Bowen Wang recently graduated from the MSc Applied Linguistics program at University of Oxford. His research interest lies in L2 reading and vocabulary acquisition. He currently focuses on how to facilitate word learning through reading and the development of lexical inference ability. For his master dissertation, he devised a novel text modification method and conducted case studies on the efficacy of this method in incidental vocabulary acquisition through reading. His interest in SLA research originated from his undergraduate study at Sun Yat-sen University (mainland China) when he majored in English language and literature. His undergraduate thesis investigated EFL learners’ reading strategy use and what it revealed about the nature of their vocabulary knowledge. He also conducted an independent research program on refining note-taking format by drawing on functional linguistics theory in his sophomore year, which was granted with a provincial level funding from the university.

Video Abstract

Written Abstract

This study investigates a novel text modification method called reflash. Reflash is designed to assist inferring word meaning from context, which is unique since the most widely researched text modification method, namely gloss, gives learners instant access to the word meaning. Reflash is a set of left and right arrows, attached to the target words in a text. By clicking the arrows, the learners can review the previous occurrences or preview the subsequent occurrences of the word. This study intends to answer whether learners use reflash, how they use it, and the efficacy of reflash regarding incidental vocabulary acquisition through reading.  Three low-proficiency junior-high-school learners of English as a foreign language (EFL) in China were recruited. They were asked to read texts with target words modified by reflash. Stimulated recall was used to examine the cognitive processes of their readings. Their vocabulary gain was measured through the Vocabulary Knowledge Scale (VKS). The results revealed that all the participants used reflash at least in some occasions, and they used it in different ways and for different purposes. Various factors were found to contribute to these differences. All the participants made progress on some words modified by reflash. A qualitative analysis of each case of word knowledge growth indicated that various factors affected the efficacy of reflash.


The present study was inspired by my use of the Navigation (Ctrl+F) function of Microsoft Word. By typing a word in the search bar, the user is presented with a list of excerpts where the target word occurs in the text. By clicking on each entry in the list, the user is navigated to the corresponding part of the text. I realized that it might be a good idea for language learners to have a navigation facility for every instance of the words they would like to learn when they are reading a text, because presumably, learners could have an easier time inferring the word meaning by synthesizing contextual clues from multiple occurrences than relying on any single occurrence. Based on this idea, I devised a new method of text modification. In the modified text, each occurrence of a target word is followed by a left arrow ← and a right arrow →. By clicking on the left arrow ←, the reader returns to the previous occurrence of the target word ‘in a flash’ (i.e., the hyperlink in the arrow takes the reader to the previous occurrence instantly); by clicking on the right arrow →, the reader returns to where she was reading in another ‘flash’. Hence, I named this new form of text modification ‘reflash’.

Reflash is unique since it is designed to assist learners trying to infer the meaning of an unknown word from the context. A large quantity of research on text modification has investigated in-text gloss and found that it facilitates incidental vocabulary acquisition (see section 2.1 for definition and discussion of this phenomenon) to some extent. However, gloss was designed to give learners instant access to the word meaning. Few text modification methods were designed to aid learners in inferring the meaning of unknown words. One of the few attempts to facilitate inferencing skills was made by Cobb (2006), who developed a tool called Hypertext Builder. In a “hypertext”, the reader is presented with a concordance list of every occurrence of a word in the text, simply by double-clicking the word, which is what Cobb called “story concordance” (2006, p. 647).

Cobb (2006) mentioned two advantages of providing the concordance list to the reader: 1. It tells the reader how many times the word will occur in the text and whether it is worth attending to, and 2. It highlights the collocational patterns of the word (i.e., the pattern in which other word/words co-occur with the word). However, I would argue that a drawback of using concordance lists is that readers can only see a part of the sentences where the word occurred. It is likely that a glimpse of the discrete sentences cannot inform readers about the larger contexts. This is where reflash could come into play. It is designed to help learners infer the word meaning by synthesizing the contextual clues provided by each occurrence, not only from the sentences but also from the larger contexts (e.g., paragraph, page). Hence, reflash, as a novel text modification method with its own hypothetical advantages, is worthy of study.

Literature Review

Incidental vocabulary acquisition through reading

This study investigates incidental vocabulary acquisition through reading modified texts. Incidental vocabulary acquisition is operationalized as vocabulary learned while the main purpose of the activity is not learning vocabulary but understanding a text. Note that incidental vocabulary learning is different from the implicit learning of vocabulary, which means that learners pick up words subconsciously (Ellis, 1994). Laufer and Hill (2000) criticized the confusion between incidental learning and implicit learning and queried whether any word can be learned without being ‘noticed’, which means a conscious awareness of the target language form. Contrary to incidental learning is intentional vocabulary learning such as memorizing word lists.

L1 (first language) studies provided robust evidence that most L1 words are acquired incidentally through reading and listening (e.g., Nagy, Herman & Anderson, 1985; Herman, Anderson, Pearson, & Nagy, 1987; Sternberg, 1987). Although extensive L2 (second language) studies show that intentional learning leads to better word retention than incidental learning (e.g., Hulstijn, 1992; Laufer, 2005; Schmitt, 2008), incidental vocabulary acquisition through reading is possible (Krashen, 1989) and has its role in L2 acquisition. This is because learners cannot learn every aspect of every word intentionally considering the sheer amount of words (Nagy, 1997) and “learners will also need the incidental learning which accrues from extensive input” (Pellicer-Sanchez & Schmitt, 2010, p. 32). Therefore, incidental vocabulary acquisition through reading is still a topic worth investigating.

The Condition of developing word knowledge through reading

To develop vocabulary knowledge through reading, a few conditions should be met. The first condition is noticing. According to Schmidt (1990, 2001), conscious awareness (noticing) of the target language form is necessary or at least facilitative for the input to become intake (the initial representation of knowledge in long-term memory). Nation (2001) argued that noticing requires decontextualization. For example, if a learner notices a word in reading, she identifies it as a new word, or recalls having seen it before and recognizes the discrepancy between the current occurrence and his/her previous knowledge of the word (Nation, 2001).

However, noticing alone is not sufficient; psycholinguistic research indicates that the depth and type of processing (Craik & Lockhart, 1972; Craik & Tulving, 1975) is important as well. For incidental vocabulary learning through reading to happen, learners need to take “additional elaboration strategies” to create “a memory trace”, such as inferring the word meaning from the context and relating the unknown words to known words (Laufer & Hill, 2000, p. 59).

To retain the vocabulary knowledge gained from noticing and deep processing, repetition is necessary. According to Baddeley (1990), learners retrieve what they have learned about the word every time they re-encounter it, which consolidates their memory and accelerates future retrieval.

In this dissertation, when discussing the factors involved in learners’ use of text modification and the efficacy thereof in incidental vocabulary acquisition, I categorized these factors regarding how each factor is related to noticing, processing, and repetition.

Research Questions

Considering the hypothetical value of reflash and incidental vocabulary acquisition, this study explores the use of reflash, as a text modification tool, among low-proficiency Chinese EFL learners and its efficacy in facilitating incidental vocabulary acquisition through reading. Because reflash is a novel design that has not been examined by any previous research, the primary goal of this study is to present a primitive account of whether learners use reflash, how they use it, and the factors underlying their different usage behaviors. Another goal is to examine the efficacy of reflash in incidental vocabulary acquisition.

The specific research questions are as follows:

1. While reading an English text modified by the novel text modification method reflash, do the adolescent, low-proficiency Chinese EFL learners use reflash when they encounter the novel lexical items?

1a. What factors influence the usage or non-usage of reflash among learners?

2. When learners use reflash, what are they doing after each click? What are they thinking when doing so?

3. What factors influence the acquisition of words modified by reflash (when learners use reflash)?



I sent a poster of this study to my parents and asked them to help me spread the word. Soon the parents of three children volunteered to let them participate in this study. They were two junior-high (year three) student (coded as B and T), and one senior-high (year one) student (coded as X) from Bozhou city in China.

Table 4.1 Background information of participants


Participant Year of School Age Estimated average English test score at school
B Junior high – grade 3 15 140+/150
T Junior high – grade 3 15 110-120/150
X Senior high – grade 1 16 110-120/150


The reading materials were chosen from the Oxford Bookworms Library (Henceforth OBL). The publisher of OBL devised a test called Oxford Bookworms Levels Tests (OBLTs) to help learners choose the Bookworms books that are suitable for their proficiency level. The tests were in the form of multiple-choice cloze where the examinee reads an article and choose an appropriate word for each blank.

Vocabulary Levels Test (henceforth VLT; Nation, 1983) was used in this study to measure the learners’ vocabulary size. It tests words on different frequency levels: 1k level (two versions), 2k level, 3k level, 5k level, university word list, 10k level (only the two 1k level and 2k level tests were used owing to the low proficiency of the participants). Except for the 1k level (with 39 items in each version), each level contained 18 items. For each item, a correct answer was scored as 1, while incorrect/no answer was scored as 0.

I used Vocabulary Knowledge Scale (henceforth VKS; Wesche & Paribakht, 1996) to check the participants’ prior-treatment knowledge of the target words as well as their progression during and after the treatment.  The original VKS asks learners to self-assess their knowledge of a word according to the following scale:

  • = I don’t remember having seen this word before.
  • = I have seen this word before, but I don’t know what it means.
  • = I have seen this word before, and I think it means ___
  • = I know this word. It means ___
  • = I can use this word in a sentence, e.g. ___

In this study, scale 5 (I can use this word in a sentence) was omitted because productive knowledge was not investigated in this study.


Figure 4.1 The procedure overview

As presented in Figure 4.1, the whole study comprised of five sessions. There were two pre-treatment testing sessions. In testing session 1, VLT and OBLTs were administered. Following is the performance of the participants in each test.

Table 4.2 Testing session 1 results

Participant VLT (1k level) VLT (2k level) OBLT (starter level) OBLT (level 1) OBLT (level 2)
B 90% 61% 29/30 29/30 28/30
T 70% 33% 27/30 25/30 18/23
X 85% 50% 28/30 27/30 23/30


The results of the VLT indicates that B knew the meaning of approximately 900 (1000*0.9) – 1220 (2000*0.61) words, T knew 666 (2000*0.33) – 700 (1000*0.7) words, while X knew 850 (1000*0.85) – 1000 (2000*0.5) words.

According to the official guideline, scoring 24-28 means “This level is just right for you” while scoring 0-23 means “This level is too difficult for you”. For the level 2 test, B scored 28/30, T scored 18/30 and Y scored 23/30. In other words, level 2 books were “just right” for B, almost “just right” for Y, and a bit “too difficult” for T. However, because the unknown words were modified by gloss or reflash in this study, the actual difficulty level of the modified texts were lower than the original texts. Therefore, level 2 books were chosen as the candidate reading material.

Since all the participants indicated an interest in mystery stories, I tentatively chose two level-2 books: Deadman’s Island and Dracula. Then, I entered the content of the two books together into the online Compleat VocabProfiler ( to generate a list of all the word types (except for the prepositions like “in” and connectors like “and”) used in the texts (Deadman’s Island: 703 word types; Dracula: 806 word types; combined: 1092 word types).

Then, I gave the target word list to the participants in testing session 2 and asked them to rate each word against the VKS. Words rated as VKS 1-3 were the final target words and were allocated to four conditions: reflash, gloss, gloss+reflash, nonmodification. The following table shows the text coverage for each participant after the text modification (the text in the two sessions combined, with gloss words, gloss+reflash words, and proper nouns regarded as known words).

Table 4.3 Text coverage after modification

Participant Text coverage
B 99%
T 98.5%
X 98.75%


 Since the threshold of text coverage for optimal vocabulary growth in extensive reading is at least 95%, preferably 98% (Nation, 2001), the table above indicates that the modified text was favorable for all the participants in the case of their vocabulary acquisition through reading.

In the following two reading sessions, the participants read the assigned modified text on a laptop with a running screen recording software. After reading the text in each session, they were tested on the target words via VKS. After that, they were shown the screen record for the stimulated recall on why they used or did not use reflash, why they used it in a specific way, and what they were thinking when using it. What they said were audio-recorded, transcribed and coded. A delayed surprise posttest was administered three weeks after reading session 2 using the same instruments.

Results and Discussion

Research question 1

While reading an English text modified by the novel text modification method reflash, do the adolescent, low-proficiency Chinese EFL learners use reflash when they encounter the novel lexical items?

To answer this question, I screen-recorded the reading processes of each participant and visualized the use of reflash in the following form:

Figure 5.1 Visualization of reflash use

This example means that the participant clicked the right arrow at the first occurrence of the word. When the arrow took him/her to the second occurrence, she continued clicking the arrow, arriving at the third occurrence. Then she used the left arrows to return to the first occurrence through the second occurrence. Each cycle of clicking forwards and backwards (regardless of the number of clicking) will be referred to as a reflash episode.  I found that all the participants in this study used reflash at least in some cases. Following is the number of reflash episodes for each participant in the first two sessions.

Table 5.1 Number of reflash episodes

Session 1 Session 2
B 0 3
T 16 12
X 0 3


Therefore, in answer to research question 1, the participants (three adolescent lowproficiency Chinese EFL learners) used reflash, but not always, when they encountered novel lexical items during reading. Some used it more often than the others and in a more consistent manner. This result is similar to learners’ dictionary lookup behavior, where some use the dictionary more frequent/selective than the others (Prichard, 2008).

Research question 1a

What factors influence the usage or non-usage of reflash among learners?

Whether learners look up a word first depends on whether they notice it. After noticing the unknown word, learners have to decide whether to skip it or process it, that is, using reflash. Therefore, in order to determine the factors that influence whether learners used reflash, I transcribed and coded the stimulated recall data into noticing-related and processing-related factors.

Noticing-related factors

Reading habit

Clicking arrows and jumping within the text is novel to everyone, so it is natural that one is initially reluctant to incorporate it in their extant reading habit. Participant X did not use reflash in reading session 1 because she was “not used to it”. On the contrary, some learners might be more open to novelty and more willing to give it a try. Participant T was eager to try reflash the first time she was introduced to it. She found it “fairly interesting” to jump back and forth in the text.

Some learners were not resistant to novelty but were more accustomed to reading without interference. Participant B made no use of reflash in the first session. She indicated that “I’m more used to reading through the text as fast as possible without any halt”; but because she had a habit of working on unfamiliar words after reading, she would have used reflash if she had been given time after she finished reading. Therefore, I allowed each participant to reread the text for whatever purposes in session 2.

Studies on dictionary use also show that learners do not use a paper dictionary for every unknown word because the lookup takes time and disrupts the flow of reading (Hulstijn, 1993). It is unsurprising that learners do not always use reflash for unknown words owing to the same reason. However, no studies have explicitly examined reading habit as an independent variable, which awaits future investigation.

Tolerance for unknown words

If one is less tolerant of having an unknown word and feels an urge to figure it out as soon as possible, she is more likely to use reflash. Participant B did not use reflash at first because she was patient enough to postpone working on individual words to the end of the reading. When asked why she only used reflash after the reading to infer the words’ meaning rather than during the reading, she answered, “I feel okay not knowing these words during reading, but I don’t feel like pausing reading to work on these words.” On the contrary, participant T used reflash during reading, usually when the word made its first occurrence. She said, “I feel guilty if I leave these words unknown”. The extant literature on dictionary use have not yet covered this factor.

Processing-related factors

Richness of contextual clues

When the contextual clues were very clear such that the learner inferred the meaning of the word immediately, the use of reflash was unnecessary. For instance, when being asked why she did not use reflash for “howling”, participant B said, “I can tell from ‘hear’ and ‘wolf’ that it (howling) means the cry of the wolf”. This result corresponds with the dictionary use studies in that higher inferability of words leads to less lookup (Hulstijn, 1993).

However, less rich contextual clues did not necessarily lead to more reflash use. There were cases when the contextual clues were so deprived that the learner predicted that reflash use could not help. Participant B (who only used reflash after reading) did not use reflash for “mustache” because she knew during her first reading that “each occurrence is almost the same… I’m sure I can’t figure out what it means, so I didn’t use it (reflash)”. On the contrary, for the same word, participant T used reflash because she “really wanted to know what it means”. Note that unlike B who always used reflash after reading, T used reflash during reading and she could not predict the contextual richness beforehand. This example shows that learners’ clicking behavior could be affected by the interaction between the richness of the contextual clues and the learners’ reading habit.

Criticality of the learner

Participants B and X did not use reflash for words when they already made a reasonable guess during reading. In contrast, participant T was more cautious and insistent in figuring out the accurate word meaning as soon as possible. For example, she recognized the meaning of “pleased” immediately but still used reflash because she “wanted to make sure”.

Participant T was the only one who used reflash for reflash+gloss words as she said, “because I wanted to ascertain whether the word is polysemous”. In other words, she cannot tolerate ignoring the possibility that the word has multiple meanings. In contrast, participants B and X did not use reflash for gloss+reflash words because “the meaning is already given”.

Studies on dictionary use also reveal that critical learners look up the words even if they could guess the word meaning (Hulstijn, 1993). Although reflash cannot help learners confirm whether they are right, they can check whether they are wrong by examining whether their guess fits into each occurrence of the word.

To sum up, learners’ clicking behaviors of reflash involved various noticing-related and processing-related factors. Some processing-related factors overlap with dictionary use (i.e., inference ability, richness of contextual clues and criticality of the learner); two new noticing-related factors (i.e., reading habit, tolerance for unknown words) emerged from the data.

Research question 2

When learners use reflash, what are they doing after each click? What are they thinking when doing so?

As seen in 5.2.2, the learners used reflash to infer the word meaning or doublechecking their extant guess. After clicking the reflash arrow for a word, they were navigated to where the word previously or subsequently occurred. They read or simply had a glimpse of the text around the word. What they reread ranged from as short as one word before or after the target word, to as long as the whole paragraph (up to 100 words).

However, they reported that they did not have any rules in mind other than going with their intuition before deciding whether to reread a few words or the entire paragraph.

In the word of B, “It’s just the feeling; I feel right when reading this much.” Combining the data from the stimulated recall and the screen record, I found that they tended to read more (e.g., the whole sentence or paragraph) when trying to infer the word meaning; they tended to read less when double checking their extant inference. When asked how they inferred the word meaning, all participants gave the same answer – “feeling”.

After clicking the reflash arrow and reading the text where the word previously/subsequently occurred, the learners either kept clicking to view more occurrences of it or clicked back to the starting point. They usually stopped clicking when they had a reasonable guess of the word meaning or when they could not find helpful clues from the first few clicks. For instance, for the word “secretary”, T found few clues from the first four occurrences, so she assumed that “what comes later must not be helpful either” and decided to give up this word. For another word “photograph”, she ended her reflash use after four clicks because she succeeded in guessing that it was a synonym of “photo”. B used reflash for “poster” but her clicking did not cover all its occurrences because she found “the contexts were almost the same every time it occurs”.

In a nutshell, after clicking the reflash arrow, the learners read or glimpsed the text around the target word (length ranging from one word to a whole paragraph), trying to infer the word meaning or double-check their guess. The decision of how long to read and the way of inferring were both based on “feeling”. The number of the clicks was determined by whether they succeeded in inferring the word meaning.

Research question 3  

What factors influence the acquisition of words modified by reflash (when learners use reflash)?

I tabled the test results of each target word for each participant to see how they progressed throughout the study (See Appendix). Although the test results indicated that learners made gains on some target words after using reflash, not all reflash usages led to the acquisition of word meaning, and not all the progress could persist three weeks later. To investigate the factors that influence the efficacy of reflash, I coded the stimulated recall transcript into processing-related and repetition-related factors.

Processing-related factors

Richness of contextual clues

Usually, the contextual clues were not sufficient for the learners to arrive at the exact meaning of an unknown word, but they could provide enough information for the learner to infer the hypernym of the word. For instance, participant T was confident that “It (“secretary”) is a type of job in companies” the first time she met the word “secretary” in the text. She then used reflash to read each occurrence of the word and tried to think of as many company job positions as possible, checking whether any of them fit the context.

She made several guesses, including “assistant” and “accountant”, which was quite close to the correct meaning.  When the contextual clues were rich, and the word meaning was transparent, learners could infer the accurate or at least a reasonable guess of the word meaning. However, in this case, reflash served more as a tool for double-checking their inference.

For instance, “dressing-gown” occurred only twice but participant B had her largest gain (from 1 to 4 on VKS) on this word. She reported that “I felt that it means pajamas when I first saw it” and the use of reflash was because she “wanted to make sure”.

On the contrary, for words that were deprived of useful contextual clues in each occurrence, even the use of reflash transpired to be unhelpful. For instance, the majority of the occurrences of “castle” was in the form of “Castle Dracula” (e.g., “And somewhere high in the mountains was the Count’s home, Castle Dracula). Participant B reported that she “can only tell that it is a sort of building” even after viewing each occurrence via reflash.

Learning burden

The stimulated recall reveals that words of the same vocabulary knowledge scale did not necessarily bear the same amount of learning burden. One VKS-3 word could be of different learning burden from another VKS-3 word. A taxonomy of learning burden emerged during the process of transcript coding in this study:

  1. Words that are recognizable in a meaningful context but not in a list (R)
  2. Unknown words that share a part with or contains a familiar word (P)
  3. Words whose textual meaning is different from the meaning the learner knows (T)
  4. The learner knows the lemma but not the target form (L)
  5. unknown words that cannot be related to any prior knowledge (U)

It transpired that no type-T or type-L words were modified with reflash. Therefore, only type-R, type-P, and type-U words are discussed here.

When considering word type and VKS score simultaneously, I found that type-R words were rated as either 2 or 3 (VKS) in the pre-treatment test, type-P words were rated as either 1 or 2, while all type-U words were rated as 1. Therefore, the target words could be further divided into: Type-R + VKS-3 words; Type-R + VKS-2 words; Type-P + VKS-2 words; Type-P + VKS-1 words; Type-U + VKS-1 words.

It is reasonable to deduce that Type-R + VKS-3 words are of the lowest learning burden because the learner has already gained partial knowledge of the word and is right on the verge of acquiring the form-meaning link. For example, the word “pleased” was the only one Type-R + VKS-3 word for participant T. It occurred in session 2 but not in session 1, and it became a VKS-4 word in session 2. In the pre-treatment test, participant T confused “pleased” with “honored” because she had always read the two words concomitantly; but she immediately “remembered it meant happy” when she saw it in the text.

Type-R + VKS-2 words might be of slightly higher learning burden than Type-R + VKS-3 words. Sometimes the learner could recognize it in some contexts, and sometimes they could not. For example, participant T recognized the word “corner” in session 1.

However, when the word reoccurred in session 2, she failed to recognize it at first sight, but by returning to the session-1 occurrence via reflash, she finally recalled what it meant.

However, Type-P + VKS-2 words are not necessarily of lower learning burden than Type-P + VKS-1 words. The learning burden of a Type-P word is determined by the extent to which the familiar word part is related to the meaning of the target word. For example, for the Type-P + VKS-1 word “dressing-gown”, all participants worked out the word meaning immediately by combining the contextual clues and their previous knowledge of the word part “dress”. But for the Type-P + VKS-2 word “secretary”, participant T failed to infer the word meaning because the relationship between “secretary” and “secret” is not direct.

Generally, Type U + VKS-1 words are of the highest learning burden because there is no prior knowledge to which the learner can relate the word. For instance, participant B made progress from VKS-1 to VKS-3 on a Type-U + VKS-1 words that occurred 12 times in session 2 but the progress was not retained in the delayed test three weeks later. She could not even recognize seeing the word then. On the contrary, she acquired and retained the meaning of two Type-P words that only occurred 2 and 6 times respectively.

Hence, the learning burden (the learners’ knowledge of what is related to the word) and the learners’ knowledge of the word per se combinedly affected the difficulty of acquiring a certain word.

Repetition-related factors

Word frequency (in the text) and occurrence pattern

It is worth noticing that higher text frequency does not necessarily lead to more gains. The word “howling” occurred only twice but progressed from VKS-1 to VKS-4 for participant X. On the contrary, the word “castle” made 12 occurrences, but participant B could merely recognize the word form. This could be explained by the effect of spaced repetition (see The distance between the two occurrences of the word “howling” was neither too long nor too short (around 400 words) while the word “castle” occurred several times on each page of the first chapter of Dracula. The appropriately spaced repetition of “howling” made it more likely to be acquired than “castle”.


To sum up, this study found that low-proficiency Chinese EFL learners used the novel text modification, reflash, during or after reading the modified text. They used reflash in different ways for different purposes: 1. Using the contextual clues embedded in the word’s multiple occurrences to infer the word meaning; 2. Verifying their guess of the word meaning by checking whether it fits into the multiple occurrences of the word. To sum up, the factors affecting the efficacy of reflash include the richness of contextual clues, the learning burden, the frequency of the word in the text and the occurrence pattern. Richer contextual clues and lesser learning burden led to more efficient processing of the word, while spaced repetition contributed to better retention.


Baddeley, A. (1990). Human Memory. London, UK: Lawrence Erlbaum Associates.

Cobb, T. (2006). Internet and literacy in the developing world: Delivering the teacher with the text. Educational Technology Research and Development, 54(6), 627645.

Craik, F. I., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11(6), 671-684.

Craik, F. I., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of experimental Psychology: General, 104(3), 268-294.

Ellis, N. C. (1994). Vocabulary acquisition: The implicit ins and outs of explicit cognitive mediation. In N. Ellis (Ed.), Implicit and Explicit Learning of Languages (pp. 211-282). London, UK: Academic Press.

Herman, P. A., Anderson, R. C., Pearson, P. D., & Nagy, W. E. (1987). Incidental acquisition of word meaning from expositions with varied text features. Reading Research Quarterly, 22(3), 263-284.

Hulstijn, J. H. (1992). Retention of inferred and given word meanings: Experiments in incidental vocabulary learning. In P.J.L Arnaud & H. Bejoint (Eds.), Vocabulary and Applied Linguistics (pp. 113-125). London, UK: Macmillan.

Hulstijn, J. H. (1993). When do foreign‐language readers look up the meaning of unfamiliar words? The influence of task and learner variables. The Modern Language Journal, 77(2), 139-147.

Krashen, S. (1989). We acquire vocabulary and spelling by reading: Additional evidence for the input hypothesis. The Modern Language Journal, 73(4), 440-464.

Laufer, B. (2005). Focus on form in second language vocabulary learning. EUROSLA Yearbook, 5(1), 223–250.

Laufer, B., & Hill, M. (2000). What lexical information do L2 learners select in a CALL dictionary and how does it affect word retention?. Language Learning & Technology, 3(2), 58-76.

Nagy, W. E. (1997). On the role of context in first- and second-language vocabulary learning. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, Acquisition and Pedagogy (pp. 64–83). Cambridge, UK: Cambridge University Press.

Nagy, W. E., Herman, P. A., & Anderson, R. C. (1985). Learning words from context. Reading Research Quarterly, 20(2), 233-253.

Nation, I. S. P. (2001). Learning Vocabulary in Another Language. Cambridge, UK: Cambridge University Press.

Pellicer-Sánchez, A., & Schmitt, N. (2010). Incidental vocabulary acquisition from an authentic novel: do things fall apart?. Reading in a Foreign Language, 22(1), 31- 55.

Prichard, C. (2008). Evaluating L2 readers’ vocabulary strategies and dictionary use. Reading in a Foreign Language, 20(2), 216.

Schmidt, R. W. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2), 129-158.

Schmidt, R. W. (2001). Attention. In P. Robinson (Ed.), Cognition and Second Language Instruction (pp. 3-32). Cambridge, UK: Cambridge University Press.

Schmitt, N. (2008). Instructed second language vocabulary learning. Language Teaching Research, 12(3), 329–363.

Sternberg, R. J. (1987). Most vocabulary is learned from context. In M. G. McKeown & M. E. Curtis (Eds.) The Nature of Vocabulary Acquisition (pp. 89-105). Hillsdale, NJ: Lawrence Erlbaum Associates.

Wesche, M. B., & Paribakht, T. S. (1996). Assessing second language vocabulary knowledge: Depth versus breadth. Canadian Modern Language Review, 53(1), 13-40.