The acquisition of psychological-verb alternations in Spanish: two teaching approaches. (Language Teaching & Learning).

Abstract

This study compares the effect of a traditional approach to teaching grammar with the processing instruction approach when dealing with psychological verbs that alternate between a transitive and an intransitive use (molestarle/molestarla). The traditional approach tries to help students develop the right connections between the form of the input and its meaning emphasizing the early production of the structures. The processing instruction approach delays the production of the structure until the student has been given enough time and practice processing the input. The results, measured both by the correct interpretation and production of the structure, indicate a clear advantage for the group that was taught following the processing instruction approach.

1. Introduction

Psychological verbs have always presented a challenge to the general theory of the relationship between syntax and semantics and have been the object of numerous studies especially in the past 15 years, Belletti and Rizzi (1988); Bouchard (1992); Croft (1993); Grimshaw (1990); Pesetsky (1987) Whitley (1995). I will definee this group of verbs as those that include two participants in their semantic structure, one representing the experiencer of some type of feeling, and the other one representing the cause of that feeling. Examples in English would be like, hate, fear, frighten, attract, worry, etc.

A peculiarity of Spanish psychological verbs that has received little attention is the fact that there is a group of verbs that can enter an alternation between a state and an event reading, which affects their syntactic structures making them intransitive or transitive respectively. An example of the state-intransitive version is shown in (1). The event-transitive version is illustrated in (2)

(1) A mi madre le preocupa mi situacion To my mother 3s-DAT worries my situation My mother worries about my situation

(2) No quiero preocuparla

No want-1s worry-3s-ACC

I do not want to worry her

The first sentence shows an intransitive use of the verb preocupar with the dative pronoun le. The second example shows the transitive version with the accusative pronoun la. I will call the transitive version the preocupar-A class, for accusative, and the intransitive version the preocupar-D, for dative. Other examples of verbs that show the same type of variation in Spanish are asustar, molestar, atraer, interesar, sorprender, etc

The difference between these two structures presents a challenge even to advanced learners of Spanish. The speaker frequently has the option to decide whether the action described by the verb is to be characterized as an event or a state. The difference in meaning depends entirely on the object pronoun used to refer to the experiencer. However, the pronoun, being a grammatical item, is not a very salient feature. As VanPatten (1996) points out, learners pay more attention to lexical than to grammatical items. To complicate things even further, indirect object pronouns are usually redundant in Spanish. All this combines to make the object pronoun a feature to which learners will pay little attention. But often learners fail to understand or adequately produce the difference between event and state precisely because of a lack of attention to the object pronoun. If communicative value and saliency are good predictors of the likelihood of detection of a feature in the input, it is not surprising that the difference between these two types of verbs may go unnoticed.

The goal of this study was precisely to determine which one of two different teaching approaches was more effective in helping learners acquire the correct form-meaning relationship between my preocupar-A and preocupar-D classes of psychological verbs. The study was carried out in order to compare the effects of a traditional approach to teaching this structure compared to an approach based on processing instruction.

2. Traditional vs. Processing Instruction

This study compares the effectiveness of a traditional approach to grammar teaching with what has been called "processing instruction". In a traditional classroom, grammar teaching is organized so that a particular point is presented and explained and then students are asked to produce it almost immediately after the explanation, the lesson goes from input directly to output. The students receive the input in the form of a grammatical explanation and then have to produce the structure (output) either orally or in writing. Traditional instruction trains subjects to produce a structure by manipulating (through error correction, for example) the learner's output.

VanPatten (1196) defines processing instruction as an input-based type of grammar instruction whose purpose is to affect the ways in which learners attend to input data. It differs from a traditional approach most notably in that subjects are given a number of practice activities that engage them in exercises to readjust the way they process the input. In the processing instruction model the students are not asked to produce the structure until very late in the lesson and before they have to produce it, they are given ample opportunities to process the form in the input to assign meaning to it, starting from very simplified tasks and moving towards connected discourse.

3. Method

3.1 Subjects

The design of this study follows those of VanPatten and Cadierno (1993) and Cadierno (1995). I used two experimental groups and a control group. They were randomly assigned to a treatment. The subjects were students in advanced Spanish courses at a four-year college. One of the experimental groups was assigned to the traditional approach and the other to processing instruction. The control group received no instruction related to psychological verbs.

The groups were made up of intact classes. However, students who were native or heritage speakers of Spanish were excluded from the study. Also excluded were subjects whose score on the pretest was higher than 80%. All subjects were familiar with the vocabulary and structures under investigation, but had never received specific instruction on how to assign the correct semantic structure to each syntactic configuration. None of the syllabi for the three classes had scheduled instruction on this matter. The traditional group had 9 subjects; the processing instruction group had 11; and the control group 10. An ANOVA performed on the pretest scores of all three on both tasks revealed no significant differences (p = .0655 on production; p = .7060 on interpretation)

3.2 Instructional Packets

Two different teaching packages were designed, one for each experimental group, with minimal differences in terms of the type of exercises involved and the amount of time devoted to explanation and practice. Both approaches can be followed using the same materials just by varying the focus of the exercise.

Explanation in the traditional packet focuses on the difference between transitivity and intransitivity and specifically as it relates to psychological verbs. Immediately after the explanation, students start practicing the structure with transitive verbs only, then intransitive ones and finally both. During the course of the lesson there was revision of the basic concepts and error correction when necessary.

A traditional approach tacitly presupposes that learners will attend to the input, assign some meaning to the structure they perceive, and incorporate that structure into their interlanguage. It presupposes also that once the structure is part of the interlanguage learners will be able to analyze it and produce it when it is appropriate to the meaning they want to express. In a typical traditional exercise the students would be given a sentence with a blank that they would have to fill out providing the right object pronoun; or they would be given a sentence in English and asked to provide the appropriate Spanish translation. The emphasis of the traditional instruction lesson was on output practice.

In the processing instruction package the emphasis was on directing the subjects' attention to the structure of the input. Rather than the traditional concepts of direct and indirect object, the processing instruction group was exposed to the notion of cause and experiencer and how the nature of the cause determines the syntactic structure of the sentence. That is, they were taught to direct their attention to the input, analyze it and then assign a particular syntactic configuration to it. And rather than giving them a holistic way to interpret the sentence, they were taught to attend to one specific feature of one of the elements involved in the meaning.

From the perspective of processing instruction, learners first have to analyze the structure to make sure they make the correct mappings between form and meaning. Once those mappings are established, they will be internalized and will enable the learner to comprehend and produce the structure when needed. The processing instruction lesson would give the students a sentence and ask them to decide whether the object pronoun that it contains is correct and why; or would give them two Spanish versions of an English sentence, one containing the transitive structure and one the intransitive, and ask them which one would be the correct version and why.

The students were already familiar with the meaning and use of psychological verbs, so they did not need extensive explanations as to how to use them. The objective of the teaching segment was only to instruct them on how to consistently produce and interpret the subtle difference conveyed by the use of a direct or indirect object pronoun. Since this was a very specific task, the teaching segments were done in only one class session (50 minutes). The instructional packages were collected at the end of the lesson and no related homework was assigned. The posttest was done during the class period immediately after the instruction (two days later), and the follow-up test a week later.

3.3 Pretest, Posttest and Follow-up

Three different tests were designed and given to four students not involved in the study to compare the reliability of their equivalency. The comparison showed that the differences among them were not significant. The interpretation task consisted of 20 sentences in English accompanied by a Spanish version that was or was not a correct paraphrase. Subjects had to decide whether the Spanish paraphrase was adequate or not. The task contained 15 instances of the structures under investigation plus 10 distractors on the pretest and 5 on the posttest and follow-up (see appendix). The number of sentences that contained transitive and intransitive structures was balanced within and across the different tests. In the production task students had to complete a sentence with an object pronoun selected from a list provided. The task included a total of 15 sentences including the verbs under investigation plus 5 distractors (see appendix). Subjects were given 15 minutes to complete both tasks.

3.4 Scoring

Both sections used the same scoring method and received the same number of possible points, 15. Since the correctness of the judgment and the production depended on only one feature, there were only right and wrong answers (1 versus 0) with nothing in between. Blank answers received a score of 0. The final scores were converted to percentages.

4. Results

4.1 Interpretation task

I have mentioned that the initial difference on the pretest scores among groups was not significant. However, posttest results show a significant effect of instruction on the interpretation task in the processing instruction group, F (9.24), df (29), p = .0009. Also, the effect seems to last over time, as indicated by the results of the follow-up test, F (14.01), df (29), p = <.0001. Traditional instruction did not have a significant effect on the ability of subjects to interpret structures correctly. The small advantage of the traditional instruction group as compared to the control group was not statistically significant. As for the variable time, the gain experienced by the groups after instruction, the difference, once again, is only significant in the case of the processing group, F (9.05), df (29), p = .0008. Figure 1 shows the effect of time (pretest vs. posttest vs. follow-up) and type of instruction (traditional vs. processing) for the three groups. See issue's website <http://rapidintellect.com/AEQweb/fallp.htm>.

4.2 Production task

Again, the scores on the pretest showed that the difference among the groups was not significant on the production task, although in this case the processing group started out with a slightly higher average score (M = 67.18) compared to the control group (M = 52.70) and the traditional group (M = 54.89). A quick look at the results displayed on Figure 3 reveal that instruction was particularly decisive in the case of the processing group, but it had very little effect on the traditional group and no effect on the control group. See issue's website <http://rapidintellect.com/AEQweb/fallp.htm>.

The results of an ANOVA performed on the scores of the three groups on the posttest proved that only the processing group improved significantly, with respect to the control group, which received no instruction F (15.11), df (29), p = <.0001. The same effect is still present at the time of the follow-up test. The ANOVA performed on the time variable for the production task shows, once again, that only the gains experienced by the processing group over time were significant (p = .0005). Once again, the advantage seems to last over time.

5. Discussion

Two main conclusions can be drawn from the results of the study:

1. Only processing instruction has a significant effect on the acquisition of the correct form-meaning relationship for this type of verbs.

2. The effect of processing instruction is equally apparent on the interpretation and the production of these structures.

The first conclusion is especially remarkable if we consider that the instruction that the traditional group received concentrated almost exclusively on production. But even on that task the effect of instruction was negligible. The interlanguage of subjects in the traditional group still contained, after treatment, an incorrect representation of the mapping from semantics to syntax of this type of verbs, which is why they were unable to significantly improve their interpretation or production of these structures.

In contrast, it was clear that the subjects in the processing group acquired a way to detect the difference in the input and reproduce it in their output. In other words, processing instruction helped them restructure their interlanguage. Processing instruction differs from a traditional approach most notably in that subjects are given a number of practice activities that engage them in exercises to readjust the way they process the input. In contrast, traditional instruction trains subjects to produce a structure by manipulating (through error correction, for example) the learner's output. What is common to the two approaches is that they both combine a theoretical explanation of the structure with some kind of practice. One could argue that the advantage showed by the processing instruction group is due to the explanation rather than the practice. Subjects in the processing instruction group may have benefited more from the explanation and performed better as a consequence. Saying that would be equivalent to maintaining that whatever knowledge my subjects had gained was learned rather than acquired knowledge, using Krashen's terminology. However, the fact that subjects were given a very limited time to complete the tasks, seems to suggest that they would have to rely on acquired knowledge. Also, in a study that replicated VanPatten and Sanz (1995), VanPatten and Oikkenon (1996) maintain that monitoring is not the cause of the gains experienced by the processing instruction group. Another possibility would be to accept that second language acquisition follows a weak interface model like the one proposed by Ellis (1994). That is, subjects can access learned knowledge when they are processing input. This knowledge allows them to notice features in the input that later become acquired.

Our results, therefore, may not be conclusive enough to make any strong claims about the reason why processing instruction seems to be superior to traditional instruction. However, looking at the results of this study as well as the other two reported in Rubio (2000) it is obvious that regarding psychological-verb structures, particularly structures in which the relationship between form and meaning is rather obscure, subjects in the processing instruction groups consistently did as well as, if not better than those exposed to the traditional approach. As with most studies of this kind, the fact that intact classes were used rather than students randomly assigned to different treatments is an important methodological limitation that must be taken into account.

There is certainly a need for more studies comparing processing instruction with traditional approaches before definite claims can be made and there is certainly a special need for studies investigating the acquisition of isolated features that exhibit a difficult form-meaning relationship. It would also be necessary to compare how the effects of both types of instruction hold over a longer period of time.

Bibliography

Belletti, A. & Rizzi L. (1988). Psych verbs and theta theory. Natural Language and Linguistic Theory 10, 291-352.

Bouchard, D. (1992). Psych constructions and linking to conceptual structures. In Hirschbuler, P. & Koerner, K. (Eds.), Current Issues in Linguistic Theory. Amsterdam: John Benjamins.

Cadierno, T. (1995). Formal instruction from a processing perspective: An investigation into the Spanish past tense. Modern Language Journal, 79, 179-193.

Croft, W. (1993). Case marking and the semantics of mental verbs. In: J. Pustejovsky (Ed.), Semantics and the lexicon. Dordrecht: Kluwer Academic Publishers.

Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford University Press.

Grimshaw, J. (1990). Argument structure. Cambridge: MIT Press.

Krashen, S. D. (1985). The input hypothesis: issues and implications. London; New York: Longman

Pesetsky, D. (1987). Binding problems with experiencer verbs. Linguistic Inquiry, 18, 126-140. Rubio, F. (2000). Psychological verbs in Spanish: Structure and acquisition. Unpublished doctoral dissertation, State University of New York, Buffalo.

VanPatten, B. (1996). Input processing and grammar instruction: Theory and research. Norwood, NJ: Alex.

VanPatten, B. & Cadierno T. (1993). Explicit instruction and input processing. Studies in Second language Acquisition, 15, 225-243

VanPatten, B. & Oikkenon S. (1996). The causative variables in processing instruction: Explanation vs. structured input activities. Studies in Second Language Acquisition.

VanPatten, B. & Sanz C. (1995). From input to output: Processing instruction and communicative tasks. In Fred Eckman et al. (Eds.), Second Language Acquisition and Pedagogy. Hillsdale, NJ: Erlbaum.

Whitley, M. S. (1995). Gustar and other psych verbs: A problem in transitivity. Hispania 78, 573-585.

Whitley, M. S. (forthcoming). Psych verbs: Transitivity adrift.

Fernando Rubio, Southern Oregon University, OR

Fernando has a PhD in Spanish Linguistics from the State University of New York at Buffalo, his current teaching and research focuses on methodology and second language acquisition.