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



––––––––   1294   ––––––––


ArribaAbajo

Applied Linguistics


Prepared by Karen L. Smith62



ArribaAbajo

An Analysis of Interlanguage Development Over Time: Part II, ser and estar


Gail Guntermann


Arizona State University


A previous article (Guntermann,1992) described the use of por and para by nine Peace Corps volunteers at the end of in-country training and again after approximately a year of living and working in their assigned sites in Central America. The purpose of this paper is to describe the acquisition of the Spanish copula by the same subjects and compare the results to those of other researchers.

The por-para analysis showed that por was the marked form, and therefore more difficult to acquire, but that the two prepositions should not always be considered to be in opposition for pedagogical purposes, since substitutions between them accounted for only 40% of the inaccuracies. Certain functions were found to be acquired before others, and two individuals consistently opted for por as their preferred form, while the other seven used para as the default form. Those who were identified as «por users» tended to be less accurate in their use of both prepositions overall, a finding that coincides with at least one study of children’s L1 acquisition (Brisk, 1976), in which the weaker learners tended to overuse the feminine (marked) gender forms. Finally, it was found that there were clear increases in accuracy at each of the threshold levels of the ACTFL Proficiency Guidelines.

The present study seeks to answer similar questions about ser and estar: How are the two verbs and their functions acquired over time? To what extent are ser and estar actually in opposition for these learners? What purposes do they serve in communication within the context of the oral proficiency interview (OPI)? Does their acquisition, like that of por and para, show clear demarcations at the threshold levels of the ACTFL Guidelines? What aspects of ser and estar were learned/acquired in structured training, as opposed to the immersion on-site experience? Do the results of this study coincide with those found by previous researchers?

VanPatten (1985, 1987) and Ryan (1991) investigated the order of acquisition of ser and estar. The following sequence represents the combined results of the two studies:

  1. Absence of copula (VanPatten, Ryan)
  2. Ryan: Overuse of ser in estar locative and conditional contexts; VanPatten: ser is used for most copula functions
  3. Absence of the copula in estar conditional contexts (Ryan)
  4. Appearance of estar in progressive constructions (VanPatten)
  5. Replacement of zero copula errors by errors with estar where ser is required (Ryan)
  6. Estar replaces zero copula in conditional contexts (Ryan)
  7. Estar replaces ser in locative contexts (VanPatten)
  8. Estar appears in conditional contexts (VanPatten)

The last two stages are reversed in Ryan’s study. Both investigators used data elicited from relative beginners; VanPatten’s subjects were classroom learners, while Ryan’s were studying Spanish formally in Spain. Both researchers assumed a form to have been acquired if it was used accurately in 90% of the cases, and both examined copula selection in obligatory contexts.

In the two studies cited, ser was found to be the unmarked form, while estar occurred less frequently in the data and was acquired gradually over time. In Ryan’s data, the group accuracy rate for ser was 90%; for estar it rose from 40% at Time 1 to 7 0% at Time 3 (that is, from the first OPI to the third, during the semester abroad).

––––––––   1295   ––––––––

Because ser is both unmarked relative to estar and similar to English to be in its singular present tense form, it is acquired first (Eckman, 1987). The main difference between the results in the two studies is that VanPatten found the locative function of estar to be acquired before the conditional, while Ryan’s subjects exhibited the opposite order, although results for the two functions were very similar and no uses of estar reached 90% accuracy. Ryan also added two steps in the use of estar in conditional contexts (Stages 3 and 6) and the replacement of zero copula with estar where ser is required (Stage 5).

DeKeyser (1990, 1991) compared study-abroad students in Spain to students at a similar level (intermediate) in the United States and found that in both groups, those who demonstrated knowledge of the rules for ser+ Noun and estar for location were able to use them very accurately in oral communication. No differences were found in the degree of monitoring between the two groups, except for scant evidence of improvement in the case of estar for the students in Spain. Likewise, no notable changes took place in the use of communication strategies in Spain. The most interesting results pertained to strategic and personality differences among individuals in the study abroad program, and to the fact that even after six quarters of Spanish study at home the students in both situations demonstrated a great lack of knowledge of the rules on which to base the monitor.

Finnemann (1990) analyzed the language acquisition literature and the uses of the two copula by three individual learners and hypothesized that some learners are form-oriented while others attend more to meaning, and that these orientations interact with linguistic features, formal or semantic, in such a way that form-oriented learners opt more for the marked members of pairs in formal domains, and unmarked in semantic domains, while meaning-based learners do the opposite. In the case of ser and estar, which is a semantic contrast, meaning-based learners should be more inclined than form-based learners to experiment with estar. Further, if these strategies are consistent from one subdomain to another, it is predictable that the users of por will select estar as well.

For the present study, ser, estar, and haber were analyzed for lexical selection in relation to their functions, and also for errors in tense, mode, aspect, and agreement in person and number, as well as invented forms and uses of copula in extremely simplified or reduced utterances. These data come from twenty oral examinations administered by interviewers who were certified by the Educational Testing Service to evaluate the proficiency of Peace Corps volunteers in Spanish. The government ratings are used here, along with their equivalents on the ACTFL proficiency scale (See Table 1).

TABLE 1

SUBJECTS, BY LEVEL OF PROFICIENCY

Translation of FSI ratings to ACTFL/ETS ratings:
0 Novice 2 Advanced
1 Intermediate 2+ Advanced Plus
1+ Intermediate High 3 Superior
(S- refers to Speaking; e.g., S-1+ is Speaking 1+)



––––––––   1296   ––––––––


Overall Results

Table 2 shows the accuracy percentages for the selection of ser, estar, and haber at seven proficiency levels. For ser and haber, the subjects were highly accurate at all levels; and once estar appeared in the data with any frequency, it too was found to be selected accurately at least 80% of the time. Overall, both copula showed 90% accuracy, which is a higher rate than those found by VanPatten (70%) and Ryan (83%). Of course, these data cover a higher range of proficiency levels, but even those with a 1+ rating achieved an overall accuracy rate of 86% with ser-estar choices.

TABLE 2
CHOICE OF COPULA - PERCENT ACCURATE
CHOICE OF COPULA - PERCENT ACCURATE


ser estar Total haber
0+ 100%
5/5
100%
1/1
100%
6/6
100%
1/1
1 88%
21/24
50%
2/4
79%
23/28
50%
1/2
1+ 87%
163/187
80%
36/45
86%
199/232
86%
36/42
2 87%
79/91
93%
37/40
89%
116/131
95%
17/19
2+ 92%
95/103
94%
49/52
93%
144/155
89%
16/18
3 97%
31/32
100%
15/15
98%
46/47
88%
7/8
3+ 100%
45/45
100%
15/15
100%
60/60
100%
8/8
Overall 90%
439/487
90%
155/172
90%
594/659
88%
86/98

In accordance with the ACTFL Proficiency Guidelines and the por-para analysis, notable increases in accuracy are found at the Advanced (S-2) and Superior (S-3) levels for estar. For ser, however, there is a steady progression, with no sharp breaks at the thresholds.

After the Advanced (S-2) level, these learners actually selected estar more accurately than ser. The most obvious explanation for this is that ser, the unmarked form, was used in place of both estar and haber more often than estar was substituted for ser/haber. Since the copulas are usually taught through contrast, however, it is interesting to note that the substitution of ser for estar accounts for only 37% of the total ser errors, and estar for ser makes up 47% of the total errors with estar. Haber was replaced by the others only eleven times and was never used in place of either ser or estar. It would seem, then, that the opposition of these lexical items, like that of por-para, may be less of a problem than most teachers would predict on the basis of experience, although it should be noted that the training program emphasized intensively the functions of the various equivalents of English be.

Table 2 shows accuracy rates for copula selection only. When all types of errors are taken into account, the accuracy rates decrease (Table 3). Of 517 total entries for ser, 415 were accurate (80%), compared to 74% (140 of 188) for estar. Clearly, estar is both less frequent and more prone to a variety of errors. Overall accuracy thresholds seem to appear at the «plus» levels for ser and S-2+ for estar, suggesting that the speakers attended more to copula selection than to accuracy of form. This may be due to the importance of lexical accuracy for communication plus the heavy emphasis on ser vs. estar in instruction.



––––––––   1297   ––––––––

TABLE 3
PERCENT ACCURATE OVERALL
PERCENT ACCURATE OVERALL


ser estar haber Total
0+ 10%
1/10
50%
1/2
33%
1/3
20%
3/15
1 67%
20/30
75%
1/4
33%
1/3
59%
22/37
1+ 77%
155/201
71%
35/49
90%
35/39
78%
225/289
2 79%
75/95
68%
30/44
100%
17/17
78%
122/156
2+ 89%
93/104
82%
46/56
94%
16/17
88%
155/177
3 81%
26/32
82%
14/17
100%
7/7
84%
47/56
3+ 100%
45/45
81%
13/16
100%
8/8
96%
66/69
Totals 80%
415/517
74%
140/188
90%
85/94
80%
640/799




Results by Function

In general terms, ser normally serves as an identifier, definer, or classifier of entities and events, estar to locate entities in space and to attribute to its subject a condition that is in contrast with what the speaker views as normal or expected qualities. For purposes of this study, the uses of ser and estar have been listed according to what the speakers seemed to be doing with them as well as by type of linguistic structure. The most frequently-appearing uses of ser fall into the categories listed below, along with the means actually employed to express them.

A. Description of the Subject-Normal Qualities
(N+) copula + adjective (e.g., El jefe es inteligente.)
ser 148
estar 9 89% accurate
0 (Omission) 10

B. Definition (+ Description), and Identification
(N+) copula (+adj.) + N (+adj.) (X es una) (ciudad) (muy bonita.)
ser 146
estar 5 96% accurate
0 1

C. Impersonal expressions
ser 73
estar 0 100% accurate
0 0

An examination of the ser utterances and their contexts leads to admiration for the usefulness of ser (especially es) for creating abbreviated structures in order to manage one’s role in a conversation. The tasks at the lower levels consisted primarily of answering questions about the trainees’ own lives and interests and their future work. Because the questions typically elicited descriptions, explanations, and definitions, the trainees were able to build a great number of their answers around ser. Furthermore, in the first oral tests, the learners interacted with strangers, so that anything they said provided new information that could be expressed acceptably in rather general, nonspecific terms.

Es plus a common, generic adjective was found to serve many communicative purposes: introducing topics; providing a means to get across a general idea when specific vocabulary was not available in the interlanguage, or introducing a longer description, perhaps buying the speaker time to plan it (Example: Es bonito); taking the place of longer explanations for not doing things (Es difícil); and providing reasons for actions (Es mejor). When one explanation

––––––––   1298   ––––––––

became too complex, Es muy mal served as a concluding statement, to close the topic.

Noun +ser+noun (+adjective) constructions were used to describe, as above; to identify and describe places unknown to the interviewer (Es un poqueño pueblo); to circumlocute in attempting to define a word (housewife): ah, es muy acyupada a la casa; and to request help (...allí yo ah vizité ah ¿qué es la capital de X?)

Impersonal expressions offered these speakers many opportunities to explain their professions and work: Es necesario que yo planté árboles; to try to deal with complex explanations normally involving subjunctives: Él desea si usted si usted ah si es posible que usted ah ararenyar ah un una cena... (in a role-play situation); to explain possible plans: en dos años es posible ah que se me casé. In each case, the context is sufficient to explain the general meaning. At these early stages, it appears that the trainees launched many utterances with a ser plus adjective, but the impersonal expressions often led them into traps involving unforseen complexities. At the 1+level, they showed difficulty in finishing most constructions, so that common outcomes were es (mejor) para + infinitive; es (posible) a + infinitive; es muy (difícil) a + verb. In most cases the structural problems had been overcome by level S-2+, although the subjunctive forms were not mastered until at least S-3.

At later stages, when the volunteers were able to express themselves with more specificity, the number of impersonal expressions decreased. They also used a greater variety of expressions: es posible que became posiblemente, no estoy seguro, pero, etc.; es necesario que was replaced by hay que, necesitar + infinitive, tener que, and deber.

For estar the functions with the most tokens are as follows:

1. Location
ser 21
estar 41 65% accurate
0 1

2. Condition
ser 18
estar 56 74% accurate
0 2

3. Progressive
ser 1
estar 45 98% accurate
0 0

While the most common uses of ser appear early in the data (see Table 4), for estar the situation is different; while estar as locative occurred at the S-0+ and S-1 levels, the other two most common estar functions did not appear until S-1+, and the progressive was not used extensively until S-2. Furthermore, the substitution of ser for estar did not disappear until level S-2+ for conditions and S-3 for locatives.

TABLE 4
MAJOR SER-ESTAR-HABER FUNCTIONS
Proficiency Levels
MAJOR SER-ESTAR-HABER FUNCTIONS


Proficiency Levels


0+ 1 1+ 2 2+ 3 3+ Totals
Estar
Functions
Location (65% accurate)
estar 1 2 13 9 13 2 1 41
ser 0 1 10 3 7 0 0 21
Omission 0 0 1 0 0 0 0 1
Condition (74% accurate)
estar 0 0 17 17 11 6 5 56
ser 0 1 9 8 0 0 0 18
Omission 0 0 1 0 1 0 0 2
Progressive (98% accurate)
estar 0 0 5 7 21 5 7 45
ser 0 0 1 0 0 0 0 1
Omission 0 0 0 0 0 0 0 0


––––––––   1299   ––––––––

Ser
Functions
Qualities (89% accurate)
ser 5 8 65 33 31 6 0 148
estar 0 2 3 2 2 0 0 9
Omission 3 4 3 0 0 0 0 10
Unite
two entities
(96% accurate)
ser 0 5 41 23 36 15 26 146
estar 0 0 4 1 0 0 0 5
Omission 0 0 1 0 0 0 0 1

Among these three most common estar functions, the progressives were the most accurate, even if nine additional instances of overuse are counted as errors rather than accurate copula choices. Between the conditional and locative functions, conditionals were represented accurately more often than locatives. In fact, in these data there is an inverse relationship between order of appearance and degree of overall accuracy; although locatives appeared earliest, ser was substituted for estar in 33% of the cases, compared to 24% for the conditional. Ser was used in place of estar only 2% of the time for the late-appearing progressives. Of the three studies being compared here, VanPatten’s is the only one that shows an early appearance for progressives. It may be that his subjects studied these forms earlier, and/or that the required tasks elicited them with greater frequency than was necessary for the OPIs.

In fact, a possible explanation for the inverse relationship between appearance of a function and its accurate expression may be found in the communicative exigencies of the OPI. It is likely that the topics of the conversations required locatives and descriptions of qualities at an earlier stage than conditionals, as the trainees reported on the locations of their homes and work sites and described their lives and families in the United States. As the learners’ linguistic repertoires grew toward the end of training, and they had practiced extensively with estar, they were able to speak at greater length, offering additional information about locations, conditions, and progressive actions. The two functions at issue between the VanPatten and Ryan studies, locatives and conditionals, probably develop in a parallel fashion, with some differences among individual learners, as Ryan found at the lower levels. These data show near 100% accuracy with the conditional at S-2+ and with locatives at S-3. At S-2, neither reaches even 80% accuracy.




Stages of Acquisition

In view of the fact that the first interviews represented either previous experience with Spanish or the end of 260-300 hours of intensive language training, it is not possible to trace in detail the stages of acquisition at the lower levels of proficiency in order to compare them with those that were postulated by VanPatten and Ryan. Nevertheless, these data 1) provide support for some of the stages and not others, 2) indicate which of these were short-lived and which continued to the upper proficiency levels, and 3) identify at least one new stage in the acquisition of the functions at the higher levels (the appearance of the passive voice).

VanPatten and Ryan hypothesize that Stage 1 in the acquisition of the copulas is omission of both. Although this stage may exist in reality, VanPatten states that it is short-lived, and these data do not reflect the earliest attempts at communication. Nevertheless, omission continued even at level S-2+, and made up 50% of the total errors at S-0+, 38% at S-1, and 9% at S-l +, when lexical selection replaced it as the most common error.

Stage 2. Overuse of ser in locative and conditional contexts (Ryan). Only five such contexts are found in these data before the S-1+ level, and ser was used twice. At S-1+, ser was indeed used in 19 of 51 instances (37%), and this phenomenon continues through levels S-2 (30%) and S-2+ (22%, locative contexts only). Clearly, this is not a fleeting phase for these learners, but part of the on-going process of defining and controlling the limits of ser and estar. VanPatten’s version of this stage stated that ser is used for most copula functions. Indeed, ser was

––––––––   1300   ––––––––

overused in estar and haber contexts, and was extended beyond normal be functions, replacing several other forms and structures (e.g., impersonal expressions in place of subjunctive; origin (es de) for vivir; and ser for age, where tener is required in Spanish). Ser also played the verb role in very abbreviated or confused utterances (Sí, X es ah muy ci-ciudad; Pero ahora es vacación). Clearly, ser was greatly overgeneralized, especially at the lower levels of proficiency.

Although the strategy of leaning on es + adjective/noun often led into complex situations that resulted in grammatical and communicative breakdown, it also may have given impetus to the gradual acquisition of the subjunctive and other structures; at higher levels, the volunteers learned how to manage these structures and replaced indicatives and infinitives with subjunctive-like forms (e.g., casé for case, planté for plante, solicitera for solicitara), and finally with accurate versions.

Ser constructions, then, seem to facilitate language acquisition in at least two ways: 1) they help the speaker to maintain a role in the conversation, thus providing communicative experience and eliciting further input, and 2) they lead the learner into structures that s/he will need to acquire in order to progress. Those who are so motivated will expand their capabilities, while others continue to depend to a greater extent on simple, truncated structures such as es+adjective to fulfill a variety of purposes.

Stage 3. Absence of the copula in conditional contexts (Ryan). The Peace Corps data do not show such a stage, although it may have diminished by the end of training. For this reason, Stage 6 is also negated for this study. (Estar replaces zero copula in conditional contexts.)

Stage 4. Estar in progressive constructions (VanPatten). Progressives began to appear with frequency at S-2+. Because the structure and its use were not presented in the textbook until after most of the rest of the verb system had been thoroughly discussed and practiced, those who entered as beginners may not have dealt with it formally during training. Nevertheless, the six examples found at the S-1+ level were from five different trainees, which would imply that it was not entirely unfamiliar to the group. A better explanation may be that OPIs before the Advanced level do not elicit narration, nor is there any other common need for this structure before learners are able and willing to offer explanations that go beyond the normal.

Stage 5. Replacement of zero copula errors by errors with estar where ser is required (Ryan). For the description of qualities and for identification, this phenomenon occurred in very small numbers at four levels, from S-1 to S-2+. Some of the trainees may have passed through this phase before their end-of-training OPIs.

Stage 7. Estar in conditional contexts (Ryan). Once this function appeared in the data, at S-1+, estar was used in 63% of the cases. Although the exclusive use of estar for conditions did not begin until S-2+, it was produced with greater accuracy than for location.

Stage 8. Estar replaces ser in locative contexts (Ryan). The exclusive use of estar for locatives occurred later (S-3) than for conditions, a finding that weakly supports Ryan’s order for Stages 7 and 8.

Brief profiles are drawn here for each of the proficiency levels, in order to achieve a more thorough description of cross-sectional progress in the acquisition of the copula and shed further light on the processes.




Novice Plus (0+) N = 1

Besides the omission of ser and the nearly total lack of contexts for estar at this level, another of its defining characteristics is the presence of many extremely reduced utterances, with ser used as the verb:

Question: ¿Usted vive en el hotel?
Answer: No, es ah en ah familia.



Ser was also overused as a filler where English would require is: Me llama es.... To describe normal qualities, this speaker used a variety of forms: ser (accurate); adjectives without ser; English is; and three examples of French e’.

Estar appeared once, for location, and also in a self correction in a case where the speaker realized that ser was inadequate:

Question: ¿Cómo se llama la familia?
Answer: ...es está X.



This subject had studied Spanish in school, and must have been trying to retrieve the rules from memory.




Intermediate Low/Mid (S-1) N = 2

Ser continues to dominate, for describing qualities, identifying and asking for identifications, and indicating existence (for haber); and it is used with key lexical items to express ideas in extremely reduced form: Sí, X es ah muy ci-ciudad.

––––––––   1301   ––––––––

(Sí, X es una ciudad muy grande). Impersonal expressions also appeared for the first time at this level, although the adjectives were generally inaccurate (Es necesita..., Es posibilimente).

Although estar began to occur for location: ...mi casa está cerca ah ah X, ser was also used by the same person. Omission was the most common single error, accounting for 38% of the total errors for all three verbs. Still no effort is apparent to describe actions in progress or conditions.




Intermediate High (S-1+) N = 8

The eight subjects who reached this level by the end of training were now able to describe locations (23 tokens), conditions (26), and progressive actions (6), although ser was used in 20 of these cases. Accuracy in estar selection jumped to 80%, compared to 50% at Level S-1, and the tokens for estar increased from 4 to 45. Estar and ser were both used for haber. The number of functions expanded dramatically among these subjects, for both verbs. For example, ser was used to express reasons, origin, purpose, and temporal goal, and to talk about age (in place of tener). For estar the progressive made its first appearance. After this point, in fact, few new functions were added.

One of the most interesting characteristics of the level is the use of 23 impersonal expressions in a variety of structures and for many functions (e.g., Ella recomiende que es necesario que ah yo escribe...). Impersonal expressions enter the data shortly after descriptions of qualities and may develop from them by analogy.

Omissions still make up 13% of the errors, but the most common are selection of the wrong copula (48%). Tense, mode, aspect, and agreement in number with the subject make up 15% of the total errors at this rather unstable acquisitional stage, where great variability is to be expected.




Advanced (S-2) N = 4

At this level, which represents roughly a year of Peace Corps service for the four subjects whose proficiency ratings improved the least, the overall accuracy rate did not improve, and 42% of the total errors were still due to wrong choice. Nevertheless, progress is demonstrated by the disappearance of omission errors; the mostly accurate use of estar for location; and an increase in progressive forms, which serve a similar range of functions as in English, resulting in their overuse in Spanish. Thirty-six percent of the errors now are with mode, tense, aspect, and agreement in number. Errors with haber have virtually disappeared. The two copulas continue to vie for roles in expressing conditions. Useful expressions for managing conversations were added here (e.g., Es decir, Es que, and Es verdad).




Advanced Plus (S-2+) N = 3

In overall accuracy with ser and estar this group improved from 74% at S-1+ to 86%. Improvement is evident both in copula selection and in inflections, especially with the forms of ser to indicate mode and aspect. For ser the most common error is in locatives, but Subject 5 accounts for all but one of the tokens. This person also produced eight of fourteen impersonal expressions. Patterns of errors seem to have disappeared for the group, although they still exist for one individual.




Superior (S-3) N = 1

This subject had mastered copula selection subsequent to the previous interview. Only eleven errors were found all together, showing patterns with inflections for mode and aspect and overuse of the progressive, for which he also used ser on two occasions.

Copula tokens decreased in number to 56 as more specific and varied modes of expression were available in the interlanguage, and only six impersonal expressions were found, all followed by subjunctives, both present and past.




Superior (S-3+) N = 1

For this individual, all verb choices were accurate, and the passive voice was used four times. This volunteer was the only one to use the plural habían, which is typical among the local people. He also used accurately several new expressions, such as es decir que, es cierto, es lástima que, and estábamos al punto de....

In summary, the progress of copula acquisition observed in these data begins with numerous omissions, which are accompanied by the use of ser for identification, definition, and descriptions of general qualities. During the process of integration of those functions, impersonal expressions enter the interlanguage of most learners in great numbers, as well as a dependence

––––––––   1302   ––––––––

upon ser plus key known vocabulary to maintain a role in the conversation. Ser is extended to locative and conditional contexts, as well as many that require other verbs, such as vivir and tener. At the same time, hay appears.

Estar then begins to dominate in locative and conditional contexts, and for progressive constructions if they are needed and frequent in the input. Finally, ser is used in passive constructions at a much later time.




Conclusion

Besides shedding some new light on ser - estar acquisition and on the nature of the OPI, this study raises again the question of the effects of instruction and provides some additional information on communication strategies as they affect the acquisition of the copula.

In looking closely at the use of these verbs at various proficiency levels, it is possible to identify the functions most often required for the types of topics that are prevalent in the interviews at each level. For the lowest levels these include identification, definitions, and descriptions of qualities. Since the interlocutors are usually strangers to each other, the examinee is able to participate acceptably, though incompletely, using the most general vocabulary; because the information is all new to the listener, an information gap exists such that very basic information is adequate. The use of es plus key words is sufficient to prevent total conversational breakdown. If the learner wants to participate more fully, s/he is motivated to acquire the connectors, vocabulary, and morphology necessary to express meanings more specifically and navigate through the various complex structures that tend to follow upon es + adjective/noun. For most of the subjects the number of these structures decreased after S-1+.

The question of the effectiveness of instruction remains controversial. If any teaching program has an influence on acquisition, however, Peace Corps training should be one of the most effective. Unlike normal academic situations, class size was limited to five students, classes met four to six hours daily with study time added, and the attention and efforts of the entire staff team were devoted to seeking the most effective means possible to assure high levels of proficiency. The textbook stressed cognitive processing and emphasized the copulas from the first pages, presenting step-by-step analyses of their uses and forms and constant review and reintegration throughout, while Glasswork was devoted to personalized oral practice of the material under study. It is likely that heavy emphasis on the Spanish equivalents of be and the generally high motivation of the trainees influenced their monitoring of these items and their consequent high level of success in lexical selection compared to the subjects in the VanPatten and Ryan studies, and compared to their own much lower accuracy with por and para, which were not contrasted until very late and may not have been formally presented to the beginners at all during training. Clearly, though, the influence of instruction is limited by the degree to which learners can apply consciously learned rules, the accuracy of that learning, and the degree to which the rules are explained and explainable. (For example, the distinctions between ser and estar for locating entities are almost never explained to learners, and the separate structures for It is + adjective + verb were probably not taught. Both structures showed intersubject and intrasubject variability).

Although intensive study may have affected overall accuracy with lexical selection positively by directing the trainees’ attention to these forms, the sequence of the appearance and acquisition of the ser-estar functions shows little relation to the order of presentation and study. While estar appears first in the textbook, for conditions and locations, it is not controlled until much later. On the other hand, while ser for describing qualities appears very early and accurately in the data, this function was not formally presented until page 150. Only the progressive forms show a relationship between presentation and use, but this phenomenon may be attributable in large part to the nature of the OPI itself.

Related to the issue of the impact of instruction is the opposite question: How much «grammar» does one acquire in a more naturalistic situation? The analyses of por - para and ser - estar both show constant progress toward acquisition of both pairs while the volunteers lived at their work sites and interacted with native speakers. Nevertheless, as in DeKeyser’s study, among individuals a great deal of variability is found in fluency, accuracy, and apparent communication and learning strategies. Four examples are illustrative. (He is used here generically for all subjects, for simplification and to avoid any possible identifying references. The small number of subjects makes male-female comparisons unfeasible).

Subject 1, who openly expressed a preference

––––––––   1303   ––––––––

for aspects of the culture back home, volunteered very little information except when talking about his work. On the surface, at least, he seems to have been instrumentally motivated; that is, he learned Spanish best for purposes of his work as a Peace Corps volunteer; in the interviews he spoke with greatly increased fluency and enthusiasm when asked to role play work-related situations. This subject preferred para over por, with great accuracy, but he was the only one to use estar more than ser, especially after nine months on site, with much less accuracy. He appeared to have difficulty with language learning and spent much time with other English speakers.

Subject 2, by contrast, progressed very rapidly in both fluency and accuracy, achieving an unusually high rating of S-3+ by his eleventh month on the job. He explained that he studied Spanish intensively on his own, and indeed he was very successful at learning the appropriate forms to express his intended meanings. In the field he acquired all of the common uses of the copula, including the passive.

Subject 7 spoke a great deal but leaned heavily on es + adjective throughout. Although he achieved a rating of S-2+ after ten months on site, his form-function relations varied greatly with both ser-estar and por-para (showing a preference for por). He expressed impatience with the study of «grammar». This speaker demonstrated an orientation toward message over form, if not a meaning-based approach as defined by Finnemann (1990). (His use of ser much more than estar does not fit Finnemann’s definition).

Finally, Subject 9 spoke briefly and with extreme rapidity compared to his expression in English, leading to the suspicion that he was attempting to hide his considerable structural inadequacies by emphasizing his fluency. He reached a rating of only S-2 after thirteen months in his site, which was located in a city that offered opportunities to spend time with English speakers.

In short, while strategies and output varied considerably, some approaches were more effective than others. All learners need to «simplify» the language in the beginning -that is, use the fewest forms and rules to express the most meaning- but language acquisition researchers might more profitably view simplification in a more positive sense: learners do not seek to reduce the language, but rather to expand their ability to use it to the point where they are comfortable with their expertise. The most successful learners continually extend their knowledge base through attention to input and instruction, and effort to organize and remember it and test it through interaction with others.

The first two analyses of the Peace Corps data have begun to build profiles of individual Spanish acquisition. Further research will focus on other aspects of the process, analyzing individuals’ acquisition of other forms and functions as well as their use of repairs and other strategies, in an attempt to describe and explain the differences in achievement among volunteers in the field.




WORKS CITED

Brisk, Maria Estela. «The Acquisition of Spanish Gender by First Grade Spanish-Speaking Children». Bilingualism in the Bicentennial and Beyond. Eds. Gary D. Keller, Richard V. Teschner, and Silvia Viera. Jamaica, NY: Dept. of Foreign Languages, York College, CUNY, 1976. 143-60.

De Keyser, Robert M. «From Learning to Acquisition? Monitoring in the Classroom and Abroad». Hispania 73 (1990): 238-47.

_____. «Foreign Language Development during a Semester Abroad». Ed. Barbara L. Freed. Foreign Language Research and the Classroom. Lexington, MA: DC Heath, 1991. 104-19.

Eckman, Fred R. «Some Theoretical and Pedagogical Implications of the Markedness Differential Hypothesis». Studies in Second Language Acquisition 7 (1985): 289-307.

Finnemann, Michael D. «Markedness and Learner Strategies: Form- and Meaning-oriented Learners in the Foreign Language Context». Modern Language Journal 24 (1990): 176-87.

Guntermann, Gail. «An Analysis of Interlanguage Development Over Time: Part I, por and para». Hispania 75 (1992): 177-87.

Ryan, John Michael. «Ser and Estar: Acquisition of Lexical Meaning in a Natural Environment». M.A. Thesis. Arizona State University, 1991.

VanPatten, Bill. «The Acquisition of ser and estar by Adult Learners of Spanish: A Preliminary Investigation of Transitional Stages of Competence». Hispania 68 (1985): 399-406.

_____. «Classroom Learners’ Acquisition of Ser and Estar: Accounting for Developmental Patterns». Eds. Bill VanPatten, Trisha R. Dvorak, and James F. Lee. Cambridge: Newbury House, 1987. 61-75.







Arriba