During my MA TEFL at Reading University, nearly 20 years ago, I stumbled into a book called ‘Words in the mind’ by  Aitchison’ (1986; but latest edition: 2012). That book changed the way I teach vocabulary forever because understanding the way our brain stores, organizes and forgets the words we learn meant being able to come up with strategies to speed up and consolidate lexical learning. In this article, I intend to share some of the knowledge I acquired from that book and through many other subsequent readings (e.g. McCarthy, 1990, Eysenck,2000; Nation, 2001; Macaro, 2007) and how it can enhance L2 vocabulary acquisition. Although I intend to discuss the implications for the classroom, I will do so very concisely, reserving to elaborate on them in a future article, for reasons of space. Before discussing how vocabulary is organized in Long-term memory (LTM) one need to understand a few important facts about it.

1.Long-term memory (LTM) and Spread of Activation

As you may know, once information is learnt, it is stored in LTM, a vast neural network connecting every single piece of information we have acquired in our lives. Thus, in actual fact, our LTM makes us what we are as it contains all our emotional and sensorial experiences, every cognitive and motor skill we have learnt and, basically, all we know about the world, including lexis and grammar rules.

The LTM ‘space’ where we store lexical items is referred to as ‘Mental lexicon’. Contrary to what scientist believed in the past, any information that makes it to LTM, is stored there permanently, and forgetting does not occur due to decay of the memory trace (see below).

When we need to translate a given ‘thought’ (or ‘proposition’, as psycholinguists call it) into words, the brain fires electrical impulses which travel at very high speed through LTM’s neural pathways in search of the words that match that thought. During this process, every single word associated with that thought receives activation.

2.1 How first language words are organized in our brain

When a lexical item is stored in LTM, the brain does not place it in just any random place along our neural networks. Insight form research on the slip-of-the-tongue phenomenon and aphasia indicates that the neural connections between the lexical items in our mental lexicon are determined by specific associative mechanisms which involve the physical aspect of a word as well as the metalinguistic, semantic, sociolinguistic and emotional domain.

2.2 Physical associations

Words are associated at the ‘physical level’ based on their spelling (graphemic level) and sound (phonological representation). Thus, words that look and sound similar (alliterate, rhyme and chime with each other) are more likely to be very strongly associated. Consequently, when our brain (our Working Memory) attempts to retrieve the word ‘dog’ from LTM, for example, and activation spreads in order to ‘fetch’ it, all the monosyllabic  words starting with ‘d’ and ending in ‘g’ will receive strong activation (e.g. Doug, dig, door, etc.). Interestingly, even the anagram of ‘dog’, ‘god’ will be highly activated.

This phenomenon explain slip-of-the-tongue errors, which are basically ‘computing mistakes’ often due to processing inefficiency, whereby instead of retrieving the word we need, we retrieve a ‘near homophone’. That’s why alliterations, rhymes, para-rhymes and other phonetic devices used by prose writers and poets are so effective in reinforcing the impact of two words in their texts which are already related in terms of meaning and thereby receive greater emphasis by their phonological connection.

2.3 Semantic Association (Field theory)

Words are very strongly linked to each other, based on their meaning (Field theory). Synonyms and other words that refer to items frequently associated in real life will also receive strong activation during the retrieval process. Going back to the ‘dog’ example, words like ‘pet, ‘bone, ‘puppy, ‘tail’ and ‘bite’, amongst others, will be activated during the retrieval process, each receiving more or less activation in our brain depending on: (1) how often I will have processed (receptively or productively) those words in conjunction with the word ‘dog’ in the past; (2) how frequently, in my personal life, the items those words refer to, are associated with the notion of ‘dog’.

Semantic associations will also be affected by the connotative meaning that a specific culture of sub-culture attaches to it. Thus, whereas the word ‘fox’ is associate both in Italian and English with the notion of ‘shrewdness ‘ and consequently to the related nouns and adjectives, the word ‘chicken’ will be related to cowardice in English but to gullibility in Italian.

2.4 Linguistic context

This point sort of relates to the previous one but deserves separate treatment because it specifically refers to the linguistic contexts in which two or three given words are used in a specific language and which may differ across languages. So for instance, the word ‘dog’ will bring about different associations to an English native speaker’s brain compared to, say, an Italian native speaker’s by virtue of the linguistic context they are found in a number of set phrases/idioms. An English person will associate ‘dog’ with the phrase ‘a dog’s life’ or ‘to work like a dog’ for example; an Italian, on the other hand will associate it with the idiom ‘solo come un cane’ (‘as lonely as a dog’) or ‘fa un freddo cane’ (‘it’s freezing’ or literally: ‘it’s dog cold’).

2.5 Word-class

Words are also organized by word-class, adjectives with adjectives, nouns with nouns, etc.

2.6 Emotional and sensorial connections

Every lexical item is also strongly associated to personal experiences and memories stored in our Episodic Memory. So if we had a very traumatic experience in our life which involves a dog (being bitten or scared by one when we were small, for example) ‘dog’ will evoke strong negative emotions and words describing objects, people or feelings related to that traumatic experience will receive strong activation.

Words will also be associated with sensorial perceptions (taste, smell, images, etc.)  based on one’s life experiences.

3. The foreign language mental lexicon

In a fluent foreign language learner with a sizeable vocabulary repertoire, the way words are stored in their L2 mental lexicon will be pretty much the same, except that there is another very important association, the one between an L2 word and its L1 (and L3,L4, etc.) translation(s). So the word ‘dog’ in the brain of a speaker of Italian, French and German will be connected with the words ‘chien’ , ‘cane’, hund, etc.  Consequently, when spread of activation occurs in search for the word ‘dog’ in one language, say ‘French’, all the words in the other languages will be activated too (Parallel activation theory); all languages one speaks will be activated simultaneously with different levels of activation, with the language in use being the most activated, and the weaker language(s) being the least activated. This explains the phenomenon whereby some learners when experiencing cognitive-processing issues in the target language, will retrieve an L1 word instead of its target language equivalent.

When the foreign language learner is not fluent, there will be fewer L2-to-L2 word connections as the mental lexicon will be smaller and many of the other connections that we discussed above might not be formed as yet – since the learner might have not internalized the word-class of all the words they acquired and/or their meaning might be fuzzy. This means that when spread of activation occurs, fewer linguistic items will be activated.

The fact that in a less fluent learner with a relatively small vocabulary repertoire there are fewer and weaker connections of the kind outlined above and therefore fewer neural pathways, majorly affects recall in that the more connections we have, the more likely we are to retrieve any word we need successfully and with little cost on Working Memory efficiency. Why? Because the successful retrieval of a word depends on two factors; (a) the strength of the memory trace, that is how often we have processed that word and (b) the use of an effective cue which helps Working Memory find that information in the brain; the more the connections a word has with other information stored in LTM the greater the chances of its successful recall will be.

4. How forgetting happens

In order to better understand the implications for teaching and learning one needs to be familiar with the notion of ‘Cue-dependent forgetting’.

4.1 Cue-dependent forgetting

The reason why we often fail to retrieve a word that we learnt is usually due less to a weakening of the memory trace than to failure to find that word. The factors that determine such failure refer to the context in which that word was encoded (‘learnt’) as that very context  provides the cues crucial to its retrieval. For example: if we learn a word highlighted in red, on our teacher’s whiteboard whilst sitting near a specific classmate,the colour red, the teacher’s whiteboard and that classmate have the potential to be effective retrieval cues for that word. The absence of these three factors may prevent recall of the same word.

In the context of vocabulary learning, this implies that the more associations are created by the foreign language learner in learning a word, the more likely s/he will be to remember it, because each association will have the potential to serve as a retrieval cue.

4.2 Forgetting from consolidation

Another possible reason why we forget is that when we take in new information, a certain amount of time is necessary for changes to the nervous system to take place – the consolidation process – so that it is properly recorded. If this consolidation process is not completed we will lose the information. As I have already pointed out in my article ‘The fundamentals of vocabulary teaching’ (elsewhere on this blog), without rehearsal of the target vocabulary, 60 % of it will be forgotten within 48 hours of having ‘learnt’ it. For this reason we need to recycle the information over and over again until this information is stored permanently in LTM.

5. Pedagogic implications

In view of the way words are organized in our brain, these may be some useful teaching strategies:

• In any given lesson we ought to teach words that are as closely related as possible at semantic and grammatical level. This is often done by textbooks.
• When teaching new words, in order to facilitate their storage and recall, teachers should try as much as possible to hook them with previously learnt lexis which alliterates, chimes or rhyme with the new vocabulary. This can be turned into a game whereby students are given the task to find (under time constraints) a rhyming or alliterating word for the new target vocabulary;
• We should also ensure that, from the early stages of acquisition students are aware of the word class an item belongs to. This will provide the learner with an added retrieval cue in the recall process. For instance, students could be asked to categorize the target words into Adjectives, Nouns, Adverbs, etc. or to brainstorm as many words they learnt on the day in those categories;
• As many opportunities as possible should be found for learners to relate words, especially the challenging ones, to their personal and emotional life. For instance, whilst learning colours the students may be asked to match each colour to an emotion or physical state. Or, when learning food ask the learners to say which fruit, pastry, drink. etc. they identify with and why ( e.g. a ‘raviolo’ because I am full of goodness);
• The learners should also be involved in activities requiring them to perform more elaborate semantic associations (deep processing) between the new target vocabulary and previously learnt lexis. For instance, by asking students to create ‘lexical chains’, i.e. given two words quite far apart in meaning, learners need to produce an associative chain of lexis that links those two items logically or pseudo-logically. For example: old lady, cats, cat food, cans, aluminum, factories, pollution) This activity can be fun and does not require knowledge of complex vocabulary.
• Activities involving semantic analysis of words, such as odd one out, definitions games, sorting vocabulary into semantic categories, matching lexical items of similar or opposite meanings, should also be performed as they create further associations, although less explicit than the ones envisaged in point (5) (see http://www.language-gym.com for self-marking online examples of these);
• Teachers should be careful when teaching cognates that are graphemically or phonologically very close in the two languages. This sort of L2-cognates can be ‘tricky’ as they are so closely associated with their L1 translation that they can give rise, under processing-inefficiency conditions to the retrieval of first language form. I often experience this phenomenon (called cross-association) myself when speaking or writing in Spanish.
• Finally – this has more to do with forgetting than word storage – teachers and learners should ensure they go back/recycle the target vocabulary across as many contexts as possible and as often as possible until it has been fully acquired – especially during the two days following the initial uptake, when most of the forgetting usually occurs.

As I have already mentioned above, I will discuss the classroom implications in greater depth in a future blog post in which I will also suggest a vast array of vocabulary building activities.

Useful follow-up to this article can be found here:

(1) Ten commonly made mistakes in L2-vocabulary instruction 

(2) Thirteen steps to effective vocabulary instruction 

Please note: more on the above can also be found in the book I have recently co-authored with Steve Smith: The Language Teacher Toolkit, available for purchase at www.amazon.com