I wonder if it might be profitable to kick-start research on the language of politics by simply asking people "in the biz" to help spin particular questions in favor of one or another desired political outcome. Alternatively, it might well be possible to just look at the kind of language used in a press release as compared to the language of analysis (or as compared to the language of a competing press release dealing with the same topic from a very different perspective). How much do spin-meisters know about the language they are using, and in what way do they know it (does it just sound good to them, or do they have an explicit, deliberate strategy which they are employing)? How sensitive are listeners to the kind of language that is used? To the identity of the speaker? How accurate are people's judgments of the slipperiness of (political) language?

I also worry about tasks which ask "circle the houses that were painted" and other such questions. It seems those kinds of questions rely heavily on the listener's judgment of the kind of outcome desired by the experimenter, which may show through more readily given the dearth of information about what to do in the instructions. It would be useful, I believe, if the subject though he or she was doing a different task than simply "circling the houses that were painted", thereby making the task less confusing and minimizing the possibility that the subject was mostly trying to infer the mental state of the person who designed the experiment.

It isn’t surprising that different verbs and versions of past tense mean different things. Dr. Matlock’s research pushes beyond the intuitive and makes these subtle semantic meanings concrete. Her exploration of how language evokes mental imagery and imaginary motion provides another illustration of how humans recruit bodily experience to create abstract meanings. I found Dr. Matlock’s talk inspiring because of the rigor and creativity in her research, and her openness and curiosity during the discussion.

When Teenie mentioned presenting fictive motion sentences to people while in an fMRI scanner, I remember her saying they were looking for activation in visual area MT. That would make sense if people are using visual imagery to understand FM sentences. But I wonder about activation of the neural mechanisms we've talked more about in this class as underlying spatial cognition--say the 3 subsystems in Namara's model of spatial memory. Would the linguistic and behavioral evidence suggest any particular sorts of spatial representations (e.g., egocentric vs. allocentric, short term vs. long term)?
Also, I must say I find it really fascinating how research in cognitive linguistics is influenced by phenomenological claims--e.g., the fleeting experience of motion we sometimes have while understanding FM sentences. I wonder if first-person, introspective methods could at all help in characterizing the nature of the mental imagery occurring during language understanding.

Political language is certainly an interesting area to investigate, but I wonder whether a controlled experimental paradigm will yield meaningful results (It is worth noting that she explicitly noted that she's just begun to study this area). It seems that there are just too many variables -- past experience with politics, opinion of the political process, understanding of the leanings of the electorate, etc. -- that influence interpretation of the particular words used.

Perhaps another way to think about the study of political language might be to try to think about ambiguity in language. I have very little knowledge of what has been done in that area already, but it would be quite relevant to understanding political language. In many ways, it seems that much of politics is the ability to say something ambiguous but to try to steer people toward a particular interpretation. The example used in the study we read, concerning the use of imperfective language, is a good example. "He was taking hush money" can mean that he is still taking it, or that he has stopped taking it, and both would be factually correct interpretations. An opponent would obviously want some the former interpretation, or at least some hint of the former interpretation. Studying how people interpret ambiguous statements in different contexts might be another angle from which to approach the issue of political language interpretation.

I wasn't able to attend this talk, which is especially unfortunate because I study language, but I did have a couple of thoughts on one of the papers Matlock sent out. In the "Can grammar win elections?" paper, I believe, but wasn't surprised by, the results of Experiment 1, which showed that using the imperfective tense when describing negative actions increases the impact of those actions on people's judgments. Similar work has been done with lexical priming (Truitt & Zwaan, 1997): people are faster to produce or recognize a word (hammer) following a sentence that contains an imperfective verb (He was pounding the nail) then one that contains a perfective verb (He pounded the nail). (Actually, Jeff Elman has done related work on generating verb expectancies, primarily with Ken McRae and Mary Hare.)

As for experiment 2, I'm not convinced that their results are due to anything other than ambiguity. Their finding was that sentences with an imperfective negative outcome and a perfective positive one ("...was removing homes and extended roads.”) have a greater negative impact than sentences with vice-versa ("...removed homes and was extending roads."). There's a problem here, though: in the first case, the scope of the first verb could legitimately encompass the whole phrase, suggesting that the senator was removing both homes and roads (but only extended roads). Given that adjectives and perfective verbs often share phonological forms (as with the word “extended”), this strikes me as a confound, especially since (at least in the example given) the contrastive sentence presented no such opportunity for ambiguity. Controlling for this would have been one way to mitigate the concern, but it would have been even nicer if subjects were asked yes/no questions afterwards - I'd bet that a good number of subjects who saw the first sentence would have said "yes" if asked whether the senator removed roads.