions involving either an active verb, a verb as a dative of interest, or a verb in the reflexive passive for accidental events, are quite well aware of the implications of using one rather than another, and that the implication of using one versus another is closely tied to the notion of one’s publicly acknowledging some responsibility for one’s involvement in some adverse situation; (2) that Costa Rican Spanish speakers use non-active constructions more frequently than active ones for adverse events; (3) that there are, overall, no differences between men and women in their linguistic behavior in this regard; and (4) that three additional sources of linguistic options are available to the speaker: lexical choice, grammatical choice, and the option of qualifying the verbal construction with an added layer of meaning through explicit expressions of unintentionality. Given the range of alternatives, speakers have quite a large degree of freedom in selecting their particular response to a specific adverse situation. It is worthwhile pointing out, furthermore. that Spanish is not unique in its use of active/non-active cases for creating varying degrees of distance between actors and actions. Japanese has something that has been termed an “adversative passive construction”. The semantic function of this adversative passive is that of “connoting that the subject of the sentence was involuntarily subjected to something unpleasant” (Niyekawa-Howard 1968: 2). When used in conjunction with a causative, the construction takes on the following connotation: “because the subject of the sentence ‘was caused to’ take the action expressed by the main verb, he is not responsible for the act nor the outcome” (Niyekawa-Howard 1968: 2).
S. Rerk - Sehgwn
/
Variatron
in Spanish
wrh
usage
165
Recently, in a pragmatic treatment of the so-called passive and the impersonal si constructions in Italian, Costa (1975) argues that they are actually passives in which generic human agents are demoted and deleted. She notes that a primary function of si constructions is to reduce the prominence of the agent, and that, “si constructions are exploited pragmatically as a means of shifting responsibility off a subject. This pragmatic use is typically a reflection of politeness and formality conventions which require that in potentially embarrassing situations speakers should avoid pinning down who is responsible for what action” (Costa 1975: 120). The pragmatic coincidence of Italian si constructions with the Spanish dative of interest and reflexive passive is strikingly evident. They seem to be functionally identical. Dutch. too, makes use of variations in case morphology to relate persons to acts. In a recent discourse analysis of the subjectless “pseudo-passive” in Standard Dutch, Kirsner (1976) finds that the pseudo-passive construction is used to background the agent-like participant in an event more strongly than does the true passive. However, common to both kinds of passive is the use of a linguistic sign as “signalling the single, relatively abstract meaning high purticipont not focused, which may be roughly paraphrased as ‘the logical subject is not the grammatical subject”‘. This is to say that in neither one is the agent-like entity foregrounded as a grammatical subject, (Kirsner 1976: 389). By the expression “high participant not focused” Kirsner means the following: involved in the event “ Participant” refers to “an entity which is centrally named by the verb. an entity which purticipates in the event rather than simply 1976: 389). By “high parbeing the scene or circumstance of it” (Kirsner “one which ranks comparatively high on ticipant” Kirsner (1976: 389) means, a scale of relative agentivity of potency, more technically a scale of relutioe conrribution to the bringing about of the event named by the verb”. And the term “focused” means “foregrounded”, “in the center of attention”. These three discourse concepts seem applicable to the study of active/passive case relationships in many other languages besides Dutch, and the meaning inhernot focused”, certainly is ent in the Dutch passive, namely, “high participant equally as valid in the case of the reflexive passive and dative of interest in Spanish. Further evidence of cross-cultural tendencies towards the use of grammatical case for dissociating actors from adverse events comes from a conversational analysis of English. In a recent ethnomethodological study on attributions of responsibility, specifically blaming, Pomerantz (1978: 116) finds that “a standard form for attributing responsibility in both blaming and praising activities. is with a construction in which the candidate blamed/praised party is formulated as an actor-agent performing a blameworthy/praiseworthy action”. Blamings contrast with what Pomerantz describes as “reports of ‘unhappy incidents”‘. These reports, which precede blame allocations, are described as containing “descriptions of ‘blameworthy’ actions/events/states with no in-
166
S. Berk -Selig.wn
/
Varratron
in Spmish
wrh
usage
corporated actor-agents. Such descriptions are shaped as reports of ‘events that happened’ rather than of actions performed by actor-agents” (Pomerantz 1978: 117). Thus, the conversational structure of blamings may be considered as follows: unelaborated announcements contain references “unhappy incidents”; no actor-agents are designated _. the characterizations while requesting elaborations transforming them from “incident that happened/is agent”. Each speaker specifies an actor-agent while (Pomerantz 1978: I I7- I 18).
to objects co-joined with descriptions of wshile recipients may prescne prior speakers’ rccipicnts recharacterire the reference-events. happening” IO “action performed by actorrequesting a description of some action(s)
The conversational evidence from English blamings is strikingly similar to the findings of the present study of Costa Rican Spanish. The dialogs between mother and son/daughter, examples of which have been presented throughout this article, would seem to indicate that as a conversational subtype, blamings may be similar across many cultures. And, once again. it can be seen that grammatical case and the presence or absence of human agents are pivotal in or. as they have been referred to in this characterizing “ unhappy incidents”. article, adverse events. In conclusion, this article has sought to demonstrate that meaning lies not merely in the semantic component of a grammar, but in spheres outside of it: in the realm of grammatical choice, and in discourse. Lexical choice is but one source of variation for speakers in a given speech situation. Furthermore the findings of the present study, together with those of others cited above. hint at a possible sociolinguistic universal, namely that variations in grammatical case are employed in speech acts involving the attribution or non-attribution of blame. Specifically, it is suggested that non-active voice and agentless verb usage may be universals in discourse which can be construed as blame avoidance.
References Alonso Cortts, Narciso, 1939. El pronombre ‘se’ y la voz pasiva castellana. Valladolid: Afrodisio Aguado. Anshen, Frank, 1969. Speech variation among Negroes in a small Southern community. Doctoral dissertation, New York University. Babcock, Sandra, 1970. The syntax of Spanish reflexive verbs. The Hague: Mouton. Barron, Nancy, 1971. Sex-typed language: the production of grammatical cases. Acta Sociologica 14: 24-72. Bcrko-Gleason, J., 1975. ‘Fathers and other strangers: men’s speech to young children’. In: D.P. Dato. ed., Developmental linguistics: theory and applications. Washington. DC: Georgetown University Press. pp. 289-297. Berk-Seligson, Susan. 1971. The Whorfian hypothesis: an empirical test. Master’s thesis. University of Pittsburgh.
S. Berk
.Selrp~~/ Variarion m Spanish c;erb usage
16-l
Berk-Seligson, Susan, 1978. Phonological variation in synchronic/diachronic sociolinguistic context: the case of Costa Rican Spanish. Doctoral dissertation, University of Arizona. Bolinger, Dwight, 1969. Of undetermined nouns and indeterminate reflexives. Romance Philology 22: 484-489. Brouwer, D&l& Marine1 Gerritsen and Dorian de Haan. 1979. Speech differences between women and men: on the wrong track? Language in Society 8: 33-50. Carrasco, Felix, 1969. Sobre el formante de ‘la voz pasiva’ en espai’rol. Revista espaiiol de linguistica 3: 333-341. Cartagena. Nelson. 1971. Las construcciones pronominales de sujeto pasivo-indeterminado. Estudios filologicos 7: l45- 172. Gary. Davis C.. 1972. More on indefinite se. Hispania 55: 312-313. Cedergren. Henrietta, 1973. The interplay of social and linguistic factors in Panama. Doctoral dissertation. Cornell University. Costa. R.. 1975. ‘Functional solution for illogical reflexives in Italian’. In: R.E. Grossman, L.J. San and T.J. Vance, eds., Papers from the parasession on functionalism. Chicago Linguistic Society. pp. 112-125. Crosby. Faye and Linda Nyquist, 1977. The female register: an empirical study of Lakoff’s hypothesis. Language in Society 6: 313-322. Fasold. Ralph W.. 1968. A sociolinguistic study of the pronunciation of three vowels in Detroit speech. Washington. DC: Center for Applied Linguistics (Mimeo). Fade. Sheila Corigan de, 1971. Separating the uses of se in Spanish. Doctoral dissertation, Georgetown University. Fischer. John L.. 1964. ‘Social influence in the choice of a linguistic variant’. In: D. Hymes, ed., Language in culture and society. New York: Harper and Row. pp. 483-488. Fries. Charles C.. 1952. The structure of English. New York: Harcourt, Brace and Co. Gcertr. Clifford, 1973. ‘Toward an interpretive theory of culture’. In: C. Geertz, ed., The interpretation of cultures: selected essays by Clifford Geertz. New York: Basic Books. pp. 3-30. Goldin. Mark G.. 1968. Spanish case and function. Washington, DC: Georgetown University Press. llalliday. M.A.K.. 1967. Notes. Journal of Linguistics 3: 47. Hymes. Dell. 1964. ‘Directions in (ethno-)linguistic theory’. In: A.K. Romney and R.G. D’Andrade, eds., Transcultural studies of cognition. American Anthropologist 66(3), Part 2. pp. 6-56. Jordan, Pablo G., 1973. La forma se coma sujeto indefinido en espaiiol. Hispania 56: 597-603. Keenan. Elinor. 1974. ‘Norm-makers. norm-breakers: uses of speech by men and women in a Malagasy community’. In: R. Bauman and J. Sherzer, eds., Explorations in the ethnography of speaking. London: Cambridge University Press. pp. 125-143. Kirsner. Robert S.. 1976. ‘On the subjectless ‘pseudo-passive’ in Standard Dutch and the semantics of background agents’. In: C.N. Li, ed., Subject and topic. New York: Academic Press. pp. 385-415. Lavoc. William. 1972. Sociolinguistic patterns. Philadelphia, PA: University of Pennsylvania Press. Lado. Robert and Fdward Blansitt, 1967. Contemporary Spanish. New York: McGraw-Hill. Lakoff. Robin. 1975. Language and woman’s place. New York: Harper and Row. Langacker. Ronald W., 1970. Review of Mark G. Goldin, Spanish case and function. Language 46: l67- 185. Menn. Lise. 1972. Squib and discussion of ‘On me’. Linguistic Inquiry 3: 228-232. Siyekawa-Howard. Agnes M.. 1968. A psycholinguistic study of the Whorfian hypothesis based on the Japanese passive. Paper presented at the Thirteenth Annual National Conference on Linguistics. New York. Otero. Carlos. 1972. Acceptable ungrammatical sentences in Spanish. Linguistic Inquiry 3: 233-342.
168
S. Herk -Selrgson
/
Vuriurion
in Sponish
cerb usage
Perlmutter, D.M., 1970. Surface constraints in syntax. Linguistic Inquiry 1: 185-255. Phillips. Jr., Robert N., 1972. ‘Variations in Los Angeles Spanish phonology’. In: F.. HernandezChavez, A.D. Cohen and A.F. Beltramo. eds., El lenguaje de 10s chicanes: regional and social characteristics of language used by Mexican Americans. Washington. DC: Center for Applied Linguistics. pp. 52-60. Pomerantr., Anita, 1978. Attributions of responsibility: blaming. Sociology 12: 115- 121. Poplack, Shana. n.d. The notion of the plural in Puerto Rican Spanish: competing constraints on /s/ deletion. Unpublished ms. Ramsey, Marathon Montrose, 1894. A textbook of modern Spanish. New York: Henry Holt. Ricoeur. Paul, 1970. The model of the text: meaningful action considered as a text. Social research 38: 529-562. Sankoff. Gillian, 1974. ‘A quantitative paradigm for the study of communicative competence’. In: R. Bauman and J. Shetzer. eds., Explorations in the ethnography of speaking. London: Cambridge University Press. pp. 18-49. Sankoff. Gillian and Henrietta Cedergrcn. 1971. ‘Some results of a sociolinguistic study of Montreal French’. In: R. Darnell. ed.. Linguistic diversity in Canadian society. Edmonton. Canada: Linguistic Research. pp. 61-87. Schroten. Jan, 1972. Concerning the deep structures of Spanish reflexive sentences. The Hague: Mouton. Schtitz, Alfred, 1970. On phenomenology and social relations. Chicago: University of Chicago Press. Shuster. Janet. 1973. Grammatical forms marked for male and female in English. Unpublished ms.. Department of Anthropology. University of Chicago. Soto. Osvaldo N., 1969. Repaso de gramatica. New York: Harcourt. Brace and World. Taylor. Charles. 1973. Interpretation and the sciences of man. Review of hletaphysics 25: 3-5 I. Woehr, Richard, John Barson and Gustav0 Valadez, 1974. Espaiiol esencial: un repaso. Corte Madera, CA: Holt, Rinehart and Winston.
is an assistant professor of Spanish Linguistics at Purdue University. She is a sociolinguist with a Ph.D. in Linguistics from the University of Arizona. who focuses her research on Spanish social dialects. Her latest research project has been a study of the linguistic constraints on code-switching, looking at the case of Spanish/Hebrew bilingualism. This was carried out under a U.S. National Science Foundation NATO Postdoctoral Fellowship. Currently. she is carrying out a study on law and language. funded by a grant from the U.S. National Science Foundation. The research examines the role of the court interpreter in the Spanish/English bilingual courtroom, particularly from the point of view of speech act theory.
Susan Berk-Se&son
Recent publicurions.
1978. The phonological correlates of social stratification in Costa Rican Spanish. Lingua 46: l-28. 1978. Language and political behavior: A methodology for utilizing the linguistic component of socio-economic status. American Journal of Political Science 22: 712-741. 1980. A sociolinguistic view of the Mexican-American speech community. Latin American Research Review 15: 65- 110.