Phonological trends in the lexicon Practicum Michael Becker - - PowerPoint PPT Presentation

▶
phonological trends in the lexicon practicum
SMART_READER_LITE
LIVE PREVIEW

Phonological trends in the lexicon Practicum Michael Becker - - PowerPoint PPT Presentation

Jan 08, 2023 221 likes •760 views

Phonological trends in the lexicon Practicum Michael Becker University of Massachusetts Amherst michael.becker@phonologist.org EVELIN 2012 MIT / UNICAMP Campinas, Brazil 1 / 20 Practicum overview Practicum overview Formulating

slide-1
SLIDE 1

1 / 20

Phonological trends in the lexicon — Practicum

Michael Becker University of Massachusetts Amherst michael.becker@phonologist.org

EVELIN 2012 MIT / UNICAMP Campinas, Brazil

slide-2
SLIDE 2

Practicum overview

  • Practicum overview

Lexicon study Experimental design Building and running an experiment Where to go next

2 / 20

  • Formulating a falsifiable hypothesis
  • Lexicon study
  • Building a lexicon
  • Data exploration with regular expressions
  • Regression modeling
  • Working with audio materials
  • Recording
  • Praat work
  • Scripting and automation
  • Experimental design
  • Formulating a testable hypothesis
  • Online experiments / web interface
  • Regression modeling
  • Comparing the lexicon and the experiment
slide-3
SLIDE 3

Lexicon study

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

3 / 20

slide-4
SLIDE 4

Building a lexicon

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

4 / 20

  • List of paradigms
  • Word list
  • Custom list
  • Opportunistic data collection
slide-5
SLIDE 5

Building a lexicon

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

4 / 20

  • List of paradigms
  • Turkish: TELL (Inkelas et al. 2000)
  • Hebrew: LLHN (Bolozky & Becker 2006)
  • Russian: Usachev (2004), based on Zaliznjak (1977)
  • Others?

Not very common, very useful — why?

  • Word list
  • Custom list
  • Opportunistic data collection
slide-6
SLIDE 6

Building a lexicon

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

4 / 20

  • List of paradigms
  • Word list
  • English: CMU (http://www.speech.cs.cmu.edu/cgi-bin/cmudict),

CELEX (Baayen et al. 1995, not free)

  • French: Lexique (http://www.lexique.org/)
  • Portuguese: LABEL-LEX (http://label.ist.utl.pt/en/labellex_en.php)
  • Many others.

Googling for e.g., "Kabardian word list" usually helps. Asking around is a good idea too. You can use the word list to prepare a list of stems, and then add the other morphological category manually. It’s a lot of work, but it can help generate ideas.

  • Custom list
  • Opportunistic data collection
slide-7
SLIDE 7

Building a lexicon

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

4 / 20

  • List of paradigms
  • Word list
  • Custom list
  • If available, use a paper dictionary. Scanning + OCR can save

a lot of work. Hire research assistants to help.

  • Use corpora and/or search engines to expand your empirical

scope. In recent years, Google has become less useful for such searches.

  • Opportunistic data collection
slide-8
SLIDE 8

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • Extract the [w]-final words
  • Collect judgments
  • Coding
slide-9
SLIDE 9

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • The original file
  • A series of regular expression substitutions
  • Result: spelling + IPA (mostly)
  • Extract the [w]-final words
  • Collect judgments
  • Coding
slide-10
SLIDE 10

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • The original file

N a N aacheniano N aal N aaleniano N aba N ababá N ababalhamento N ababosamento

  • A series of regular expression substitutions
  • Result: spelling + IPA (mostly)
  • Extract the [w]-final words
slide-11
SLIDE 11

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • The original file
  • A series of regular expression substitutions

For example:

([aeiouáéíóú㘠eõ])s([aeiouáéíóú㘠eõ]) → $1z$2 ss → s

Learn more about regular expressions!

http://pt.wikipedia.org/wiki/Expressão_regular

We automated the substitutions with a Perl script.

  • Result: spelling + IPA (mostly)
  • Extract the [w]-final words
  • Collect judgments
slide-12
SLIDE 12

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • The original file
  • A series of regular expression substitutions
  • Result: spelling + IPA (mostly)

N a "a N aacheniano aaSeni"ano N aal a"aw N aaleniano aaleni"ano N aba "aba N ababá abab"a N ababalhamento ababaLam"˜ eto N ababosamento ababozam"˜ eto

  • Extract the [w]-final words
  • Collect judgments
slide-13
SLIDE 13

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • Extract the [w]-final words
  • Again, a regular expression: w$
  • No need for programming — a text editor with support for

regular expressions is good too: Notepad++ (Windows), TextWrangler (Mac) + OpenOffice/LibreOffice

  • We got a list of 5742 words — mostly nouns and adjectives.
  • Collect judgments
  • Coding
slide-14
SLIDE 14

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • Extract the [w]-final words
  • Collect judgments
  • Do we really need ALL the adjectives that end in [aw], [ew]...?
  • The monosyllables are manageable, so we take all of them.

We asked three people to supply plurals for them.

  • We want a good sample of polysyllables.
  • Randomize, and choose the sizable portion.

Excel trick: items in one column, =rand() in a second column, and sort by the random number. We asked one person to supply plurals for our sample of polysyllables.

  • Coding
slide-15
SLIDE 15

Example: Portuguese plurals

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

5 / 20

How did we get from the word-list of LABEL-LEX to a corpus of Portuguese plurals?

  • Transform from spelling to IPA
  • Extract the [w]-final words
  • Collect judgments
  • Coding
  • Some words did’t have a plural → excluded
  • 0 = faithful, 1 = alternating, .5 = optional
  • [malis], [abRilis] coded as faithful
  • Items with one than one rating → averaged
slide-16
SLIDE 16

Dealing with text files

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

6 / 20

Text is the bread and butter of computing.

  • Text file vs. binary file
  • Plain text editors
  • Unicode
  • Regular expressions
slide-17
SLIDE 17

Lexical statistics

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

7 / 20

  • Descriptive statistics
  • Inferential statistics
  • Limits of logistic regressions
slide-18
SLIDE 18

Lexical statistics

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

7 / 20

  • Descriptive statistics
  • What is the average alternation rate for each size? for each

final vowel? for combinations?

  • Can be done in Excel/OpenOffice (pivot tables, subtotals)
  • Even better: R (xtabs, aggregate)
  • Commercial programs: SPSS, Stata
  • Become an expert in Excel and/or R
  • Visualization
  • Inferential statistics
  • Limits of logistic regressions
slide-19
SLIDE 19

Lexical statistics

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

7 / 20

  • Descriptive statistics
  • What is the average alternation rate for each size? for each

final vowel? for combinations?

  • Can be done in Excel/OpenOffice (pivot tables, subtotals)
  • Even better: R (xtabs, aggregate)
  • Commercial programs: SPSS, Stata
  • Become an expert in Excel and/or R
  • Harald Baayen: Analyzing Linguistic Data
  • Keith Johnson: Quantitative Methods in Linguistics
  • Andrew Gelman & Jennifer Hill: Data Analysis Using

Regression and Multilevel/Hierarchical Models Can be found online.

  • Visualization
  • Inferential statistics
  • Limits of logistic regressions
slide-20
SLIDE 20

Lexical statistics

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

7 / 20

  • Descriptive statistics
  • What is the average alternation rate for each size? for each

final vowel? for combinations?

  • Can be done in Excel/OpenOffice (pivot tables, subtotals)
  • Even better: R (xtabs, aggregate)
  • Commercial programs: SPSS, Stata
  • Become an expert in Excel and/or R
  • Visualization

0% 20% 40% 60% 80% 100%

mono iamb trochee

7 265 45 8 6 17 37 2

faithful intermediate alternating

slide-21
SLIDE 21

Lexical statistics

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

7 / 20

  • Descriptive statistics
  • Inferential statistics

Regression:

  • making predictions
  • confidence in predictions

Example: logistic regression for the Portuguese [w]-final words

β SE(β) z p(>|z|)

(Intercept) 0.10 0.38 0.27 mono vs. iamb 3.21 0.49 6.60

<.0001

iamb vs. trochee 3.52 1.24 2.85

<.005

lax 2.67 0.57 4.64

<.0001

high 0.20 0.29 0.69

>.1

  • Limits of logistic regressions
slide-22
SLIDE 22

Lexical statistics

  • Practicum overview

Lexicon study

  • Building a lexicon
  • Example: Portuguese

plurals

  • Dealing with text files
  • Lexical statistics

Experimental design Building and running an experiment Where to go next

7 / 20

  • Descriptive statistics
  • Inferential statistics
  • Limits of logistic regressions

A logistic regression is impossible when some predictor leads to a categorical distinction (e.g., some alternation always happens/never happens when...)

  • In R: bayesglm in the arm package
  • Decision trees (in R or other programs)
slide-23
SLIDE 23

Experimental design

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

8 / 20

slide-24
SLIDE 24

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Recruiting participants
  • Randomization
  • Fillers/distractors
  • Instructions
  • Respect, demographic questions, feedback
slide-25
SLIDE 25

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • The hypothesis guides the creation of the items.

e.g. monosyllables and polysyllables, words with [e] and words with [E]

  • Try to factor out the things you are not interested in, like

consonants e.g. compare [dew] and [dEw] to test a vowel effect

  • More = better.

50 people responding to 1 item each is much better than 50 people responding to the same item. (why?)

  • Choosing a task
  • Recruiting participants
  • Randomization
  • Fillers/distractors
  • Instructions
  • Respect, demographic questions, feedback
slide-26
SLIDE 26

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Binary forced choice
  • Scalar forced choice
  • Binary judgment
  • Scalar judgment
  • Production task (orthographic)
  • Production task (auditory)
  • Recruiting participants
  • Randomization
  • Fillers/distractors
  • Instructions
  • Respect, demographic questions, feedback
slide-27
SLIDE 27

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Recruiting participants
  • On campus, online, in the field, in class...
  • Amazon’s Mechanical Turk
  • Each participant must respond independently to get valid

results.

  • Randomization
  • Fillers/distractors
  • Instructions
  • Respect, demographic questions, feedback
slide-28
SLIDE 28

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Recruiting participants
  • Randomization
  • Randomizing the order of items
  • In a forced choice task, random order of choices
  • Why?
  • Fillers/distractors
  • Instructions
  • Respect, demographic questions, feedback
slide-29
SLIDE 29

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Recruiting participants
  • Randomization
  • Fillers/distractors
  • Need to make sense given the task
  • How many?
  • Randomized with the target stimuli
  • Instructions
  • Respect, demographic questions, feedback
slide-30
SLIDE 30

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Recruiting participants
  • Randomization
  • Fillers/distractors
  • Instructions
  • Make them short. People don’t read them anyway.
  • Give them in the same language as the experiment.
  • Maybe the ideal experiment doesn’t have instructions at all...
  • One or two practice items.
  • Respect, demographic questions, feedback
slide-31
SLIDE 31

Wug-test

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

9 / 20

  • Creating items
  • Choosing a task
  • Recruiting participants
  • Randomization
  • Fillers/distractors
  • Instructions
  • Respect, demographic questions, feedback
  • Consult your institution’s policy on experiments with humans.
  • Never lie to your participants.
  • Don’t give participants any motivation to lie.
  • Suspicious participants → run, then throw the data out.
  • Ask for feedback at the end.
slide-32
SLIDE 32

Artificial grammars

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

10 / 20

  • Reminder: artificial grammar inspired by English voicing
  • Task: training and testing
  • Presentation of novel concepts
slide-33
SLIDE 33

Artificial grammars

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

10 / 20

  • Reminder: artificial grammar inspired by English voicing

monosyllabic training iambic training Training 10 stop-final monos 10 stop-final iambs "mip "mibni t@"gep t@"gebni "stut "studni g@"Sut g@"Sudni 5 sonorant-finals: "muN-ni, n@"Ãol-ni Testing 10 stop-final monos 10 stop-final monos "gaIp "gaIp "klet "klet 10 stop-final iambs 10 stop-final iambs f@"Ùop f@"Ùop b@"git b@"git 10 sonorant-finals: "pler, Z@"taIm

  • Task: training and testing
  • Presentation of novel concepts
slide-34
SLIDE 34

Artificial grammars

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

10 / 20

  • Reminder: artificial grammar inspired by English voicing
  • Task: training and testing
  • Train: give the participant examples of the pattern to learn
  • Test: did the participant apply the pattern to items they

haven’t seen before?

  • Holdout condition: did the participant apply the pattern to a

kind of item they haven’t seen before?

  • Compare two groups of participants
  • Presentation of novel concepts
slide-35
SLIDE 35

Artificial grammars

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

10 / 20

  • Reminder: artificial grammar inspired by English voicing
  • Task: training and testing
  • Presentation of novel concepts
  • Easiest concepts to present: concrete objects
  • Affixes: plural, dual vs. plurals, feminine(?)
  • How do you present a novel adjective? Novel verb?
  • Affixes: comparative? perfective?...
slide-36
SLIDE 36

Others kinds of experiments

  • Practicum overview

Lexicon study Experimental design

  • Wug-test
  • Artificial grammars
  • Others kinds of

experiments Building and running an experiment Where to go next

11 / 20

  • Tasks
  • Lexical decision
  • Confusability
  • Morphological knowledge
  • etc. etc.
  • Make your own (you are allowed!)
  • Modalities
  • On paper
  • Keyboard/mouse
  • Microphone
  • Button box
  • Eye tracking, ERP

, fMRI, ultrasound...

slide-37
SLIDE 37

Building and running an experiment

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

12 / 20

slide-38
SLIDE 38

Audio stimuli

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

13 / 20

  • Preparing materials
  • Recording
  • Chopping up — Praat scripting
  • Converting to mp3 (http://www.macroplant.com/adapter/)
slide-39
SLIDE 39

Audio stimuli

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

13 / 20

  • Preparing materials
  • Presentation of materials to the consultant
  • Randomizations
  • Repetitions
  • Recording
  • Chopping up — Praat scripting
  • Converting to mp3 (http://www.macroplant.com/adapter/)
slide-40
SLIDE 40

Audio stimuli

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

13 / 20

  • Preparing materials
  • Recording
  • Sound booth ≻ quiet room ≻ outdoors
  • Decent equipment
  • Chopping up — Praat scripting
  • Converting to mp3 (http://www.macroplant.com/adapter/)
slide-41
SLIDE 41

Audio stimuli

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

13 / 20

  • Preparing materials
  • Recording
  • Chopping up — Praat scripting
  • Mark pauses, add labels, save labeled intervals
  • Find Praat scripts and help online
  • Converting to mp3 (http://www.macroplant.com/adapter/)
slide-42
SLIDE 42

Online experiments

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

14 / 20

  • Hard to set up the first time, easy to run every time
  • Provides access to remote participants
  • Programs that can help:
  • Experigen (Becker & Levine 2010)
  • Ibex Farm
  • Webexp
slide-43
SLIDE 43

Paper experiments

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

15 / 20

Slower and more labor-intensive, but lower barrier to entry

  • Just prepare a document, print it out, give to participant.
  • Items still have to randomized (manually?)
  • More advanced option: a script that creates a L

A

T EX file.

  • Responses need to be typed up.
slide-44
SLIDE 44

Results

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

16 / 20

  • Descriptive statistics
  • Inferential statistics
  • Portuguese nonce words again
slide-45
SLIDE 45

Results

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

16 / 20

  • Descriptive statistics
  • For the population, not for individual participants
  • Inferential statistics
  • Portuguese nonce words again
slide-46
SLIDE 46

Results

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

16 / 20

  • Descriptive statistics
  • Inferential statistics
  • The regression is your friend.
  • ANOVA: an older, more restricted kind of regression. Not

necessary anymore, still used by many people.

  • Florian Jaeger’s slides, lab blog.
  • Portuguese nonce words again
slide-47
SLIDE 47

Results

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

16 / 20

  • Descriptive statistics
  • Inferential statistics
  • Portuguese nonce words again

(Linear) regression for the experimental results

β

SE(β) t p-value (Intercept) 4.47 0.10 42.91 mono vs. trochee 0.35 0.09 3.80

<.0005

mono & trochee vs. iamb 0.55 0.11 4.77

<.0001

lax 0.43 0.09 4.63

<.0001

low

−0.40

0.26

−1.53 >.1

mono vs. trochee:low 0.72 0.31 2.35

<.05

slide-48
SLIDE 48

Lexicon vs. experiment

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

17 / 20

  • Predictions from the lexicon
  • Correlation
  • Model comparison
slide-49
SLIDE 49

Lexicon vs. experiment

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

17 / 20

  • Predictions from the lexicon
  • Correlation

1 2 3 4 5 6 7 3 4 5 6

lexicon model predictions mean participant response mono iamb trochee

  • Model comparison
slide-50
SLIDE 50

Lexicon vs. experiment

  • Practicum overview

Lexicon study Experimental design Building and running an experiment

  • Audio stimuli
  • Online experiments
  • Paper experiments
  • Results
  • Lexicon vs.

experiment Where to go next

17 / 20

  • Predictions from the lexicon
  • Correlation
  • Model comparison
  • Which factors improve the fit?
  • Measured with a χ2 test
slide-51
SLIDE 51

Where to go next

  • Practicum overview

Lexicon study Experimental design Building and running an experiment Where to go next

  • Resources
  • References

18 / 20

slide-52
SLIDE 52

Resources

  • Practicum overview

Lexicon study Experimental design Building and running an experiment Where to go next

  • Resources
  • References

19 / 20

  • Skills to learn
  • Working with text files
  • Working with sound files
  • Experimental design
  • Descriptive statistics
  • Inferential statistics
  • Automation: Praat scripting, some other scripting language

(Perl, Python, Javascript, etc.)

  • Theory: John McCarthy’s books (Doing OT, reader), Bybee
  • Quantitative methods: Johnson, Baayen, Gelman & Hill, Jaeger’s

slides, journal articles

  • Statistics class in the psychology department
  • Internship with a cognitive psychologist or experimental linguist

(design, run experiments, stats)

slide-53
SLIDE 53

References

  • Practicum overview

Lexicon study Experimental design Building and running an experiment Where to go next

  • Resources
  • References

20 / 20

Baayen, R. Harald, Richard Piepenbrock & Leon Gulikers (1995). The CELEX Lexical Database (CD-ROM). Linguistic Data Consortium, University of Pennsylvania, Philadelphia, PA. Becker, Michael & Jonathan Levine (2010). Experigen - an online experiment platform. Available at https://github.com/tlozoot/experigen. Bolozky, Shmuel & Michael Becker (2006). Living Lexicon of Hebrew Nouns. Ms. University of Massachusetts, Amherst. Available at http://becker.phonologist.org/LLHN. Inkelas, Sharon, Aylin Kuntay, John Lowe, Orhan Orgun & Ronald Sprouse (2000). Turkish Electronic Living Lexicon (TELL). Website, http://socrates.berkeley.edu:7037/. Usachev, Andrei (2004). Fully accented paradigms from Zaliznjak’s (1977) grammatical dictionary. http://dict.buktopuha.net/all_forms.rar.