Iranian University Students’ Preference for Verb Centered VS. Construction Cues to Sentence Structure

Abstract

The present study sought to determine whether Iranian EFL learners majoring in English were more influenced by the type of verb or argument structure patterns in their interpretation of English structures. To answer this question, 108 English majoring participants took part in a sorting task which was designed to reveal the participants’ preference in sentence interpretation. The results of a Chi square test indicated that the said participants were decidedly more inclined to sort the structures (i.e., through interpreting them) by relying on the argument structure patterns around which the structures were configured rather than paying attention to the structures’ verb types.

Key Words: Verb, Argument Structure, Processing,

Introduction
It has been argued that the lexical representation of a verb specifies the number and type of arguments related to the participants described by the verb (Chomsky 1965). This characteristic of the verb is generally referred to as its sub-categorization frame argument structure. According to this view the verb is the best predictor of sentence interpretation. To make this point more lucid, consider the lexical representation of the verb give. The lexical presentation of this verb specifies that it requires three arguments: a subject, a direct object, and an indirect object as in:
1- John gave a book to Kim.
The argument structures of this verb may be changed through a transformation that would have no effect on the overall meaning of the structure as in:
2- John gave Kim a book.

Then, one may assert that a particular verb may in fact be manifested through different argument structure patterns. This phenomenon has given way to a series of psycholinguistic studies where the main aim has been to see whether in the interpretation of a structure the processor is more biased toward the verb or the argument structure. For example, Healy and Miller (1970) found that English-speaking participants were more likely to group together structures having the same verb than the structures having the same subject argument. Thus, Healy and Miller concluded that the verb is the main determiner of sentence meaning.

More recently, however, the observation that a particular verb may occur in many more argument structure patterns than was generally assumed (Goldberg, 1995) has somewhat complicated the prediction that the verb is the main predictor of the sentence meaning. For example, the verb kick, which is traditionally assumed to be a transitive verb, can occur in various arguments structure configurations:
3- Pat kicked the wall.
4- Pat kicked Bob black and blue.
5- Pat kicked the ball into the stadium.
6-Pat kicked the ball.
7-Pat kicked her foot against the chair.
8- Pat kicked Bob the ball.
9-Horses kick.
10-Pat kicked his way out of the operating room.

These sentences designate a variety of event types, ranging from simple transitive for structure (3), to caused change of state in structure (4), to caused motion in (5), to attempted action in (6), to transfer in (8), and to motion of subject in (10). Thus, it may be concluded that in non- transformational structures, the verb kick denotes a different sense in the event that it constructs.

Hypothesis
The present study is geared toward answering the question of whether sentence interpretation is more a function of the type of verb or the types of argument structure within which the sentence is configured. Given, the contradictory views and findings on this topic, this researcher opted for a non-directional hypothesis with the following null hypothesis:
H0: There is no difference in the parser’s preference for verbal or argument structure cues of the sentence.

The significance of the study
It hardly needs to be mentioned that should the findings of the present study provide evidence for processing preferences based on verb type or construction type, the pedagogical overtones of the findings would affect both teaching activities and syllabus design and material development for English books used in the university. More specifically, by focusing either on the type of verb or the construction patterns related to it, both material developers as well as the instructors can facilitate the process of learning L2 in the classroom.

Methodology
Participants
A total of 108 junior and senior students from Azad University, Shiraz, took part in the study. Given that Azad University is a private institute one may assume that the participants’ socio-economic make up were of almost the same composure. Furthermore, in the light of the fact that the greater majority of the university’s students consisted of female students, the number of male participants stood at 19. Due to the relatively low number of male participants, no attempt was made to introduce ‘gender’ as a moderator variable.

The junior and senior participants were all selected from English-major fields that included ‘Teaching EFL’ and ‘Translation’. In completing the first two years of their university studies, the participants had taken a great variety of English proficiency courses in all the four skills and were either in the midst or had actually completed courses in ‘linguistics’, ‘discourse’, ‘contrastive analysis’, and other courses, of the more technical nature. Hence, it may be asserted that the participants had attained a sufficient base not only in the use but also the usage of English as a foreign language. Thus, the participants were endowed with the necessary knowledge base to process the structures with either the verb centered or the construction motivated strategies and that their choice for either strategy could not be attributed to a lack of the required knowledge base, rather their choice of strategy could be directly related to the operations cognitive mechanisms.

Stimuli
By crossing four verbs with four construction types, sixteen English sentences were created (see Appendix A). The participants were asked to use a sorting paradigm following Healy and Miller (1970) to complete the task of interpreting the sentences. This design yielded four sets of sentences with the same verbs (throw, take, get, and slice), and there were four sentences each with the following argument patterns: transitive, intransitive, caused motion, and resultative constructions. No content words other than the main verbs were repeated through the stimuli set. All of the names used in constructing the stimuli were of the same gender to avoid introducing an irrelevant factor.

Procedure
The participants were tested as a group. Each participant was given one shuffled set of stimulus cards to sort (i.e., each shuffled set contained 16 cards similar in size with the structures printed on each). The participants were then asked to sort the sentences (i.e., 16 cards) into four piles while paying special attention to the meaning of each sentence as well as their structures. After the sorting task was completed, each pile would contain four sentences that fell in the same overall class of meaning. Thus, sentences considered to be closer in the overall meaning would be placed in the same pile. This sorting paradigm was anticipated to shed light on the determining factor employed by the participants in the interpretation of the sentence. It was further anticipated that the participants might utilize one of the following strategies in their sorting activity:
a- Form piles based on verb type (i.e. each pile would be made up of four similar verbs), or
b- Form piles according to the argument structure of the sentences.

Results
As will be recalled, a total of 108 participants took part in this study. The sorting task administered to the participants was designed to register their preference for the effect of the argument structures of the sentences of the task versus the effect of verbs of the sentences. The task was composed of 16 separate sentences with four different verbs and argument structures. It was anticipated that in their endeavor to sort the sentences into four sets, the participants would either take into consideration the type of the sentences’ verbs or argument structures.

The effect of verb type on sorting
Only 19 of the 108 participants relied on the type of the verb structures to sort the sentences. Thus the sorting activity of these participants was based on the verb type of the 16 sentences (i.e., four groups with four sentences having the same verb in each group).

The effect of argument structure type on sorting: the rest of the participants, 89 in all, sorted the sentences into four groups according to the type of argument structure of the sentences.

Hence, in effect argument structure tendencies for sorting the Sentences outperformed verb centered preferences by a proportion of almost 4.5 to 1.

The verb VS. argument structure sorting preference of the
participants on the task

Inferential Statistical Analysis
In order to test the null hypothesis of this study (i.e., there is no significant difference in the participants’ preference for verb type or argument pattern of structures) one has to resort to inferential statistics. The particular statistical formula suitable for testing the null hypothesis is the Chi square. The justification for choosing this particular formula is that the constructs with which the present study is concerned are operationalized as nominal variables. Consequently, whereas in comparing two means of the interval type the t-test is used; with variables of the nominal type the chi square formula is used to compare their means. The following table shows the results of the Chi square formula:

Comparison of the means obtained for the participants’
sorting preferences.

x2 = (O-E)2 /E = 181.50

The observed f column represents the frequencies of the participants’ sorting preferences (i.e., 19*4=76 and 89*4=348 for the frequency of the number of times the participants sorted by paying attention to the verb types and argument structure types, respectively). The total number of sorting preferences was also multiplied by 2 (i.e., 108*4/2). In other words, the expected f column represents the figure that would be expected if the participants ‘preference for sorting was determined purely by chance (i.e., if the participants had displayed no sorting preference).

Finally, the obtained value for x2 (the Greek symbol used for the test) was calculated to stand at 181.50. This value is way above the critical values of 3.84 and 6.63 for the 0.05 and 0.01 levels of significance, respectively. Hence, there is substantial statistical evidence to reject the null hypothesis; and to conclude that there was a definite preference on the part of the participants for argument structure patterns over verb type in their sorting task.

Summary of findings
As the Chi square test in the preceding section revealed, the statistical evidence for rejecting the null hypothesis was substantial even at the 0.01 level of confidence. Thus, the findings of the study indicated that at least for these participants the argument structure pattern of the sentences was of a much higher importance than verb type in their sorting task. To take this argument to a higher level, one may conclude that the argument structure pattern of a structure plays a much higher significance in the interpretation of English structures by Iranian EFL learners of advanced standing than the type of verb used in the same structures. The above findings have colossal implications for material developers and the teachers of EFL classes in Iran.

Discussion
The participants of the study exhibited three different sorting strategies in their performance. A summary of these strategies is offered hereunder.

Verb -Centered Strategy
Out of a total of 108 participants who took part in the study, only a very small group numbering 19 in all sorted the structures based on the type of the verb encountered. In other words, in their decision to group structures having a common base, the criterion that was deemed most essential to them was that the structures were made of the same verb.

Construction-based performance of the first type
Of the remaining participants, 37 sorted the groups in such a fashion that each group contained four different argument structures patterns. Simply stated, to these participants the lexical aspect of the verb was of no significance; rather the type of role associated with the verb was the determining criterion in their sorting activities. Quite surprisingly, not only had these participants become aware of the significance of the argument structure patterns to their sorting activity, they also made sure that each sorted deck of cards contained the four different patterns.

Construction- based Performance of the second type
The remaining participants, a group of 52 individuals, embarked on a construction-based activity which was different from the previously mentioned group of participants. More specifically, the strategy used by these participants consisted of placing four constructions of exactly the same type within each sorted group. For example, the sorting activity of these participants was based on placing, say, 4 transitive structures in each group. Before discussing the behavior of each group, the following table is intended to clarify the sorting strategies exhibited by each group.

Sorting strategies displayed by each group

Key: V=Verb, A= Argument Structure Pattern

Discussion:
In this section the strategies sought in sorting the structures are discussed with the aim of explaining them.

Sorting by attending to verb type
As was mentioned previously, only a total of 19 participants sorted the structures by attending to verb type. In this type of processing the constituent elements of the structure like the verb ‘are’ of utmost importance to the processing of the whole structure, hence the label data-driven processing or bottom-up is applied to this type of processing (Chastain, 1988). In this type of processing the reader/listener attends to the individual words and structures in the text itself, using these to build up an interpretation of the whole text or structure. Thus, only approximately one out of five participants in this study resorted to a data-driven mode of processing which in effect indicates that to these participants the emphasis was on the language of the structures in their comprehensions of the sentences. In other words to these participants, meaning resided in the structures’ linguistic input and that no attempt was made on their part to derive the structures’ meaning by activating their prior or world knowledge in arriving at the meaning.

Sorting by attending to different argument structures
Given that each verb type was associated with four different argument structures that denoted different events and relationship in the external world, those participants that categorized according to different argument structures as their criterion were in effect categorizing based on their knowledge of the external world and the relationship thereof whereas the previous group of participants based their decision on purely linguistic criteria (i.e., verb type).

There were two different types of conceptually driven or top-down strategies detectible for the participants. However, before entering in to any discussion of the two types, a general definition of conceptually driven or top-down strategies is in order. Top-down processing reflects a direction of thought that begins its operation from a knowledge base to work on specific pieces of information (Chastain, 1988). In reading for example, the top-down reader not only possesses a general knowledge base about the world, but also some knowledge about reading in general (Chastain, 1988). Thus, equipped with these bases of knowledge, rather than relying merely on the linguistic input provided by the text, he/she makes sense of the text.

From this point onward, this discussion of the findings is devoted to the two types of top-down strategies displayed by the participants of this study, namely sorting by attending to argument structures a) as a group, and b) on an individual basis.

Sorting by attending to the argument structures as a group
As will be recalled, one of the ways of sorting the structures was based on grouping the four cards based on the properties of the four different argument structure types. The following scheme is helpful to understand the strategy undertaken by these participants:

Tables sorting strategy utilized by group 2

As the table shows, these participants sorted each set by placing the four types together with their different argument structure manifestations in each sorted deck of cards. Hence, to these participants the main criterion in sorting was the role relationships denoted by each verb. In other words, a top-down processing strategy was utilized in that the participant had to resort to their prior world knowledge about role relationships in order to process in this fashion.

Moreover, in assigning the different role relations to each verb, the whole variety of such relationships was considered as the sorting criterion (i.e., all possible argument structures related to a verb were considered). In this light, these participants, sorting strategy maybe referred to as purely top-down. Thus, to these participants, each verb had found its rightful place in the schema network of the participants. For this reason, one may assert, that this group of participants had succeeded in associating linguistic elements, in this case the verb, with their other bases of knowledge, and in so doing, had culminated their pragmatic competence.

Sorting by attending to different types of argument structures
The third group of participants sorted each category based on the particular type of argument structure which was deemed as the criterion in sorting.

Table sorting by attending to different types of Argument Structure.

In this particular type of sorting strategy, the participants’ main criterion in sorting was a top-down view of the argument structures associated with the verb. However, unlike the previous group, the sorting strategy was limited to associating a particular argument structure to each verb encountered rather, than going down the whole range of argument structures associated with each verb. Hence, to these participants each verb was only associated with one type of argument structure, and hence in their sorting of the 16 structures into 4 similar sets, it was this preconceived relationship between the type of the verb and the related argument structure which was the determining factor, for example the participant may have through top-down or conceptually driven strategies have come to associate the verb, throw to the argument structure, say, the transitive.

Participant may have come to associate the same verb to the other 3 remaining argument structures. Hence, in his/her sorting activity, the participant well aware of the connection between the different type of argument structures, or external world relationships, and one specific verb starts off on a top-down processing mode with inklings of bottom -up processing when he/she associates the same verb with 4 different relationships in the real world, or bottom-up processing.

Then, to conclude the participants of this study embarked up on three different processing strategies when faced with sorting strategies. In the first group, 19 of participants, the processing was purely bottom-up since the criterion for sorting was based on a purely linguistic notion, the verb, with no external characteristics of relationship attached to the constituent. The second group of participants, numbering 37, embarked on a purely top-down sorting strategy since the main criterion to them in classifying structures of the same type as the role relationships that they had conceptually come to associate with each verb.

As for the performance of the third group of participants, 52 in all, the primary emphasis was finding a balance between the different types of argument structures and their linguistic manifestation through the verb. This type of processing although takes a great amount of impetus from the top-down processing made in the initial stages, later display some signs of bottom-up processing strategy; hence one may refer to this type of strategy as semi-top-down strategy.

Conclusion
The results of this study were quite different from those obtained from native speakers in that the native speakers were most attentive to verb type whereas participants of this study were more affected by argument structure pattern of the structure.

One may argue that the reason for the participants’ preference for argument structure pattern of the structure over verb type may be that the students who took part in this study were all majoring in translation and teaching; and as such may through the linguistics courses they had previously taken become aware of the importance of argument structure patterns to sentence making. Nevertheless, the participants’ preference for argument structure patterns so overwhelmingly outperformed their preference for verb type that this researcher is led to believe that there might be other factors at play.

Hence, for future research studies this researcher proposes the same task to be administered to other target populations, say, freshmen (i.e. more target populations of lower English proficiency.) The proposed line of research is essential to establish whether the argument structure pattern is in fact such a great factor in EFL learners’ interpretation of sentences. Should the findings of future studies also confirm the emphasis placed on argument structure pattern at the expense of verb type; then, one may think about the pedagogical implications of this phenomenon.

REFERENCES

Bencini, G.M.L., & Goldberg, A.E (2000). The contribution of argument structure constructions to sentence meaning.Journal of Memory and Language, 43, 640-51.

Chastain, K. (1988). Developing second language skills.Orlando, Fl.: Harcourt Brace Jovanovich.

Chomsky, N. (1965). Aspects of the theory of syntax.Cambridge, MA: MIT Press.

Cook, V.J. (1991). Chomsky’s Universal Grammar. London: Blackwell.

Filmore, C., & Kay, P. (1999). Regularity and idiomaticity in grammatical constructions: The case of Let Alone. Language, 64, 501-538.

Goldberg, A. (1995). Constructions: A construction grammar approach to argument structure. Chicago, IL: Univ. of Chicago Press.

Goldberg, A. (1997). Relationships between verb and construction. In M. Verspoor & E. Sweetser (Eds.), Lexicon and grammar (pp. 383-398). Amsterdam: John Benjamins.

Healy, A., Miller G. (1970). The verb as the main determinant of the sentence meaning. Psychonomic Science, 20, p. 372.

Jackendoff, R. (1997). Twistin’ the night away. Language, 73, 534-559.

Karimi, S. (1997). From deep structure to minimalism.Majalle-ye Zabanshenasi, Vol. 11, No.2.

Levin, B., & Rappaport Hovav, M. (1995). Unaccusativity at the syntax-lexical semantics interface. Cambridge, MA: MIT Press.

Michaelis, L., & Lambrecht, K. (1996). Toward a construction-based theory of language function: The case of nominal extraposition. Language, 72, 215-247.

Partee, B. H. (1965). Subject and object in modern English.New York: Garland.

Pinker, S. (1989). Learnibility and cognition: The acquisition of argument structure. Cambridge, MA: MIT Press.

Rappaport H.M., Levin, B. (1998). Building verb meanings. In M. Butt & W. Geuder (Eds.), The projection of arguments: Lexical and compositional factors (pp. 97-134). Stanford, Ca. CSLI Publications.

Sepassi, F. (2002). Intelligence-related constrains on information processing to serve as a platform to enhance the description adequacy of the competition model. Unpublished PhD dissertation, Shiraz University, Iran.

Appendix A: