Learners ’ Incidental Vocabulary Acquisition : A Case on Narrative and Expository Texts

This study was intended to determine whether or not the genre of a reading text affects the incidental vocabulary acquisition of L2 learners while reading. To this aim, 40 Iranian EFL students whose vocabulary knowledge was within a limited range (already determined by Nation’s Vocabulary Levels Test) were divided into two groups of 20 each for the reading sections. The Narrative Group comprised the participants who read the narratives, and the Expository Group were those who read the expository texts. Three types of vocabulary tests (i.e., Form recognition, Meaning translation and Multiple-choice items) were administered after the reading sessions to assess the incidental vocabulary gains of the participants. Overall, this study demonstrated the relative superiority of expository texts over narratives in terms of enhancing readers' incidental acquisition of unknown words. It is argued that depending on the genre of a text, readers will invest processing resources with different depths and varying degrees of cognitive elaboration for the task of comprehension.


Introduction
It is generally believed that most vocabulary, in first, second or foreign languages, is acquired incidentally.That is, vocabulary acquisition occurs as a by-product of reading and listening activities not explicitly geared to vocabulary learning (Huckin and Coady, 1999, p. 183).In this respect, the role of extensive reading as a prolific source of incidental second language vocabulary acquisition becomes evident.Through extensive reading, according to Swanborn and de Glopper (2002, pp. 95-96), "New word meanings are derived and learned even though the readers' purpose for reading is not the learning of the new vocabulary."Of course, the process of acquiring the meaning of unknown words as a by-product of extensive reading depends on many factors that have been the focus of attention of many researchers.Central to this are factors such as the resources and procedures applied in meaning inference (Nassaji, 2003), the influence of reading task and learner factors (Hulstijn and Laufer, 2001), the effect of dictionary or glossing support (Knight, 1994), or the types of texts (Gardner, 2004).
Along with the previous studies, this paper is intended to investigate, from a text analysis point of view, the effect that the two most common text genres to EFL learners, namely narrative and expository, may have on their potential incidental vocabulary acquisition.In other words, the question that this paper commits itself to address is which one of these two text types is more conducive to incidental learning.Rather than relying on intuition, teachers and L2 reading material designers can shape their instructions and materials based on the insights they receive from the findings of studies of this type.

Review of Literature
The literature on incidental vocabulary acquisition has witnessed a great diversity of, and sometimes controversial, views held by various researchers.These controversies principally arise from the complex nature of incidental learning and its application in L2 learning enterprise generally and incidental vocabulary acquisition in particular.Therefore, it seems in order to have an in-depth investigation of these concepts, followed by some discussion on the empirical research conducted in the field.

Types of Vocabulary Learning
Incidental and intentional learning have been naively taken to be used interchangeably and have become indistinguishable from implicit and explicit learning, respectively (For a recent study see Hunt and Beglar, 2005).Here, we firstly distinguish incidental from intentional learning and then we deal with incidental vs. implicit and intentional vs. explicit distinctions which are more complicated and subject to many misunderstandings.
The distinction between incidental and intentional learning, according to Ellis (1999), is based on the distinction between focal and peripheral attention.To him, "intentional learning requires focal attention to be placed deliberately on the linguistic code (i.e., on form or form-meaning connections)," while "incidental learning requires attention to be placed on meaning (i.e., message content) but allows peripheral attention to be directed at form " (pp. 45-46).Therefore, any learning, whether intentional or incidental, can only take place with some degree of attention (Schmidt, 1994, p. 198).By the same token, Hulstijn (2003, p. 357) claims that intentional or incidental learning requires some attention and noticing.However, attention is deliberately directed at committing new information to memory in the case of the former whereas the involvement of attention is not deliberately geared to an articulated learning goal in the case of the latter.Gass (1999) introduces factors that are involved in learning vocabulary in a schematic representation that also captures the difference between incidental and intentional learning.The continuum that she portrays, as shown in Figure 1 below, suggests that words are more likely to be learned incidentally if (a) there are recognized cognates between the native and the target languages, (b) there is significant L2 exposure, or (c) other L2 related words are known.
Insert Figure 1 Here.Reider (2003), relying on Schmidt's (1990) and Ellis's (1994b) definitions of the concept of 'consciousness', attributes all these confusions regarding the difference between implicit and incidental learning to the inconsistent use and unclear status of this term noted by various researchers in the literature.Based on the interpretation of the term consciousness, Reider is said to argue the types of incidental learning that can take place.That is, if we equate consciousness with intentionality, then the absence or presence of consciousness will lead to incidental and intentional learning.In a similar vein, if we consider consciousness as awareness, then we will have explicit learning in the presence of consciousness and implicit learning in its absence.
According to these definitions, the term implicit will be equated with 'non-consciousness' in the sense of 'unawareness', while incidental will be equated with 'un-intentional' (without any restrictions as to the role of awareness).Reider (2003, p. 28) finally concludes that "incidental learning as being composed of implicit learning processes (which happen without the learner's awareness) and/or of explicit learning processes (which take place without learning intention but nevertheless involve online awareness and hypothesis formation)."(See the diagram below for more clarification).
Insert Figure 2 Here.The definition presented above was, however, not satisfying because the concepts under discussion were clearly distinct from what is involved in a genuine implicit learning.DeKeyser (2003, p. 314) argues that automaticity is the result of a learning process, not a characteristic of the learning process itself.By replacing intentionality or automaticity with the construct of awareness, DeKeyser defines implicit learning as learning without awareness of what is being learned.DeKeyser, nevertheless, warns us to distinguish implicit learning from two concepts that are often confused with it, namely inductive learning and implicit memory.The former, defined as moving from particular to general, can itself be implicit or explicit.Implicit memory and implicit learning are two independent concepts, and although implicitly acquired knowledge tends to remain implicit and explicitly acquired has the tendency to be explicit, the latter can become implicit in the sense that learners can lose awareness of its structure over time.

Narrative vs. Expository
There is a long tradition of research into the differences between expository and narrative texts.Generally, Zwaan (1994, cited in Yun Dai andWang, 2007, p. 335) suggested that the reader mentally represent and process texts differentially, depending on the genre of the text involved and their related expectations and schemas.For example, expository passages are viewed as more difficult to process than narrative passages (Zabrucky and Moore, 1999).Furthermore, expository texts are less cohesively organized by temporal and causal connections, thus demanding more explicit logical inference.Narrative texts, on the other hand, are typically more ambiguous and open to different interpretations than expository texts thus inviting personal participation and meaning interpretations.
Overall, empirical evidence indicates that for most students, expository reading poses a greater challenge than does narrative reading (Taylor and Beach, 1984).Although many factors may contribute to the difficulty students experience with expository reading, the four most commonly cited are text structure, conceptual density and familiarity, vocabulary knowledge, and prior knowledge.

The Effect of Text Types on Incidental Vocabulary Acquisition
It seems that certain kinds of texts facilitate incidental vocabulary acquisition.According to Huckin and Coady (1999, p. 188), texts which are personally interesting to the learners are more conducive to incidental vocabulary acquisition.
In a study examining the effect of topic familiarity on text comprehension and second language incidental vocabulary acquisition, Pulido (2007, p. 189) reached the conclusion that although topic familiarity does affect text processing, comprehension and lexical gain, it does not affect the role that text comprehension plays in incidental vocabulary acquisition.
Textual elaboration, by contrast, appears to have an unclear effect on incidental vocabulary learning.Chung (quoted in Urano, 2000, p. 8), in this regard prepared five different versions of a reading passage: unmodified baseline, simplified, lexically elaborated, structurally elaborated, and lexically and structurally elaborated.Three vocabulary tests were applied involving form, meaning, and delayed meaning recognition.The results showed no significant effects of textual elaboration on reading comprehension or incidental vocabulary learning.
In a similar study on Japanese learners of English, Urano (2000) investigated the effects of two different types of input modification, simplification and elaboration on second language comprehension and incidental vocabulary acquisition.Contrary to Chun's study, Urano (2000, p. v) stated that lexical elaboration triggers incidental vocabulary acquisition while simplification does not.
However, it can be readily observed that none of these studies has touched upon the issue of genre or text type in its fullest sense.That is, they were either dealing with a change in the grammatical and lexical structure (Chung, 1995;Urano, 2000) or the topic (Pulido, 2007).
Recently, by explicitly stating the concept of genre, Gardner (2004) analyzed the lexical differences between narrative and expository reading materials used in upper elementary education and explored how these differences could affect children's potential acquisition through reading.He came to the conclusion that children's narratives tend to utilize a greater proportion of General High Frequency words than their expository texts.Gardner (2004, p. 24) argues that this is because narratives place fewer lexical demands on children in general.
In addition to the factors mentioned above, the importance of depth of processing which can be quite operative in terms of L2 vocabulary learning tasks has been emphasized by Laufer & Hulstijn (2001, pp. 543-544) with the notion of involvement, consisting of (i) a motivational component, comprising the need to determine a new word's meaning, and (ii) a cognitive component, comprising search (e.g., dictionary look up) and evaluation (e.g., evaluating whether the information obtained from the dictionary applies to the verbal and non-verbal context).To this end, as the methodology section below reveals, some deeper steps into data collection were taken.First, we did a text modification which had hardly been done in the previous studies.Second, some fitting balance between familiar and unfamiliar words was made in order to adjust successful guessing.Third, some crucial and frequent words to understanding the texts were replaced by some other less ordinary ones to test the effect of comprehension under the influence of this strategy.
Since studies concerning the topic, cited above, have paid little attention to the crucial effect of genre on incidental learning of new lexical items, examining the effect of the commonest text genres found in EFL students (i.e, narrative and expository) seems to be a much needed line of inquiry and an area which has not yet been thoroughly elucidated.Therefore, this paper is intended to shed some light on this subject.

Participants
The initial participants of the study were male and female Iranian EFL students (freshmen and sophomores) enrolled at Shahid Chamran University of Ahwaz.To assess the vocabulary level of the students and make sure of the homogeneity, a modified version of Nation's (1990Nation's ( , 2001) ) Vocabulary Levels Test (Schmitt et al., 2001) was administered.Forty students whose vocabulary knowledge was within a limited range were chosen for attending the reading sessions.These students were later divided into two groups (n = 20) in the reading sections.The participants who read the narratives were referred to as the Narrative group, and those reading the expository texts are known as the Expository group.

Pre-testing
The Vocabulary Levels Test (Schmitt et al., 2001) was used to assess the subjects' vocabulary knowledge in order to gain a homogenous group in terms of their vocabulary knowledge.Once the tests were completed and returned, the researchers corrected them by hand.Each section of the test, which corresponds to one of Nation's frequency levels, was marked out of 30 since there were 30 questions per section, and then converted to percentages.In this way, the researchers were able to ascertain the predicted vocabulary knowledge for each of Nation's five levels (except the 10000 word).
Having scored the Vocabulary Levels Tests (VLT), the participants were divided into two equivalent (in terms of their vocabulary level) groups, namely Narrative group (those reading the narrative text) and Expository group (those reading the expository text).

Text Preparation
The texts chosen for this study were two pairs (Note 1) of (narrative/expository) texts.Pair I was adopted from a Ph.D. thesis investigating processing and learning of expository and narrative texts (Eng, 2002).The expository/narrative texts of this pair have the same theme, the same target words and almost the same length.What makes them different is the very genre in which they are written.Although generating an expository version out of a narrative text or vice versa might make the discourse of the text unnatural, this approach was adopted, as it was in Eng (2002), in order to control factors such as topic and vocabulary items of the text that have already been proved to be influential in incidental vocabulary acquisition (Pulido, 2007, p. 166).However, to eliminate the effect of unnatural discourse, another pair was also considered for this study.The texts of Pair II are three narrative and three expository paragraphs each on a different subject and taken from two different books.The texts of both of these pairs were later modified to control many factors that could influence vocabulary learning.
In order for learning to take place, there should be a fitting balance between known and unknown words.If the text is too difficult, successful guessing will be hard to achieve.To this aim, the words of the texts were counted.Then, 5% of the words of the text were considered as the target words.These words were selected according to two factors, namely how crucial these words were to the comprehension of the text and their frequency of occurrence.Not surprisingly, those words that were most crucial to the understanding of the text and also were repeated frequently in the text were selected.Then, these words were replaced by substitute words (see section 3.1.3below).However, we were not yet sure that the participants knew all the remaining (i.e., 95%) words.To come to terms with this, the texts intended for our final experiment were given to the participants two weeks prior to the experiment, and the participants were instructed to scan the texts quickly and circle any word that seemed alien to them.Then, the texts were once again modified in terms of replacing the non-target unknown words with their synonyms or, if replacing synonyms was not possible, they were deleted.

Pre-knowledge
One of the problems in designing vocabulary-learning experiments is controlling for pre-knowledge of the target words.When participants already have some L2 knowledge, it is hard to rule out the possibility of having (partial) knowledge of the target words that are used in the experiment.To solve this problem, it was decided that the spelling of the target words should be changed.These words which are called substitute words in Waring and Takaki's term (2003, p. 136) should not be confused with nonsense words as they are sometimes referred to in the literature.As in Waring and Takaki (2003), the substitute words were also checked by five learners, who were not part of the experiment, to ensure that that they could pronounce them fairly well so that it would not slow their reading.Implausible words, and words difficult to pronounce, were discarded.

Post-testing
In this study, our focus was on the very early stages of learning new words in an L2 since the students cannot be expected to go beyond this level after reading the texts for the first time.To this end, three tests were adopted from Waring and Takaki (2003).The tests include: 1) a Form recognition test; 2) a Multiple-choice (prompted recognition) test; and, 3) a Meaning translation (unprompted recognition) test.The three tests were extensively piloted with a group of eight subjects of similar ability and background.These subjects were not part of the main study.The aim of the piloting was to confirm that the tests contained enough words and the text was not too long and could be read in about one hour at a reasonable reading speed.(For more on these tests see Waring and Takaki, 2003).

Vocabulary Levels Test
In order to determine the effect of text types on EFL learners' incidental learning of new words, the first step was to obtain a homogenous group of subjects in terms of their vocabulary knowledge.This was done by administering a modified version of Nation's (1990) Vocabulary Levels Test (VLT).Of course, VLT has been subjected to many modifications.The reliability of VLT has been increased by carefully discarding items with lower inter-item correlations and creating longer 27-item forms (Beglar and Hunt, 1999, p. 135).In terms of validity, the test has been recently validated by Schmitt et al. (2001).

VocabProfile
In order to make sure that the vocabulary level of the reading texts corresponds to the participants' vocabulary level (already determined by VLT), VocabProfile (VP) was used.VocabProfile is a computer program that performs lexical text analysis.It takes any text and divides its words into four categories by frequency: (1) the most frequent 1000 words of English, (2) the second most frequent thousand words of English, i.e. 1001 to 2000, (3) the academic words of English (the AWL, 550 words that are frequent in academic texts across subjects), and (4) the remainder which is not found on the other lists.

Data Analysis
Scores from the post-reading tests were subjected to a t-test to determine whether or not there are significant differences between each group's performances in terms of the specific genre of the text they have read.In each case, the independent variable was the genre of the text (i.e., narrative or expository) while the dependant variable was the students' incidental acquisition of the meaning of the target words.

The Overall Results
The following tables demonstrate the overall achievements of the participants from the two reading sessions.As shown, except for the Form recognition test, the scores obtained from the other two tests (i.e., Meaning translation and Multiple-choice) indicate a significant difference between the lexical gains of narrative readers and those of expository ones: participants of Expository Group had a significantly higher performance than those in the Narrative Group on the Meaning translation and Multiple-choice tests.
Insert Tables 1 and 2 here!

Mental Models of Text Comprehension
Since any incidental vocabulary acquisition from reading is intricately interwoven with and related to text comprehension, as this has been repeatedly pointed out by Pulido (2007), it seems in order to start the discussion from an understanding of overall text comprehension as well as the mechanisms involved in this process and then use this understanding as a basis in order to discuss the main dimensions of incidental vocabulary acquisition from a genre analysis point of view.
When readers set out to comprehend a text, they construct a coherent mental representation of the events, actions and states present in the written text.In cognitive psychology, these mental representations are known as mental models (Kintsch, 1974) or situation models (van Dijk and Kintsch, 1983).Successful text comprehension has been equated with the construction of a coherent situation model.Readers construct and update their mental model of the textual meaning by means of interacting bottom-up and top-down processes throughout the reading process.In this process, the existing model serves as a basis for the interpretation of newly read information and is in turn continually tested and updated by this new information.In fact, readers try to alleviate any probable discontinuities in their mental model in order to maintain textual coherence.

Text Processing in Mental Models
One of the significant implications of situation models is when L2 learners encounter an unknown word in a text.Their mental model of the textual meaning will exhibit a discontinuity with regard to this unknown word.If this discontinuity becomes crucial for the text comprehension, the reader will normally stop and attempt to infer the meaning of that unknown word and bridge the gap in his/her mental representation of the text.However, sometimes this discontinuity becomes marginal to the readers and learners will continue constructing their mental representation of the text with the aid of factors (other than vocabulary) that contribute to the construction of the mental model.Various factors have been claimed to influence the construction of a coherent situation model among which are reader-related factors such as reading skill, fluency in the language of the text, motivation, and goals, as well as text-related factors such as text structure, vocabulary, and genre (Zwaan andBrown, 1996 cited in Roloff, 1999, pp. 14-15), and these factors may interact in such ways that it is very difficult to account for the appropriateness of this representation without considering at least some aspects of these reader and text-based factors.

The Effect of Genre on the Construction of a Mental Model
According to previous studies conducted on text types, systematic differences in how people respond to different types of texts, particularly narrative and expository texts, have been observed (e.g., Einstein et al., 1990;Zwaan, 1994).Therefore, it is sensible to claim that a reader attempts for constructing a mental model of the text and finally comprehending it is significantly dependent on the genre of that text.In fact, the genre of a text causes readers to allocate their processing resources in specific ways that meet the constraints of that given genre.
Needless to say, most work in this respect has been done on stories, which are the typical manifestations of situation models (i.e., experiences), accounts in conversations or news reports.Many other genres such as expository text may be representations of general knowledge, and not personal experiences, hence are not based on mental models of events referred to.However, according to van Dijk ( 2006), like all modes of discourse, the expository talk and text also manifest another kind of mental model: context models-subjective representations of the communicative situationwhich control how we formulate discourse so that it is adapted to the communicative situation.Therefore, with the adoption of mental model in its new van Dijkian sense, we speak about the construction of mental models of expository texts although mental model in this very sense is applicable to all discourses as well.

Incidental Vocabulary Acquisition While Constructing a Mental Model of a Text
In addition to the claim that genre of a text has a great contribution to the way readers decide to process a text and finally form a coherent mental model of it, it is justifiable to state that if readers face a discontinuity in reading due to the presence of an unknown word in a text, the genre of that text has a great contribution to their decision for discovering the meaning of that unknown word.In fact, in texts which display an easy-to-construct state of affairs due to their very genre, readers are less likely to attempt discovering the meaning of every single unknown word because they rely on inferencing higher level text organization such as propositions rather than individual words.However, readers are likely to take refuge in finding the meaning of the unknown words in texts from which a mental model is difficult to construct and their top priority regarding text comprehension becomes 'trying to know', or at least 'guess', the meaning of individual words rather than inferencing propositions.Interestingly, this is in line with the literature on expository versus narrative distinction where there is a consensus that readers spontaneously generate causal bridging inferences when reading narrative text (e.g., Klin, 1995;Suh and Trabasso, 1993; both cited in Wiley and Myers, 2003, p. 110).However, there is considerably less evidence that similar inferences are necessarily drawn from expository texts (ibid).This can also be attributed to the way readers of the present study processed the texts.Having been exposed to the macro-structurally rich narrative texts, readers of Narrative Group (i.e., those reading the narratives) were more disposed of inference making as a central tool for text comprehension; hence instead of focusing on individual words they emphasized attaining the global thematic information through inferential strategies.Thus, their scores on the vocabulary tests were relatively lower than those of the readers of Expository Group.The readers of this Group, however, compensated for the lack of a unified global structure of their texts by trying to get the meaning of the individual unknown words in order to construct a coherent representation of the texts (see Table 1 above).

Text Type and Incidental Vocabulary Acquisition: a Textual Perspective
From the perspective of incidental vocabulary acquisition, an important aspect to bear in mind here is that when trying to overcome a discontinuity in reading the text, readers of expository texts will usually invest a great deal of their processing resources on the word level, and less on higher order text integration (Zubrucky and Moore, 1999, cited in Eng, 2002, p. 17).They will continue doing this until sufficient coherence is ensured.This is because expository or informational texts generally contain more unfamiliar vocabulary and concepts and fewer ideas related to personal experience as well as a variety of structures which are usually less cohesively organized by temporal and causal connections (Dai and Wang, 2007, p. 336).Zubrucky and Moore (cited in Eng, 2002, p. 17), specifically maintain that since expository texts tend to be less familiar and less predictable to readers, relatively more attention is needed for lower level processes, leaving fewer resources available for higher level text organization.However, narrative texts typically involve agents, actions, and event sequences occurring through a timeline, and are linked together by causes, reasons, motives, goals and plans and have a more prototypic organization that is well learnt through frequent exposure since childhood.Therefore, comprehending a narrative, unlike expository texts, will entail the investment of processing resources on higher level text organizations (i.e., propositions).In sum, narratives can be claimed to invite relational processing or processing directed toward understanding global and thematic information whereas expository texts invite individual item processing, directing readers' attention to the details of the passage.
The above discussions regarding the differences between narrative and expository texts are vividly reflected in the following excerpts taken from the texts used in our study.

Excerpt 1: Narrative
There is a talk around here lately about salmon territory and hydroelectric dams.I remember I used to stop the car by the Pacific Northwest's Monroe River and watch the salmon journey upstream during spawning season.In those days, thousands of spawning Coho and Sockeye would swim up the Monroe River.People wondered for centuries how salmon could find their way home up a web of rivers to an often tiny stream bed.We now know salmon can detect odors; each stream is said to contain a particular "bouquet" of smells that mark themselves on the salmon before they leave for the ocean, which helps them find their way back.I am no biologist, but this theory makes sense to me.The aroma of pine can elicit many memories of my family home and the brief happiness of Christmas time.

Excerpt 2: Expository
A large number of hydroelectric dams have been built in the Pacific Northwest over the past 70 years to generate electricity.These power resources are operated by Pacific Power, whose profits produce income taxes that flow to the government to pay for public services in the Pacific Northwest.If power production declines, or the cost of generating power increases, then the government would receive less money from the electricity sales and therefore have less money to pay for public services.The Monroe River is representative of many river systems that produce power and salmon in the Pacific Northwest.The river and its adjoining streams provide spawning and living areas for Coho and Sockeye salmon.
As seen in these two excerpts, the narrative text involves agents (e.g., 'I', 'people', 'Coho' and 'Sockeye') actions (e.g., 'stop the car' and 'swim up'), and event sequences (e.g., 'used to' and 'in those days') occurring through a timeline, and are linked together by causes, reasons, motives, goals and plans (e.g., 'people… wondered how salmon could find their way' or 'this makes sense to me').These characteristics, as mentioned earlier, are known to be shared virtually in any narrative.Therefore, the saliency of such prototypic characteristics in narratives directs the attention of the readers to propositions and seeks thematic information by spontaneously generating causal bridging inferences for text comprehension.However, since there is no such universally unique textual characteristics to be shared in all expository texts, readers of these texts attempt to rely on seeking the meaning of individual unknown words than to risk processing higher level textual organizations that might not be as vividly discernable as those of narratives.
Nevertheless, according to what has been discussed so far, one counterargument might consider the nature of each test to be adversely affecting the very way readers process a text.That is, one might ask if the test items of narrative texts were designed on micro-structural units (i.e., lower level textual organization) of the text and those of the expository ones were on the macro-structures (i.e., higher level textual organization), would the readers still process the texts in the way mentioned above?Such an argument is, of course, valid as far as we assume that the tests used for the two genres are intended to tap different recall of the texts and that the subjects are aware of these tests.However, the three types of tests used for each Pair tap exactly the same knowledge of the target words in both narrative and expository texts.That is, the focus of all the three types of tests in both genres was on a lexical level (spelling of words in the Form recognition test, their meaning in form of a definition in Multiple choice test, and their L1 equivalent(s) in the Meaning translation), and the test items were all selected regardless of their role in the construction of the micro or macro-structures of the text but with regard to their frequency of occurrence and the contribution they had to the overall understanding of the text.As with the participants, they were not informed in advance that there were going to be tests of vocabulary after reading the texts; otherwise, there was no incidental learning happening at all.

Evidence from the Present Study
The results of the present study lend support to the aforementioned discussions in that Meaning translation and Multiple-choice meaning recognition test both indicate that it was the expository texts which significantly invited the participants to invest more processing resources on word-level textual organization and consequently achieved higher scores on these tests.The relatively poor lexical recall of subjects who read the narrative texts, however, can be attributed to the almost full and consistent availability of all the macro level information they needed to form an adequate representation of the text.The inconsistent results of the Form recognition test, on the other hand, may signal the equal processing investment on the superficial text level regardless of the genre of the text.That is, the quality of vocabulary knowledge this test was intended to measure seems to be independent of the genre of the text.Or, we may consider these results to be the consequence of the methodological approach adopted in this study for the elimination of the effect of pre-knowledge of the words (i.e., our use of identical substitute words instead of the target words).Since we used the same word forms in both genres, it could have been intuitively expected that the results of the Form recognition test would be similar.However, we are still reluctant to draw any hasty conclusion based on the inconclusive results of the Form recognition test.

Conclusion
This paper was intended to investigate, from a text analysis point of view, the effect that the two most common text genres readily available to EFL learners, namely narrative and expository, may have on the learners' potential incidental vocabulary acquisition.In other words, the question that this paper committed itself to address was which one of these two text genres is more conducive to incidental learning.The results of this study were clearly suggestive of the paramount and fundamental role of genre of the texts in incidental vocabulary acquisition processes.
Overall, this study demonstrated the relative superiority of expository texts over narratives in terms of enhancing readers' incidental acquisition of unknown words and thus pointed to the need for repeated encounters with expository texts if teachers or material designers want to improve the L2 learners' incidental learning of new words.
What this study aimed to discover was the effect of genre on the very early stages of learning new words in an L2, and we did not attempt to study the multiplicity of other aspects or levels of L2 word knowledge (e.g., the noticing of collocations, colligations or patterns within text).The adoption of such an approach is both pedagogically and theoretically justified.In fact, when talking about incidental vocabulary acquisition, we should not expect the learners to go to levels beyond recognizing the form of those words or providing an L1 translation for it, and this seems to justify the suggestion of many researchers that incidental learning should be followed up by intentional learning.Note.This was because there were 6 target words in this Pair.

Table 1 .
Achievements of participants from reading Pair I (scores of each test were out of 5) Note.This was because there were 5 target words in this Pair.

Table 2 .
Achievements of participants from reading Pair II (scores of each test were out of 6)

Narrative Group Number of subjects = 20 Test type Mean Std. Deviation Difference between Means
* ( =.05)