A contrastive study into the lexico-grammatical reactances of two intransitive verbal categories – Lucy Chrispin

For an enlarged version, click on the poster or download file:

Abstracts page

Bios page


1) This poster is based on a finished draft of a chapter –  are there any general questions or critique that stand out?

2) Are there any other statistical analyses to use on the results (besides a Fisher’s exact test on lexical aspect)?


  • Emily Powell

    Hi Lucy, -great poster. It made something that I find difficult to understand a lot clearer!
    I’d be interested to know how this chapter fits in with your other chapters and I’m guessing this would be useful to know if someone was reading this as a chapter too.

  • Lucy Chrispin

    Hi Emily,

    Thanks for your comment and glad you could make (some) sense of it. My thesis as a whole is looking at intransitive constructions, and is split into three micro-studies. The first one aimed to increase the understanding of Behaviours as a category, and identified its lexico-grammatical reactances with real-language data (in a very similar process to this one). The chapter presented here is the second micro-study, which has identified lexical aspect to be a prominent feature in differentiating behaviours and material processes. I was interested in researching the lexical aspect further for the final micro-study, and so the following chapter investigates these intransitive constructions as a whole, to understand how close the relationship is between a grammatical construction and the lexical aspect that is construed.

  • Catherine Laing

    Really nice poster design! You’ve managed to make it very visual and the results/conclusions are clear. The method and results sections are particularly good. It would be useful to see a ‘key message’ of why this research is important. You do this to some extent in the Background section, but I think you could go a step further by simplifying the language and putting it in bold. It would also be useful to see some clarification of key terms and some examples to illustrate.

  • Lucy Chrispin

    Thanks for this feedback Catherine, it’s really useful to get an alternative opinion and I’ll take everything you’ve said on board!

  • Gerard O'Grady

    Hi Lucy, I guess my sole comment is to do with the graphics. It is hard to see if there is a difference between behavioural and intransitive material processes. It would have been nice if this could have been highlighted. Behaviourals are a very odd category and it is great that someone is brave enough to investigate it. Your examples from IFG are a very useful reminder of the danger of relying on semantic differences and not grammatical reactances. The five processes that you have chosen sneeze, mediate, frown converse and stare are not examplars of behavioural processes that Halliday said are near material such as dance or sing so it might be worth checking if they pattern like mental processes, mediate and stare, like verbal processes converse and differently in relation to the other two.

  • Kate Kavanagh

    Hi Lucy, thanks for the poster. I found it really clear and complete. When I was looking at the graphs I was wondering which results were significant and which not. From the bullet points that follow I infer that only the Lexical Aspect distribution showed significant difference, but this could perhaps be made more explicit. Good stuff! :)

  • Andy Buerki

    Hello Lucy,

    What a nice piece of work, with two very notable results it seems (clear difference in lexical aspect and different from expected behaviour in perfective aspect). To your questions:
    1) this is a tiny detail, but your that-complement measure, is described as ‘ability to take -that complement’. How was this ‘ability’ measured in the concordance lines? If it was measured by the actual occurrence of a that-complement, I would probably adjust the reactance label to ‘occurrence of that complement’. If, on the other hand, you looked at each concordance line and decided based on intuition whether a that-complement could have occurred (regardless of whether it did or not), the ‘ability to take’ would be right, but of course that would be somewhat removed from direct corpus evidence. It looks like there were no actual that-complements at all in any case (as expected), and probably none possible, so this makes little difference.
    2) I think chi-squared/Fisher’s exact tests on each of the reactances is probably a very sensible way of testing for differences. What you don’t get if you do it this way is an overall picture. That is to say, whether across all of the reactances, it is possible to say that behavioural and Intransitive material processes differ. However, since you’ve already shown clearly that they can distinguished on lexical aspect (and not so much in other ways), that overall picture does not particularly add much. There is another, more involved procedure that could be applied: a binary regression. This would be looking at whether, based on each of your reactances as predictors, it is possible to predict (with better accuracy than change) whether a verb is of the behavioural or intransitive material process type. In one go, you can then measure how good an overall prediction is, and also how good (or useless) each of the individual predictors are. However, the results are very clear as they are, so I’m not sure that this more involved regression analysis (neat as it would be) would be worth your while.

  • Harry

    This poster was so clear and well written that even a layman like myself could develop a basic understanding of the topic! Thanks Lucy!

  • Lucy Chrispin

    Hi Gerard and Kate,

    Thanks very much for your presentation comments, and Gerard for your patterning advice – this is something I will look into.

  • Lucy Chrispin

    Hi Andy,

    Thanks very much for this detailed feedback. I completely agree with your comments on that complements – I think occurrence is much better. Even if they are instances that, based on intuition, are unable to take the complement, I see problems in intuitive-based rather than corpus-based claims.

    As for the binary regression, I’ll definitely look into this and hopefully include to support the results. Thanks.

  • Debbie Cabral

    Hi, Lucy!
    WOW! These findings are so interesting! Seeing the results and realising you were able to test the theoretical claim must have been really exciting! Thanks for the poster! It is very well done!

  • Kateryna

    Hi Lucy,
    Your research is interesting and elegant. It looks consistent and complete.
    When I was reading your poster, a questions came to my mind. What is the basis of the study’s lexical categories (state, activity accomplishment, etc.)? I mean: are they taken from a particular theory?

  • Kateryna

    Lucy, have you counted the number of occurrences (the frequency) in various categories? I wonder if logistic regression can be used when the value of the independent variable (reactences) is discrete.

  • Lucy Chrispin

    Hi Kateryna, thanks for these comments. Those categories are rather grounded in the theory of lexical aspect (expression of temporal structures by lexical items), with contributions from Smith (1997), Vendler (1967), Declerck (1991). I have used a summary by Van Rompaey (2013) who reviews this research very well.

    I have got the frequencies of each categories, and so I’ll definitely look into logistic regression, thanks.

  • Kateryna

    Thanks for your answer, Lucy!

    I think I did not express myself clearly. I meant that if your variables are frequencies, than logistic regression might not be suitable, for which the variables have to be either categorical or continuous. Frequency is a discrete variable.

  • Lucy Chrispin

    Hi Kateryna, right I think I see what you mean. I have frequency tables, but I also have the larger dataset of categorical variables.. this would perhaps still be discrete though right? If you wouldn’t mind, this might be easier for me to email you with at some point just to confirm if that’s ok.

  • Kateryna

    Categorical variables are nominal like ‘yes/no’, ‘female/male’, etc. The list of categories you are using in the study also will be a nominal variable. Just wanted to draw your attention to the fact that when using logistic regression, make sure that you do not use frequencies:)

  • Lucy Chrispin

    Yes thanks K! I follow now and I do have this data so I will give it a go :)

Leave a Reply

Your email address will not be published. Required fields are marked *