Is each person’s lexicogrammaticalsystem unique? An ......Is each person’s lexicogrammaticalsystem unique? An experimental study on linguistic individuality Dr Andrea Nini...

Is each person’s lexicogrammatical system unique? An experimental study on linguistic individuality

Dr Andrea [email protected]

UK Cognitive Linguistics Conference 2020University of Birmingham

29th July 2020

Michael Cameron Colin Murphy

mailto:[email protected]://www.andreanini.com

AuthorshipIdentification

Jenny NichollLast seen 30th June 2005.

After her disappearance a series of text messages were sent from her phone.

Linguistic analysis showed that the later messages were more likely to have

been sent by her ex-lover, David Hodgson.

Hodgson was convicted of Jenny’s murder 19th February 2008 in part because of the linguistic evidence.

School Law School Lawyer

School Electrician Manager

different life experiences

in-group variation that influences perception (personality, cognitive abilities, …)

à different texts à different exemplars

entirelyabsolutelyaltogethercompletely

entirelyextremely

fullyperfectly

thoroughlytotallyutterly

criticalagreecentral

acceptableunderstand

respect…

Tony Blair

3M words of interviews, speeches, PM question time

Maximisers

Speechtotally agree, perfectly entitled,perfectly obvious

Newspaperperfectly simple,perfectly fair

Parliamententirely endorse,entirely share

Tony Blairentirely understand

Mollin, Sandra. 2009. “I entirely understand” is a Blairism: The methodology of identifying idiolectal collocations. International Journal of Corpus Linguistics 14(3). 367–392. doi:10.1075/ijcl.14.3.04mol.

Words that go together. Measuring individual differences in native speakers’ knowledge of collocations

Dąbrowska, Ewa. 2014. Words that go together. Measuring individual differences in native speakers’ knowledge of collocations. The Mental Lexicon 9(3). 401–418. doi:10.1075/ml.9.3.02dab.

blatant lie clear lie

conspicuous lie distinct lie

recognizable lie

double production enlarge production

boost production extend production

redouble production

peculiar remark queer remark

unnatural remark weird remark

odd remark

40 items

% of correct responses

~linguistic

experiencelinguistic

proficiency IQ

Words that go together. Measuring individual differences in native speakers’ knowledge of collocations

Dąbrowska, Ewa. 2014. Words that go together. Measuring individual differences in native speakers’ knowledge of collocations. The Mental Lexicon 9(3). 401–418. doi:10.1075/ml.9.3.02dab.

80 adults Distribution of backgrounds roughly following general UK populationü Grammatical comprehension testü Vocabulary testü Test of print exposureü Nonverbal IQ test

30 adults

28% 98%74%

Test/retest reliability: 80%

meanmin max

Words that go together. Measuring individual differences in native speakers’ knowledge of collocations

Dąbrowska, Ewa. 2014. Words that go together. Measuring individual differences in native speakers’ knowledge of collocations. The Mental Lexicon 9(3). 401–418. doi:10.1075/ml.9.3.02dab.

% of correct responses

~linguistic

experiencelinguistic

proficiency IQ

blatant lie clear lie

conspicuous lie distinct lie

recognizable lie

double production enlarge production

boost production extend production

redouble production

peculiar remark queer remark

unnatural remark weird remark

odd remark

data for 62 participants à 1891 pairs

Words that go together. Measuring individual differences in native speakers’ knowledge of collocations

Dąbrowska, Ewa. 2014. Words that go together. Measuring individual differences in native speakers’ knowledge of collocations. The Mental Lexicon 9(3). 401–418. doi:10.1075/ml.9.3.02dab.

80%

blatant lie clear lie

conspicuous lie distinct lie

recognizable lie

double production enlarge production

boost production extend production

redouble production

peculiar remark queer remark

unnatural remark weird remark

odd remark

Words that go together. Measuring individual differences in native speakers’ knowledge of collocations

How much inter-speaker variation is there between two random speakers?

How much intra-speaker variation is

there within the same speaker?

intra-speaker variation

inter-speaker variation

8 participantsAll academics in the humanities with PhDAll native speakers of British English

Questionnaire

Collocation test

Grammatical alternations test

Test on collocations25 sentences

Into each of the following sentences, please insert the word which you think would most naturally fill the gap. It is important that you insert only individual words into each space. Please do not leave any sentence space blank. For this task there is no correct or incorrect answer and so you should choose the answer which seems most natural to you.

1. Even among the largest and most successful of companies, there is still ____________ competition.

2. The Madison banks collapsed after sustaining ____________ losses on real estate loans.

a. majorb. substantialc. significant

British Academic Written Corpus (BAWE)

Test on grammatical alternations27 sentences divided in 12 categories

The author relies on speculation to make his point, _________ providing hard evidence to back up his argument.

a. rather than

b. instead of

British Academic Written Corpus (BAWE):

rather than V 544

instead of V 171

rather than providing 5

instead of providing 4

Test on grammatical alternations27 sentences divided in 12 categories

• rather than / instead ofPreposition (7)

• take N off / take off NParticle-verb alternation (5)

• which / that / ∅ / who / whomRelative pronouns (3)

• give N N / give N to NDative alternation (2)

• s-genitive / of-genitiveGenitive alternation (2)

• half / half ofDeterminer (2)

• e.g. Ikea’s website / the Ikea websiteProper name (1)

• help V / help to VHelp (1)

• try to V / try and VTry (1)

• e.g. spray water on the grass / spray the grass with waterLocative alternation (1)

• sort of / kind of / type ofSKT (1)

• e.g. not affected / unaffectedNegation (1)

Controlling for factors that influence variation

8 participantsAll academics in the humanities with PhDAll native speakers of British English

Questionnaire

Collocation test

Grammatical alternations test

at least 2 weeks

Collocation test

Grammatical alternations test

“another questionnaire”

Changed the order of questions

Changed the order and as much lexical material as possible

Grammatical alternations test Collocations test

Wilcoxon test p = 0.0007 Wilcoxon test p = 0.0005

Results

Higher levels of similarity for grammar test than collocations test

Cohen’s d = 1.8 with p < 0.05 power is 0.99 even with this small sample size!

Grammatical alternations test Collocations test

Sanity checkResults

No two questionnaires are identical

Same person tests more similar to themselves than random pairs

Results are consistent with usage-based approaches to lexicogrammar

Results have massive implications for forensic linguistics

Could it be a memory test? Anecdotal evidence from debriefing points to weak memory effect

Conclusions

Larger replication in the works

Is each person’s lexicogrammatical system unique? An experimental study on linguistic individuality

Dr Andrea [email protected]

Michael Cameron Colin Murphy

We would like to thank Prof Ewa Dąbrowska for sharing the Words that Go Together data set