Topic-Comment Frames in HPSG Gert Webelhuth Georg-August University - - PowerPoint PPT Presentation

Topic-Comment Frames in HPSG

Gert Webelhuth

Georg-August University G¨

• ttingen

November 24, 2006

Gert Webelhuth Heidelberg

English Fronting

Gundel 1985, 88 There are two different types of OSV sentences in English: Topicalization of Focus:

◮ primary stress falls on the sentence-initial object

(1) a. What do you like? b. beans I like Topicalization of Topic:

◮ initial object receives a high pitched accent ◮ primary stress falls on some other constituent in the sentence

(2) a. How do you feel about beans? b. Beans i like

2

Gert Webelhuth Heidelberg

German behaves differently

Fronted objects need not carry any topic or focus pitch accent: (3) a. Mich friert. b. Mich hat etwas gestochen. c. Mich hat heute jemand vom Finanzamt angerufen

3

Gert Webelhuth Heidelberg

Overview of this Talk

The Main Goal: to make the difference in form-function correspondence between the 3 English and German fronting constructions displayed above expressible in HPSG. Difficulties:

• 1. There is as yet no theory of discourse and information

structure in HPSG that is sufficiently comprehensive to capture the usage generalizations of fronting in English and German (good beginnings by Kordula de Kuthy!).

4

Gert Webelhuth Heidelberg

Analytic strategy

• 1. to associate a semantic representation format with the

syntactic machinery of HPSG that allows direct reference to properties

• 2. to design the semantic representations in such a way that it

can capture both

2.1 the truth-conditional content of signs, and 2.2 the way this content is structured into topic and comment and background and focus.

More concretely:

• 1. choice of semantic representation format: lambda-DRT, plus
• 2. a watered-down variant of Krifka 1992’s theory of

topic-comment articulation

• 3. incorporation of ideas of Sailer, de Kuthy, Jacobs, B¨

uring, and

• thers

5

Gert Webelhuth Heidelberg

Incorporating lambda-DRT into HPSG

The structure of loc

2 6 6 6 4 loc cat cat cont " me type type # 3 7 7 7 5

i e d a-type c-type:

»in type

• ut type

–

type i = individual, e = eventuality, d = DRS

6

Gert Webelhuth Heidelberg

Subtypes of me

con var atom appl:

»func me arg me –

abstr:

»lam var body me –

drs cond me:

ˆ type type ˜ 7

Gert Webelhuth Heidelberg

Subtypes of me

simple-drs:

»univ list-of-var conds list-of-cond –

merged-drs:

»drs1 me drs2 me –

drs Abbr.:

1 + 2 =abbr

2 4 merged-drs drs1

1

drs2

2

3 5

predication:

2 4 pred pred arg0 atom . . . 3 5

neg-drs:

»op ¬ drs drs –

cond-drs:

2 4 drs drs

• p

⇒ drs2 drs 3 5

cond

8

Gert Webelhuth Heidelberg

Examples of contents

Fido:

2 6 6 6 6 4 word phon ˙ fido ¸ ss " loc " cont " type i fido # # # 3 7 7 7 7 5

◮ fido is a constant of type i. ◮ see Muskens 1994, A Compositional Discourse Representation

Theory.

9

Gert Webelhuth Heidelberg

Examples of contents

Fido:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 word phon ˙ fido ¸ ss 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 loc 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 cont 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 abstr type 2 6 6 6 4 c-type in " in i

• ut d

#

• ut d

3 7 7 7 5 lam P body 2 6 4 appl func P arg fido 3 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 10

Gert Webelhuth Heidelberg

Examples of contents

barked:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4

word phon ˙barked¸ ss 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 loc 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 cont 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 abstr type 2 6 6 6 6 6 4 c-type in i

• ut

2 6 4 c-type in e

• ut d

3 7 5 3 7 7 7 7 7 5 lam x body 2 6 6 6 6 6 6 6 6 6 6 6 6 4 abstr type 2 6 4 c-type in e

• ut d

3 7 5 lam e body 2 6 4 simple-drs univ

• conds ˙e:barked(x)¸

3 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5

3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 11

Gert Webelhuth Heidelberg

Semantic composition

ˆ s|l|cont

1

˜ ˆ s|l|cont

2

˜ h s|l ˆ cont apply( 1 , 2 ) ˜ i

apply( 1

" type " in

3

• ut 4

# #

, 2

ˆ type

3

˜) = 2 6 6 4 appl type

4

func

1

arg

2

3 7 7 5

apply( 2

ˆ type

3

˜, 1 " type " in

3

• ut 4

# #

) = apply( 1 , 2 ).

12

Gert Webelhuth Heidelberg

Binding of the event variable

2 6 4s 2 6 4loc " cat

1

cont 2 # nloc 3 3 7 5 3 7 5 2 6 6 6 4 phrase s 2 6 4loc " cat

1

cont ec( 2 ) # nloc 3 3 7 5 3 7 7 7 5

ec(

2 4type|out∗

»in e

• ut d

– λx1. . . xnλeD

3 5) = 2 4λy1. . . yn

B @ e’ + λx1. . . xnλeD(y1. . . yn,e’) 1 C A

3 5 13

Gert Webelhuth Heidelberg

Existentential closure of barked

2 6 6 6 4

word phon ˙barked¸ ss 2 6 6 4loc 2 6 4cont 1 λxe e:barked(x) 3 7 5 3 7 7 5

3 7 7 7 5 "

phrase ss h loc h cont ec( 1 ) i i

# 14

Gert Webelhuth Heidelberg

Existentential closure of barked,2

2 6 6 6 6 6 4

word phon ˙barked¸ ss 2 6 6 6 6 4 loc 2 6 6 6 6 4 cont 2 6 6 6 6 4 type i(ed) term λxe e:barked(x) 3 7 7 7 7 5 3 7 7 7 7 5 3 7 7 7 7 5

3 7 7 7 7 7 5 2 6 6 6 4

phrase ss 2 6 6 6 6 4 loc 2 6 6 6 6 4 cont 2 6 6 6 6 4 type id term λy e’ e’:barked(y) 3 7 7 7 7 5 3 7 7 7 7 5 3 7 7 7 7 5

3 7 7 7 5 15

Gert Webelhuth Heidelberg

Fido barked

Function-argument application:

2 4

word phon ˙fido¸ ss h loc h cont 1 λP.P(fido) i i

3 5 2 6 6 6 4

word phon ˙barked¸ ss 2 6 6 4loc 2 6 4cont 2 λy e’ e’:barked(y) 3 7 5 3 7 7 5

3 7 7 7 5

Subj H

"

phrase ss h loc h cont apply( 1 , 2 ) i i

# 16

Gert Webelhuth Heidelberg

Fido barked

apply(λP.P(fido), λy

e’ e’:barked(y)

) =

2 6 6 6 6 6 6 4 appl func λP.P(fido) arg λy e’ e’:barked(y) 3 7 7 7 7 7 7 5

which reduces to e’ e’:barked(fido)

17

Gert Webelhuth Heidelberg

Topic-comment structures

Krifka 1992:

◮ The content of a sentence is structured into a topic and a

comment.

◮ Both topic and comment may themselves be structured into

background and focus to signal that the topic or comment is considered in contrast to a set of salient comparable alternatives in the context of utterance. The truth-functional contribution of a sign is obtained from its information structure as follows:

• 1. For background and focus: apply the background to the focus.
• 2. For the information structure as a whole: apply the comment

to the topic. This means that the background of the comment contains bound variables “representing” the focus of the comment and the topic!

18

Gert Webelhuth Heidelberg

Examples of informationally structured expressions

General structure: istr: < comm, top > cont: comm(top). Example: [top Fido] [comm barked] istr: < λxi.bark(x), fidoi > cont: bark(fido). Example: [comm fido] [top barked] istr: < λPid.P(fido), λxi.bark(x) > cont: bark(fido).

19

Gert Webelhuth Heidelberg

Examples of informationally structured expressions, 2

The comment and the topic may each be structured into background and focus: istr: < < comm bgcomm, foccomm >, < top bgtop, foctop > > cont: [(bgcomm(foccomm)](bgtop(foctop)) Example: [ [contr−top sandy] [comm [ncontr likes] [contr jill ]]] istr: <<λzλx.likes(x,z),jill>, <λvi.v,sandy>> Truth-functional content: [λzλx.likes(x,z)](jill) = λx.likes(x,jill) [λvi.v](sandy) = sandy [λx.likes(x,jill)](sandy) = likes(sandy,jill).

20

Gert Webelhuth Heidelberg

Adding features to loc to create an information structure

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 loc cat cat cont me ist 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 istr comm 2 6 6 6 6 6 4 comm-str t-bind list-of-var c-bind list-of-var bg me foc list-of-me 3 7 7 7 7 7 5 top list-of- B B B @ 2 6 6 6 4 top-str c-bind list-of-var bg me foc list-of-me 3 7 7 7 5 1 C C C A 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 21

Gert Webelhuth Heidelberg

Some types and constraints

top-str comm-str:

ˆ t-bind list-of-var ˜

top-comm-str:

2 4 c-bind list-of-var bg me foc list-of-me 3 5

The Contrast Constraint In every feature structure of type top-comm-str, there is a bijective relationship between the c-bound variables inside the value of bg and the members of foc The Topic Constraint In every feature structure of type istr, there is a bijective relationship between the t-bound variables inside the value of comm|bg and the members of top

22

Gert Webelhuth Heidelberg

Subtypes of istr

ncontr-comm-istr contr-comm-istr comm-istr ncontr-top-istr contr-top-istr top-istr mixed-istr istr

23

Gert Webelhuth Heidelberg

Constraint on comm-istr

comm-istr →

ˆ top ˜

Example:

2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind c-bind bg fido foc

• 3

7 7 7 5 top

• 3

7 7 7 7 7 5 24

Gert Webelhuth Heidelberg

Constraint on top-istr

top-istr →

ˆ top ne-list ˜

Example:

2 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙

1

¸ c-bind bg

1 var

foc

• 3

7 7 7 5 top * 2 6 4 c-bind bg fido foc

• 3

7 5 + 3 7 7 7 7 7 7 7 7 7 7 5 25

Gert Webelhuth Heidelberg

Constraints on words and phrases

The mixed-istr Constraint If the istr of a phrase is of type mixed-istr, then either (A) the phrase has one daughter whose istr is of type top-ist and another

• ne whose istr is of type comm-ist or (B) the phrase

has a daughter whose istr is of type mixed-istr or

ˆ

... istr top-istr˜

ˆ

... istr comm-istr˜ (A)

ˆ

... istr mixed-istr˜

. . .

ˆ

... istr mixed-istr˜. . . (B)

ˆ

... istr mixed-istr˜

26

Gert Webelhuth Heidelberg

The Phrasal Uniformity Constraint If the istr of a phrase is of type τ, for τ a subtype of comm-istr or top-istr, then the istr of each daughter likewise is of type τ.

ˆ

s|l|istr τ˜

ˆ

s|l|istr τ˜

ˆ

s|l|istr (τ ∧ comm-istr) ∨ (τ ∧ top-istr)˜

27

Gert Webelhuth Heidelberg

Principles for computing the ISTR of phrases

General case:

2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind list c-bind list bg β1 foc list 3 7 7 7 5 top list 3 7 7 7 7 7 5 2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind list c-bind list bg β2 foc list 3 7 7 7 5 top list 3 7 7 7 7 7 5 2 6 6 6 6 4

comm 2 6 6 6 4 t-bind append-corr-dtr-lists c-bind append-corr-dtr-lists bg apply(β1,β2) foc append-corr-dtr-lists 3 7 7 7 5 top append-corr-dtr-lists

3 7 7 7 7 5 28

Gert Webelhuth Heidelberg

Principles for computing the ISTR of phrases, 2

Special case 1: both daughters have a c-bound background:

2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind c-bind ˙ X1 ¸ bg X1 foc ˙ F1 ¸ 3 7 7 7 5 top

• 3

7 7 7 7 7 5 2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind c-bind ˙ X2 ¸ bg X2 foc ˙ F2 ¸ 3 7 7 7 5 top

• 3

7 7 7 7 7 5 2 6 6 6 6 4

comm 2 6 6 6 4 t-bind c-bind ˙Z¸ bg Z foc ˙apply(F1,F2)¸ 3 7 7 7 5 top

• 3

7 7 7 7 5 29

Gert Webelhuth Heidelberg

Principles for computing the ISTR of phrases, 3

Special case 2: both daughters have a t-bound background:

2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ X1 ¸ c-bind bg X1 foc

• 3

7 7 7 5 top ˙ T1 ¸ 3 7 7 7 7 7 5 2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ X2 ¸ c-bind bg X2 foc

• 3

7 7 7 5 top ˙ T2 ¸ 3 7 7 7 7 7 5 2 6 6 6 6 4

comm 2 6 6 6 4 t-bind ˙Z¸ c-bind bg Z foc

• 3

7 7 7 5 top combine(T1,T2)

3 7 7 7 7 5 30

Gert Webelhuth Heidelberg

Principles for computing the ISTR of phrases, 4

Special case 3: Head-filler phrases

2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind list c-bind list bg β1 foc list 3 7 7 7 5 top list 3 7 7 7 7 7 5 2 6 6 6 6 6 6 6 6 6 4 s 2 6 6 6 6 6 6 6 6 6 4 l 2 6 6 6 6 6 4 istr 2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind list c-bind list bg β2 foc list 3 7 7 7 5 top list 3 7 7 7 7 7 5 3 7 7 7 7 7 5 n|s D ˆ s|l|i|c|bg V ˜ E 3 7 7 7 7 7 7 7 7 7 5 3 7 7 7 7 7 7 7 7 7 5

F H

2 6 6 6 6 4

comm 2 6 6 6 4 t-bind append-corr-dtr-lists c-bind append-corr-dtr-lists bg apply(λV .β2, β1) foc append-corr-dtr-lists 3 7 7 7 5 top append-corr-dtr-lists

3 7 7 7 7 5 31

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE

Text: A red light you stop the machine. A green light you ignore. Assumptions: [S [top A greenfoc light]i [comm you ignorefoc ti ]] Desired topic-comment structure:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 t-bind ˙ R(id)d ¸ comm R(λy e’ e’:ignore(hear,y) ) top * λN B B B B @ u light(u) green(u) + N(u) 1 C C C C A + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 32

Gert Webelhuth Heidelberg

The Fronting Construction

front-ph = ⇒

»phon|accs A1 ⊕ high-pitched ⊕ A2 s|l|istr top-istr – ˆ s|l|istr comm-istr ˜

F H

ˆ

hd-fill-ph˜

Constraint The nuclear accent of an expression is contributed by a subexpression whose istr is of type comm-istr.

33

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 2

[t]:

2 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 6 6 6 4 t-bind c-bind bg " q1 type q # foc

• 3

7 7 7 7 7 7 5 top

• 3

7 7 7 7 7 7 7 7 5 34

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 3

[ignore]contr−comm:

2 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 6 6 6 6 6 6 4 t-bind c-bind ˙ T0q(i(ed)) ¸ bg T0 foc * λQλxλe.Q(λy e:ignore(x,y) ) + 3 7 7 7 7 7 7 7 7 7 5 top

• 3

7 7 7 7 7 7 7 7 7 7 7 5

[ec(ignore)]:

2 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 6 6 6 6 6 6 4 t-bind c-bind ˙ T1q(id) ¸ bg T1 foc * λQλx.Q(λy e’ e’:ignore(x,y) ) + 3 7 7 7 7 7 7 7 7 7 5 top

• 3

7 7 7 7 7 7 7 7 7 7 7 5 35

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 4

[ec(ignore) t]:

2 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 6 6 6 6 6 6 4 t-bind c-bind ˙ T1q(id) ¸ bg T1(q1) foc * λQλx.Q(λy e’ e’:ignore(x,y) ) + 3 7 7 7 7 7 7 7 7 7 5 top

• 3

7 7 7 7 7 7 7 7 7 7 7 5

[you]ncontr−comm:

2 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind c-bind bg hear foc

• 3

7 7 7 5 top

• 3

7 7 7 7 7 5 36

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 5

[you ec(ignore) t]:

2 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 6 6 6 6 6 6 4 t-bind c-bind ˙ T1q(id) ¸ bg T1(q1)(hear) foc * λQλx.Q(λy e’ e’:ignore(x,y) ) + 3 7 7 7 7 7 7 7 7 7 5 top

• 3

7 7 7 7 7 7 7 7 7 7 7 5

[light]ncontr−top:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ Rid ¸ c-bind bg R foc

• 3

7 7 7 5 top * 2 6 6 6 6 6 6 4 c-bind bg λz light(z) foc

• 3

7 7 7 7 7 7 5 + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 37

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 6

[green]contr−top:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ M(id)(id) ¸ c-bind bg M foc

• 3

7 7 7 5 top * 2 6 6 6 6 6 6 6 4 c-bind ˙ M2(id)(id) ¸ bg M2 foc * λPλv B B @ green(v) + P(v) 1 C C A + 3 7 7 7 7 7 7 7 5 + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 38

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 7

[a]ncontr−top:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ D(id)((id)q) ¸ c-bind bg D foc

• 3

7 7 7 5 top * 2 6 6 6 6 6 6 4 c-bind bg λMλN B B @ u + M(u) + N(u) 1 C C A foc

• 3

7 7 7 7 7 7 5 + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 39

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 8

[green light]:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ Oid ¸ c-bind bg O foc

• 3

7 7 7 5 top * 2 6 6 6 6 6 6 6 6 6 6 6 6 6 4 c-bind ˙ M2(id)(id) ¸ bg M2 B B @λz light(z) 1 C C A foc * λPλv B B @ green(v) + P(v) 1 C C A + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 5 + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 40

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 9

[a green light]:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 4 t-bind ˙ R(id)d ¸ c-bind bg R foc

• 3

7 7 7 5 top * 2 6 6 6 6 6 6 6 6 6 6 6 6 6 4 c-bind ˙ M2(id)(id) ¸ bg λN B B @ u + M2 B B @λz light(z) 1 C C A (u) + N(u) 1 C C A foc * λPλv B B @ green(v) + P(v) 1 C C A + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 5 + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 41

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 10

[a green light you ec(ignore) t]:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 comm 2 6 6 6 6 6 6 6 6 6 4 t-bind ˙ R(id)d ¸ c-bind ˙ T1q(id) ¸ bg T1(R)(hear) foc * λQλx.Q(λy e’ e’:ignore(x,y) ) + 3 7 7 7 7 7 7 7 7 7 5 top * 2 6 6 6 6 6 6 6 6 6 6 6 6 6 4 c-bind ˙ M2(id)(id) ¸ bg λN B B @ u + M2 B B @λz light(z) 1 C C A (u) + N(u) 1 C C A foc * λPλv B B @ green(v) + P(v) 1 C C A + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 5 + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 42

Gert Webelhuth Heidelberg

Derivation of A GREEN light you IGNORE, 11

[a green light you ec(ignore) t]:

2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 t-bind ˙ R(id)d ¸ comm R(λy e’ e’:ignore(hear,y) ) top * λN B B B B @ u light(u) green(u) + N(u) 1 C C C C A + 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 43

Gert Webelhuth Heidelberg

Conclusion

What we (hopefully) have: an HPSG syntax that

◮ compositionally derives discourse representation structures ◮ stratified into topic-comment and background-focus

What this framework can (hopefully) be used for: to capture the rich dependencies between

◮ sentence form ◮ sentence meaning, and ◮ information structure.

44

Gert Webelhuth Heidelberg

The End

Gert Webelhuth Heidelberg