Vocabulary denial and the false god of structuralism in Ofsted’s 2021 Curriculum Research Review for languages

ABSTRACT

This paper reviews the vocabulary section of Ofsted's Curriculum Research Review of Languages (2021). It begins with a reality check and observes that while the review talks confidently of learners achieving expert levels of knowledge and performance, current learner levels are nowhere near expert. In terms of vocabulary knowledge, learners at GCSE know only 5–10% of the 8000 or 9000 words needed to be expert. It considers criteria for an effective vocabulary curriculum from Milton and Hopwood's (2022) Vocabulary in the Foreign Language Curriculum: Principles for Effective Instruction. None of these criteria is adequately covered. The review is unrepresentative of vocabulary learning research. Guidance on vocabulary size, for example, is missing but language proficiency is very much a function of vocabulary size. This is a disastrous oversight. It advocates teaching a small lexicon of overwhelmingly frequent words, insufficient for anything but the lowest levels of performance in the CEFR. In this review, it is as though a large lexicon were of no importance to language proficiency. It is a falsehood to call this a review of research at all. It gives a false veneer of respectability to current, structuralist teaching practice, which teaches very little vocabulary, and which has led directly to historically low levels of attainment in British schools

A reality check

It is necessary to begin this evaluation of the vocabulary section in Ofsted’s 2021 Curriculum Research Review for languages (OCRR) with a reality check. The report asserts that the process of learning a foreign language at school is one that leads successful learners to become ‘expert’ in the target language, and the term ‘expert’ is used repeatedly in relation to the requirement of the curriculum. For example, the review frames the learning process as, ‘Planned and purposeful progression in the curriculum: from novice to expert learner’ (OCRR: 14). The prism of vocabulary learning allows the idea of the ‘expert’ language user to be tied to a measurement of vocabulary size, and the current attainment of GCSE learners can be placed in relation to this. Proficient users of a modern European language will probably need something over 5000 words to reach C1 ‘Advanced’ level in the Council of Europe's (n.d.) Common European Framework of Reference (CEFR). To become genuinely expert at CEFR C2 ‘Mastery’ level, they will likely need 8000–9000 words, and many expert users will have substantially more (Laufer and Ravenhorst-Kalovski 2010; Nation 2006). It is currently estimated, however, that learners take and pass GCSE examination in French with 500–800 words (David 2008; Milton 2006); that is, with between 5 and 10% of the attainment of genuinely ‘expert’ language users. The latest Subject Content guidelines from the Department for Education (DfE 2022) limit vocabulary teaching – teaching content, not learner attainment – to 1200–1700 words; in other words, perhaps 15% of the vocabulary required to be ‘expert’. Milton and Hopwood (2021) show how current textbooks for French GCSE teach a vocabulary of this size. However, progress in a foreign language is so dependent on growing a large vocabulary, that Alderson (2005) asserts that language proficiency is largely a function of vocabulary size. At present, therefore, school learners in the UK, unless they learn the language outside the bounds of the current school curriculum, can never become expert users of their foreign language. They can never even attain the CEFR A2/B1 levels for GCSE that textbook writers, examiners and the DfE currently think are the goals of learning (for example, Curcin and Black 2019; Department for Education and Science 2005; Woore 2020). Current learners with knowledge of only 500–800 words can never progress beyond A1, the CEFR’s lowest level.

The standard for vocabulary input in the English curriculum for modern languages has fallen substantially over the years. Milton and Hopwood (2021) point to a 40% decline in the vocabulary content of key textbooks for teaching French at GCSE level in the last 10 years, while Milton (2013b) calculated a decline of 75% or more in both the vocabulary loading of teaching materials and the vocabulary attainment of learners at GCSE/O level¹ between 1978 and 2006. The progress and achievement of learners in the UK in terms of vocabulary is now just a fraction of learners elsewhere (Milton and Alexiou 2009). To put this point into perspective, Rodusaki and Alexiou (2021) observe that learners of a foreign language in Greece routinely achieve at age 8, after two or three years of learning, levels of vocabulary knowledge that learners in UK only achieve at age 17 or 18, after 11 or 12 years of study, and only if they specialise in learning a language.

This reality check is needed because the level of attainment expected of learners using the curriculum will determine much of the content. A curriculum designed to teach learners within the framework of CEFR A1 attainment will be very different and draw on a different background, to a curriculum designed to take learners to C2 and ‘expert’ levels of language use. A curriculum designed to raise learners to ‘expert’ levels of language use will need to input thousands of words in the foreign language which are absent from the current provision. A review such as the OCRR would have to lead a massive process of curriculum renewal, not least in its vocabulary content, if learners were to have any chance of to becoming ‘expert’ through study at school.

The aim of the OCRR

What, then, is the purpose of the OCRR? Is it to begin a serious attempt to raise the standard of language teaching in the UK by helping to define new goals for the curriculum, where learners will learn thousands of words and become ‘expert’? Or does it simply reflect a lack of awareness concerning the realities of language learning in schools? It should be a serious and well-informed attempt to help organise the process of learning a foreign language in school that takes the learner from no knowledge to a high level of communicative competence. A curriculum is, or should be, about setting the goals of learning, usefully defining the content of learning, and helping to plot a route through the content to achieve the goals. The OCRR’s stated goals (p. 1), however, are vague and all-embracing:

The review explores the literature relating to the field of foreign languages education. Its purpose is to identify factors that contribute to high-quality school languages curriculums, assessment, pedagogy and systems [and] … examine how languages are taught in England’s schools.

The problem with unclear and unfocussed aims is that these will produce an unfocussed review and unclear guidance. To provide effective guidance to curriculum designers, the writers need to be knowledgeable and in command of the research background, so that a large and disparate body of materials can be made coherent. To be useful, current practice in schools should not be just examined; it needs to be critically appraised especially in the light of what good research shows about how effective it is. If high-quality curriculums, and their characteristics, are part of this evaluation, then they should be identified, including those outside the UK where EFL practice has much to inform language teaching here. The OCRR needs to be very clear about the message it intends to deliver; however, its goals are really not framed to suggest that the content can usefully provide guidance to inform curriculum design. What guidance, then, should be given to curriculum writers about the vocabulary of a language? Milton and Hopwood (2022) address this question directly, and their guidance includes:

1. The number of words to be learned. – Vocabulary size relates, usually strongly, to overall language level and attainment. The curriculum writer will need to know how many items the learners will need to learn in order to reach the communicative goals of the curriculum.

2. The number of words to be taught. – This is not the same as the words that are learned. There is a difference between vocabulary input and vocabulary uptake. The curriculum writer will need to plan for this so teaching materials contain the right number of words for learners to achieve their goals.

3. The rate of word learning. – The rate at which words can be learned in school is usually fairly predictable. The curriculum designer will want to understand this to ensure that sufficient learning time is made available. Failure to do this means the curriculum is undeliverable.

4. The selection of words to be taught. – The curriculum designer will need to strike a balance of frequent structural vocabulary with the less frequent content and topical vocabulary to achieve a lexicon that can be used for communication. If the balance is wrong, then learner will be unable to use their language in communication.

5. The nature of learning. – There is a lot to teach if learners are to master vocabulary for communication, a lot more than learning a word form and its translation. This spectrum of knowledge develops in relation to vocabulary size and the curriculum designer will need to plan content to reflect this.

What guidance does the review provide?

In the case of vocabulary, the OCRR comprises a single page (p. 9), followed by a guidance (p. 10) that includes features that are not part of that review. Can a review with advice this minimal, give useful guidance for something as complex and important as curriculum design for vocabulary? What issues, and what research, does this review consider and what does it omit?

Vocabulary size goals – the number of words to be learned

The OCRR vocabulary section (p. 9) begins by asserting that a ‘wide vocabulary’ correlates with other aspects of language ability. After the opening sentence or two, issues of vocabulary size are ignored. It is extraordinary that a review of research into the teaching of foreign language vocabulary can omit any reference to what a ‘wide’ vocabulary is, or to explanations of the relationship between vocabulary size and performance. The research which relates vocabulary size to performance and language levels such as the CEFR hierarchy is very extensive and goes back more than 30 years. Vocabulary size and foreign language attainment do not just correlate; they are, very often, strongly linked (for a review, see Wang and Treffers-Daller 2017). Vocabulary size is regularly reported to be the most important factor in explaining variance in performance in the four skills (for example, Stæhr 2008; Milton 2013a) and in determining the scale of achievement possible in a foreign language (for example, Alderson 2005). Learners report that learning a lexicon of the right size is the most difficult task they face and that they find it more important to comprehension than other factors such as grammatical knowledge (see, for example, Laufer and Sim 1985). The very least the OCRR should have done to help guide curriculum design is to present the information that demonstrates normative vocabulary size scores in relation to levels of attainment. This is crucial to curriculum design. If learners taking GCSE are really expected to attain CEFR A2 then a learning target of about 3000 words is probably required. If the goal is for learners to achieve CEFR B1, then something nearer 4000 words might be a realistic target. Expert users at C2 will need 8000–9000 words as explained at the start of this paper. This will inevitably be a substantial part of any effective curriculum. It helps in setting goals for attainment and progress, but these are large numbers which require effective planning. The OCRR gives no hint of this but it should be central to any research review in this area. This is an omission that is so huge that it must be deliberate. OCRR’s decision to omit the extensive research which gives the numbers included here presents a fiction that the few hundred words currently learned at GCSE, is a ‘wide’ vocabulary and can justify the current, minuscule, vocabulary content of the curriculum. It cannot be stressed too strongly that research shows repeatedly that the ‘wide’ vocabulary associated with high levels of foreign language performance comprises thousands of words, not a few hundred words.

Vocabulary teaching and learning

Effective vocabulary planning in the curriculum has to come to terms with the relationship between words that are taught and the words that are actually learned as part of a course of study. It would be naïve to assume that all the words that are included in the curriculum are actually learned although, it seems, teachers often assume something like this occurs (Milton and Hopwood 2022). Research suggests that in effective vocabulary learning environments, on average, about 50% of the words that are input are then recalled in tests of knowledge (for example, Vassiliu 2001). Able learners can learn rather more. In poor and ineffective environments – and sad to say, the UK is such an environment – uptake can be only about 20% of the words which are taught (Milton 2011). This has implications for the volumes of vocabulary teaching that have to be planned for, in order to achieve the curriculum’s vocabulary attainment goals. If CEFR A2 and B1 are the targets for GCSE learners, then attainment goals of 3000–4000 words are needed, and this implies an input of 6000–8000 words. This is a lot of words and requires a lot of learning. However, nowhere in the OCRR does this kind of consideration feature. Where, as noted above, input is currently limited to 1200–1700 words, it requires a major change to the current curriculum that the OCRR should address. Because the review avoids almost all reference to vocabulary size goals, it can also ignore tackling the issue that taking learners to ‘expert’ levels will entail such a change.

Vocabulary and teaching time

A curriculum, no matter how well-researched, can only be useful if it can be delivered. A crucial component of delivery is the time available in class to teach the subject. This is important not only for vocabulary teaching but for all other elements of the curriculum. With vocabulary, however, research gives the curriculum designer a good idea of how quickly words can be acquired by learners through normal school learning. It is possible, then, to calculate, on a principled basis rather than by mere guesswork, the teaching time that is needed to deliver the curriculum. None of this information is included in the OCRR. In effective learning environments, 3–4 words per contact hour are generally acquired. In poor learning environments, uptake may be limited to only 1–2 words per contact hour (Milton and Meara 1998). Calculations of uptake of French vocabulary in the UK puts uptake at less than one word per contact hour in Years 8 and 9 of secondary school (Milton 2006). To give an idea of what this means, even in an effective learning environment, a learner might need 2000–3000 hours of learning to reach the vocabulary levels associated with expert users. In the UK classroom, time allocated for GCSE language is probably only about 10% of this, and this might be divided between two languages. The classroom time available for teaching language in UK is small compared with language teaching elsewhere and this is one of the factors which may help explain UK learners’ low levels of attainment (Milton and Meara 1998). The OCRR appears to recognise this (p. 5). However, the OCRR does not illustrate the teaching time available elsewhere and which might guide curriculum requirements in UK. The Goethe Institute (Goethe Institute 2022), for example, suggests up to about 500 classroom hours might be needed to take learners to B1 level in German. In Saudi Arabia, nearly 600 classroom hours are made available to bring learners to the cusp of A2 and B1 level, and to knowledge of 2200 words, in English (Ministry of Education in Saudi Arabia 2016). These suggest two or three times the timetable provision that is available in the UK to take learners to an equivalent level of knowledge.

Research also shows that, in UK, even the time that is available for teaching is poorly exploited. The rate at which vocabulary is taught per classroom hour is a fraction of that elsewhere. The OCRR should have explained that effective vocabulary loading is usually something like 8–12 words per classroom hour, but in the UK can be as small as 3 words per contact hour (Milton and Hopwood 2021, 2022). Yet, the kind of information which could inform the construction of an effective curriculum for schools is entirely absent from the OCRR. Learners in Britain and are not uniquely unable in language learning; they are, however, the victims of a curriculum that appears uniquely low in teaching time and uniquely low in vocabulary content. If taking learners to ‘expert’ level really is the goal of the curriculum, then it is beyond believable that the enormous change in teaching time required has just been overlooked.

Vocabulary selection and frequency

Over half the OCRR section on vocabulary (9) is devoted to issues that might broadly be termed the selection of vocabulary. How should curriculum designers choose which words they want to teach, rather than how many words, which – as noted above – is not considered at all. The OCRR is, at best, half-hearted about anything other than a frequency criterion for selection. It includes the observation that the most frequent 2000 words in a language can provide, generally speaking, about 80% coverage of most texts. These most frequent words are important, therefore, and frequency should be a criterion in word selection. So far, so good, but a little learning is a dangerous thing; the context of this observation needs to be understood if it is not to be misinterpreted. It can be misused, for example, to justify a teaching lexicon that comprises almost exclusively these frequent words (DfE 2021, 2022). The origins of this observation go back to the creation of the earliest corpora and frequency lists where the concern was that the most frequent words were typically under-represented in teaching texts, which would thus undermine the teaching of a communicative lexicon (Milton and Benn 1933). Analysis of recent textbooks (Alexiou and Konstantakis 2009; Milton and Hopwood 2021; Tschichold 2012), however, suggests that the criterion of frequency is now over-applied, which has an equally damaging impact on the learning of the lexicon. The latest GCSE Subject Content specification for languages in England (DfE 2021, 2022), which adopts an almost exclusively frequency-based approach to vocabulary selection, has been criticised because it excludes a huge proportion of the lexicon which is needed for communication (Milton and Hopwood 2021). Other parts of the OCRR, nonetheless, reinforce the frequency bias without acknowledging this. Words that are polysemous and which have a wide range are recommended for teaching. These words are often frequent and it is their qualities of opacity in meaning and polysemy which allow them to feature in a wide variety of structures and forms (Milton and Hopwood 2022).

Less frequent words are dismissed with the comment that they ‘can be useful’ (OCRR: 11). It is suggested that these words might really be relevant only for ‘the individual learner and the kinds of situations they are likely to need language for’ (p. 11). Curriculum designers are invited to ‘consider carefully which topic-based vocabulary (other than high-frequency words) they teach’ (p. 12) as though this material were better avoided. The OCRR guidance is clear that these words are not a priority in learning. It cannot be stressed too strongly how misleading these comments are, and for two reasons. One is that they misrepresent the scale of the lexicon that is needed to be an expert language user or even to be simply communicative. If an ‘expert’ user needs at least 8000–9000 words, the overwhelming majority of these words will be infrequent words beyond the 2000 word range. These words are not an optional part of being an effective language user; they are essential. It is a myth that learners can be communicatively expert with a lexicon of fewer than 2000 almost exclusively frequent words (as proposed in DfE 2021 and the final GCSE Subject Content, DfE 2022) and this is demonstrated in Hu and Nation (2000). No amount of substitution and circumlocution can overcome the lack of words, and the lack of the right words, needed for communication. Is Ofsted, therefore, implying in the OCRR, that communication is no longer a goal of language learning in school? The advice to curriculum designers should be that substantial volumes of less frequent words must be included, appropriate to the CEFR goal of language learning, as well as the high-frequency words. This advice is contradicted.

The second, and related, reason is that less frequent words are typically the content and subject words of language. Language and communication are always about something; they have content. These less frequent words are needed for that content. The most frequent words tend to be structure and function words which, of course, are generally useful in the contribution they make to the structuring of language. Usually, however, they cannot convey the specific information that communication is about. Milton and Hopwood (2022) give an example that may help illustrate this. They hypothesise the learner of French may want to ask for a cup of tea. Neither cup (tasse) nor tea (thé) falls within the most frequent 2000 words of the Lonsdale and LeBras (2009) frequency dictionary of French. There are only two items of food and drink in the top 1000 words in this list, eau and fruit, and only two more in the top 2000 words, poisson and café. No amount of invention, no strategies, are generally good enough to make up for the absence of this content vocabulary where it is needed, and it is needed in almost everything we say and write. I once arrived in Outer Mongolia at a hotel where everything was written in Chinese and no one spoke English. I wanted a cold beer, but I did not know the words for this. After inventive use of my Chinese – with much circumlocution and some acting out – I was served an omelette.

The balance of frequent and infrequent words in the curriculum for basic-level learners has received research attention, but this is nowhere referred to in the OCRR. Effective materials generally have something like a 50/50 split of frequent and infrequent vocabulary (Milton 2009). It is a balance that allows learners to acquire structure, to learn the words needed for communication, and to begin to be communicative within the thematic areas which are taught. Materials with a heavy bias towards frequent words appear much less effective (Milton 2009, 2011). Curriculum writers need to understand this kind of requirement and ensure an effective balance of frequent and infrequent words is presented. The selection criteria suggested in this review give a very strong steer away from this and are completely inadequate for the construction of a curriculum for communication or for ‘expert’ level performance.

Vocabulary selection and thematic content

The flip side to advocating a curriculum that is heavily weighted to frequent vocabulary is that less frequent vocabulary will have to be excluded, and the lexicon itself will have to be comparatively small. A wide range of themes and topics will generally require a lot of this less frequent vocabulary, so avoiding thematic content becomes a necessary part of a frequency-based vocabulary curriculum. Not surprisingly, then, the OCRR is not encouraging about a thematic organisation of the curriculum. It simply states: ‘Many schemes of work and textbooks are based around topic areas like hobbies and holidays. Themes and topics can be practical ways of organising vocabulary’ (p. 9). The review acknowledges that exam specifications often use themes but suggests that: ‘Other ways of organising vocabulary lists are also possible, such as by part of speech (noun, verb and so on) or alphabetical order’ (9). It is hard to avoid the impression that OCRR is giving a very clear steer away from themes and thematic vocabulary. The curriculum advice given is that frequent words must be prioritised but that thematic vocabulary needs only to be considered (p. 10) and are not a priority, therefore. As a principle for curriculum design, this seems nothing less than absurd. The choice of words is what gives language meaning, so it is surprising that a review of this kind might seriously suggest that the meaningful content of the curriculum should be discarded and replaced by decontextualised lists of words organised by part of speech. Had the OCRR reviewed more widely and accurately the research and background to the effective teaching of vocabulary, it would have noted the many advantages to thematic organisation of the curriculum.

First among these advantages is the evidence that shows how effective a wide choice of thematic content is (for example, Vassiliu 2001), and how removing this is damaging to language learning (Milton 2011). Also, thematic organisation of teaching materials also allows for an effective, 50/50 balance of frequent and infrequent vocabulary to be taught and this promotes vocabulary knowledge across the frequency bands, which produces a lexicon suitable for communication. Further, a good choice of topical and thematic content can provide the subject matter for meaningful language use. A wide variety of themes allows for language content that is interesting and on topics that are relevant to learners and thus both motivating and useful. A feature of the research into UK textbooks for teaching French in school, which teach a small vocabulary around a very limited selection of topics, is the conclusion that the materials are dull, repetitious and demotivating (Häcker 2008; Milton and Hopwood 2021; Tschichold 2012).

It cannot be stressed too strongly that effective communication needs much more vocabulary than the 2000 most frequent words referred to in the OCRR and which is the prescribed vocabulary content of teaching materials in the DfE’s latest GCSE Subject Content for languages (DfE 2022). If expert users know 8000–9000 words in their foreign language, then the overwhelming majority of this knowledge will be infrequent words and will cover a wide range of themes and topics. If the curriculum does not aim to teach such words - remember the OCRR is quite clear these words are not a priority - how are they to be learned? The high frequency and anti-thematic priority put forward by the OCRR ensures learners will never become ‘expert’ language users through learning in school.

Vocabulary and learning

The lexicon needed for communication is not just large; it is also complex. There is, generally, a lot of knowledge associated with each word the lexicon contains. The OCRR, therefore, introduces Daller et al.’s (2007) three-dimensional framework of vocabulary knowledge, comprising the dimensions of breadth, depth and fluency, which might have brought order to this complexity. It is not clear why this analysis is introduced in the middle of the review rather than at the beginning, where it might have brought better organisation. It is not systematically applied. Had it been systematically applied, the OCRR would have had to address vocabulary breadth or size, the first of the three dimensions. The need to teach word forms receptively and productively is mentioned, and both orally and in writing. However, as noted above, the number of words needed to become an ‘expert’ user is not discussed at all. A principal feature of the OCRR, then, is the guidance that the curriculum should prioritise, overwhelmingly, only frequent words, and verbs. The lexicon which is taught, then, has to be small. A small vocabulary size means that the goal of learning is for learners to achieve only elementary levels of performance. A much larger vocabulary, including thousands of less frequent and content words, is needed to be ‘expert’. If a large vocabulary were expected, then these thousands of less frequent words would be prioritised too.

The OCRR is much more interested in the second of the dimensions, that of vocabulary depth. Vocabulary depth is usually characterised as how words link together in the lexicon; their associations, their collocations, their subtleties of use and restrictions on use. Vocabulary depth can be characterised as the number of word links, where the greater the depth of knowledge then the greater the number of links between words (as in Wolter 2005). This is important because a condition of having a lot of links between words is to have a lot of words to link (Vermeer 2004). The dimensions of knowledge are not separate components but are so closely connected that it can even be suggested that they approach collinearity (Fitzpatrick and Milton 2014). The OCRR is unaware of this. Developing a good depth of knowledge and good fluency is absolutely dependent on developing a large vocabulary of thousands of words. But, of course, the goal of developing a large vocabulary size is quite clearly excluded by OCRR’s curriculum guidance.

The OCRR proposes the curriculum should include knowledge of morphological derivation; knowledge of semantic networks, lexical sets and word families; knowledge of antonyms; knowledge of synonyms; knowledge of shades of meaning and context-related usage. The choice of these items is not explained or justified. Why is knowledge of the rules for morphological derivation included, for example, but the rules for grammatical inflection omitted? There is not sufficient time to go into this here, but this is a huge oversight in terms of the development of a foreign language lexicon. If there is a common feature to the qualities of depth that OCRR includes, it is that they are characteristic of ‘expert’ language users. The rules for derivation, for example, are so complex and difficult that Hilton (2019) places their mastery at the very end of the learning process; it is part of the fourth of her four propositions. A substantial vocabulary of several thousand words is required before this kind of knowledge can be systematically addressed because only at that level can such a system be exemplified in the learner’s lexicon. It is suggested that there is a threshold of about 5000 words before this type of knowledge can begin to emerge (Martinez 2013). This certainly cannot be done with a small number of only highly frequent words and it is quite inappropriate for current learners before GCSE who only know a few hundred words. There is a contradiction, therefore, in OCRR’s guidance where a small lexicon prioritising highly frequent items is required but, simultaneously, a very large lexicon of many thousands of words is also required for knowledge of the qualities of lexical depth it proposes. The guidance OCRR proposes is confused and undeliverable. This contradiction and confusion is not a feature of the literature or research in vocabulary learning. It is entirely a product of OCRR’s lack of knowledge and understanding of vocabulary and the vocabulary learning process.

This discussion raises a misconception in the OCRR about the nature of developing a foreign language lexicon. The impression is distinctly given in relation to the growth of semantic networks and lexical sets, that the learner is acquiring a foreign language lexicon that is separate and different from the first language lexicon and with no overlap of knowledge. The same impression is given by the OCRR’s treatment of phonics and the reading process (p. 10). The OCRR refers to these networks and sets emerging with the growth of vocabulary knowledge. As Henriksen (2008) shows, however, this kind of conceptual knowledge is transferred across languages: the moon comes out at night, it shines and is associated with romance in most languages. Overwhelmingly, learners are not learning new clusters or sets; they are relabelling the lexical entries in existing ones. This has implications for the curriculum because it transfers the focus of teaching from lexical depth to one of breadth.

The OCRR, then, does not give coherent guidance on the construction of a vocabulary curriculum. It lists some, but only some, of the elements which a curriculum should contain. It includes other things, like the prioritisation of verbs, which really should not be there at all. And, incomprehensibly, it omits many more that are crucial for success, such as teaching a lexicon of an appropriate size and quality, and the teaching of grammatical inflections for items in the lexicon. Its priorities suggest a profound misunderstanding of the research literature, and no appreciation of successful practice in other countries. Its guidance cannot lead to the creation of a successful curriculum calculated to raise learners to the ‘expert’ levels the OCRR expects.

Other features

In addition to these considerations, there are other features of the OCRR, and not necessarily just in the vocabulary section, which should be considered in relation to vocabulary.

In the OCRR’s summary section on vocabulary (p. 10), the teaching of verbs, in particular, is strongly advised in the early stages of learning, but no evidence is referenced in support of this. Milton and Hopwood (2022) list this among what they refer to as ‘British vocabulary myths’. Milton (2009) considers, but is inclined to dismiss, the idea that a verb focus can promote language learning. What evidence there is suggests that knowledge of nouns and adjectives is more important to comprehension than the knowledge of verbs. The noun focus that is observed in the language production of low-level learners (Marsden and David 2008), is a feature of learners with a very small vocabulary. The low-level learners in the Marsden and David study probably knew 100–200 words, so low they could scarcely string two words together in original production. It is a feature of the literature on learners with very small lexicons that the words they contain tend to be concrete words, like nouns, rather than abstract words (Alexiou and Konstantakis 2009; Cameron 2001). Marsden and David’s more advanced learners had a rather larger vocabulary and could produce longer utterances, which might have included verbs. The acquisition of verbs is not driving the improvement in level; it is a reflection of the growth of the lexicon as a whole, and that is what is driving the improvement. Yet again, this is an illustration of how the OCRR fails to understand the importance of a large lexicon to learning a language, and how a large lexicon is learned.

A further point relates to the multiple references to ‘memory’ that are made in the OCRR as a whole. Memory is important in vocabulary learning. Memorisation of word forms is the starting point of learning a lexicon and we know that differences in memory will affect language learning (Meara et al. 2001). It does not seem to me, however, that the OCRR helps teachers or curriculum designers by introducing concerns about the limitations of short-term memory without really explaining what this is and why it might be important. The short-term memory that interests researchers in assessing vocabulary learning is rather different to the understanding of teachers, who use this as a reason for limiting the vocabulary taught. Teachers imagine that too much vocabulary will overwhelm the memories of learners. The OCRR should probably have included, then, the research that shows that prioritising vocabulary learning, and increasing vocabulary input, seems to benefit language learning (Vassiliu 2001). The volumes of vocabulary input that associate with high uptake should have been considered (Milton 2011; Milton and Hopwood 2021). It would have been useful to point out that intensive courses of language study can routinely introduce 60 words a day, every day, to move learners to communicative ability really quickly, but without overwhelming their memories. The OCRR’s overconfident use of the terms ‘short-term memory’ and ‘long-term memory’, like the references to ‘cognitive psychology’ found throughout the review, display little understanding.

A third point is to note the way the OCRR, including the vocabulary section, pays no heed to the range of individual variation to be found in learners. Learners do not all set about the process of learning in the same way and they do not make uniform progress. One of the least satisfactory elements of the current curriculum, and of the way the OCRR is slanted, is the way it implies constraints on the progress and attainment of the most able learners. Where a successful curriculum presents something like 6000 words to learners at A2 level so that, on average, 50% of these words will be learned, the other 50% is not wasted. This is a resource that the more able learners can tap into and the best learners will learn many, sometimes even most, of these words. A good curriculum allows these learners to make the most of their talents as well as making provision for the rest of the range of ability. The limitations on the vocabulary content of the curriculum in the OCRR, make that impossible. Milton (2008) shows how reducing the size of vocabulary in the curriculum for French has severely limited the performance of the most able learners at GCSE. They attain a fraction of the knowledge they would have done under earlier curriculums. Why is Ofsted, in the OCRR, so keen to limit the attainment of the most able learners?

There are a couple of final points which have to be made in relation to the quality of the OCRR, and the section on vocabulary in particular. One is to question where the information it contains comes from. The references are not representative of research in the area of vocabulary and suggest that the writers have consulted none of the principal texts on the subject (for example, Milton 2009; Nation 2001; Schmitt 2008 and 2010), nor read extensively around the subject. I am not the first person to question the integrity of the Ofsted research review process (see, for example, Schools Week 2021). There are no books on vocabulary teaching and learning referred to at all. A good book on vocabulary can bring, amongst other things, coherence to a very complicated area of language knowledge. Coherence is something that is missing in the OCRR section on vocabulary; rather, it seems to be an assemblage of ideas partially drawn – and only partially drawn – from a limited and unsystematically chosen set of sources. There also appears to be reason to doubt how carefully the OCRR has considered the literature it has used. The three-dimensional model of lexical space (Daller et al. 2007, mentioned above) is an example of this. The model is attributed to a secondary source, Milton (2013a) on vocabulary and the four skills, but if this is important enough to include here, why has the original treatment not been sourced? Furthermore, the ‘dimensions’ are misnamed as ‘components’, which gives a very different connotation to the idea and not one the authors intended. Even in the secondary source, these are clearly referred to – 17 times – as ‘dimensions’ and never as ‘components’. Furthermore, the OCRR ignores completely the repeated references in Milton (2013a) to the way a large vocabulary size is crucial in determining proficiency in the four skills. These issues so undermine the credibility of the OCRR that its conclusions, of themselves, cannot be believed.

Discussion

This paper began with a reality check which pointed out the vast disparity between the ‘expert’ levels of language performance that the OCRR seems to suggest learners can, and will, attain at school, and the actual levels that learners achieve. As noted above, learners of French at GCSE have been found to have only a few hundred words of target language knowledge and are, effectively, uncommunicative. This paper began by asking the question whether the goal of the OCRR is to raise the standard of language teaching in England by helping to define new goals for the curriculum or whether the writers really have no idea of the level of language achieved in schools. Does the guidance provided here direct a process of curriculum renewal to raise standards, or does it cement in place the historically low levels of foreign language attainment now apparent in UK schools? The answer is surely the latter, and the OCRR’s review of vocabulary research suggests limited understanding both about the nature of the lexicon, about language itself, and the about the way languages are learned for communication. Its guidance could not be better calculated to ensure learners never progress beyond CEFR A1 level.

At the heart of this is the OCRR’s guidance that a curriculum should prioritise only highly frequent words. This affirms, but without any justification from research evidence, that current learning of only a few hundred words is sufficient for communication on a wide range of topics. The latest Subject Content proposals, it will be remembered, require that the vocabulary content of the curriculum is limited to between 1200 and 1700 words overwhelmingly drawn from the most frequent 2000 words (DfE 2022). A consequence of prioritising this small number of frequent words is that all other words are, necessarily, deprioritised. If these words are not a priority, then why should anyone teach them? So, textbooks now routinely teach a small, predominantly frequency-biased lexicon of about 1700 words (Milton and Hopwood 2021). Learners, then, learn a small vocabulary and can never progress beyond CEFR A1 level, which is a consequence of such little knowledge. This frequency priority for vocabulary means the goal of the curriculum will be, and can only ever be, A1 level.

It is entirely misleading to suggest, therefore, as the OCRR does, that curriculum designers should consider teaching elements of the lexicon such as subtleties of meaning and usage in communication, where learners are at A1 level and will struggle to communicate at all. This type of knowledge and ability can only emerge with the growth of a much larger lexicon of thousands of words. If A1 is the goal of learning, and remember it is the bottom level of the CEFR hierarchy, then the curriculum needs other things. Guidance should include a range of vocabulary across the frequency bands, a range of thematic choices for interest and motivation, rules for grammatical inflection which are so important to recognition of word form, and a range of prepositional structures from within the lexicon being taught. These are features of successful foreign language teaching outside the UK. None of this emerges from the OCRR’s guidance. Because the goal of learning is A1 level, the OCRR can avoid making any mention of classroom time for learning. The current and minimal timetable provision for language learning in schools, like setting small vocabulary goals, fits with low expectation in progress and A1 level attainment. The OCRR appears to recognise that the standards of learners are low compared to learners in other countries and that teaching time is a factor in this. However, it is utterly at a loss in knowing what to do about it. But, if the curriculum teaches A1 level vocabulary content and devotes A1 level time to learning, then the learning outcome will be A1.

The OCRR, however, asserts that learners will attain ‘expert’ levels of knowledge and language usage; presumably C1 or C2 level in the CEFR. Here, the curriculum will need to be very different. For that to come about, there cannot be a frequency priority in the choice of vocabulary. To be ‘expert’ the curriculum will need to organise the learning of some 8000 or 9000 words. Overwhelmingly these will be infrequent. The teaching of the most frequent words is essential, of course, but so is the teaching of thousands of less frequent words. All of this vocabulary is a priority and for the OCRR to suggest that it is not, is entirely wrong. Principal texts on vocabulary learning (for example, Milton 2009; Nation 2001; Schmitt 2008 and 2010 – the principal texts which, as noted above, are never mentioned in this review) all repeat vocabulary learning goals of this order and never suggest learners can be communicative, across a wide range of topics, with only a few hundred, highly frequent, words. As part of the teaching of a lexicon of this scale, the guidance can include a greater focus on elements of vocabulary depth, as a sizable lexicon develops. The OCRR, however, gives no clue as to how these ideas should be sequenced, in line with the growth of vocabulary size, for learning to be effective. There are other elements of curriculum guidance that will be needed if the goal is for learners to attain ‘expert’ levels of performance. Principal among these will be teaching time sufficient to include all of this vocabulary. I have already noted that other countries routinely expect and deliver double or even triple the number of learning hours available in UK, to take learners just to A2 and B1 levels of performance. If the goal is for learners to become genuinely ‘expert’ then even more timetable hours are required. Other countries are able to produce sensible learning time guidelines to manage the delivery of an effective curriculum. The OCRR does not do this. A final point might be to note that if the curriculum is intended to guide learners to ‘expert’ levels of performance then all school teaching materials will need to be rewritten since they currently teach to a much lower level. The OCRR gives no idea as to how this might be done.

It is actually hard to avoid issues of vocabulary size in any review of vocabulary and language learning because it is so central to the understanding of the learning process and there is so much of this research. And yet, it is avoided in the OCRR, so there must surely be a reason for this. The review seems to cherry-pick ideas and elements of research to support an argument to avoid vocabulary learning, rather than truthfully representing the whole field of research, which repeatedly suggests a large vocabulary should be prioritised. I think there are two ideas behind the production of a review of this kind.

One reason is that this ‘review’ of the research evidence for vocabulary gives an appearance of respectability to current teaching practice and maintains the curricular status quo. Current teaching practice is to minimise vocabulary teaching and avoid thematic content (DfE 2021). The OCRR repeats this nostrum despite its manifest inadequacy. It is thus a pretence that current practice is a product of the best and latest research and will lead to teaching that is as effective as it can be. Rather, this is an approach that is failing the learners. By avoiding embarrassing issues of the size of vocabulary taught, it can maintain a pretence that the standard of learning is much higher than it really is. Learners at school, even at A level, do not routinely progress beyond elementary levels of performance, while the OCRR, unbelievably, talks in terms of these learners becoming ‘expert’. It is also failing the teachers who on the basis of the OCRR, quite perversely, are likely to be rewarded for delivering, in class, bad practice leading to poor learning.

The second reason is that it fits with a structuralist approach to language and language teaching and learning. A structuralist approach is one where it is believed that communication relies on structural knowledge. Within this approach, Wilkins (1972) suggests the lexicon is only important insofar as it is needed to illustrate the grammar of language. It is an approach to teaching which dropped out of fashion almost everywhere some 50 years ago because it was both theoretically unsound and ineffective when put into practice (Milton 2009). Van Ek and Trim (1990: 1), for example, explain the growth of communicative methodologies, and their own Council of Europe Waystage and Threshold A2/B1 level materials back in the 1980s, as an attempt to, ‘convert language teaching from structure-dominated scholastic sterility’. Milton and Hopwood (2022) argue that, despite its much-highlighted deficiencies, structuralism is an approach that has taken over the curriculum and language teaching in UK schools. It has become central to a particularly British culture of language teaching where the goal of the curriculum has become the teaching of structural knowledge, rather than of a communicative language and the vocabulary required for this. It is a culture which does not see the confusion between having enough vocabulary to give examples of grammar, which might be possible with a relatively small lexicon, and having enough vocabulary for independent communication, which is impossible without a large lexicon. It is an approach that does not understand that illustrating language use is not the same as using a language for communication.

Within this paradigm, it makes perfect sense to highlight the importance of verbs in vocabulary teaching. The presence of verbs allows rules for grammatical inflection and morphemic derivation to be taught even if this will divert effort from learning a communicative lexicon. The teaching of something like phonology under the guise of ‘phonics’ also fits with this approach and is promoted as best practice even when, as noted in Oftsed’s own inspectors’ reports (Wardle 2021a), this practice can result in poor progress. Teaching phonology is teaching the structure of a sound system. The same inspector’s report, it should be noted, demonises the practice of teaching words through topics in primary school, something that will almost certainly lead to rather better progress. It is a paradigm that sees no problem in excluding most of the vocabulary from the curriculum in order to lay greater emphasis on the teaching of grammatical concepts (DfE 2021; Teaching Schools Council 2016).

The recently revised Subject Content for teaching language in school devotes 85% of its description to grammar (DfE 2021; see also DfE 2022). Wardle (2021b) explains – and he is Ofsted’s lead inspector (HMI) for languages – that learning more words does not promote fluency. Nothing could be further from the truth, of course, but this idea seems to have become a tenet of belief even in the languages inspectorate. Thus it is, that a culture seems to have developed which rewards structural accuracy in language assessment rather than communicative ability (Milton and Hopwood 2022). Milton and Hopwood (2022) describe this as a culture of vocabulary learning myth that suggests minimal investment in vocabulary learning even though it is crucial for language learning. This culture can produce a training content for teachers that encourages the avoidance of anything other than frequent vocabulary and includes the mistaken advice that the unknown words can be avoided through strategy use and circumlocution (Woore 2020). This would be laughable advice if it were not so serious; a generation of language learners no longer has a realistic opportunity to learn a foreign language at school. It is a culture that means a decline of 75% in the standard of vocabulary attainment by learners can pass entirely unnoticed by Ofsted.

This culture is particularly misleading and damaging, as the reality check at the start of this paper shows. It has driven down the content and standard of language teaching, while at the same time making it possible to assert that standards are rising. Milton (2013) quotes Anthony Adonis as Schools Minister in 2006 saying, ‘continued progress in exam performance is real – it is not the result of dumbing down of standards – and the roots of this success lie in a fundamental shift in the quality of teaching in our schools’. He was referring to a rise in the proportions of students passing language exams and to a higher proportion of pupils gaining the top grades and the accusation in the press (reported in Milton 2013b) that this was a product of declining standards. Between 1990 and 2006, however, Milton records a decline in teaching time and a 75% decline in the vocabulary attainment of learners. The culture of language teaching in UK may like to pretend that standards are rising, but the cold, hard light of vocabulary measurement is quite clear in showing that attainment has declined massively and is now low. Even the OCRR notes that pupils report their language attainment compares unfavourably with their European counterparts (p. 5).

As a product of its culture, therefore, the OCRR fails to identify what an effective vocabulary curriculum should look like. It talks in vague terms of learners becoming ‘expert’, but without defining what this means, for example, in terms of the CEFR. It also, and this is almost the only piece of clear guidance it provides, prioritises only frequent vocabulary. I have laboured this point repeatedly in this paper because the consequences of this guidance are not generally appreciated. A small lexicon comprising only frequent words guarantees that even the most able and motivated learner can never progress beyond elementary levels of language performance. The guidance which could really take learners to ‘expert’ levels is absent in the OCRR. Guidance in vocabulary should be providing something precise and detailed which can help define the content of learning for syllabus and materials writers, and for teachers, so teaching is appropriate, coherent and equivalent across all schools and colleges and among all examination bodies. The OCRR should be clear what the goal of learning is, in terms of CEFR communicative levels. It should define the size of vocabulary, and the spread of vocabulary across the frequency bands, which can deliver the goals. It should probably advise on the choice of themes and topics which can best deliver the content for the age and other features of the target learners. It will probably need to define a core set of topics and vocabulary list around which materials can be built so, for example, examining bodies can produce assessment systems that are as fair to all learners as can reasonably be made. It needs to define the time expected for teaching, and intermediate goals for attainment to check progress. None of these is present. The literature on vocabulary learning, and effective practice in other countries, are both rich in the materials which can inform a well-structured vocabulary curriculum. None of this is evident in the OCRR.

Creating an effective vocabulary learning curriculum is a large and complicated business. The research which should underpin its creation is also large and complicated. The guidance which should emerge from this, if it is to be useful, is also likely to be large-scale, detailed and complex. Why, then, would Ofsted think it is a good idea to try to compress all this into less than two pages? Its coverage could only ever be, at best, partial and superficial and is always likely to be misleading. It is hard to escape the conclusion that the small scale of this review, its almost total lack of useful guidance, and its lack of engagement with the literature are deliberate choices. A research review with these qualities can appear to endorse current GCSE Subject Content (DfE 2021, 2022), and current teaching practice, where only minimal vocabulary teaching is a requirement. It can create a false veneer of legitimacy for a curriculum design that is unsupportable both in its content and its outcomes. It can do this, and it does do this, by omitting any of the research evidence that challenges current approaches. A more substantial and less tendentious review would have exposed these approaches as deeply flawed and damaging to language learning. Make no mistake, this is a very destructive document for the teaching of languages in schools. By misrepresenting research in vocabulary, it promotes a curriculum that is inadequate in content, it supports poor teaching practice, and it ensures low levels of learner attainment.

Conclusion

The vocabulary element of the OCRR is a review in name only. It is tendentious and limited in its coverage. It appears as a product of its structuralist culture, doing little to fulfil its goals of exploring the extensive literature relating to vocabulary learning and teaching, and identifying the features of effective vocabulary curriculums. It makes no effort to challenge existing curriculum practice, despite the fact this is so manifestly inappropriate and unsuccessful. It does not promote the features of good practice, to be found elsewhere and that might usefully be copied. It does not even begin to identify the vocabulary content of a curriculum needed for communication. It represents instead existing bad practice as though it were good practice. It cannot possibly lead to the development of a communicative lexicon. It air-brushes out of existence the issue of vocabulary size, which is so important to language learning that language learning must fail without it. It promotes an approach to language that has little to do with learning a language for communication and can never help language learners to become ‘expert’. It appears instead to be dedicated to the false god of abstract structural learning and the structure-dominated ‘scholastic sterility’ which, even 50 years ago, was criticised as undermining successful language teaching and learning.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The General Certificate of Education ‘O level’ examinations were taken by English pupils typically at age 16 up until 1987. They were then replaced by the General Certificate of Secondary Education (GCSE) examinations.