Explore the semantic changes that various English words have gone through:
Language change (semantic change)

Developed by the Humanities E-learning Team for The University of Manchester


Building on the first exercise in Part II, where you looked at style (attention to speech), explore similar data in more depth:
Methods (statistical testing)

Developed by the Humanities E-learning Team for The University of Manchester



Approaches to the study of language change

Which methodological approach do these (made up) studies use to investigate language change: apparent or real time?

A. A team of researchers investigates the use of high-rising terminals by a group of 20 people in Dunedin, NZ. They first recorded them in 1990, then again in 2000 and compared these recordings.

B. A postgraduate student from the University of Fribourg collected traditional interview data from 30 people in a small Swiss-German village. He then compared his results on consonant clusters with maps in the Dialect Atlas of Switzerland compiled some 50 years earlier.

C. Two researchers from SUNY Buffalo collected data on [au]-raising in Buffalo from 4 generations of speakers. Speakers below 40 raised [au] four times more often than older generations. They conclude that this might be language change in progress.

D. In 1971, a postgraduate student investigated the use of glottal stops in a neighbourhood in Glasgow. She went back to this neighbourhood in 2001 and conducted the same study again with different speakers.

E. An Honours student conducted 5-minute interviews with 60 members of his family, which had all come together for a weekend to celebrate a wedding. She looked into the use of post-vocalic [r] by 3 generations in her family, found young females used it the least and concluded that language change is in progress.



Style: Effect of the interviewer

Here is some data testing the effect of interviewer on use of four features of African-American Vernacular English in the speech of two African-American women in a small town in Texas. It compares their speech talking to first another African-American, and second a White fieldworker on the project (data from Cukor-Avila & Bailey 2001, Tables 4, 5 and 8). “Samantha” = “Sam”; “LaShonda” = “LaSh”.

The four variables are:

  • absence of possessive –s, e.g. “John mama” instead of “John’s mama”;
  • absence of third person singular –s in the present tense, e.g. “she take” instead of “she takes”;
  • absence of the copula, e.g. “she nervous” instead of“she is nervous”;
  • use of be followed by an imperfective form of the verb for habitual events, e.g. “we be chasing them”.

Reference: Cukor-Avila, Patricia and Guy Bailey 2001. The effects of the race of the interviewer on sociolinguistic fieldwork. Journal of Sociolinguistics 5. 254–270.


    1. Which variables seem to pattern together for an individual speaker?

      a. Can you provide a coherent linguistic or social motivation that might underlie this sub-grouping?


      b. Is there some reason why one or more variables might be patterning differently from the others?

    2. Do both speakers seem to treat the variables in roughly the same way? If not, how do they differ?



Style: Effect of the topic

Here is some data from Schilling-Estes (2004, Tables 2 and 3). It tracks variation in the course of an extended interview/conversation between two young Southern American men, Alex and Lou. Lou identifies as Lumbee (Native American) and Alex identifies as African American and part Cherokee (Native American). Their use of non-constricted /r/ and monophthong variants of (ay) are tracked during the interview across different topics. Some topics were recycled in the conversation; this is indicated as, e.g. Civil War 1 and Civil War 2. Race relations are discussed several times from different perspectives.

(You can do this exercise without reading the article, but you’ll get a lot more out of it if you make yourself a little bit familiar with who Alex and Lou are.)

Reference: Schilling-Estes, Natalie 2004. Constructing ethnicity in interaction. Journal of Sociolinguistics. 8. 163–195.

  • Schilling-Estes claims that the data indicate that towards the end of the interview there is “decreasing personalization” of the topic of race, and that this “is accompanied by increasing alignment with respect to topics” (p. 181). To what extent do you think her data supports this generalization?
  • What could Schilling-Estes do to strengthen her claim? What kinds of quantitative tools or qualitative data would strengthen it?



Apparent-time change: Quebec French question formation

The following data comes from work by Elsig & Poplack (2006) on developments in the way speakers of Quebec French do question formation.

Historically, there were four options for questioning in Quebec French:

  • inversion of subject and verb
  • use of a question particle -tu
  • no syntactic change, but question intonation
  • prefixing a declarative with the question sequence est-ce que

In the figure below, you can see the probability of the four variants in a corpus of twentieth-century Quebec French. Speakers were divided into two groups: those under 35 years and those over 35 years.

Reference: Elsig, Martin & Shana Poplack (2006) Transplanted dialects and language change: Question formation in Québec. University of Pennsylvania Working Papers in Linguistics. 12, 2. 77–90.

    1. If all you know about a speaker is their age, what can you say about the likelihood of their using the different question formation strategies?

      a. Drawing on the apparent-time data here, which form appears to be on the increase?

      b. What does the pattern for est-ce que tell us Real time meets apparent time

Look at the table above. This adds real-time data to the variable of question formation in Quebec French that we looked at in apparent time in the previous exercise. (Remember that values above 0.51 mean the variant is favoured in that condition; values below 0.49 mean the variant is disfavoured in that condition. n.s. in the table means that a factor was not significant for that particular variant.)
The apparent-time data indicated that the -tu strategy is on the increase among younger speakers, and that est-ce que is exclusively used by older speakers.

    1. What variant(s) was the most typical strategy for forming questions in:

      a. seventeenth-century Quebec French?

      b. twentieth-century Quebec French?

    2. What factors most strongly constrain (favour/disfavour) inversion and intonation strategies in both centuries?

      a. To what extent do the constraints remain the same over time? How do they change?

      b. What effect does the frequency of a verb have on how likely a question strategy will be used?

    3. Comment on the Style effect for all the variables.

      a. What seems to be going on with respect to style?

      b. To what extent can you use Style as a way of linking this real time data with the apparent time trends we looked at last week?