Our paper on familiar collocations in Broca’s aphasia is published!
Bruns, C., Varley, R., Zimmerer, V.C., Carragher, M., Brekelmans, G., & Beeke, S. (2018). “I don’t know”: a usage-based approach to familiar collocations in non-fluent aphasia. Aphasiology. DOI: 10.1080/02687038.2018.1535692
When you talk to someone with Broca’s aphasia, you will find that they predominantly produce single words when trying to get their message across. Despite this non-fluent language output, however, you may also hear short phrases such as “I went to”, “I don’t know” or “it’s alright”. These phrases are a feature of everyone’s language ‘inventory’ – we use them all the time when we converse. It is interesting that individuals with Broca’s aphasia have difficulty to produce grammatical sentences, but that such familiar phrases seem to be preserved. Familiarity of phrases is important to my PhD work, and to this paper.
The first take away point from our paper is: As the result of aphasia, speakers rely more on common word combinations than their conversation partners.
This is what our quantitative, frequency-based analysis suggests. We analysed videotaped conversations from 9 individuals with Broca’s aphasia and their regular conversation partners. One concept that tells us about familiarity of word combinations is ‘collocation strength’: the degree to which the words of a combination are associated. For example, the words in “don’t know” occur much more frequently together than “don’t drive”, which makes the former more familiar - so collocation strength is a useful variable to characterise how familiar or common a particular phrase is. We used the FLAT, an automated tool to analyse the collocation strength of well-formed two- and three-word combinations (read more about the FLAT here). Especially at the level of three-word combinations, individuals with aphasia used phrases that are more familiar than the phrases used by their conversation partners.
However, the frequency-based analysis does not tell us about the meaning or function of certain phrases. The second part of the paper addresses this point. This was an entirely data-driven analysis that evolved from the results of the first analysis: The phrase “I don’t know” was available to the majority of both individuals with aphasia and their conversation partners. It stands out from other combinations because the words are very strongly associated (i.e., it’s a very common phrase that might even be processed as one ‘big word’).
The take away point from the second part of our paper is: Individuals with aphasia appear to use familiar phrases appropriately in conversations, and they can be a resource for constructing turns.
We applied a rating scheme to all “I don’t know” utterances that appeared in the two speaker groups. Conversation partners served as the control group. We looked at phonetic forms (e.g., “dunno” vs “I don’t know”), syntactic variations (e.g., isolated “I don’t know” vs “I don’t know what time it is”), and most importantly, conversational functions (e.g., using the phrase to signal lack of knowledge vs to indicate that you are still thinking). While the phrase was used for a number of different functions in the aphasic group, suggesting that it can help to construct turns in aphasia, we identified more multifunctional “I don’t know” utterances in the CP group. Interestingly, in both groups, we found patterns that showed how the phrase can be adjusted to deal with difficulties arising from aphasia.
We often debated on whether this paper should be published as two separate articles, i.e. one paper with the frequency-based analysis only, and one qualitative “I don’t know” paper. However, I am glad we stuck to the combination of both: In my view, this reflects best that the frequency-based analysis led to the qualitative, interactional linguistic exploration of the most common phrase. This combination is compatible with a usage-based framework, where a linguistic form and its semantic-pragmatic function are paired. For example, we can use “I don’t know” to signal a lack of knowledge, but sometimes, we are using the same phrase to indicate that we are still thinking, or to avoid disagreeing with someone – technically, these are different constructions (form-function pairings).
I am very pleased with how the peer reviewing process went. One reviewer made me re-think the way I described the methods section which was a very useful thing to do, for instance in preparation of my PhD viva. A second reviewer commented on the potential role of the linguistic complexity of residual word combinations in aphasia, an interesting question for future research. Moreover, my examiners in my viva pointed out that the conversation partners might not represent a ‘typical’ control group after all (remember that they were engaged in conversations with individuals with language difficulties). What I like most about this paper is that it highlights the value of combining a frequency-based with a qualitative analysis, to study usage patterns of familiar collocations from different angles. Finally, it shows the importance of usage frequency beyond the single word level when studying aphasic language processing.
-Claudia