The description of information structure (Topic-Focus
Articulation, TFA) within a dependency framework
Eva Hajicova, Charles University, Prague
1. WHERE TO REPRESENT: UNDERLYING STRUCTURE VS. SURFACE
STRUCTURE
an interface level: representation of the meaning=09
- tectogrammatics - on which the irregularities of the
outer shape of sentences are absent (including synonymy
and at least the prototypical layer of ambiguity)
'tectogrammatics' differs from most other
predicate-argument levels, and partly also from the
standard view of Logical Form, in several ways, three
of which appear to be among the most relevant:
(i) the tectogrammatical level is conceived of as one
of the layers of language structure, in which no
devices created in the metalanguage of logic are
present,
(ii) the core of sentence structure is viewed upon as
based on the dependency relation (i.e. on pairs of
heads and their complementations), and
(iii) on this level, the articulation (bipartition) of
the sentence into its topic and focus is overtly
represented (where 'topic' is a linguistic counterpart
of the presystemic notion of 'known' or 'given'
information and 'focus' is that part of the sentence
structure that conveys 'new information')
an analysis of focus is not a restricted, specific task,
but, rather, a necessary ingredient of any systematic
account of sentence structure in general
Chomsky's (1969) view has much in common with the Prague
tradition ('imported' to the USA by M.A.K.Halliday in
1967), similarly as the views of S. Kuno, of E. Prince, or,
even more recently, those of E. Engdahl and E. Vallduv=A1
most of the other recent views (M. Rooth, J. Jacobs, M.
Krifka, and others): primarily based on the analysis of
focus sensitive operators, rather than on that of the
sentence structure
(this difference, however, is in fact not as crucial as has
often been assumed)
(a) ARGUMENTS FOR REPRESENTING TFA ON AN UNDERLYING LEVEL:
SEMANTIC RELEVANCE OF TFA
(A) contextual appropriateness (no difference in truth
conditions:
(1)(a) Columbus discovered America.
Cz.: Kolumbus objevil Ameriku.
(b) (It was) COLUMBUS (who) discovered America.
Cz.: Ameriku objevil Kolumbus. or: KOLUMBUS objevil
Ameriku (emphatic, "subjective order")
(B) presupposition versus allegation:
(2)(a) My son (didn't) cause(d) our defeat.
(b) Our defeat was (not) caused by my son.
allegation: that kind of entailment that is triggered by
the affirmative assertion, whereas the corresponding
negative assertion triggers neither the allegation, nor its
denial; only the existence of the speaker's son is
presupposed by (a), and only the defeat by (b); in the
reversed case, the given piece of information is just an
allegation
in Czech (and also in German and other Continental
languages): passivization not necessary, cf. (2)(c):
(2)(c) Na=A8i por=A0=91ku nezp=96sobil m=96j syn.
Our-Acc defeat-Acc did-not-cause my-Nom son-Nom
only the TFA differs, the syntactic structure of the two
sentences is identical in all other respects
(C) truth conditions proper:
(i) focalizers:
(3)(a) Jim only put salt on the potatoes.
(b) Jim only put SALT on the potatoes.
readings of (3)(a): 1. Jim did not do anything else than put
salt on the potatoes, with just the subject in Topic
2. Jim put nothing anywhere else than salt on the
potatoes, with the verb in Topic,
3. Jim put salt nowhere else than on the potatoes,
where also 'salt' belongs to Topic
(3)(b) lacks this ambiguity; here only 'salt' (or 'put
salt') can belong to the focus of the focalizer.
the "free" word order of Czech is determined by TFA,
distinguishing the equivalents of (3)(a):
(3')(a1) Jim jen dal s=96l na brambory.
(3')(a2) Jim dal jen s=96l na brambory.
(3')(a3) Jim dal s=96l jen na brambory.
(3')(b1) Jim na brambory dal jen s=96l.
(b2) Jim na brambory jen dal s=96l.
(with the unmarked prosodic contour, with the
intonation centre at the end)
(4)(a) John only introduced Bill to SUE.
(b) John only introduced BILL to Sue.
(5)(a) Mary always takes John to the movies.
(b) Mary always takes JOHN to the movies.
explicit if-clauses in construction with adverbs of
quantification (ex. from Partee):
(6)(a) Always, if John saw a street musician, he gave
him a shiny coin.
(b) Always, if John had a shiny coin, he gave it to
a street musician.
(7)(a) John always gave a street musician a shiny coin.
(b) John always gave a shiny coin to a street musician.
or
(b') John always gave a STREET MUSICIAN a shiny
coin.
(ii) certain quantifiers:
(8)(a) Many men read few books.
(b) Few books are read by many men.
(8')(a) Mnoho lid=A1 (Nom.) =87te m=A0lo knih (Acc.).
(b) M=A0lo knih (Acc.) =87te mnoho lid=A1 (Nom.).
(9)(a) Londoners are mostly in Brighton.
(b) In Brighton there are mostly Londoners
(D) implicit quantifiers (or focalizers?):
(10)(a) English is spoken in the Shetlands
(b) ENGLISH is spoken in the Shetlands.
(c) English is spoken only in the Shetlands.
(10')(a) Anglicky se mluv=A1 na Shetlandsk=98ch ostrovech.
(b) Na Shetlandsk=98ch ostrovech se mluv=A1 anglicky.
(c) Anglicky se mluv=A1 jenom na Shetlandsk=98ch ostrovech.
(11)(a) One smokes in the hallway.
(b) In the hallway one smokes.
Kuno's exhaustive listing:
which structures invite exhaustive listing: the relevant
item in the Focus of the sentence, with the verb in the
Topic (as a contextually bound element), or the verb has
a 'general' meaning (e.g. 'carry out', 'perform', ...) or
the verb is semantically close to the topic, i.e. it 'adds'
only a little to the Topic (as e.g. in (10) above)
(E) some other cases:
(sign in London underground: Halliday 1970)
(12) (a) Dogs must be carried.
(b) DOGS must be carried.
why-questions:
(13)(a) Why did Clyde marry Bertha?
(b) Why did Clyde MARRY Bertha?
(c) Why did CLYDE marry Bertha?
(b) ARGUMENTS FOR A DEPENDENCY-BASED UNDERLYING STRUCTURE:
immediate-constituents-based approach (Chomsky and others):
topic (focus) assumed to be a single constituent
in the general case: not so (see Haji=87ov=A0 and Sgall 1975)
(14) (We are not going for a fortnight to Mallorca,) but for
a weekend to our grandmother.
(with the two adverbials, or PPs, in the focus, the topic
being deleted)
(15) (What happened to Jim?) A burglar INJURED him.
(with both the subject and the verb in the focus, the topic
consisting only of the pronominal object)
two dimensions of the dependency tree:
horizontal: TFA
vertical: dependency
advantage: if sentences differ only in TFA, they have the
same vertical dimension relations, but differ in
horizontal demension
in the minimalist theory: the "bottom of the tree" vs.
what is "above" the VP (if subject-internal VP)
however: TFA is not a theory dependent issue!
2. WHAT TO REPRESENT: WHICH MAIN DISTINCTIONS ARE TO BE MADE
(i) the (elementary) notions of contextually bound (CB) and
non-bound (NB) nodes of the tree:
the linguistic counterpart of the pragmatic notions of
'given' and 'new' information, but determined rather by
grammatical and prosodic properties (e.g. 'strong' and
'weak' pronominal forms, specific morphemes or syntactic
constructions, the word order, the intonation pattern of the
sentence - see below)
a CB element prototypically occurs in the topic (T) of the
sentence, but (if more deeply embedded) it can also occur in
its focus (F);
an NB element occurs prototypically in F but there may be
embedded NB elements in T as well, see the use of green and
with coloured spots in (16), where the boundary between
T and F is indicated by a slash:
(16) (Mary has her wardrobe full of nice summer blouses).
The green one with coloured spots / looks like a spring
meadow in full bloom.
the relevance of the distinction between CB and NB elements
for anaphora resolution (Haji=87ov=A0 et al, 1995)
(17) You have been listening to our night programme. The
piano concertos by Chopin were played by Richter. We
will devote to him also our next programme.
(18) You have been listening to our night programme. The
piano concertos by Chopin were played by (Mrs)
Kramperov=A0. We will devote to him also our next
programme.
(19) You have been listening to our night programme. The
piano concertos by Chopin were played by (Mrs)
Kramperov=A0. We will devote to her also our next
programme.
(20) You have been listening to our night programme of the
works of F. Chopin. Chopin's piano concertos were
played by S. Richter. We will devote to him also our
next programme.
two observations:
- (19) seems to offer a smoother context for deciphering the
reference in the last sentence than (18): pronominal
reference to (a mental image of) an object occurring in
F of the previous sentence
- a comparison of these examples with (20), where him seems
to preferentially refer to an item in T: indicates that
a CB element in T (Chopin, in our case) establishes
a preferential candidate for pronominal reference.
(ii) communicative dynamism as the underlying word order
(21)(a) It is JIM who does linguistics on Sundays.
(b) It is JIM who on Sundays does linguistics.
(21')(a) Lingvistice se o ned=88l=A1ch v=88nuje JIM.
(b) O ned=88l=A1ch se lingvistice v=88nuje JIM.
(22)(a) Jim does linguistics on Sundays.
(b) On Sundays, Jim does linguistics.
the order of elements in the relative clause in the English
cleft-construction in (21a) and (b) corresponds to the order
exhibited by (22a) and (b), respectively, and the semantic
effect is the same, but:
- in (21a) and (b) both linguistics and on Sundays are in
T (as indicated by the prosodic contour of the clefting
structure)
- in (22a): on Sundays in F, in (22b) on Sundays in T
(iii) the definition of Topic and Focus can be anchored in
the concepts of contextual boundness and of CD
in Sgall et al. (1986, Ch. 3): this definition is based
on characterizing the verb itself and any of its immediate
dependents as belonging to T iff they are CB (with
a specific proviso for cases where all these items are CB)
3. HOW TO REPRESENT TFA:
in my talk: informally
in Petkevi=87's talk: formal tools
a description of TFA presenting an explicit account of
insights known from older tradition in Czech linguistics:
our formal framework for handling the syntax of TFA was
outlined in Sgall, Haji=87ov=A0 and Bene=A8ov=A0 (1973) and then
formulated in Haji=87ov=A0 and Sgall (1980), Sgall, Haji=87ov=A0 and
Bur=A0=A4ov=A0 (1980), Sgall, Haji=87ov=A0 and Panevov=A0 (1986) and
Petkevi=87 (1987; in press)
the topic-focus structure of a sentence is so understood
that topic and focus constitute two parts of the (structure
of the) sentence, with the topic being optional (absent at
least in the sentences corresponding to some thetic
judgments) and the focus being present in every sentence
in the tectogrammatical (i.e. underlying) representation
(TR) of the sentence:
(A) topic and focus are divided by a cut through one of the
edges of the dependency tree
(i) in the prototypical case this is the edge between the
main verb and the 'most dynamic' of its daughter nodes that
belong to the topic
(ii) two secondary cases:
(a) the boundary divides the verb from the 'least
dynamic' of its daughter nodes that belong to the focus;
thus both in (i) and in (ii)(a) the verb is adjacent to the
boundary;
(b) if the main verb and all its daughter nodes
are parts of the topic, then the focus is more deeply
embedded, being constituted by one or more nodes
subordinated to the rightmost of the verb's daughter nodes
(where 'subordinated' is the transitive closure of
'dependent')
we do not work with an implicit mood operator (e.g. ASSERT,
see Jacobs 1984; cf. Krifka ms. 1993) assuming that its
correlate may be present in the output of the
semantico-pragmatic interpretation
(B) the left-to-right order of the nodes corresponds to the
scale of communicative dynamism (CD), the degrees of which
are relevant for quantifier scopes
4. Illustrations how this framework works:
(i) for the primary interpretation of the distribution
of the scopes of quantifiers:
(8)(a) Many men read few books.
(b) Few books are read by many men.
(8')(a)
read.Pres.Declar
Actor Obj
man.Plur book.Plur
Gener Gener
many few
(8')(b)
read.Pres.Declar
Obj Actor
book.Plur man.Plur
Gener Gener
few many
(23)(a) It was JOHN who talked to few girls about many
problems.
(b) It was JOHN who talked about many problems to few
girls.
(23')(a)
talk.Pret.Declart
Addr Obj Actor
girl.Plur problem.Plur John
Gener Gener
few many
(23')(b)
talk.Pret.Declart
Obj Addr Actor
problem.Plur girl.Plur John
Gener Gener
many few
(ii) for the presence of presuppositions:
(24)(a) From Chicago they flew to Boston.
(b) To Boston they flew from Chicago.
a difference between falsity and inappropriate use:
(24a) appropriate either for such a context as (with flew
being NB, see (24'a)):
At first, they stayed a couple of days in Chicago
or for such as when answering the question
Where did they fly to from Chicago?
(with the verb being CB, see (24''a), and thus triggering
a presupposition)
(24b) may follow such an utterance as
I met them last week in Boston
(with the verb NB, see (24'b)
or it may answer a question such as
From where did they fly to Boston?
(with the verb CB, see (24''b), presupposing that they flew
(from somewhere) to Boston.
(24')(a)
fly.Pret.Declar
Direct.1 Actor Direct.2
Chicago they Boston
(24')(b)
fly.Pret.Declar
Direct.2 Actor Direct.1
Boston they Chicago
(24'')(a)
fly.Pret.Declart
Direct.1 Actor Direct.2
Chicago they Boston
(24'')(b)
fly.Pret.Declart
Direct.2 Actor Direct.1
Boston they Chicago
in linearized notation:
(24')(a) (Chicago)Dir.1 (they)Actor / fly.Pret.Declar
(Dir.2 Boston)
(24')(b) (Boston)Dir.2 (they)Actor / fly.Pret.Declar
(Dir.1 Chicago)
(24'')(a) (Chicago)Dir.1 (they)Actor fly.Pret.Declar /
(Dir.2 Boston)
(24'')(b) (Boston)Dir.2 (they)Actor fly.Pret.Declar /
(Dir.1 Chicago)
5. Systemic ordering of kinds of complementation
a fundamental ordering of the kinds of complementation,
i.e. of 'arguments' and 'adjuncts':
'systemic ordering' (SO) understood as partly
determining the communicative dynamism in individual
sentences:
if a complementation A precedes another one, B,
under SO, then the order A B is unmarked in the hierarchy of
CD (and, in prototypical cases, also in the surface word
order); the marked order B A (with B less dynamic than A)
can occur only if B is contextually bound.
asymmetry in (25)(a) and (b):
in (25a) the 'from'-group is in the topic in some
readings, in the focus in others;
in (25b) the 'to'-group can only be in the topic
(25)(a) They flew from Chicago to Boston.
A B
(b) They flew to Boston from Chicago.
B A
a similar relationship can be found with other pairs of
complementations:
(26) Manner - Direction
(a) They went by car to a river. (Means - Dir.2)
A B
(b) They went to a river by car.
B A
(27) Objective - Means
(a) Jim dug a ditch with a hoe. (Objective - Means)
A B
(b) Jim dug a DITCH with a hoe.
A B
(28) Manner - Locative
(a) Ron cannot sleep quietly in a hotel. (Manner - Loc)
A B
(b) In a hotel Ron cannot sleep quietly.
B A
SO differs from one language to the other: it appears that
for some of the main complementations of English the scale
of SO is as follows:
complementation: illustration:
Actor subject of active verb
Addressee indirect object
Objective direct object
Origin (Source) Iout of woodI
Effect Imake into a canoeI
Manner 'quietly'
Directional.1 'from Boston'
Means 'with a hoe'
Directional-2 'to Chicago'
Locative 'in a hotel'
Czech, and probably also German, differ from English
in that the positions of Objective and Effect in these
languages are more to the right, after most of the adverbial
complementations
consequences: in the lexical entries - 'frames' with ordered
complementations; in the process of generation this order is
obeyed when generating the focus part of the sentence
REFERENCES:
Chomsky N. (1969), Deep structure, surface structure and
semantic intepretation. In: Chomsky (1972), 9-48.
Chomsky N. (1972), Studies on semantics in generative
grammar, Haag:Mouton.
Engdahl Elisabet and Enric Vallduv=A1 (1994), Information
packaging and grammar architecture: A constraint-based
approach, to appear in Engdahl Elisabet (1994), Integrating
information structure into constraint-based and categorial
approaches, DYANA-2 Report R.1.3.B, CLLI, Amsterdam.
Haji=87ov=A0, E. (1973), Negation and topic vs. comment,
Philologica Pragensia 16,81-93.
Haji=87ov=A0, E. (1984), Presupposition and allegation
revisited, Journal of Pragmatics 8, 155-167.
Haji=87ov=A0 Eva (1995), Surface and underlying word order. In:
Prague Linguistic Circle Papers, Vol.1, Amsterdam: John
Benjamins, 113-124.
Haji=87ov=A0 Eva, Hoskovec Tom=A0=A8 and Petr Sgall (1995),
Discourse modelling based on hierarchy of salience, Prague
Bulletin of Mathematical Linguistics 64, 5-24.
Haji=87ov=A0 Eva, Partee H. Barbara and Petr Sgall (in prep.),
Topic-focus articulation, tripartite structures and semantic
content.
Haji=87ov=A0 E. and P. Sgall (1975), Topic and focus in
transformational grammar, Papers in Linguistics 8, 3-58.
Haji=87ov=A0 E. and P. Sgall (1980), A dependency-based
specification of topic and focus, SMIL 1/2, 93-140.
Halliday M. A. K. (1967), Notes on transitivity and theme
in English, Journal of Linguistics 3, 1967, 37-81, 199-244;
4, 1968, 179-215.
Jacobs, J. (1983), Fokus und Skalen, Zur Syntax und Semantik
von Gradpartikeln im Deutschen. T=81bingen: Niemeyer.
Jacobs, J. (1991), Focus ambiguities, Journal of Semantics
8 (1&2), 1-36.
Kratzer, A. (1991), The Representation of Focus. In: A. von
Stechow & D. Wunderlich, eds., Semantics, de Gruyter,
825-834.
Krifka Manfred (1992), A compositional semantics for
multiple focus constructions, Linguistische Berichte,
Sonderheft Informationsstruktur und Grammatik, ed. by J.
Jacobs, 17-53.
Krifka, M. (1994), Focus and operator scope in German. In:
P. Bosch and R. van der Sandt (eds.), Focus and Natural
Language Processing, Vol. I: Intonation and Syntax, Working
Paper 6, IBM, Heidelberg (133-152).
Krifka, Manfred (to appear) "Focus and/or Context: A Second
Look at Second Occurrence Expressions" to appear in H. Kamp
and B. Partee (eds.), Proceedings of the Workshop on
"Context Dependence in the Analysis of Linguistic Meaning",
Prague and Bad Teinach, 1995, University of Stuttgart
Working Papers.
Kuno S. (1972), Functional sentence perspective, Linguistic
Inquiry 3, 269-320.
Partee, B. (1991), Topic, focus and quantification. In: S.
Moore, and A. Wyner, eds., SALT I, Cornell Univ. Dept. of
Mod. Lgs. and Ling.
Petkevi=87, V. (1987), A new dependency based specification
of underlying representations of sentences, Theoretical
Linguistics 14, 143-172
Petkevi=87, V. (in prep.), An extended dependency-based
specification of linguistic description
Pfeiffer, O. E., P=96=87ek, M. and P. Sgall (1994), Die
Thema-Rhema-Gliederung im Deutschen und ihre automatische
Analyse. In: Computatio Linguae II, ed. by U. Klenk, Franz
Steiner Verlag, Stuttgart, 148-164
Prince E. (1981), Toward a taxonomy of given-new
information. In: P. Cole, ed., Radical pragmatics. New York:
Academic Press.
Rooth, M. (1985), Association with Focus, Ph.D. diss.,
UMass, Amherst.
Rooth, M. (1992), "A Theory of Focus Interpretation",
Natural Language Semantics, 1, 75-116.
Selkirk, E. (1995), Sentence prosody: intonation, stress
and phrasing. In: Handbook of Phonological Theory, ed.
J.A.Goldsmith. Cambridge, Mass. and Oxford, UK: Basil
Blackwell, 550-569.
Sgall Petr, Haji=87ov=A0 Eva and Eva Bene=A8ov=A0 (1973), Topic,
focus, and generative semantics. Kronberg/Ts.:Scriptor.
Sgall Petr, Haji=87ov=A0 Eva and Eva Bur=A0=A4ov=A0 (1980), Aktu=A0ln=A1
=87len=88n=A1 v=88ty v =87e=A8tin=88. [Topic-focus articulation in Czech].
Prague: Academia.
Sgall Petr, Haji=87ov=A0 Eva and Jarmila Panevov=A0 (1986), The
meaning of the sentence in its semantic and pragmatic
aspects. Prague: Academia and Dordrecht: Reidel.
Steedman, Mark (1994), Remarks on intonation and 'focus',
Draft version. In: P. Bosch and R. van der Sandt (eds.),
Focus and Natural Language Processing, Vol. I: Intonation
and Syntax, Working Paper 6, IBM, Heidelberg (185-204).
Vallduv=A1 Enric (1993), Information packaging: A survey.
Research Paper HCRC/RP-44, Human Communication Research
Centre, University of Edinburgh.
APPENDIX
TFA and the means of its expression
1. Certain phenomena of English cannot be described without
taking into account their relationships to topic-focus
articulation (TFA). This concerns first of all the placement
of the phrasal stress, which certainly belongs to the domain
of the means expressing TFA. Not only the placement of the
intonation center (sentence stress, indicated by capitals in
the examples below) on the focus (or focus proper), as in
(1), belongs to these means, but also the opposition of the
presence or absence of a secondary (phrasal) stress
(indicated by italics) e.g. on the verb of (2).
(1)(a) The bell is RINGING.
(b) The BELL is ringing.
(2)(a) Jim was swimming in the POOL.
(b) Jim was swimming in the POOL.
If we understand well the arguments of Selkirk (1984; 1995),
the functional distinction between (2)(a) and (b) may be
illustrated by questions which can be answered by the given
sentences. The questions are understood here as representing
each a class of contexts in which the sentence can be
appropriately used, although in an actual dialogue this
sentence - without a deletion or pronominalization - would
constitute a redundant answer, since (a part of) the topic
can be deleted.
(2')(a) What was Jim doing?
(b) Where was Jim swimming?
It is true that the appropriateness for a certain
contextual position is not a matter of grammar. However, as
was argued in our previous publications, TFA is relevant not
only for pragmatics (with the 'given/new' opposition), but
also for semantics in the narrow sense (truth conditions);
this concerns examples with specific quantifiers, such as in
(3), but also without them, see (4) and (5).
(3)(a) John talked to many girls about few PROBLEMS.
(b) John talked about few problems to many GIRLS.
(4)(a) Staff behind the COUNTER.
(b) STAFF behind the counter.
(5)(a) I do linguistics on SUNDAYS.
(b) On Sundays I do LINGUISTICS.
Furthermore, the semantic opposition between the (a) and (b)
examples cannot be accounted for just by referring to the
dichotomy of 'given' and 'new' (which again certainly is
beyond grammar). We are in a similar situation here as with
grammatical tenses, the distribution of the morphemic means
of which is not described in grammars with a direct
reference to the time axis and time intervals; grammarians
have found good reasons to speak about tenses (present,
past, future,...) and their meanings as mediating between
morphemic items and issues of time. What is needed in the
domain of TFA is exactly such a mediating (interface) level,
or the respective items of this level - notions such as
topic (T), focus (F), contextual boundness (CB) and
communicative dynamism (CD).
Along with accentuation, deletion itself is to be
described in a grammar, and cannot be described without
reference to TFA. Otherwise a grammar could not state that
the speaker may utter just (2'')(a) instead of (2)(a), and
(2'')(b) instead of (2)(b).
(2'')(a) Swimming in the pool.
(b) In the pool.
This is not to say that it is always possible to delete the
whole topic part of a sentence (e.g. contrastive topic does
not get deleted, and there may be some grammatical
restrictions preventing the deletion of some parts of
topic). On the other hand, it holds that if something can be
deleted in the outer shape of a sentence, then it belongs to
the topic (of the sentence or of a clause; more precisely:
to the contextually bound elements), with strictly limited
exceptions such as those concerning coordination deletion,
cf. (6).
(6) We were offered red and white WINE.
2. To present another illustration of phenomena which
require TFA to be taken into account in grammatical
description, we recall certain specific issues of word
order, especially in presence of such focus sensitive
particles as only.
Example (7)(a) is at least three ways ambiguous, as for
the focus of the particle (which, in our understanding, is
here identical to the focus of the sentence; this is the
prototypical case with such particles).
(7)(a) They only travelled from London to EDINBURGH.
(b) They travelled only from London to EDINBURGH.
(c) They travelled from London only to EDINBURGH.
(d) They travelled only to EDINBURGH from London.
(e) ??They travelled only to Edinburgh from LONDON.
(f) They only travelled to Edinburgh from LONDON.
In one of the readings of (7a), the focus contains
everything that follows the particle; in the second, it
contains the two prepositional groups (this is also
the reading of (b)); in the third, identical with that of
(c) and of (d), only the to-group belongs to the focus.
If the description is to mark (e) as odd,2 then
reference is to be made to the fact that only the order of
from - to (i.e. the order of Directional-from and
Directional-to), rather than the reversed one, is
acceptable if both the groups are included in the focus.
This is what is predicted by the hypothesis of
'systemic ordering' (see Sgall, Haji=87ov=A0 and Panevov=A0, 1986,
Ch. 3; Haji=87ov=A0 and Sgall 1987): there is a certain ordering
of the kinds of complementation that underlies the
communicative dynamism (CD, deep word order); if A precedes
B under this ordering, then the order A B is unmarked in the
scale of CD (and, in prototypical cases, also in the surface
word order). The marked order B A (with B less dynamic than
A) can occur only if B is contextually bound (belongs to the
topic). Thus (7a) and (7b) have from London in topic on some
readings, in focus on others; (f) has from London in focus
and to Edinburgh in topic on all readings.
3. Also a complete account of the distinction between
active and passive in English requires to refer to TFA,
since, as is well known, passivization is often used to
bring the surface word order into harmony with CD (to start
the sentence with its topic even when this is the
Objective), as in (8):
(8) (Jim wrote a new book.) The book was reviewed by
HOPKINS.
In English, passivization compulsorily triggers
the word order shift, but a comparison with German or
Czech shows that without such a shift the truth conditions
of the active sentence and of its passive counterpart
coincide:
(8') (Jim hat ein neues Buch geschrieben.) Das Buch ist
von HOPKINS rezensiert worden. =3D3D Das Buch hat HOPKINS
rezensiert.
This is not to claim that there is (otherwise) a full
synonymy of active and passive. Sentences with sentential
adverbials such as willingly, deliberately, with pleasure
show that in the general case the truth conditions of the
two diatheses differ:
(9)(a) The cannibals willingly ate the missionaires.
(b) The missionaires were willingly eaten by the
cannibals.
This opposition concerns also languages such as German
or Czech, in which active and passive can be characterized as
quasisynonymous, in the sense that normally (in the absence
of such a sentential adverbial) passivization does not
change the truth conditions (not affecting TFA).
4. Along with sentence intonation and word order, for the
description of which a reference to TFA is necessary with
respect to English, in other languages further phenomena are
connected with this requirement. One of such points concerns
the opposition between weak and strong pronouns in Czech and
several other languages (e.g. French): the weak
form can only occur as contextually bound (mainly in the
topic, see (10a)), whereas the strong forms may be found in
focus (see (10b) and also in contrastive topic (see (10c)),
or after a preposition:
(10)(a) Tu knihu jsme mu koupili. (Lit.: The book we-have
him bought.)
(b) Tu knihu jsme koupili jemu. (Lit.: The book we-have
bought him-strong.)
(c) Jemu jsme koupili knihu, j=3DA1 jsme koupili obraz.
(Lit.: Him-strong we-have bought book, her we-have
bought picture.)
5. A further point concerns such languages as Japanese or
Tagalog, where the topic is grammatically characterized by
a specific morpheme. With negation, the position of the
negative particle indicates focus e.g. in German, Russian,
Navajo. Thus, it is not possible to describe the grammar of
these languages without referring to TFA.