Transcript Grammatical Noriegas

That vexed problem of choice
reflections on experimental design and statistics
with corpora
ICAME 33
Leuven 30 May-3 June 2012
Sean Wallis, Jill Bowie and Bas Aarts
Survey of English Usage
University College London
{s.wallis, j.bowie, b.aarts}@ucl.ac.uk
Outline
• Introduction
• Definitions
• Refining baselines and the ratio principle
• Surveying ‘absolute’ and ‘relative’ variation
• Potential sources of interaction
• Employing alternation analysis
• Objections
• Conclusions
Introduction
• Research questions are really about choice
– If speakers had no choice about the words or
constructions they used, language would be invariant!
• Lab experiments
– Press button A or button B
• Corpus
– Speakers may choose construction A or B
• But they can only actually chose one, A, at each point
• We have to infer the other type, B,
counterfactually
• Identifying alternates is often non-trivial
Mutual substitution
• Mutual substitution A  B
– Given a corpus, identify all events of Type A that
alternate with events of Type B, such that A is
mutually replaceable by B, without altering the
meaning of the text.
• Replacement
– B replaces A if B increases, and vice-versa
• p (A)+p (B)+... = 1
• Freedom to vary
• p (X)  [0, 1]
– Ideal: eliminate invariant Type C terms
Mutual substitution
• Mutual substitution A  B
– Pronoun who/whom
• A = whom
• B = who
Mutual substitution
• Mutual substitution A  B
– Pronoun who/whom
• A = whom
• B = who (objective)
– But whom is limited to objective case
• C = who (subjective)
• We therefore limit alternation to Objects
– If whom is used ‘incorrectly’ as a Subject, it has an
additional constraint (social disfavour)
True rate of alternation
• True rate of alternation
– If A  B
F (A)
• p (A | {A, B}) =
F (A)+F (B)
True rate of alternation
• True rate of alternation
– If A  B
F (A)
• p (A | {A, B}) =
F (A)+F (B)
• Proportion (fraction) of all cases that are Type A
– we use p (A) as a shorthand for p (A | {A, B}) if the
baseline {A, B} is stated
True rate of alternation
• True rate of alternation
– If A  B
F (A)
• p (A | {A, B}) =
F (A)+F (B)
• Proportion (fraction) of all cases that are Type A
– we use p (A) as a shorthand for p (A | {A, B}) if the
baseline {A, B} is stated
• Contingency tables
IV
DV A
Total
B
condition 1 f1(A) f1(B) f1(A)+f1(B)
condition 2 f2(A) f2(B) f2(A)+f2(B)
Total
F (A) F (B) F (A)+F (B)
probability
p1(A)
p2(A)
p (A)
True rate of alternation
• Shall/will alternation over time in DCPSE
1
p
baseline = {shall, will}
0.8
0.6
0.4
0.2
0
1955
1960
1965
1970
(Aarts et al., forthcoming)
1975
1980
1985
1990
1995
True rate of alternation
• Shall/(will+’ll) alternation over time in DCPSE
1
baseline = {shall, will, ’ll}
p
0.8
0.6
0.4
0.2
0
1955
1960
1965
1970
(Aarts et al., forthcoming)
1975
1980
1985
1990
1995
True rate of alternation
• Logistic ‘S’ curve assumes freedom to vary
– p (X)  [0, 1]
1
p
t
0
True rate of alternation
• Logistic ‘S’ curve assumes freedom to vary
– p (X)  [0, 1]
1
p
shall/(will+’ll)
shall/’ll
t
0
– as do Wilson confidence intervals
Refining baselines
• Over-general baselines
– conflate opportunity and use
– ‘normalisation’ per million words
• implies that every word other than A is Type B!
• is this plausible?
B A
• ‘Art’ of experimental design
– refine baseline by narrowing dataset
• reduce and eliminate non-alternating Type C cases
• optionally: subdivide where different constraints apply
– different baselines test different hypotheses
• cf. shall / will / ’ll
Refining baselines
• Tensed VPs per million words, DCPSE
Total:
constant over
time
140,000
120,000
Diachronic
variation:
within text
categories
100,000
80,000
60,000
(Bowie et al., forthcoming)
Total
prepared sp
assort spont
legal x-exam
parliament
commentary
b interviews
b discussions
telephone
informal f-to-f
formal f-to-f
20,000
0
Synchronic
variation:
between text
categories
LLC
ICE-GB
40,000
The ratio principle
• Simple algebra
– any sequence of ratios can be reduced to the ratio
of the first and last term:
F (modal)
F (tVP)
F (modal)


F (tVP)
F (word)
F (word)
The ratio principle
• Simple algebra
– any sequence of ratios can be reduced to the ratio
of the first and last term:
F (modal)
F (tVP)
F (modal)


F (tVP)
F (word)
F (word)
– we saw that the ratio tVP:word varies
synchronically and diachronically in DCPSE
• we can eliminate this variation by simply focusing on
modal:tVP
• use tensed VPs as baseline for modals
The ratio principle
• Simple algebra
– any sequence of ratios can be reduced to the ratio
of the first and last term:
F (modal)
F (tVP)
F (modal)


F (tVP)
F (word)
F (word)
– we saw that the ratio tVP:word varies
synchronically and diachronically in DCPSE
• we can eliminate this variation by simply focusing on
modal:tVP
• use tensed VPs as baseline for modals
– this baseline is not a strict alternation set
• we have not eliminated all Type C terms
‘Absolute’ and ‘relative’ variation
• Changes in core modals over time in DCPSE
p (modal | tVP)
p (modal | modal tVP)
0.30
0.04
0.25
0.03
0.20
0.15
0.02
0.10
0.01
0.05
0.00
0.00
can
could may might must shall should will would
(Bowie et al., forthcoming)
Left axis:
absolute change
as a proportion
of tensed VPs
Right axis:
relative change
as a proportion
of set of modals
Employing alternation analysis
• Simple grammatical interaction
– Independent and dependent variables are
grammatical
• mutual substitution concerns the dependent variable
Employing alternation analysis
• Simple grammatical interaction
– Independent and dependent variables are
grammatical
• mutual substitution concerns the dependent variable
– Numerous examples in Nelson et al. 2002
• e.g. clause table: mood  transitivity
IV
DV
exclamative
interrogative
Total
montr
ditr
CL(montr, exclam) CL(ditr, exclam)
CL(montr, inter)
CL(ditr, inter)
…
…
CL(montr)
CL(ditr)
Total
CL(exclam)
CL(inter)
…
CL
• not alternation, but survey: could be refined
Employing alternation analysis
• Repeating choices: to add or not to add
– e.g. repeated decisions to add an attributive AJP to specify
a NP head: the tall white ship
• A = add AJP
• B = don’t add AJP (and stop)
Employing alternation analysis
• Repeating choices: to add or not to add
– e.g. repeated decisions to add an attributive AJP to specify
a NP head: the tall white ship
• A = add AJP
• B = don’t add AJP (and stop)
– Sequential analysis: examine p (A | {A, B}) at each step
0.25
p
Conclusion:
decision to add
an AJP becomes
successively
more difficult
0.20
0.15
0.10
0.05
0.00
0
1
(Wallis, forthcoming)
2
3
4
Employing alternation analysis
• Grammatically diverse alternates
– Biber and Gray (forthcoming) investigate evidence for
increasing nominalisation
• A = nouns that have been derived from verb forms
– This paper reports an analysis of Tucker’s central prediction
system model and an empirical comparison of it with two
competing models. [1965, Acad-NS]
• B = verbs that could be nominalised
Employing alternation analysis
• Grammatically diverse alternates
– Biber and Gray (forthcoming) investigate evidence for
increasing nominalisation
• A = nouns that have been derived from verb forms
– This paper reports an analysis of Tucker’s central prediction
system model and an empirical comparison of it with two
competing models. [1965, Acad-NS]
• B = verbs that could be nominalised
– Could just use clauses as baseline
• But this is little better than words
– Better option is to enumerate types
• analysis
• prediction
• comparison
• analyse
• predict
• compare
Employing alternation analysis
• Grammatically diverse alternates
– Biber and Gray (forthcoming) investigate evidence for
increasing nominalisation
• A = nouns that have been derived from verb forms
– This paper reports an analysis of Tucker’s central prediction
system model and an empirical comparison of it with two
competing models. [1965, Acad-NS]
• B = verbs that could be nominalised
– Could just use clauses as baseline
– Better option is to enumerate types
• analysis
• prediction
• comparison
• analyse
• predict
• compare
– Examine cases: is alternation possible?
Objections
• If this is such a good idea, why isn’t
everybody doing it?
• Three main objections are made:
 alternates are not reliably identifiable
 baselines are arbitrarily chosen by the
researcher
 different constraints apply to different terms
(no such thing as free variation)
Alternates are not reliably identifiable?
• Identifying alternates can be difficult
– phrasal vs. Latinate verbs
Alternates are not reliably identifiable?
• Identifying alternates can be difficult
– phrasal vs. Latinate verbs
• Strategies:
 enumerate cases from bottom, up
• find Type B cases for each Type A
Alternates are not reliably identifiable?
• Identifying alternates can be difficult
– phrasal vs. Latinate verbs
• Strategies:
 enumerate cases from bottom, up
• find Type B cases for each Type A
put up tolerate
?position
build, make
display, project
sell
propose
increase
accommodate
finance
4
3
3
2
2
2
1
1
1
put up with it [S1A-037 #1]
put your feet up [S1A-032 #21]
shacks put up without any planning [S2B-022 #118]
put up two… trees [on the screen] [S1B-002 #157]
put the plant up for sale [W2C-015 #8]
put [a motion] up [S1B-077 #127]
put up the poll tax [W2C-009 #3]
we could put up the children [S1A-073 #197]
put up the money [W2F-007 #36]
Alternates are not reliably identifiable?
• Strategies:
 enumerate cases from bottom, up
• find Type B cases for each Type A
Alternates are not reliably identifiable?
• Strategies:
 enumerate cases from bottom, up
• find Type B cases for each Type A
 refine baseline from top, down
• start with verbs, eliminate non-alternating Type Cs
– Copular verbs
– Clitics
– Stative verbs
• are dynamic verbs the upper bound for alternation
with phrasal verbs?
Alternates are not reliably identifiable?
• Strategies:
 enumerate cases from bottom, up
• find Type B cases for each Type A
 refine baseline from top, down
• start with verbs, eliminate non-alternating Type Cs
– Copular verbs
– Clitics
– Stative verbs
• are dynamic verbs the upper bound for alternation
with phrasal verbs?
– combine strategies:
• identify stative verbs lexically
• a few verbs are stative and dynamic
– check in situ
Baselines are arbitrary?
• Is there such an ‘objective’ baseline?
– No, but optimum baselines identify where
speakers have a real choice: Type A vs. Type B
• Baselines are a control
– Experimental hypothesis:
• the ratio of Type A to the baseline is constant over
values of independent variable
– Baseline cited as part of experimental reporting
• Indeed we can experiment with baselines
– e.g. does the present perfect correlate
more with past-referring or
present-referring VPs?
Comparing baselines
• Does the present perfect correlate more with
past-referring or present-referring VPs?
Comparing baselines
• Does the present perfect correlate more with
past-referring or present-referring VPs?
present
present perf
present non-perf
Total
LLC
ICE-GB
2,696
2,488
5,184
33,131
32,114
65,245
35,827
34,602
70,429
present perf
other TPM VPs
Total
2,696
2,488
5,184
18,201
14,293
32,494
20,897
16,781
37,678
Total
past
LLC
ICE-GB
Total
(Bowie et al., forthcoming)
Comparing baselines
• Does the present perfect correlate more with
past-referring or present-referring VPs?
present
present perf
present non-perf
Total
LLC
ICE-GB
2,696
2,488
5,184
33,131
32,114
65,245
35,827
34,602
70,429
present perf
other TPM VPs
Total
2,696
2,488
5,184
18,201
14,293
32,494
20,897
16,781
37,678
Total
past
LLC
ICE-GB
Total
– Present perfect correlates more with
present-referring VPs
(Bowie et al., forthcoming)
d% = -4.455.13%
f’ = 0.0227
c2 = 2.68ns
d% = +14.925.47%
f’ = 0.0694
c2 = 25.06s
Different constraints apply in each case?
• Speakers choices are influenced by multiple
pressures
– to talk about a single ‘choice’ is misleading
– there is no such thing as free variation
• We are not attempting to infer “the reason” for
a particular speaker decision
– we are attempting to identify statistically sound
• patterns
• correlations
• trends
– across many speakers
Different constraints apply in each case?
• Does one or more of these multiple constraints
represent a systematic bias on the true rate?
 Yes = try to identify it experimentally
 No = ‘noise’
• Can focus on subset of cases to restrict
different influences
– e.g. limit shall / will by modal semantics
• This objection is misplaced:
– freedom to vary
= grammatical and semantic possibility (potential)
= not that choices are free from influence
A competitive ecology?
• Not everything is a binary choice
– but the same principles apply
100%
100%
p
p
Meanings of THINK
Complementation patterns of HOPE
hoping that / Ø
80%
80%
‘cogitate’
60%
60%
40%
40%
hoping to
‘intend’
20%
20%
quotative
hoping for
interpretive
0%
0%
1920s
1960s
(Levin, forthcoming)
2000s
1920s
1960s
2000s
Conclusions
• Researchers need to pay attention to questions of
choice and baselines
– This does not mean that an observed change is due to a
single source
• Minimum condition: baseline is a control
– statistics evaluate difference from this control
• is it a good control?
• Alternation studies: baseline is opportunity for
making choice under investigation
• Word-based baselines should only really be used for
comparison with other studies
– we should not make statements about choice
unless we investigate that question
Conclusions
• ‘Alternation’ can be interpreted
– strictly
• all Type As and Type Bs identified and cases checked
– generously
• small number of Type Cs permitted
– Alternation is semantically bounded but
grammatical analysis helps identify cases!
• We may try different experimental designs,
modifying baselines and subsets
– many more novel experiments are possible
• experimental assumptions
should always be clearly reported
References
ACLW: Aarts, B., J. Close, G. Leech and S.A. Wallis (eds.) (forthcoming).
The Verb Phrase in English: Investigating recent language change with
corpora. Cambridge: CUP.
Preview at www.ucl.ac.uk/english-usage/projects/verb-phrase/book.
•
Aarts, B., J. Close and S.A. Wallis. forthcoming. Choices over time:
methodological issues in investigating current change. ACLW Chapter 2.
•
Biber, D. and B. Gray. forthcoming. Nominalizing the verb phrase in academic
science writing. ACLW Chapter 5.
•
Bowie, J., S.A. Wallis and B. Aarts, forthcoming. The perfect in spoken English.
ACLW Chapter 13.
•
Levin, M., forthcoming. The progressive verb in modern American English. ACLW
Chapter 8.
•
Nelson, G., S.A. Wallis and B. Aarts. 2002. Exploring Natural Language.
Amsterdam: John Benjamins.
•
Wallis, S.A. forthcoming. Capturing linguistic interaction in a grammar:
a method for empirically evaluating the grammar of a parsed corpus.
Statistical postscript
• Type Cs make statistical tests less sensitive
– What happens to confidence intervals as we add
to F (A)+F (B) = 100 alternating cases?
0.25
eN/100
F (A)
0.2
95
80
60
0.15
Including Type Cs
makes statistical
tests conservative
40
0.1
20
0.05
5
0
100
1,000
10,000
Tests assume
freedom to vary
(F (A)+F (B) = N )
N