Dellar, 2024: Whither Teachng Lexically?

I was up early on Friday morning and during what was supposed to be a quick browse of social media, I came across a video of Hugh Dellar talking about memorisation. I was invited to look at several other videos by Dellar, and down the rabbit hole I went.

He’s a busy lad. He’s got the Lexical Lab website, hundreds (sic) of videos on YouTube, Facebook, etc., he gives teacher training courses and conference presentations, he runs a summer programme, and he publishes three different coursebook series. In 2024, Dellar gave a number of presentations on “Better Listening Outcomes” (IATEFL, Brighton, TESOL Spain; an article in a German ELT magazine); a session on “Performing Memorization in Class” at the Polish IATEFL conference; and a session on “The Taboo of Translating” in Astana, Kazakhstan. I’ll use these talks to examine what progress Dellar has made recently to hone his Lexical Approach to ELT.

Listening, Part 1

In his presentations on “Better Listening Outcomes”, Dellar makes two main points

1. Listening comprehension depends on how much of the target language a person knows (who knew!)
2. Successful listening depends mostly on bottom-up processing.

On the first point, Dellar says:

I'm sure you've often played an audio in class and you stop and ask them, “So what did you get?” The students just kind of make this weird imagined version of what English sounds like to them and sort of go agla, …blah, …gaggle, because basically that's what they're hearing. They're hearing this kind of blur because they don't have the language and they don't have the decoding abilities yet. …………….. What distinguishes good listeners from not-so-good listeners is the simple fact is that the good listeners know more of the language.

This lead Dellar to the question

If language is at the heart of skills development and listening development, what language should we be prioritising?

His answer is:

We should prioritise language that students at any particular kind of level most need in order to be able to have predictable, normal, everyday conversations. This means thinking about word frequency data. Words are frequent because they're frequently used. And by focusing more on frequent words than non-frequent words, you're helping students read and listen better because this is the language they're more likely to encounter.  

In his talk on “Memorisation“, Dellar deals more fully with the language that needs prioritising. He says that at the B2 level on the CEFR scale,

you need 3,000 to 4000 key words, which means not just memorising those words but knowing how they interact with each other and how they interact with grammar. So once you start thinking about the collocations and the chunks and the groups of words together, you’re already thinking about tens of thousands of words.

Furthermore, Dellar adds,

to understand a reading text, you need to understand 98% of the words of that text, and for that, you need a working knowledge of around 15,000 words, which again involves knowledge of the multiplicity of uses.

Dellar then points out that if teachers use a coursebook series to work through from Beginner to Advanced, then even if 50 new words are introduced in every unit of all 5 coursebooks in the series, drum roll........

they will only cover 3,000 words by Advanced level. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

And just memorising things isn’t enough. You need to be basically working on moving language from your short-term memory (taps right side of the front part of his head) into your kind of long-term, internalised, proceduralised, automatic memory (taps back of his head). A bit like when you’re learning to drive. First you have to think about it all - put foot on clutch, look in mirror, indicate, change gear - thinking about it all, thinking “I’m never going to get the hang of this”. But after a while it all becomes automatic.     

Discussion

Dellar’s 2024 work does nothing to address the “Big Question” that has remained unanswered since he first unveiled his lexical approach more than twenty years ago: “How can teachers adequately teach all this language?” It just doesn't add up! If knowing the 3,000 to 4000 key words in English entails knowing “tens of thousands of words" ("collocates, lexical chunks, and the grammar that goes with them"), and if to understand a reading text, you need a working knowledge of all the various uses of around 15,000 words, how can you justify basing your ELT courses on using a succession of Outcomes coursebooks, which, by advanced level, have only covered 3,000 words? There’s surely something basically wrong with an approach which, by its own criteria, fails to teach students most of what they need to know.

The problem is made worse by Dellar’s idiosyncratic selection of lexis, colocations and lexical chunks and Dellar's 2034 output gives no signs of progress in providing coherent criteria for such a selection process. I remember Mchael Swan pointing out that “memorising 10 lexical chunks a day, a learner would take nearly 30 years to achieve a good command of 10,000 of them” and wondering how Dellar might go about the task of selecting from among them.

Rafatbakhsh & Ahmadi (2020) list the most frequent idioms used in contemporary American English, in the academic, fiction, spoken, newspaper, & magazine genres. There’s also the 505 multiword items (take place, at all costs, etc.) in the PHRASal Expressions List (PHRASE List; Martinez & Schmitt 2012) which gives 10% of the 5000 most frequent word families, with 95% of the 505 made up of the 1000 most frequent words in English. While Dellar often mentions the importance of frequency as a criterion for selection, I can find no mention of any of these most frequent multiword items and idioms in any of Dellar's 2024 articles and presentations, or in the 200 + items which make up his “One Minute English” video series, which include It does my head in, budge up, wouldn’t / might not go amiss, scrub up well, up sticks, and throw a sickie.

Apart from a certain wierdness in Dellar's choices, the fundamental flaw in Dellar’s approach to ELT is. I suggest, explained by his reliance on explicit teaching, which results in far too little time being devoted to giving students opportunities to learn for themselves.

(Just by the way, Dellar’s description of how learners listen – “moving language from your short-term memory into your long-term, internalised, proceduralised, automatic memory” is mistaken, both in terms of where and how it occurs, and the tired analogy of driving a car blithely ignores the fact that language learning is on a completely different level of complexity.)

Listening, Part 2

In classroom listening activities Dellar advises teachers to focus on bottom-up processing. Having listed the ‘sub-skills’ which the CEFR says learners should have (listening for gist, specific information, detailed understanding and for implications, and listening as a member of a live audience) Dellar asks “How do we teach them?” He says:

The dominant way of thinking about all of this has been SCHEMA THEORY, which stresses what’s called top-down processing. This emphasises students’ prior knowledge and predictions / expectations about what will be said. Often this means that before we ‘do’ our listening in class, we get students to predict content from pictures, context, etc. Now, this is all well and good, but read deeper in the literature on the field and problems soon start emerging, as the following quote makes clear:

“For complex social and psychological reasons, [learners] are less sure they have grasped the topic being spoken of, the opinion being expressed about it, and the reasons for the speaker wanting to talk about it. They are less sure of the relevance of their own experience in helping them to arrive at an interpretation. On top of all that they are less sure of the forms of the language… for all these reasons learners are less able to bring to bear top-down processing in forming an interpretation and hence are more reliant on bottom up processing.” (Brown quoted in Jenkins, 2001 OUP)

What Brown focuses on is the idea of BOTTOM-UP PROCESSING. In short, this says that what is important is HEARING individual sounds, decoding words, decoding chunks, decoding sentences and so on, and that it is through the process of doing this that learners build up a mental picture of what is being discussed. If you accept this – and I do – there would seem to be some profound implications for teaching listening.

The first profound implication is that getting students to predict or use their previous knowledge, i.e. activating their schemata isn’t necessary.

In the real world, language in use can be very unpredictable indeed – and the only way to deal with this is to listen to it all and understand it all.

The second profound implication is that once they learn enough language, students will do what proficient English users do:

The listener recognises the sounds and the words as they're hearing the language being spoken, and they know the meaning of them when they hear them. …. What's going on is they're not just hearing sounds and words that they know and understand, but they're able to process those things in real time. …… What you're doing is, as you're hearing the sounds and the words and you're remembering them and you're processing them automatically in real time, you're slowly building up an understanding from the bottom up. As you're going along, you're understanding the sound, the word, the chunk, the sentence, the next sentence. And as you're going along, you're remembering what you've heard already, and you're adding the understanding of what you're then encountering to what you've already understood, and you're building up this kind of meaning.”

In the rest of his texts on listening, Dellar discusses the challenges of recognizing words in continuous speech and suggests strategies to help students, such as drilling words in isolation and in common chunks, and introducing 'understanding fast speech' exercises (a new feature of the third edition of the Outcomes series, he adds). He stresses the importance of repetition in developing automatic recall of language chunks; he suggests “allowing students time to co-construct meaning” and share understanding during listening exercises; and he invites teachers to explore his online school and website, and to avail themselves of free samples of materials from the new Outcomes series.

Discussion

First, as Rost (2001) points out, the processes of learning to listen and listening to learn are different. While learning to listen involves enhancing comprehension abilities, listening to learn involves “creating new meaning and form linking and then repeating the meaning and form linking, which helps the learners to be ready for paying more attention to the syntax and lexis of the language through listening (p.90). Rost concludes “the optimal goal of L2 listening development is to allow for the L2 to be acquired through listening, not only to allow the learner to understand spoken messages in the L2” (p. 91).  

Second, Dellar is wrong to claim that “reading deeper” about listening comprehension reveals that top-down processing is unnecessary. On the contrary, there is near unanimity among SLA scholars that both top-down processing and bottom-up processing are involved in listening comprehension. Combinations of top-down processing with bottom-up processing of information from the stimulus itself are used. Linguistic knowledge and world knowledge interact in parallel fashion as listeners create a mental representation of what they have heard (Hulstijin, 2003). Dellar interprets Brown as claiming that it’s unhelpful for teachers to encourage top-down processing through preparatory schemata-building exercises. In fact, Brown says no such thing. He says that learners – particularly those with a relatively low level of proficiency – find top-down processing difficult; he certainly does not say, or imply, that teachers should not scaffold learners’ encounters with aural texts by using top-down processing strategies with them.

Third, Dellar’s description of “what a listener actually does” reveals little evidence of being familiar with “the literature on the field” as he calls it. There are, in fact, various and sometimes mutually contradictory accounts of the listening process (Krashen’s and Susanne Carroll’s, for example). One common view is that the language comprehension process has three stages: perception, parsing and utilization. In the perception stage, the spoken message is encoded; in the parsing stage, the words in the message are unconsciously examined in terms of the learner’s interlanguage grammar and transformed into a mental representation of the combined meaning of the words. If the listener fails to parse the message, various outcomes ensue, such as ignoring that bit and attending to the next, trying to work out the meaning, asking for help, or giving up. (Many scholars see failure to parse part of a text as the trigger for most of the development in L2 learners’ interlanguage.) The third stage is the utilization stage, in which listeners use the mental representation of the sentence’s meaning to respond. If the sentence is a question, they may answer; if it is an instruction, they may obey, etc. Dellar’s assertions about the role of bottom-up processing are badly-informed and misleading.   

The L2 learning Process

In his presentation to teachers at the Scholarium conference in Kazakhstan this year, Dellar discussed the “taboo” of using translation in the ELT classroom. This being 2024, the novelty of challenging the translation taboo has long gone (it wasn’t even a “taboo” in 2018 when he wrote his first version of the talk), but in a lengthy run up to extolling the virtues of translation, Dellar talked about the SLA process and the role of Schmidt's (1990, 2001, 2010) construct of noticing.

He began by explaining that for the first edition of his Outcomes coursebooks, he included a feature called “Patterns” which focused on “a lexico-grammatical pattern that wasn’t strictly grammar, but that were definitely beyond single words”, such as the examples below (note that lists like this are the hallmark of Dellar’s work)

Mongolia is known as ‘the land of the horse’.

Shanghai is known as ‘the Paris of the East’.

Aubergines are also known as eggplants.

The area is known for its oysters.

The village is well known for its leather goods.

This rare species of shark is known to inhabit fresh water.

Very few details are known about this rare species.

Dellar wanted students “to notice pattens in these uses of known and to notice how this word known is used in different contexts”. He then explained Schmidt’s construct of Noticing. To quote Dellar,

Schmidt says you can’t actually learn things [formal aspects of the L2 like grammar] until they’ve been noticed. I don’t think it’s always true, but a lot of the time I think it’s true. You can’t properly learn something until you’ve noticed what it is that you’re trying to learn.

Noticing is thus “a key idea in ELT”, and a key part of Dellar’s account of how people learn a foreign language.

To learn a foreign language, there’s not many steps you need to go through. You need to hear or see the language; you need to be somehow led to understand the language, whether that’s through explanation, bad drawings, people hopping around the floor making weird noises , or translation; you need to notice the way the language works with the words and the grammar working together; you need to do something with the language, use it in some way; and you need to keep repeating these stages over time. 

That’s it.

How, then, asks Dellar, can we get our students to notice and how do we know if they’ve noticed?

As a teacher you can say “Look! Notice!” And the students say “Yeah, alright teacher, Noticed!” And you go “All right! Great! Noticing has been achieved!” Obviously this is not an effective way of achieving noticing. So we tried this: [Dellar shows the sentences below on a slide]. 

It’s hardly the same thing!

Hardly an instant solution then!

It’s hardly surprising people are concerned about it.

Hardly a day goes by without hearing one of these stories.

I hardly know anyone who agrees with it.

There’s hardly any funding available for research into it

Having put these sentences on display, students are asked to do the following:  

1. Write the sentences in your language.
2. Translate them back into English.
3. Compare your English to the original.

Dellar comments

Translating back and forth between languages like this forces noticing in a way that nothing else does (from printed version in 2018).  

 I can only guess at precisely what "patterns" in those sentences using the word "hardly" Dellar wanted his students to “notice”. He didn't expalin either in the 2018 blog post or the 2024 presentation.  

Discussion

Dellar’s 2024 description of the steps you need to go through in order to learn a foreign language is a garbled version of the one given in the Dellar & Walkley (2016) book Teaching Lexically, where six steps in learning “any given item of language” are described. The book doesn’t define “item”, but we can assume it means contextualised lexis and that it includes collocations and lexical chunks. In any case, Dellar’s 2024 "four steps to learning" make up an infantile description which explains nothing about the process of SLA.

In their book, Dellar & Walkley use Hoey’s Lexical Priming theory in conjunction with Schmidt’s Noticing hypothesis in an incoherent  account which fails to resolve the contradictions between them - Hoey says L2 learning is basically an unconscious process, while Schmidt says conscious “noticing” is of paramount importance. Dellar relies more on noticing, but it's not Schmidt's construct of noticing, although Dellar implies that it is. Schmidt took great care to carefully define the construct of “noticing”, and twice subsequently revised it in response to critics, ending up by saying it was the same as Gass’ construct of apperception. Dellar claims to be using Schmidt's construct, but actually, he uses it in its canonical dictionary meaning “to become aware of something that has caught one's attention”. For instance, some in the audience watching Dellar's home videos might notice that Dellar's bookshelf has "Teaching Lexically" prominantly dispayed. This use of the word noticing strips Schmidt's construct of all his efforts to make it represent a complex explanation of how input becomes intake. Furthermore, Schmidt early on conceded that learners don't notice grammar in the input because abstract rules are nor there to be noticed. By the same token, most of the "patterns" that Dellar is so desperate for his students to notice are beyond the scope of what Scmidt refers to as noticing. I think it takes Dellar to reduce Schmidt's 30+ years of brilliant work to this: "You can’t properly learn something until you’ve noticed what it is that you’re trying to learn".

Nor does Dellar seem to have made any progress in tackling the problem of the teachability of lexical chunks which have been raised by so many scholars, including Boers and his colleagues. Dellar assumes that drawing attention to features of language and making students aware of collocations, co-text, colligations, antonyms, etc., by giving students (repeated) exposure to carefully-chosen written and spoken texts, using drills, concept questions, input flood, bottom-up comprehension questions, and so on, will allow the declarative knowledge taught to become fully proceduralised. Quite apart from the question already discussed of how many chunks a teacher can realistically be expected to treat so exhaustively, there are good reasons to question the assumption that such instruction will have the desired result. 

For reasons touched on above, there is, at last, increasing interest being shown by course designers in the generally accepted view of SLA scholars that L2 learning is predominantly an unconscious process of implicit learning. Let me give just three quotes:

Nick Ellis (2005) says:

the bulk of language acquisition is implicit learning from usage. Most knowledge is tacit knowledge; most learning is implicit; the vast majority of our cognitive processing is unconscious.

Whong, Gil and Marsden’s (2014) review of a wide body of studies in SLA concludes:

“Implicit learning is more basic and more important than explicit learning, and superior.  Access to implicit knowledge is automatic and fast, and is what underlies listening comprehension, spontaneous speech, and fluency. It is the result of deeper processing and is more durable as a result, and it obviates the need for explicit knowledge, freeing up attentional resources for a speaker to focus on message content”.

Finally, here’s Catherine Doughty:

…. (iv) explicit, declarative information is only helpful in improving performance in cases where complex tasks involve few and obvious variables; and (v) implicit practice at the relationships underlying the algorithms is beneficial. In sum, the findings of a pervasive implicit mode of learning, and the limited role of explicit learning in improving performance in complex control tasks, point to a default mode for SLA that is fundamentally implicit, and to the need to avoid declarative knowledge when designing L2 pedagogical procedures (Doughty 2003, p. 298).

This has led to the growth of TBLT and Dogme-based courses, which involve students in tasks where they learn by doing things in the target language, not by being told things about the language. I have argued for the efficacy of strong versions of TBLT on many occasions, and together with Neil McMillan, Marta González-Lloret, Peter Skehan, Rose Bard, Roger Gilabert and others, we continue to refine the courses. A great deal more ground can be covered if we teachers allow learners to "learn by doing", implicitly dealing with all the input coming at them as they work through pedagogic tasks, and we resrtict ourselves to scaffolding their work, making short, timely interventions aimed at speeding up their progress.

There is nothing in the SLA literature to support spending the majority of classroom time explicitly teaching the kinds of things Dellar recommends. The impossibility of covering the ground necessary for proficiency by adopting Dellar’s approach is apparent - even to him, it seems - and yet Dellar continues to insist that students need to be led through the materials of his coursebooks, where most of the time the book and the teacher tell them about the target languguage on the mistaken assumption that all this explict teaching will somehow result in students getting the procedural knowledge they need to use the target language with confidence and fluency.

While I was looking through Dellar’s output, I came across the 2018 post in the Lexical Lab blog which he re-worked for his 2024 Kazakhstan presentation. A reader who signed himself Sam left a comment which began:  

Due to my experience, I have to disagree with the article. It seems to me that speaking, learning and living a second language happens in a new different place in the mind

Dellar replied:

With all due respect, the idea that a second language happens in a different part of the brain to a first has no basis in research. When learning a second language, the new information is added to the old, and whilst that obviously means to old information about L2, it also obviously means L1.   

Well that's not true. First, Sam was talking about the mind, which is a theoretical construct, endowed by some with different types of modules. Second, pace Dellar, with regard to the brain, the view that a second language happens in a different part of the brain to a first has quite a big basis in science. For example there was an article in Nature (Kim, et al, 1997) "Distinct cortical areas associated with native and second languages", which reported on an investigation of two brain areas associated with second language acquisition, Wernicke and Broca. It found that if the second language is acquired after the person has grown up, it will be stored in a new Broca's area dedicated to this new language. There are many reports of studies affirming that L2s learnt by adults are stored in a different brain area, while 2 native languages (learnt during early childhood) are stored next to each other in the same area. But it's still a topic that's disputed and needs better measurement tools to try and deal with the enormous complexity of the workings of the brain.

For most SLA scholars what's more important is the well-established finding that implicit and explicit language learning, memory and knowledge are separate processes and systems, and their end products are stored in different areas of the brain, thus further questioning the claim that "practice" transforms declarative knowledge into procedural knowledge.

References

Dellar, H. & Walkley, A. (2016) Teaching Lexically. Delta.

Doughty, C. J. (2003) Instructed SLA: Constraints, Compensation, and Enhancement. In C. J. Doughty, & M. H. Long (Eds.), The Handbook of Second Language Acquisition (pp. 256-310). Blackwell.

Ellis, Nick. (2005). At the interface: Dynamic interactions of explicit and implicit language knowledge. Studies in Second Language Acquisition. 27, 305 - 352.

Kim, K., Relkin, N., Lee, KM. et al. (1997) Distinct cortical areas associated with native and second languages. Nature 388, 171–174

Martinez, R. & Schmitt, N. (2012) A phrasal expression list. Applied Linguistics, 33,3,299-320.

Rafatbakhsh. R & Ahmadi A. (2020) The Most Frequent Idioms Used in Contemporary American English: A Corpus-based Study. Applied Research on English Language, 9, 2, 205-228.

Rost, M. (2001) Teaching and Researching Listening. Longman.

Schmidt, R. (1990) The role of consciousness in second language learning. Applied Linguistics 11, 129–58.

Schmidt, R. (2001) Attention. In P. Robinson (Ed.), Cognition and second language instruction (pp.3-32). Cambridge University Press.

Schmidt, R. (2010) Attention, awareness, and individual differences in language learning. In W. M. Chan, S. Chi, K. N. Cin, J. Istanto, M. Nagami, J.W. Sew, T. Suthiwan, & I. Walker, Proceedings of CLaSIC 2010, Singapore, December 2-4 (pp. 721-737). Singapore: National University of Singapore, Centre for Language Studies.

Swan, M. (2006) Teaching Grammr – Does it work? MET, 15,2,5-13.

Whong, M., Gil, K.-H., & Marsden, H. (2014). Beyond paradigm: The ‘what’ and the ‘how’ of classroom research. Second Language Research, 30, 4, 551-568.