Verbal Working Memory as Emergent from Language Comprehension and Production

Steven C. Schwering, & Maryellen C. Macdonald. (2020). Verbal Working Memory as Emergent from Language Comprehension and Production. Frontiers in Human Neuroscience, 14, 68. https://doi.org/10.3389/fnhum.2020.00068

This review paper discusses current models of verbal working memory. Verbal working memory is the ability to temporarily hold aspects of language in mind (speech sounds, meaning-based information, grammar). Traditionally, verbal working memory is considered to be separate from long-term memory, that is, the information or knowledge stored in the brain over an extended period. The key issue considered in this paper is how long-term knowledge of the language influences verbal working memory.

There are two ways that long-term language knowledge might impact verbal working memory:

1.   Redintegration. Imagine you are asked to remember 3 words: ant, chair, birthday. A short time later you’re asked to recall the words and you realize you’ve forgotten the first word. You kind of remember that it was a short word for an insect, maybe starting with the “a” sound. After searching your long-term memory, you remember that the word was ant. In this way, you have used knowledge held in long-term memory to fill in the gaps in your verbal working memory. This is called redintegration.

2.   Integrated Language Model. According to more integrated accounts, verbal working memory is the temporary activation of long-term memory (rather than a separate memory store). It might be that when a word is encountered, full knowledge of the word is activated and available in working memory. This full activation is the rich emergent view. It could also be that only some of the information is activated from long-term memory whereas other information is held in working memory. This more restrictive view is the limited emergentaccount. 

There is growing evidence in favour of integrated language models. For instance, it is easier to repeat nonwords containing familiar sound sequences or familiar parts of words (e.g., -ing). Consider also the nature of language. In most sentences we say and hear, living things occur earlier than non-living things. This experience impacts the accuracy of word list recall both in terms of the likelihood of remembering a word and the order it appeared in the list. In line with the rich emergent account, memory for an item in a list and its order cannot be easily separated. Instead, it is important to understand how different aspects of a linguistic representation influence each other and are integrated during language processing.  

Viewing verbal working memory from an integrated perspective has implications for clinical tools and research moving forward. For instance, tools used to measure verbal working memory may be tapping higher language skills (encoding, maintenance, and ordering) beyond memory span or capacity. Language use (turn-taking) may draw on domain-general resources (attention, cognitive control), in addition to more language-based knowledge (understanding and planning speech). Finally, knowing that language production and comprehension involves interactions of all aspects of a linguistic representation will inform future research questions such as how interference among similar words affects verbal output or how language experience influences comprehension difficulty.

Blogger: Theresa Pham is a student in the combined SLP MClSc/PhD program, supervised by Dr. Lisa Archibald.