Please print out each of the following two sections separately, and turn them in as separate documents on Tuesday. (Note that these should be typed, not hand-written)


Imagine this scenario.  You live in a post World War III world in which the US is run by a despotic dictator who blames Conservative Christians and Mormons for the war. He hunts them down like the Lamanites did the Nephites.  However, many do not confess their beliefs even under horrific torture.  The government contacts you because not only are you a brilliant linguist, but you are an infamous, treacherous, ex-Mormon who dresses poorly to boot.  Your mission is to design an experiment that will be given under the guise of being a simple linguistic experiment.  However, the true purpose  behind it is to determine who the Mormons are.   You will take advantage of the fact that Mormons probably process certain words differently from the general population. Remember that you want to hide the true nature of the experiment.

  1. Use one of the experimental paradigms discussed in class.  Don't invent a new one.

  2. List 10 of critical test words you would use.  Why would you choose them over others?

  3. What other words would you include?  List 10.  Why would you include them?

  4. How would the experiment be carried out?

  5. What would you measure? How is this related to the purpose of the study?

  6. What statistic would you use to determine the statistical significance of your study?

  7. What would the issues of internal and external validity be for this study?


These questions deal with several aspects of the word trick. Please feel free to use whatever resources and corpora you'd like, other than another human being :-). Make it as long or short as you'd like, as long as you adequately cover each topic.  For each question, please state briefly which corpus/resource you used, and how you obtained the data. Good luck.

  1. Has there been an overall increase or decrease in frequency of this word over time? How would you measure this? What problems would/do you have in getting this data, and what limitations are there on the data that you've obtained?

  2. Have there been changes in meaning over the past 200 years? (Use the Oxford English Dictionary and at least one another source.) What kind of evidence does the regular OED interface give? What kind of evidence did you get from the other source?

  3. What are 3-4 of the most frequent words that are morphologically-related to T-R-I-C-K? Are there any of these that have arisen in the last 200-300 years? (Note: look just for derivationally-related words, not inflectionally-related. For example, if the word were light, you'd look for lightly, backlight, lightning, etc, but not lights or lighted/lit).

  4. Is there much difference in frequency between genres/ registers? How do you measure this? Any explanation for the differences?

  5. Any difference in meaning between the genres / registers. How can you measure this?

  6. Does the frequency in the Ensign (see vary from that of a standard corpus? Would you expect it to? What are the problems in measuring this?

  7. Compare the frequency in the US, Britain, Australia, and India. Any differences? How did you measure this?

  8. How many different "word senses" are there for this word, both as a noun and as a verb (i.e. different "meanings")? Where did you get this data? Are there any meanings that seem to be missing, or that could/should be combined?

  9. Do the corpus data provide evidence for all of the different meanings of the word, and are there meanings of the word that should show up in corpora, but that don't in other sources?