Recordings

Note: Ethical question on recordings (book says OK, but US / IRB issues)

What used for?

  • pronunciation:
    • ethnic influence
    • foreign learner speech
    • first language acquisition
  • vocabulary
    • frequency of usage
    • regionalisms
  • Conversational analysis

Practical considerations:

  • Audio only or audio and video?
  • Regular tape recorder or digital recorder?
  • How much time do you need?
  • External microphone
  • Making copies
  • Identifying speakers
  • Quiet place, no kids running around, traffic

New technologies:

  • Recording from Internet (e.g. BBC or regional English): Replay A/V
  • For transcripts only, can sometimes get these on Web: e.g. CNN, NPR, movie scripts
  • PodCasting, etc

General problems:

  • you may not get enough instances of what you are looking for
  • observer paradox
  • technical difficulties
  • very time consuming to go through and find things/transcribe

Acoustic Phonetics using spectrographs to make precise measurements. VOT in English

Available speech corpora (so you don’t have to do your own :-)

Linguistic Data Consortium: Membership, collects corpora, used by programmers, speech recognition-transcribed orthographically, phonetically, time stamp. Examples:

Observations
  • "The collection of data without manipulating it"
  • "Simply observe ongoing  activities, without making any attempt to control or determine them"

Can use in: (give examples)

  • psycholinguistics
  • L1
  • L2
  • literacy
  • accent
  • dialects / accents
  • sociolinguistics (e.g. * of -z dropping; that man say he sick; 1-10-57-71% deletion)
  • language and gender
  • conversational analysis

Examples:

  • Labov, Montreal-observe whether French or English used; age, sex, place of people, topic of conversation
  • Observing people’s speech in grocery store. No observer’s paradox, but may be unethical
  • African Americans say they speak with less marked features in the workplace, but observation found otherwise
  • Boys vs girls: Give kids puppets to play with, play session taped. Boys use assertive aggressive and girls use cooperative language

Advantages:

  • "Holistic" view of language
  • Little "observer's paradox"; observe non-standard forms
  • Easy to administer
  • If TV / radio / Internet, etc, fairly easy to get data

Disadvantages:

  • Have to guess at age, socioeconomic status, education level
  • Ethics of hidden recordings
  • Need lots of data to get the "linguistic nuggets" that you want

General methodology:

  • Different environments yield different data (e.g. slips of the tongue)
  • Make notes on environmental variables (time of day, number in group, what doing, etc)
  • To what degree should the observer be part of the group?
  • Keep numbers small, but can pool subjects
Case studies

When use:

  • Longitudinal studies (e.g Genie)
  • Individual circumstances
  • Complex phenomena

Can use in: (give examples)

  • L1, L2
  • Speech impediments / aphasia-stroke / speech therapists
  • Language and education (literacy), (e.g. teachers talk less to minority children)

Advantages:

  • (Similar to observations)

Disadvantages:

  • Date from just one person / can it be generalized?
  • You and/or subject gets tired of project
  • Objectivity

General methodology:

  • Hopefully someone you know and feel comfortable with
  • Treat the subject with respect
  • Create baseline and measure from there
Experiments

When use:

  • "Control as many variables as possible and manipulate one (or more) to find out what effect that variable has on the overall result"
  • Have explicit hypothesis, and then attempt to prove (or disprove) that hypothesis
  • When subjects can't give judgments

Can use in: (give examples)

  • L1 acquisition: infants and phoneme recognition
  • L1 acquisition: toddlers and passives
  • Sociolinguistics (Labov and /r/ in New York City department stores)
  • Education (e.g. different teaching methods; 8 AM vs 9 AM classes)
  • Psycholinguistics (Dichotic - one ear vs other)

Advantages:

  • Provide precise data for/against a given hypothesis (e.g. dichotic)
  • With enough detail when published, can be replicated

Disadvantages:

  • Can results be generalized outside of experiment (e.g. language learning)
  • Often, quite hard to set up experiment correctly, and to account for all variables
  • Some people don't make good research subjects (e.g. dichotic for children)

General methodology:

  • Want to change just a few variables (e.g. two different teaching methods; class time; teachers)
  • Best to have a "control group" (e.g. language learning)
  • Distribute groups evenly (e.g. language learners: M/F, but men 15-25 and women 45-55)
  • Double blind study eliminates experimenter bias (e.g. language learning)
  • Internal validity: Is the measure reliable? (E.g. bathroom scales that read differently each time)
  • External validity: Are the results relevant to situations outside of the experiment? (E.g. test involving BYU linguistics students, who have extensive language experience abroad)
  • Large enough sample size
  • Run a pilot study first
  • Right equipment (e.g. one-ear listening test, or response time on the computer)

Examples:

  1. Pragmatics: Making inferences

  2. Psycholinguistics: Bottom-up vs top-down processing

  3. Reading: Eye movement

  4. Reading: Word recognition