Adverb Position and Information Structure in Processing English

1
Adverb Position and Information Structure in
Processing English

• Britta Stolterfoht, Lyn Frazier Charles
  Clifton, Jr.
• University of Massachusetts, Amherst

Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
2
Overview

• Information Structure and Word Order
• Experimental evidence for information-structural
  constraints on word order variations
• Information Structure and Adverb Placement
• Information-structural constraints dependent on
  adverb placement
• Adverb Position in Processing English
• A self-paced reading study
• 4. Discussion

Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
3
1. Information Structure and Word Order
Information-structural constraints on order
variations in languages with flexible word
order ? change of word order affects
focus-background articulation (e.g.,
Abraham, 1992 Höhle, 1982 Haider, 1993, 2000
Steube, 2000)
SOV (1a) ..., dass die Tante die Nichten begrüßt
hatF that the aunt the niece welcomed has
... that the aunt welcomed the
nieces OSV (1b) ..., dass die Tante die Nichten
begrüßt habenF that the aunt the niece
welcomed have ... that the nieces
welcomed the aunt
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
4
1. Information Structure and Word Order
Experimental evidence for information-structural
constraints on word order variations Bader
Meng (1999) Grammaticality Judgements
SOV (1a) ..., dass die Tante die Nichten begrüßt
hatF 91 (2a) ..., dass sie die Nichten
begrüßt hatF 89 OSV (1b) ..., dass die Tante
die Nichten begrüßt habenF 35 (2b)
..., dass sie die Nichten begrüßt habenF
55
Stolterfoht (2005) Event-Related Brain
Potentials (ERPs) ? Correlate of focus structural
revision (right-central negativity)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
5
1. Information Structure and Word Order
Experimental evidence for information-structural
constraints on word order variations
Kaiser Trueswell (2004) for Finnish Weskott,
Hörnig, Féry, Kern, Fanselow, Kliegl (2005)
for German Eye-movements in visual world
paradigm OVS signals Given-New ordering ?
anticipatory fixations of new referent
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
6
2. Information Structure and Adverb Placement
Adverbs as diagnostic in the analysis of phrase
structure(e.g., Emonds, 1976 Platzack,
1983) Mapping Hypothesis (Diesing, 1992
Kratzer, 1994)
(3) a. ..., weil sie immer Briefe aus Europa
beantwortet. since she always letters from
Europe answers '..., since she is always
engaged in answering letters from Europe.'
b. ..., weil sie Briefe aus Europa immer
beantwortet. since she letters from Europe
always answers '...,since she never
leaves a letter from Europe unanswered.'
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
7
2. Information Structure and Adverb Placement
Adverbs and topicality (Meinunger, 1995)
(4) a. ... als er wieder rauskam war auf einmal
der HUND verschwunden. when he again
out.came was of once the dog
disappeared '... when he came back out,
all of a sudden the DOG had disappeared.' b.
... als er wieder rauskam war der Hund auf einmal
verSCHWUNden. when he again out.came
was the dog of once disappeared
'... when he came back out, all of a sudden the
dog had disapPEARed.'
? DPs to the left of boundary-marking adverbs
occupy specifiers of AgrPs with the feature
topic
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
8
2. Information Structure and Adverb Placement
Topic position above sentential adverbials in the
GermanMittelfeld (Frey, 2000) Sentential
Adverbials (SAs) are ? Evaluatives
(erstaunlicherweise, 'amazingly')? Evidentials
(offensichtlich, 'obviously ? Epistemics
(wahrscheinlich, 'probably') SAs are
characterized as the boundary between given
andnew information (Haftka, 1995, 2003) base
position above all other arguments and adverbial
classes (Frey Pittner, 1998)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
9
2. Information Structure and Adverb Placement
Topic position above sentential adverbials in the
German Mittelfeld (Frey, 2000)
(5) Ich erzähl Dir mal was von Otto. 'I
will tell you something about Otto ') a.
Nächstes Jahr wird Otto wahrscheinlich seine
Kollegin heiraten. Next year
will Otto probably his colleague
marry b. Nächstes Jahr wird wahrscheinlich
Otto seine Kollegin heiraten. Next
year will probably Otto his
colleague marry Next year Otto
probably will marry his colleague.
aboutness topic expression about whose
referent(s) the sentence predicates or makes a
judgement (Reinhart, 1981, 1995)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
10
2. Information Structure and Adverb Placement
Topic position above sentential adverbials in the
German Mittelfeld (Frey, 2000) Test
non-referential expressions
(6) Während des Vortrags hat keiner anscheinend
geschlafen. During the talk
has nobody apparently slept

• aboutness topics must have identifiable
  discourse referents for addressation
• (Jacobs, 2001 Reinhart, 1981, 1995)

Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
11
2. Information Structure and Adverb Placement
Germanic languages (Bobaljik Jonas, 1996) Two
landing positions for subjects in German and
IcelandicSpecAgrSP and SpecTP(sentential)
adverbials attach to TPSpecAgrSP is linked to
topicality One landing position for subjects in
English and DanishAll DPs move to SpecAgrSP,
not sensitive to information-structural status
of DPs
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
12
2. Information Structure and Adverb Placement
One landing position for subjects in English and
Danish All DPs move to SpecAgrSP, not sensitive
to discourse status of DPs (Svenonius, 2002)
(6) a. Have any students probably read the book?
b. Have probably any students read the
book? c. I think that probably some students
have read the book.

• possibility of CP-recursion in English adverbs
  can be attached to AgrP

Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
13
2. Information Structure and Adverb Placement
Conclusion Languages with flexible word order
are sensitive to information-structural
constraints with regard to the order of
arguments (SOV vs. OSV) the placement of
adverbials (AdvS vs. SAdv) A language like
English with a relatively fixed word order is not
sensitive to information-structural constraints
with regard to the order of arguments (SVO)
the placement of adverbials (SAdv)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
14
3. Adverb Position in Processing English
Non-referential expressions in English
(7) a. The envoy said that presumably the king
defeated the knights. b. The envoy said
that the king presumably defeated the knights.
c. The envoy said that presumably no king
defeated the knights. d. ? The envoy said
that no king presumably defeated the knights.
Do adverbs convey information-structure
constraints in a fixed word order language like
English?
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
15
3. Adverb Position in Processing English
(7) a. The envoy said that presumably the king
defeated the knights. b. The envoy said
that the king presumably defeated the knights.
c. The envoy said that presumably no king
defeated the knights. d. The envoy said
that no king presumably defeated the knights.
Hypotheses H1 If English does not have a
specific position for topics, no difference
with regard to the referential status of the
DP should be found (no interaction of order and
subject type)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
16
3. Adverb Position in Processing English
(7) a. The envoy said that presumably the king
defeated the knights. b. The envoy said
that the king presumably defeated the knights.
c. The envoy said that presumably no king
defeated the knights. d. ? The envoy said
that no king presumably defeated the knights.
Hypotheses H2 If English provides a specific
position for topics, a non- referential
subject preceding the adverb should be highly
marked (interaction of order and subject type)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
17
3. Adverb Position in Processing English
Self-paced Reading Study Materials24 items with
12 different sentential adverbials
Evaluatives surprisingly, amazingly,
unfortunately, fortunately Evidentials
evidently, obviously, apparently, supposedly
Epistemics presumably, possibly, probably,
certainly Independent variablesORDER adverb
early vs. adverb lateSUBJECT TYPE referential
vs. negative Participants52 undergraduate
students of the University of Massachusetts
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
18
3. Adverb Position in Processing English
Self-paced Reading Study
ProcedurePresentation --- ----- ---- ----
---------- --- ---- -------- --- -------- The
envoy said that ---------- --- ---- -------- ---
-------- --- ----- ---- ---- presumably the king
defeated the knights. Taskchoosing paraphrases
after every sentenceThe envoy assumed that the
knights won. The envoy assumed that the knights
lost. Four presentation lists with 24
experimental sentences and 88 fillersentences
(each participant saw one version of experimental
items)
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
19
3. Adverb Position in Processing English
Results for Region 1 reading times in ms
SUBJECT TYPE ORDER TYPE x ORDER all ps gt .10
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
20
3. Adverb Position in Processing English
Results for Region 2 reading times in ms
SUBJECT TYPE p1 lt .01 p2 lt .01 ORDER p1 lt .10
p2 lt .05 TYPE x ORDER p1 lt .05 p2 lt .10
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
21
3. Adverb Position in Processing English
Results for Region 2 planned comparison
negative-late with Other comparisons nega
tive-early F lt 1.0 p1 lt .02 p2 lt
.03 referential-early p1 lt .01 p2 lt
.004 referential-late p1 lt .001 p2 lt .007
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
22
3. Adverb Position in Processing English
Results for choosing paraphrases correct
ORDER p1 lt .10 p2 lt .10 SUBJECT TYPE TYPE x
ORDER Fs lt 1.0
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
23
3. Adverb Position in Processing English
Results for choosing paraphrases response
times in ms
SUBJECT TYPE ORDER TYPE x ORDER all Fs lt 1.0
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
24
4. Discussion
Processing difficulties (longer reading times) in
the critical region for sentences with a
non-referential subject preceding the sentential
adverbial. H1 If English does not have a
specific position for topics, no
difference with regard to the referential status
of the DP should be foundH2 If English
provides a specific position for topics, a non-
referential subject preceding the adverb
should be highly marked
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
25
4. Discussion
Evidence against the assumption that English has
only one landing position for subjects which is
not sensitive to information-structural status of
DPs Outlook Further evidence for influence
of information structure Contextual
information? Experimental evidence for topic
position in German?
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
26
THANK YOU and William Evans for his help with
collecting the data !
Linguistic Evidence 2006 Tübingen, 2. 4.
Februar
27
Anderson (2004)
(8) a. A climbing expert scaled every
cliff. ambiguous b. The climbing expert
scaled every cliff. unambiguous c. A
different climbing expert scaled every
cliff. unambiguous

• preference for surface scope
• no effect of ambiguity

Linguistic Evidence 2006 Tübingen, 2. 4.
Februar