1. A complex, integrated, incremental, process perspective (CIIP) of vocabulary knowledge and acquisition

Writing my last academic book (Schmitt, Reference Schmittin press) gave me a great chance both to think about what I learned in my 36+ years in applied linguistics and second language (L2) vocabulary studies, but also to reflect on how the field might be pushed forward. Despite the many advances in the years since Paul Nation published Teaching and learning vocabulary in Reference Nation1990, there are still major gaps in our understanding of vocabulary knowledge, acquisition, and use. Some examples include issues concerning the mechanics of vocabulary acquisition (e.g., the relationship between receptive and productive mastery of vocabulary), the most effective vocabulary pedagogy (e.g., what kind of activity is necessary to improve a learner’s ability to use a word,Footnote ¹ i.e. help to move a word from declarative knowledge to proceduralized proficiency and fluency), and vocabulary assessment (e.g., how does performance on vocabulary tests relate to performance in real-world usage in the four skills?).

I feel a major stumbling block to answering quite tricky questions like these has been a relatively crude, black-and-white view of vocabulary knowledge – that a person either knows or doesn’t know a word, or that a person has either learned that word or hasn’t. This is often inferred from a correct/incorrect response to a test item. While such a dichotomous knows/doesn’t know and learned/hasn’t learned view has certainly allowed us to research and understand much about vocabulary knowledge and acquisition, we have virtually reached its limitations.Footnote ² We now seem to be treading water in vocabulary research to some degree. True, there are continuing advancements, but these are rather small incremental ones. The field needs a bit of a jolt to start making more substantial gains again.

There are (at least) four characteristics of a more nuanced view of vocabulary. None of them are new or rocket science, but to my knowledge they have never been presented before as a unified approach:

• • Vocabulary knowledge is Complex: Knowledge of vocabulary is not a single homogenous construct, but rather consists of a constellation of various aspects of word knowledge, such as spelling, pronunciation, meaning, grammatical characteristics, collocations, and stylistic characteristics.

• • Vocabulary knowledge is Integrated in acquisition and use: Vocabulary knowledge consists of many aspects, but these are interrelated to a very large extent. When one aspect is enhanced, it usually facilitates the enhancement of other aspects.

• • Vocabulary learning is Incremental: Because vocabulary is complex, it is not acquired in a binary “learned/not learned” manner. Rather, it is mastered incrementally over a (usually long) period of time. Thus, the knowledge of vocabulary can only be understood as some point along a continuum of mastery, rather than simply as being “known.” In fact, the typical state of knowledge for most words, for most L2 learners (and even L1 speakers), is partial mastery.

• • Vocabulary learning is a Process: Because vocabulary is complex with many integrated components, each of which is mastered incrementally over time, vocabulary acquisition is a process.

My shorthand umbrella term for a more nuanced conceptualization of vocabulary knowledge taking these points into account is a Complex, Integrated, Incremental Process perspective of vocabulary acquisition (CIIP for short – pronounced /sɪp/). Although none of these four characteristics are new or startling insights, it nevertheless always surprises me how seldom teachers, materials writers, and scholars treat vocabulary learning as a process rather than a one-off learning event. So while CIIP characteristics might be implicitly understood, I feel their pedagogical implications have seldom been fully realized and taken on board. A CIIP perspective entails repeated exposure and practice points, different types of instruction for various stages of the developmental continuum, assessments which tap into degrees of mastery along the continuum, and curriculums designed for sustained engagement with lexical items over the extended length of time it takes to build truly meaningful vocabularies.

To give an idea of how a CIIP approach might be useful in advancing the field, in the sections below I will discuss its implications for three key pedagogical issues: (1) discerning the benefits and limitations of various vocabulary activities, (2) examining how different kinds of input facilitate (or not) the various aspects of word knowledge, and (3) examining what various test formats do and do not tell us about learners’ vocabulary knowledge.

2. What aspects of lexical knowledge do various vocabulary activities actually promote?

There are a wide variety of activities designed to develop vocabulary knowledge, and I think many people tend to think of them as an undifferentiated whole, more or less interchangeable and simply promoting “vocabulary knowledge.” But from a CIIP perspective, no single activity can possibly address all of a word’s myriad characteristics. That is, there is no such thing as a “general vocabulary activity.” Rather, each activity will address some word knowledge aspects and not others. Similarly, they might be better at promoting declarative knowledge of words more than their fluency in use, or vice versa. Therefore, instead of just using any old activity to encourage vocabulary learning, it is much better to understand what each activity does and does not do. We can then use this understanding to target the learning of particular word knowledge aspects with relevant activities, at appropriate points along the learning continuum.

Let us look at several different activities to illustrate this. I will start with word pair/paired associates learning, where target words are listed with some method of illustrating their meaning, e.g., translation: cat = gata (Spanish). Substantial research shows this activity is highly effective, with learning rates of 34 and 84 words per hour being reported, with low rates of forgetting (based on form-meaning tests). However, many teachers and textbook writers seem to dislike word pairs, believing that they are boring and methodologically antiquated, and that words should be learned in context. I think this misunderstands what word pairs can and cannot do, and how they should be used. The skeptical practitioners would probably be right if word pairs were the only activity used to learn vocabulary, and were left unsupported. Unfortunately, this is often the case. Students are given words to memorize as homework, then are quizzed on them at the beginning of the next class. But then the words are never used again, and the students soon learn to game the system and memorize the words just for the quiz and then forget them.

But this is not a failure of word pair learning, but rather a failure to understand how it fits into an integrated vocabulary program. Word pairs are typically used to establish an initial understanding of the form-meaning link, which is the essential first step toward learning a word. They are very good for this. But a CIIP perspective highlights that learning is an incremental process, not only in terms of the need for consolidation of previously learned knowledge, but also in terms of gaining different aspects of word knowledge from subsequent engagements. I feel that there is no reason why learners cannot start the learning process on their own through word pair learning as homework. But this fragile, decontextualized, initial form-meaning learning must then be followed up and contextualized in classroom activities, such as in readings, in explicit exercises, in word games, and in other activities like being used in example grammar sentences. In other words, learners can start the learning process with the form-meaning link, and then teachers can provide contextualized exposures in the classroom which can then add other aspects of knowledge like the word’s grammatical characteristics, its collocations, and its level of formality. This entire contextualized enhancement should be much easier if learners already have some initial idea of the new words to be practiced in class.

A related point is that many practitioners do not like L1 translations. I feel they have a place. Learning a word is a long-term process, and L1 translations can quickly jumpstart that process, which must be useful. But again, that initial learning must be followed up by using the word in more contextualized L2 uses. This will help to build the word knowledge aspects which word pairs do not focus on.

On the other end of the activity spectrum is extensive reading. While this is widely promoted for language learning in general, it is also useful for vocabulary acquisition. But maybe not in the way most people might think. We know that readers can and do become familiar with new lexical items from reading, but also that the uptake is relatively low, with many repetitions necessary to reach even a meaning recognition level of mastery. So extensive reading is not the best way to begin learning substantial numbers of new lexical items. Explicit activities are much better for this. Rather, extensive reading offers an excellent way of enhancing knowledge of partially known items. Seeing these items repeatedly in many different contexts is exactly the kind of contextualized exposure which can facilitate the learning of contextualized aspects of word knowledge like collocation, derivatives, and stylistic appropriacy. If we view vocabulary learning as a process, then it makes sense to have learners introduce themselves to as many new items as possible through explicit study (such as through word pair study), and then have as much subsequent contextualized exposure as possible to both consolidate the initial the initial form-meaning link, and to “fill out” their knowledge of the item (such as through extensive reading). Thus we see that different activities can be most useful at different stages of the acquisition process.

Looking at some other activities, crossword puzzles can be interesting for learners. In terms of vocabulary learning, they are mainly useful for enhancing learners’ productive knowledge of a word’s spelling, and also for reinforcing its form-meaning link. But they do not enhance any other word knowledge aspect. Likewise, derivational grids are helpful in learning a word’s various derivational forms (e.g., noise → noisy, noisily) but enhance no other word knowledge aspects except spelling. From these examples, we see that we should think of specific vocabulary activities as addressing specific lexical learning needs. It follows that there are no “good” or “bad” vocabulary activities. Every activity enhances certain word knowledge aspects, and thus any activity’s value must be established in terms of whether those aspects are the ones the teacher wishes to focus on at that point in time. Only by understanding what any particular vocabulary activity does and does not do, can its most effective place in an overall integrated vocabulary teaching program be best established.

3. What aspects of lexical knowledge do various kinds of input really promote?

Just as vocabulary activities enhance some word knowledge aspects and not others, we can examine various kinds of vocabulary input in the same way. Let us consider the lexical input that learners receive from classroom teacher talk and language textbooks.

Most language teachers do give explicit attention to vocabulary, but research suggests that it is often minimal. For example, Folse (Reference Folse2010) observed only 4.8 cases per 50-minute class of either teacher-led or student-led explicit focus on vocabulary in an American IEL program. He found that many of the teacher interventions were relatively trivial, and not done in a way that would facilitate remembering or recycling. There seemed to be the mentality of a “one-shot” approach to the vocabulary episodes: a quick defining exchange, and then moving on as if no follow-up was necessary. Some teachers give vocabulary little explicit attention at all, believing it will be “picked up” incidentally from general language exposure in the classroom (e.g., in Sweden; Bergström et al., Reference Bergström, Norberg and Nordlund2022).

In terms of which aspects of word knowledge are taught by teachers, Horst et al. (Reference Horst, Collins and Cardoso2009) found that meaning was the aspect most taught (72%), while form was focused upon in 25% of the cases, and usage only 3%. All of the usage interventions commented on collocation; other aspects of use were not addressed at all. It is not surprising that meaning is the aspect most addressed, as teachers naturally comment on unknown words, especially if they feel these are critical to comprehension of the current text. Teaching meaning in itself is no bad thing, but the lack of opportunities to learn about other word knowledge aspects must certainly hamper learners from developing fuller knowledge about the items. Moreover, the overriding emphasis on meaning betrays a focus on teaching only new or unknown words.

Moving to L2 language textbooks, I find the situation very discouraging. While I (and many others) advocate a more principled approach to vocabulary teaching (e.g., a more thoughtful selection of target words, more repetition of those words), I fully realize that teachers are busy and often do not have the time or energy to fully implement such an approach. This makes language textbooks crucial, as I strongly suspect that most language teaching around the world is largely (or almost completely) dictated by textbooks. Thus, textbooks will drive most vocabulary instruction, or the lack thereof. Unfortunately, as I have commented before in this journal (Schmitt, Reference Schmitt2019), most language textbooks lack a systematic approach to vocabulary. My further research into this issue when writing my latest book only reinforced my unease about this.

One major problem with vocabulary in L2 textbooks is that there is simply not enough of it which is effectively targeted. Several scholars have analyzed the vocabulary in language textbooks and found that they did not supply enough vocabulary to reach the stated proficiency level of the textbook (often some CEFR level), even if students learned all of the vocabulary available in the textbooks. Also, textbooks can be very inconsistent in the average number of new words presented per unit, with a great deal of variation between one unit and another.

A further problem is that most textbooks lack any obvious systematic approach to the selection of the vocabulary that they do present. I think it is fair to say that most textbooks use a reading passage as the anchor for the lesson. This makes sense because if we want to talk, listen, read or write, it needs to be about something, and so we need some carrier content. But the target vocabulary tends to be drawn from whatever words appear in these reading passages, rather than being selected for usefulness or frequency of occurrence. Overall, several textbook analysis studies indicate that the vocabulary included in textbooks appears to be quite haphazard and mainly dependent on the personal preferences of the writer(s).

A related problem is that content tends to be very loosely connected at best, if it has any relationship at all. Publishers regularly require textbook writers to focus on “high interest” topics to motivate students, which is fine in itself, but these topics tend to jump from one to another, rather than developing a sustained theme. For example, four consecutive topics in one textbook are (1) Family and friends (e.g., with target words aunt, brother), (2) Sports and hobbies (athletics, basketball, books, chess), (3) School subjects (biology, English, math), and (4) Clothes (boot, top, jeans). Unrelated topics like these, and the loose general themes that are often used to supposedly link them together, lead to very limited (or often no) repetition of the target words. One would think that the recycling of new vocabulary would be a natural characteristic of teaching materials, especially at the beginner level. However, a range of research has found minimal recycling in various texts, with words in textbooks occurring just once or twice being more the rule than the exception. Essentially, traditional linear course structures are not focused on deepening and consolidating previously taught (but imperfectly learned) material because the focus is always on new things (topics, vocabulary, grammar features, skills, strategies, etc.), leading to a “one shot” approach to vocabulary teaching.

A last point concerns the types of vocabulary activities in textbooks. Research shows that they almost exclusively focus on the form-meaning link. Brown (Reference Brown2011) looked at nine General English textbooks at three proficiency levels, finding form and meaning received by far the most attention at all three levels. Grammatical functions and spoken form also received some attention, but other aspects received little or none. In my latest book, I reviewed three EFL textbooks that were in common use in early 2025, finding that Brown’s results still basically held, with form and meaning by far the most addressed aspects, although one noticeable feature in all three textbooks was the attempt to contextualize the target vocabulary by practicing it in contextualized sentences or dialogues, or in activities highlighting the four skills.Footnote ³

The inescapable conclusion from all this is that most language textbooks do not provide enough vocabulary, the right vocabulary, enough repetition of the vocabulary taught, or exercises which expand vocabulary knowledge to enable learners to build their lexicons to the point to where effective language use is possible. To remedy this, we need to get publishers to require their authors to have a greater focus on vocabulary in their textbooks, with a more principled approach to vocabulary selection and substantial recycling. However, I am not sure how to pressure publishers into doing this, other than having book customers/purchasers demanding it.Footnote ⁴

4. How to better interpret vocabulary test results

The CIIP perspective probably has no greater implications than in vocabulary measurement. In no other realm has the “knows a word” vs “doesn’t know a word” dichotomy been more widely applied than in the interpretation of vocabulary tests. When an examinee (L1 or L2) answers a vocabulary item correctly, that result is typically interpreted as demonstrating that the target word is “known.” Likewise, in a vocabulary acquisition study, when a post-test vocabulary item is answered correctly, the result is usually interpreted as showing that the word has been “learned.” Traditionally, there has been little additional elaboration of the degree of knowledge or learning. Happily, this is slowly changing, with more careful description of the degree of form-meaning link knowledge in terms of meaning-recognition, form-recognition, meaning-recall, and form-recall (see Footnote 2).

Perhaps it is good to start this discussion by saying there is no “best” or “one-size-fits-all” vocabulary assessment. All test formats have their strengths and limitations, and I do not think there will ever be a test format that can capture all the complexity of vocabulary knowledge, or show fully where a word lies on the learning continuum. The key to selecting or creating an appropriate vocabulary test is clearly understanding at the very beginning exactly what you want to find out about your learners’ vocabulary knowledge. It could be many things: e.g., how many words were learned from a word list at the meaning-recognition level, how many words a learner can understand when reading, what range of words were produced in a learner’s essay, does a learner know the most common idioms at a meaning-recall level, or how fluently can a learner use their vocabulary when speaking?

No test format can measure all these things, so it is important to use a test suitable for one’s specific purpose. This means we should not merely speak of “vocabulary tests” in general but rather specify those tests in terms of the different levels and types of vocabulary knowledge they describe. Different test formats will be more appropriate for measuring different points along the learning continuum. Form-recognition and meaning-recognition formats can show initial learning of the form-meaning link. Form-recall and meaning-recall formats can show more developed mastery of this link, but still only address these two types of word knowledge. To measure more advanced vocabulary knowledge, we need to use test formats which somehow capture more contextualized word knowledge aspects like derivations, collocations, and stylistic appropriacy.

I also think that there is a problem in how the results from test formats are interpreted. Recognition and recall formats are often seen as indicating receptive and productive vocabulary knowledge respectively, but meaning-recall actually indicates receptive, not productive knowledge. When listening or reading, a person hears/sees a word form and then must recall its corresponding meaning. Only form-recall is productive; that is, when trying to express a meaning in speech or writing, one must be able to recall the corresponding word form. To be more precise, I think that the terms receptive and productive should be reserved for describing how vocabulary is employed in the four skills, i.e., as usage-based definitions of mastery. Certainly, on the face of it, form-recognition and meaning-recognition formats do not demonstrate the ability to understand a word when encountered in listening or reading, because in listening and reading, no options are present. Therefore, the test format is quite different from how vocabulary is encountered in the real world.

In fact, I see this as a major gap in our assessment knowledge. We have almost no idea of what recognition tests actually tell us about the degree of vocabulary knowledge a learner possesses, or the stage of acquisition that a target word is at. This is despite selective-response vocabulary tests being widely used, to the extent that the multiple-choice format is probably the most widely used format globally (e.g., Vocabulary Levels Test, Updated Vocabulary Levels Test, Vocabulary Size Test, Peabody Picture Vocabulary Test). Maybe recognition tests do indicate some ability to comprehend target words receptively, but maybe they do not; we simply do not know. Until future research demonstrates the relationship between recognition test scores and the ability to understand words receptively in real-world use, I recommend that recognition tests are reported as merely indicating form-recognition or meaning-recognition levels of mastery, which I would interpret as the initial stages of the vocabulary learning trajectory.

With recall formats, it is easier to make the argument that they may indicate receptive or productive knowledge. Receptive knowledge entails knowing a word well enough to extract communicative value from speech or writing. Productive knowledge involves knowing a word well enough to produce it when it is needed to encode communicative content in speech or writing. Meaning-recall and form-recall formats are at least a clear first step toward measuring such skills-based abilities. However, a problem with form-recall formats is that they prompt the target words. It is not yet clear how well these formats indicate the ability to think up appropriate words on one’s own without prompting, i.e. in the free speaking and writing that is typical of real-world use. Again, we have very little idea of how to interpret these formats in these terms.

Ultimately, we need assessment formats which inform about ability to comprehend/produce vocabulary which is embedded in longer stretches of discourse and thus is constrained and dependent on that discourse. For example, receptive knowledge could be measured by learners reading a passage and then being tested on how well they understood the words in the text, and the overall meaning expressed by those words. Or production could be measured by analyzing the vocabulary produced in a task designed to elicit target lexical items (e.g., describing a picture). But for now, I think the CIIP implications for test formats are clear. First, we need more research into vocabulary test formats to discover what they do and do not tell us about examinees’ lexical knowledge. Developing more skills-based vocabulary test formats would also be useful. Second, to the extent of our current knowledge, we must be much more precise when reporting what current vocabulary test results do and do not mean.

5. Conclusion

Current research continues to advance our knowledge of vocabulary acquisition and use. It just has not yet been able to really grapple with some crucial pedagogical issues, e.g., how to promote vocabulary learning beyond the initial form-meaning link, or how to measure this more advanced learning. To address these trickier (but very interesting) issues, we need to start conceptualizing vocabulary learning as moving along an acquisition continuum, and start thinking about how we can help learners move their vocabulary knowledge along various points this continuum. I hope that the CIIP perspective can be a useful tool in developing this more sophisticated conceptualization.