Professional Documents
Culture Documents
Managing Editors.
ELiSABET ENGDAHL, University of Wisconsin
JAAKKO HINTIKKA, Florida State University, Tallahassee
ST ANLEY PETERS, Stanford University
Editorial Board:
EMMON BACH, University of Massachusetts at Amherst
JOAN BRESNAN, Stanford University
JOHN LYONS, University of Sussex
JULIUS M. E. MORAVCSIK, Stanford University
PATRICK SUPPES,Stanford University
DANA SCOTT, Carnegie-Mellon University, Pittsburgh
VOLUME 25
GREGORY T. STUMP
Department of English, University of Kentucky
THE SEMANTIC
VARIABILITY
OF ABSOLUTE
CONSTRUCTIONS
DORDRECHT/BOSTON/LANCASTER
library of Congress Cataloging in Publication Data
Stump, Gregory T. (Gregory Thomas), 1954-
The semantic variability of absolute constructions.
PREFACE xiii
ACKNOWLEDGEMENTS xvii
August, 1984 G. T. S.
The University of Kentucky
ACKNOWLEDGEMENTS
xvii
xviii ACKNOWLEDGEMENTS
infinitive form:
(26) Having done either of' the above his turn is ended
unless he uses all tiles on his rack ••• or if he
accumulates 9 points on one stack
(Instructions to stack-ominos 1M , Pressman Toy
Corporation)
a) time, e.g.
My task having been finished, I went to bed;
(p.154)
b) cause, e.g.
The rain having ruined my hat, I had to get a
new one; (p.155)
c) condition and exception, e.g.
As yet few have done their full duty, present
company excepted; (p.155)
d) attendant circumstance, e.g.
He entered upon the new enterprise
cautiously, his eyes wide-open; (p.156)
e) manner proper, e.g.
He put on his socks wrong side out; (p.157)
or f) concession, e.g.
Granted the very best intentions, his conduct
was productive of great mischief. (p.157)
by a relation of:
e) concession, e.g.
He is the more to be admired, that, labouring
under such a difficulty, his verses are so
numerous, so various, and so harmonious;
(p.1137)
f) means, manner, e.g.
Brownie ••• was lanky and limp--so limp that
Mother said he was becoming unnaturally
elongated, pulling that cart up those hills;
(p.1139)
g) hypothetical similarity, e.g.
He paused as if expecting her to answer;
(p.1138)
and h) purpose, e.g.
We planted a hedge for preventing the cattle
from straying. (p.1139)
a) cause, e.g.
All our savings gone, we started looking for
jobs; (p.762)
b) time, e.g.
Cleared, this site will be very valuable;
(p.759)
and c) circumstance, e.g.
A case in both hands, Mabel stalked out of
the house. (p.759)
Table
The Logical Roles Played by Free Adjuncts and Absolutes
According to Five Traditional Grammars
Curme
( 1931) x x x x x x
Kruisinga
(1932) x x x x
Jespersen
(1940 ) x x x x
Visser a
(1972) x x x x (x)b
Quirk, et al.
(1972) x x x
aExcluded are 'hypothetical similarity' and 'purpose'; none of Visser's
examples of expressions bearing these relations are free adjuncts in the
sense assumed here.
b,(X)' indicates a logical role assumed for free adjuncts but not for
absolutes.
3. Plan of Discussion
[±Adjunct] [±Perfect]
[±Plural]
Table 2
Partial Lexical Entries for Four Verbs
All sentences and verbal phrases will have their main verb
enclosed in brackets, as will temporal abstracts in the
subclass [-Adjunct]. If a is a basic expression of any
verbal category, it will be of the form [S], for some S.
Rules combining a sentence, temporal abstract, or (basic or
derived) verbal phrase y with another expression will retain
or delete the brackets in y according as the main verb of
y is or is not to be the main verb of the resulting
expression. Thus, the rule 335 combining an intransitive
verb phrase Y with a verb-phrase adverb will retain the
bracketing in Y:
The main verb of the bare verb phrase is the same as the
main verb of the adverbially modified verb phrase in this
example. On the other hand, the rule 327 joining a verb
with a sentential complement ¢ will delete the bracketing in
OBJ(heo) = himo;
OBJ(John) = John;
Footnotes
as in (35)-(37).
13. Chapter II and parts (1) and (ii) of Chapter III are an
elaboration of Stump (1980a).
(42) S18. If a t: P MOD ' ¢ t: Pt " and ljJ <: Pt' [-Tense],
then F 18 (a, ¢,ljJ) E P t , [-Tense], where
F 1S (cx,¢,ljJ) is CONWIB(¢, RB«(y',ljJ».
(49) can'(D(cb)(p»(AJobn-touches-the-ceiling')
----------=----
everyone, t, 55,0
Under the assumed semantics for modals, the fact that each
of the a sentences in (9)-(12) entails the truth of its
strong adjunct must be taken as evidence that these adjuncts
are outside the scope of the accompanying modal verb; thus,
the gross syntactic analysis for a sentence such as (9a)
must be one in which the adjunct is 'added on' to the main
clause only after the modal has already been added, as in
the informal analysis in (56).
(65) K(L)("'x-being-a-master-of-disguise')
(ABill-uould-fool-everyone')
drunk
alone
asleep
sick in bed
naked
walk(ing) to the store
ly(ing) on the beach
carry(ing) a load of over 10 lbs.
driv(ing) this sort of car
wear(ing) his new outfit
stand(ing) on a chair
notic(ing) the snake
sit(ting) in the garden
hear(ing) this song
in his maroon suit and white shoes
on a stage
be(ing) a bastard
be(ing) silly
be(ing) arrested
be(ing) awarded the Medal of Honor
(76) [continued]
American
be(ing) American
fat
be(ing) fat
(79) be-intelligent'(dogs')
(100)
3.2.3. Be 3 E PIVs/PREDi
The be3 in be a bastard, be a hero, be polite, be silly, be
obnox10us, and so on, is the 'active be' postulated by
MODALITY AND FREE ADJUNCTS 77
Partee (1977:305-310).20 As she shows, expressions
resulting from the combination of be 3 with some predicative
expression are syntactically and semantically very different
from normal copulative phrases.
(i) They behave like active predicates in that they
may occur in the progressive:
(105 ) Be polite!
John is trying to be polite.
John is deliberately being polite.
We persuaded John to be polite.
( 113)
R(Xs,x k )]
(116 ) [continued]
( 118 ) [continued]
be1 e: BIVi/PREDs
be2 e; BIV i /PREDi
be3 e; Brvs /PREDi
be4 e; B1VS /PREDs
Though the semantic details of the distinction between
stage-predicates and individual-level predicates are quite
complex, the relevance of this distinction to the
weak/strong bifurcation of free adjuncts is clear, as I
shall now show.
84 CHAPTER II
drunk
alone
asleep
sick in bed
naked
walking to the store
lying on the beach
carrying a load of over 100 lbs.
driving this sort of car
wearing his new outfit
standing on a chair
noticing the snake
sitting in the garden
hearing this song
in his maroon suit and white shoes
on a stage
being a bastard
being silly
being arrested
being awarded the Medal of Honor
(76) [continued]
(135 )
(139 )
MODALITY AND FREE ADJUNCTS 89
As this brief discussion shows, the proposed analysis
of free adjuncts in modal contexts not only correctly
predicts that as- and with- phrases may serve as weak
adjuncts; it also reveals an intimate semantic connection
between as and be, and between with and have.
4. Chapter Summary
Footnotes
(i) Had Mary seen the morning paper, she would have
known all about the fire.
(ii) If Mary had seen the morning paper, she would have
known all about the fire.
to the individual level, so that 'be1 with T' and 'have T'
will both have properties of individuals as their
intensions. (In the present fragment, however, the
intension of 'have T' (unlike that of 'be1 with T') is,
specifically, a property of objects.)
McCawley (1983) argues that with-phrases having a
possessional meaning are derived from augmented absolutes
through the deletion of the underlying subject and the main
verb have; thus, the adjunct with green eyes derives from
(something like) (i) under his analysis.
1. Preliminaries
now denotes i;
moment denotes the characteristic function of the
set of unit sets of members of T;
pastC~) is true iff there is some nonempty
interval i' such that i" < i' < i, where i" is the
denotation of ~;'
pres(~) is true iff the denotation of 4 is a
subset of i;
102 CHAPTER III
(21) [continued]
(23) [continued]
------------
John, T [claim] that Mary is hitting Bill, lVi, 21
I
[claia] that Mary is hitting Bill, IVS, 27
~
[claim], IVs/t (Mary) [is] hi tting Bill, t, 60
I
(Hary) [is] hitting Bill, TAB, 58
I
(Mary) [is] hitting Bill, TAB, 56
I
(Mary) [is] hitting Bill, t, 19
(33) Next year, John will claim that Mary is his wife.
John, T
-----------
[discover] that Mary is walking home, lVi, 21
I
[discover] that Mary is walking home, IVs, 27
~
[discover], IVs/t (Mary) [is] walking home, t, 60
I
(Mary) [is] walking home, TAB, 58
I
(Mary) [is] walking home, TAB, 56
I
(Mary) [is] walking home, t, 19
AT(t 1 , Kary-is-walking-home')])])]
TENSE AND FREE ADJUNCTS 111
--------
I
[claim] that Mary will hit Bill, lVs, 27
--------
(Joho) [clau.s] that Mary hit Bill, t, 19
AT(t 1, Hary-hits-Bill')])])]
(56) a.
(= today')
b. AtVt,[[day'(t,) & At 2 [todaY'(t 2 ) + [t, < t2
& At 3 [[t, < t3 & t3 < t 2 ] + todaY'(t 3 )]]]] &
t ~ t,]
c. AtVt,[[day'(t,) & At 2 [today'(t 2 ) + [t 2 < t,
& At 3 [[t 2 < t3 & t3 < t 1 ] + todaY'(t 3 )]]] &
t ~ t,]
(57) John discovered that Mary had been here during the
past summer.
T61. If a E: PMTA ,
PTAB , and a,~ translate as
~ £
AT(t, John-is-happy')]]
-----------
I
(John) [was] asleep yesterday, TAB, 61b
yester~)
MTA, 63
[saw] Mary in the morning, TAB, 61b
______________
I in the morning, (John) [saw] Mary, TAB, 57
yesterday, MTA I
TA (John) [sees] Mary, TAB, 56
I
(John) [sees] Mary, t, 19
~
John, T [see] Mary, lvi, 21
I
[see] Mary, IV s , 28
~ Mary,
[see], TV s T
TENSE AND FREE ADJUNCTS 121
AT(t" Mary-sings')]]
c. Atvt,[[t < t, & M(t,t,)] & [PRESet,) &
AT(t" Mary-sings')]]
d. AtVt,[[t, < t & M(t,t,)] & [PREset,) &
AT(t" Mary-sings')]]
ATCt" Mary-sings')]]
b. Atvt,[[t ~ tl & ~momentCtl)] & [preset,) &
ATCt" Mary-sings')]]
c. At Vt, [ [t < t, & M( t , t, ) ] & [preset,) &
AT(t" Mary-sings')]]
d. AtVt,[[t, <t & M(t,t,)] & [preset,) &
ATCt 1, Mary-sings')]]
AT(t" Mary-sings')]]
b. AtVt,[[t ~ t, & ~ment(t1)] & [FUT(t,) &
AT(t" Mary-sings')]]
c. AtVt,[[t < t, & MCt,t,)] & [FUT(t, ) &
AT(t" Mary-sings')]]
d. AtVt 1 [[t 1 < t & MCt,t,)] & [FUT(t 1 ) &
ATCt 1 , Mary-sings')]]
a. before translates as
AptAtvt,EEt < t, & M(t,t,)] & pt{t,}]
b. after translates as
AptAtVt,EEt, < t & M(t,t,)] & pt{t,}]
c. when translates as AptAtVt,Et = t, & pt{t,}]
d. while translates as
AptAtvt,EEt ~ t, & ~DOment(t,)] & pt{t,}]
T68. If a E: PTSC' B E:
PTAB ' and a,B translate as
a',S', then F68(a,B) translates as
Ct'(~AtEpast(t) & S'(t)]).
128 CHAPTER III
(92) [continued]
(100 ) John left when Jane said that Mary was leaving.
Rule (92) above can only assign one translation to the when-
clause in (100), namely that in (101).
(103 ) John left at the time at which Jane said that Mary
was leaving.
since (100) and (103) are ambiguous in exactly the same way,
i t is tempting to derive the when-adverbial in (100) by a
rule of unbounded extraction; this is in fact exactly what
Geis proposes. In the fragment developed here, however,
TENSE AND FREE ADJUNCTS 131
movement rules as such are not employed; unbounded
dependencies are instead captured by rules of variable
binding. Accordingly, a rule of variable binding is here
proposed to account for unbounded dependencies in temporal
adverbial clauses such as that in (100):
------------- I
at-that-time4'
MIA
(Jane) [said] that Mary was leaving,
TAB, 57
(Jane) [says] that Mary was leaving, TAB, 56
I
(Jane) [says] that Mary vas leaving, t, 19
132 CHAPTER III
-----------
(Mary) [wasJ leaving at-that-ti.e 4 , TAB, 61b
AT(t, Jane-says-that-Hary-was-leaving')JJ
(111) when Jane said that Mary was leaving, TA, 67,4
~
when, (Jane) [said] that Mary was leaving at-that-time4'
TSC I t, 60
[as in (106)]
(112) when Jane said that Mary was leaving, TA, 67,4
~
when, (Jane) [said] that Mary was leaving at-that-ti.e4'
TSC I t, 60
[as in (107)]
I
when Jane said that Mary was leaving, (John) [leaves],
TA, 67,4 TAB, 56
I
[as in (112)] (John) [leaves],
t, 19
(100) John left when Jane said that Mary was leaving.
( 119) John left while Jane said that Mary was leaving.
-------------
(John) [will] leave when Mary sings, TAB, 61b
------------
at-that-time4' MTA (Mary) [sings], TAB, 58
I
(Mary) [sings], TAB, 56
I
(Mary) [sings], t, 19
---------
when Mary sings, TA, 69
yesterday,
MT A, 63---------------
(Jane) [arrived] at-that-time4' TAB, 61b
____________
I at-that-tiae4 , (Jane) [arrived], TAB, 57
yesterday, MTA I
TA (Jane) [arrives], TAB, 56
I
(Jane) [arrives], t, 19
Note that the iteration of rule (63) (by which the main
tense adverbs yesterday and at-that-time4 are added on
successively in the derivation of the before-clause in
(138» is critical for this derivation.
Since rule (104) never figures in the derivation of
while-clauses, sentence (136) receives an analysis somewhat
different from (138), namely (140). In (140), rule (92)
rather than rule (104) is used to generate the temporal
adverbial clause. By this analysis, (136) is assigned the
interpretation of (141).
TENSE AND FREE ADJUNCTS 139
(140) while Mary was singing yesterday, (John) [left],
I t, 60
while Mary was singing yesterday, (John) [left], TAB, 61a
I
MTA, 63 ----------(John) [left], TAB, 57
I
(John) [leaves], TAB, 56
while Mary was singing yesterday, I
~ TA,68 (John) [leaves], t, 19
while,
TSC
---------
(Mary) [is] singing yesterday, TAB, 6'b
yesterday,
MTA, 63
I
yesterday, TA
(Mary) [is] singing, TAB, 56
I
(Mary) [is] singing, t, 19
-------------
Jane, T [say] that Mary arrived yesterday, lVi, 21
I
[say] that Hary arrived yesterday, IVs , 27
~
---------
[say], IVs/t (Mary) [arrived] yesterday, t, 60
I
(Hary) [arrived] yesterday, TAB, 61b
---------
(Jane) [says] that Mary arrived at-tbat-time4' t, 19
-----=------
[say] that Mary arrived at-that-time4' IV s , 27
I
Jane, T [say] tbat Mary arrived at-tbat-time4 yest~rday,
Iy1, 21
[say] that Mary arrived at-that-time4 yesterday, IYS, 27
-----------
(Mary) [arrived] at-tbat-time 4 yesterday, TAB, 61b
, .4. 5. 1. When.
( 171) i1
~
I I
-
~
"" ~
i3
--- .............
i2
It was raining in New Or leans at i 1•
We were in New Orleans at i 2 •
( 173)
(174) The balloon broke when Lydia was playing with it.
(175 )
I I I
( 176)
( )
(183 )
)
------------~-----------
John averages 60 mph at i 1 •
John is in Kansas throughout i 2 , averaging 55 mph.
1.4.5.2. While
1.4.5.3. Before
1 .4.5. 4. After
Table 3
Five Categories and Their Corresponding Logical Types
Category Type
t t
TAB <i,t>
TA <i,t>
MTA «s,<i,t»,<i,t»
TSC «s,<i,t»,<i,t»
-----------
I
when John arrived, (Mary) [left], TAB, 61a
c
~g
I I
d h
e0 i
2.1. A-ABSTRACTS
(211) [continued]
I -------------
today, MTA, 63
today, TA
wearing that new outfit, TAB, 1,4
I
wearing that new outfit, PRPL s
(232) A
Vt, [AT(t" Vx s[ R(s i
x, x4) &
leaYing~ansas-City'(xs)])]
(233)
This is clearly a deficiency in (231) as a translation for
(223), since an important part of the logical relation
inferred to hold between the adjunct and the main clause in
(223) is temporal in nature: the event picked out by the
adjunct in (223) is almost inevitably inferred to precede
170 CHAPTER III
(235) Apt At[K(L 6 )( .... vt 1 [M 9 (t,t,) & AT(t 1, Vxs[R(x s , x~) &
leaving-Kansas-City'(xs)]»))( .... [pt{t}])]
(236) Vt[K(L 6 )( .... Vt,[M9 (t,t,) & AT(t" Vxs[R(x s , xt) &
leaving-Kansas-City'(x s )]) )( .... [at-noon'(t) &
[past(t) & AT(t, JobD-reaches-St.-Louis')]]»)
(239)
leaving Kansas City in the morning,
(John) [reached] St. Louis at noon, t, 60
I
leaving Kansas City in the morning, (John) [reached] St. Louis at"noon,
-------
TAB, 61a
~
leaving Kansas City in (John) [reached] st. Louis at noon, TAB, 61b
the morning, MIA, 11,6,9
I
leaving Kansas City in
the morning, TAB, 61b
at noon, ~lTA, 63
I
at noon, TA
I -
(John) [reached] St. LoU1S,
TAB, 57
I
~ (John) [reaches] St. Louis,
in the morning, leaving Kansas City, TAB, 56
i1TA, 6::; I TAB,1,4
(John) [reaches] st. Louts,
in the mornin!!., leaving Kansas City, t, 19
TA PRPL s
---------
(261) when she is figuring her taxes, (Jane) often
[uses] a calculator, t, 55,'
------------
Jane, T wben sbe, Cigured ber, taxes, (sbe,) octen
[used] a calculator, t, 60
I
wben sbe, Cigured ber, taxes, (abe,) octen [used]
a calculator, TAB, 64a
---\\------------
often, when she 1 Cigured (she1) [used]
RFA her, taxes, TA, 67,0 a calculator, TAB, 57
~ I
when, (she,) [figured] her, (she ) [uses]
TSC taxes at-that-timeO' t, 60 a calculator, TAB, 56
I I
(she,) [figured] her, taxes (she1 ) [uses]
at-that-timeO' TAB, 6,b a calculator, t. 19
~
at-that-ti.eo' (she,) [figured] her 1 taxes, TAB, 57
MTA I
[ figures] her1 taxes. TAB. 56
I
(she1) [figures] her 1 taxes, t. '9
Sentence (255) is analyzed as in (265) and translates
as in (264).
-----------
RFA paat su.mer, TA, 67,0 TAB, 57
I
when, (it) [rained] at-tbat-tt.eO (our roof) [leaks],
TSC during the past summer, t, 60 TAB, 56
I I
(it) [rained] at-that-tt.eO (our roof) [leaks], t, 19
during tbe past su.mer, TAB, 61b
I ------------
during tbe past summer,
MTA, 63
(it) [rained] at-that-tt.eO'
~ 61b
at-that-tt.eo'
during the past su.aer,
TA MTA I
(it) [rained],
TAB, 57
(it) [rains], TAB, 56
I
(it) [rains], t, 19
Sentence (266), in which the adverb often isn't
explicitly restricted by a set-level time adverb, is derived
by means of rule (259) as in (267), and is thus assigned the
translation (268), in which 16 is a variable over sets of
time intervals (whose value is inferred).
----------
I
(Jane) often [uses] a calculator, TAB, 65,6
---------
for a walk at"ter be ate supper, t, 55,0
----------
after heO ate supper, TAB, 61a
wben ben was in (hea) always [went] for a walk
Kansas City, MTA, 63 after heO ate supper, TAB, 64b
~_---;7r- ____
alwayS;
RFA
after beO ate supper,
TA, 67,0 I
(heO) [went] for a walk,
TAB, 57
(heo) [goes] for a walk, TAB, 56
I
(he o) [goes] for a walk, t, 19
The feasibility of this analysis of relative frequency
adverbs provides further justification for the category TAB
of temporal abstracts; moreover, temporal abstracts will be
indispensably employed in the analysis of a certain variety
of generic sentence in section 4.
The foregoing analysis of the interpretation of
relati ve frequency adverbs affords a very clear
understanding of the difference between strong and weak
adjuncts observed in sentences (241 )-(244), as will be seen
in the following subsection.
(276) Vt[K(L 6 )(A Vt ,[M g (t,t,) & AT(t" Vxs[R(x s , x~) &
lying-on-tbe-beacb'(x s ) ]» ])(A[past(t) &
AT ( t , J obn-sllOkes-his-pipe') ] ) ]
-------
t, 60
being a sailor, (John) sometimes [smokes] a pipe, TAB, 61a
In such read ings of sentences li.ke (283), the adj unct does
not restrict the interpretation of the frequency adverb,
even though it is weak. The important generalization is
that a weak adjunct ~ restrict a main clause frequency
adverb, while a strong adjunct may not.
Nonpresent tense sentences having frequency adverbs
restricted by weak adjuncts are straightforwardly derived in
the proposed system. Sentence (284), for example, is
assigned the analysis in (285) and the translation (286).
(284) Carrying over 1500 Ibs., our truck often made the
bridge shake.
often,
______,T----------_
the bridge shake, TAB, 64a
4. A Generalization Operator
So far, two environments have been examined which motivate a
distinction between two varieties of free adjuncts: modal
sentences and sentences with relative frequency adverbs. In
this section, a third environment motivating a distinction
between strong and weak adjuncts is discussed. This third
environment is exemplified in sentences (290)-(293).
(312) have-four-legs'(Fido')
have-four-legs'(Rover')
have-four-legs'(Spot')
have-four-legs'(Rex')
In order for G'(P) to hold for some kind xk, it need not
always be the case that all objects realizing xk have
property P; Dogs have four legs, for example, is not
falsified by the existence of some three-legged dog. But
G'(P) 'cannot be true of xk unless enough of the objects
realizing xk have P. So how many are 'enough'? Carlson
demonstrates at length that 'statements of frequency or of
absolute number of occurrences cannot be stated generally
for all generic sentences' (1979:59): while Lions are
.ammals applies to all lions, Lions give milk to their young
applies only to mother lions; Lions attain a weight of
several hundred pounds applies only to lions which reach
physical maturity; Lions attack zookeepers applies only to
lions in captivity; and so on. Ultimately, the question of
when a property ceases to be an accidental property of
certain objects and becomes an essential property of the
kind they realize is epistemological in nature, and thus one
which the grammar of English should not be expected to
answer.
TENSE AND FREE ADJUNCTS 195
Carlson also finds that any predicate (headed by a
verb) interpretable as a stage-predicate may also be
interpreted as a predicate of objects and kinds. He
suggests that this too is the result of generalizing the
application of predicates from 'lower' to 'higher' entities;
observing several stages of Fido chasing cars, one can
generalize and say that Fido (the object) chases cars--or
observing stages of several dogs engaged in the same
pursui t, one can say that dogs (the kind) chase cars.
Accordingly, a second generalization operator G taking
intensions of stage-predicates into extensions of
individual-level predicates is posited:
(314) S22. If a E Prvs, then F22 (a) E Prvi, where F22 (a)
is a.
<316 )
(318 )
(319) G,(AAXO(G(AAXS(xs_barks-wben_xs_sees-a-
mailman'])(xo)])(dogs')
(320)
TENSE AND FREE ADJUNCTS 197
(321) vt[Vt,[t = t, & vt 2 [t 2 = t, & [PRES(t 2 ) &
AT(t 2, see-a-mailman'(x s »]]] & [PRESet) &
AT(t, bark'(x s »]]
Here, the when-clause is interpreted as picking out a single
interval. But when (32') occurs within the scope of a
generalization operator, as in (318) and (319), it picks out
a number of distinct intervals: for any object 0, the truth
of (322) implies that there are a certain number of distinct
stages satisfying (323);
(333)
---------
Fido, T
(Fido) [barks] when he sees a mailBan, t, 55,0
I
after heO finished dinner, (heO) [a.okes] a cigar,
~A, 67,3 TAB, 56
after, (heO) [finished] dinner (heo) [smokes] a cigar,
TSC at-that-time 3 , t, 60 t, 19
I
---------
(heO) [finished] dinner at-that-time 3 , TAB, 61b
a. 'Multiple-interval':
b. 'Single-interval':
---------
I
when cats are widespread, (dogs) [are] rare, TAB, 66a
---------
(340) when he is drunk, (John) [is] unavailable, t, 55,0
-------=-
when heO is drunk, (heO) [is] unavailable, TAB, 66a
-------------
(beO) [is] drunk at-tbat-time4 , TAB, 61b
in first gear,
TA, 9,9
I
-----------(the truck) [makes] tunny noises, TAB, 58
I
(the truck) [makes] fUnny noises, TAB, 56
in Cirst gear, I
TAB, ',4 (the truck) [makes] fUnny nOises, t, '9
being a businessman,
~(John) [s.okes] cigars, TAB, 58
MTA, ",6,9 I
I (John) [s.okes] cigars, TAB, 56
being a businessman, I
TAB, 2,4 (John) [smokes] cigars, t, 19
---------
John, T [saoke] cigars, lVi, 22
I
[s.oke] cigars, IVs, 28
--------
I
lying on the beacb, (.lobo) [smoked] Cigars, TAB, 668
5. Chapter Summary
Footnotes
1. The reason for defining the past tense predicate in
this way will be taken up in section 1 of Chapter IV.
i' i
~ ~
0(
( ------
-------------~~---------------
it
i
)
------- ------
(17) (15) is true at i iff (16) is true at i'.
i' i
I
---------------~---------------
in
i' i
~ ~
( »
[i' is a past interval relative to i.]
To see how the extended now theory accounts for these facts,
one must consider the semantic distinction between the three
sorts of adverbs listed in (22)-(24), and the way in which
adverbs enter into the interpretation of perfect sentences.
The adverbs in (22)-(24) are regarded as expressions of
the category MTA of main tense adverbs, and thus as denoting
functions from properties of time intervals to sets of time
intervals. They do, however, have three distinct sorts of
denotations. At a given index <w,i>, the adverbs in (22) in
general denote functions from properties to sets containing
both past intervals and extended nows relative to i; for
example, the denotation of today is that of (28), as has
already been seen (see (59a) in Chapter III, section '.3.2).
t =- t,]
These truthconditions are captured by the translation (34)
for (32) (where yesterday" is (29»:
AT(t" John-is-on-the-train')])(t)]
i
~
I I I I I )
i
~
I I I
i
~
I I
I I
~
i'
Adverbial infinitives:
(45) [continued]
Gerunds:
Participial absolutes:
(6') [continued]
----------
(John) [has] been on the train, TAB, 70
I
(John) [has] been on the train, TAB, 62
----------
I
(John) [has] been on the train since noon, TAB, 61b
.
s~nce noon, MTA (John) [has] been on the train, TAB, 62
----------------
(John) [has] been on the train yesterday. TAB. 70
I
(John) [has] been on the train yesterday. TAB, 61b
----------
I
(John) [was] on the train since noon, TAB, 61b
---------
(John) [has] been on the train, TAB, 70
I
(John) [has] been on the train, TAB, 62
----------
I
(John) [will] have been on the train today, TAB, 61b
----------
I (John) [has] been on the train, TAB, 70
today, I
TA (John) [has] been on the train, TAB, 62
(80) [continued]
------
I
having been on it2' (John) [will] know
exactly whyit2 derailed, TAB, 61a
--------
I I
having been on it2' (John) [knows] exactly
TAB, 4,4 derailed, t, 19
---------
having been on the train since noon,
(John) [knows] exactly why it derailed, t, 55,2
--------
having been on it2 since noon, (John) [knows] exactly
~TAB, 61b why it2 derailed, t, 19
----------
(88) having been on the train yesterday,
(John) [knows] exactly why it derailed, t, 55,2
The fact that the time adverb and the tense in (91)
characterize the same interval is reflected in the
translation assigned to (91) by the proposed fragment:
The main tense adverb since noon in (93) joins with the
perfect to characterize the interval i'; in particular, it
specifies a set of extended nows (relative to i) having i'
as a member. (Because it does denote a set of extended
nows, this particular adverb cannot join with the past tense
to characterize i; if it did, a contradiction would result.)
The fact that the time adverb in (93) joins with the perfect
to characterize a particular interval is captured in the
translation assigned to (93) (where, again, since-noon" is
(31» :
(100 )
------
(Mary) [had] decided to go back to 3chool, t, 60
I
working at the P03t office during the past summer,
(Mary) [had] decided to go back to school, TAB, 61a
vt 2 [t 2 =- t1 &
AT(t 2 , Mary-decides-to-go-back-to-school')]])])]
( 105)
working at the post office during the past summer,
-----
(Hary) [decided] to go back to school, t, 60
I
working at the post office during the past summer,
(Hary) [decided] to go back to school, TAB, 61a
(113 ) Having been reading the book, John knew all the
answers.
i i"
I I I I
~
i'
[ i c i ' : i " i LJ
F
the set of
inertia worlds I )
assigned to I
<w,i> I
wn I
i"
-
i
w
--
~
I I
"'-"""
i'
Augmented adjuncts:
(128 ) [continued]
Augmented absolutes:
(143)
(147) [continued]
---------
t, 60
crossing the street, (John) [was] hit by a car, TAB, 61a
-----------
I
being asleep, (John) [was] unaware of Mary, TAB, 61a
AT (t, John-is-unaware-of-Mary')])]
_. Chapter Summary
Yet, under the assumption that time is dense (and that for
all w1 e: Inr«w,D), Inr«w,D) = Inr«w1,D», the adjuncts
in (i) and (ii) have equivalent translations in my analysis:
(1) His mother being a doctor, John would know the way
to the Med Center.
(2) The water being a little cold, the children must
stay on the beach.
(3) His arm being in a cast, Bill might not be asked
to participate.
Jane, T
-------------
with her1 hair braided, (she1) [must]
resemble Mary, t, 60
I
with her1 hair braided, (she 1 ) [must] resemble Mary, TAB, 58
I
with her 1 hair braided, (she 1 ) [must] resemble Mary, TAB, 56
I
with her 1 hair braided, (she 1 ) [must] resemble Mary, t, 18
l
[must] , with her 1 hair brt~~,·de10d, M(Sahrey1)t[re1SgembleS]
MOD , ,
~ir braided, TAB, 5b
---------
(38) with her children asleep,
(Mary) often [watches] TV, t, 55,'
I
often with her, children asleep, (she,) [watches]
RFA TA, 9,9 TV'I TAB, 58
with her, children asleep, (she,) [watches] TV, TAB, 56
. ~TAB, 5b I
~ ~ (she,) [watches] TV, t, '9
ber, children, T asleep, ADJ s
Jane,
T ------------
with her 1 hair braided, (she 1 ) [reminds]
us of Mary, t, 60
I
with her 1 bair braided, (she 1 ) [reminds]
us of Mary, TAB, 66a
~
with her, hair braided, (she,) [reminds] us of Mary,
TA, 9,9 TAB, 58
I
with her 1 hair braided,
I
[reminds] us of Mary,
~AB, 5b TAB, 56
In the latter case, in which the time adverb joins with the
perfect, the adverb is within the scope of tense; this is
not so in the former case, however.
Now consider sentence (50) again. (50), like (51), is
ambiguous. On one interpretation, (50) entails that John
led the project during the most recent summer relati ve to
the speaker's interval; on the other interpretation, it
entails that John led the project during the most recent
summer relative to some past interval. These two
interpretations can be represented as in (54) and (55).
Vt 2 [t 2 ~ t, &
AT(t 2 , we-make-a-lot-oC-progress')]])])]
with John leading the project during the past summer, (we) [had]
made a lot of progress, t, 60
I
with John leading the project during the past summer, (we) [had] made a
lot of progress, TAB, 57
I
with John leading the project during the past summer, (we) [have] made a
lot of progress, TAB, 70
I
with John leading the project during the past summer, (we) [have] made a
lot of progress, TAB, G1a
~
with John leading the (we) [have] made a lot of progress, TAB, 62
project during the past ~
summer, hTA, 11,6,9 [have], (we) [make] a lot of progress, TAB, 56
I TAb/ITAb I
with John leading the (we) [make] a lot of progress, t, 1~
project during the past summer,
(61) with John leading the project during the past summer, (we) [made]
a lot of progress, t, 60
I
with John leading the project during the past summer, (we) [made] a lot of
progress, TAB, 61a ~
with John leading the project (we) [made] a lot oC progress, TAB, 57
--------
during the past summer, MTA, 11,6,9 I
I (we) [make] a lot of progress, TAB, 56
with John leading the project I
during the past su.mer, TAB, 61b (we) [make] a lot of progress, t, 19
during the past summer, tHA, 63 with John leading the project, TAB, 5b
I ~
during the past summer, TA John, T leading the project, PRPL s , 45
THE SEMANTICS OF ABSOLUTES 289
(62) Vt[K(L 6 )(A Vt ,[M 9 (t,t,) &
[during-the-past-summer'(t,) &
AT(t" Vxs[R(x s , John') &
Ing(~[lead-the-project'(xS)])])]])(~[past(t) &
AT(t, ve-make-a-lot-of-progress')])]
4. Chapter Summary
in Chapter IV.
Thus, the overall analysis of absolutes advocated in
this chapter closely parallels the analysis of free adjuncts
set forth in Chapters II-IV. In particular, inference is
assumed to play the same kind of role in the interpretation
of absolutes as in the interpretation of free adjuncts. For
example, absolutes of category MTA, like free adjuncts in
this category, are assigned intensional logic translations
containing the variables Land M, which represent semantic
indeterminacies that can only be resolved by means of
language users' inferences. Such inferences, I claim, are
what is responsible for' the relation of causation or
explanation that is felt to link the absolute with its
superordinate clause in (63), for the relation of temporal
succession in (64), and so forth.
Footnotes
correct, since the former but not the latter can appear in
the usual diagnostic environments for stage predicates;
furthermore, the former expressions serve as weak adjuncts,
the latter as strong adjuncts--
cannot be maintained.
Finally, section 5 is devoted to the theoretical
implications of the proposed analysis. In particular, I
shall address the followinb central issue: Is the proposed
analysis consistent with Montagovian compositionality? That
is, can the meaning of an absolute construction be regarded
as a function strictly of the meanings of its parts and of
the manner in which they are combined, or must language
users' inferences be reckoned as a third potential
determinant of the meaning of such a construction? My
discussion of this issue will necessarily be very general
and somewhat speculative; nevertheless, some significant
conclusions regarding this and other matters do seem to
follow from the account of free adjuncts and absolutes set
forth here.
Table 4
The Formal Characteristics and Category Membership
of Weak and Strong Adjuncts and Absolutes
Free Adjuncts Absolutes
The relation between the strong adjunct and the main clause
of this sentence is semantically indeterminate in at least
two ways. First, the temporal relation between the adjunct
and the main clause is open: (7) can be inferred to mean
either that John is thirty-five now, or that he will be
thirty-five in 1988. Temporal considerations aside, the
relevance of the adjunct to the main clause in (7) is not
strictly specified by the semantics of English. One would
almost surely infer that the adjunct in (7) provides the
explanation for the truth of the main clause; note, however,
that in sentence (8), the very same adjunct would most
likely be inferred to bear an adversative relation to the
main clause.
(9 ) Vt[K(L)(~vt1[H(t,t1) &
(11) Picassos;
be asleep
be in the barrel
be ready to leave
etc.
(26) You loyter heere too long, being you are to take
souldiers up.
(Shakespeare, Henry IV, Part 2, 1597, cited
by Visser (1972:1218)-)- - - -
2.2.2. Instantaneousness
heron while she was walking home or, somewhat less probably,
after she got home; in (42), it seems quite natural to infer
a relation of succession.
meanwhile therefore
then hence
subsequently thus
afterwards consequently
beforehand thereby
furthermore still
moreover anyway
in addition nevertheless
besides nonetheless
also instead
(69) Once her hair was braided, Mary put on her makeup.
When her hair is braided, Mary always resembles
Jane.
If her hair had been braided, Mary might have won
the lookalike contest.
on the other hand, i f the adjunct carrying the box under his
INFERENCE AND LOGICAL ROLE 331
arm is substituted into (70a-c), it is not assigned the same
three logical roles; instead, it is interpreted as
expressing an 'attendant circumstance' in all three cases.
If the subtle dependencies illustrated in these examples
were somehow built into the transformation converting
adverbial clauses into absolute constructions (so that
crossing the street would, for example, derive from a while-
clause in context (70a), from an after-clause in (70b), and
from a by-phrase in context (70c», the resulting rule would
be so highly conditioned by semantic and pragmatic factors
as to be radically unlike any transformation which has been
seriously entertained. But to the extent that the
transformationalist approach failed to account for these
dependencies, it would have failed to provide a credible
account of the semantic variability of absolute
constructions.
Another reason for doubting the feasibility of the
transformational approach to absolute constructions is the
fact that such constructions cannot always be plausibly
paraphrased by adverbial subordinate clauses; as was pointed
out in Chapter I, Kruisinga (1932:275) has observed that
'the use of free adjuncts in English is greatly promoted by
the almost complete absence of conjunctions that can
introduce a subordinate clause expressing attendant
circum stances.' For example, the adjuncts in (71)- C7 4)
cannot be paraphrased without rather elaborate
circumlocution; these adjuncts can't simply be regarded as
'reduced' versions of full adverbial clauses.
(74) Carrying the box under his arm, John entered the
bank.
(81) She clapped her hands like a child, her lucid eyes
sparkling.
(OSI, 246)
INFERENCE AND LOGICAL ROLE 333
(82) ••• a female matador lies across the horse's back,
her breasts bared, an espada still clutched in her
lifeless hand.
(NY 6/30/80, 57)
(93) The loss to him was two hundred and fifty pounds,
the equivalent of the savings of four good years--
a good year being one in which the net income of
the croft, including the government subsidies he
gets for his sheep and cattle, is about a hundred
and fifty pounds.
(CL, 33)
INFERENCE AND LOGICAL ROLE 335
(94) Zelinsky once tried to answer the question of
where the South ends by taking a horse-mule census
within the territory in doubt--the theory being
that farmers who used mules to pull their plows
obviously lived in the South.
(NY 6/16/80, 107)
:>..t[past(t) &
AT (t, John-reads-the-commission's-report')]
5. Theoretical Implications
Footnotes
( iii)
1. Intensional Logic
(i) t'
0 t (read: t' is included in t)
(ii ) i'
0 i
( iii) k, 0, st
CD e
De,E,W,I =E
Di,E,W,I = I
D
t,E,H,I = {A, {A}} (where the empty set is identified
with falsehood, the unit set with
truth)
Do,E,W,I = o'(E)
Dst,E,I-J,I = st'CE)
(k', 0', and st' are functions connected with the respective
sorts; k'(E), o'(E), and st'CE) are all subsets of E, and
are to be taken as the set of kinds, of objects, and of
stages, respectively. k', 0', and st' must be so structured
that k'(E) n o'(E) = k'(E) fl st'(E) = o'(E) n st'(E) = A,
and that k'(E) U o'(E) U st'(E) = E.)
Di',E,W,I = I
= Sa' ,E,W,!
D[a',b'],E,W,I = {x: x E Da',E,W,I or x E Db',E,W,I}
[It will have been noted that kinds, objects, and
stages are here treated as sorts of entities, while
intervals are treated as things of a distinct type from
e n tit i e s • I n Do wt y' s (1 9 7 9 : 326) w0 r d s , ' the p rim a r y
moti vat ion for sorting is to allow certain variables
and constants to range over the whole domain of
entities, as well as allowing other variables to range
over only a part of it.' Since it is important that
there be variables ranging over kipds, objects, and
stage:;;--for exall)ple, the variable Xl over individuals
in AX1VXs[R(xs,x 1 ) & asleep'(x s )] must be of the sorted
type [o,k]--these must be treated as different sorts of
entities; but no expression of the fragment requires
that there be variables ranging over intervals and
entities, so that the type of intervals can be treated
as distinct from that of kinds, objects, and stages.
Some denotations of type (a,b> and (s,a> (a,b £
Type) are allowed to be partial functions so that the
set of possible denotations of sorted types (a' ,b'> and
(s,a'> (a',b' E SType, a' ®
a, b' ®
b) can
simply be characterized as a certain subset of the set
of possible denotations of types (a,b> and (s,a>; as
will be seen below, this does entail some complications
in the definitions of the denotations of certain
expressions.]
The language La of intensional logic is assumed to have
denumerably many variables and an infinite number of
constants of each sorted type. Given any nonnegative
integer n and any a' E SType, vn a' is the nth variable of
sorted type a'; Con a , is the set 6f constants, Var a , the set
354 APPENDIX
contain p.
U) If a E: HEa and u is a variable of sorted type b',
then [A uaJ !} wig is that function h with domain
Db,E,W,I Cwhere b' b) whose total
subfunction has domain Db',E,W,I and is such that
®
. [ J!} wig'
for a 11 x E.: Db',E,H,l' h ( x ) 1S a , where g ,
is the .a-assignment like g except for the
possible difference that g'Cu) is x;
(4) If a ,
E.: fJlE<a b> and S E.: ~1Ea' then [aCS)] .Qwig is
[a] !} wigc [s] .Q wig);
(5) I f a,S E: MEa' then [[0: = S]] Dwig is
A FRAGMENT FOR FREE ADJUNCTS AND ABSOLUTES 357
{A} iff [ex] .Qwig and [8] .Qwig are defined and
identical;
A iff [ex] .Qwig and [8] .Qwig are defined and
nonidentical;
and otherwise undefined:
(6) I f q"lj! E: MEV then [-'q,] .Qwig is
{A} iff [¢].Q wig is A;
g' is as above;
and otherwise undefined;
similarly for Au~;
2.1. SYNTAX
BTA =- [-Adjunct].
( iii) PTAB n [+Adjunct] = [-Weak] U [+Weak] ;
[-Weak] n [+Weak] = A.
(iv) Pt U PTAB = [-Tense] U [+Tense];
[-Tense] [l [+Tense] = A·,
Bt =- [-Tense].
(v) PTAB = [-Perfect] U [+Perfect];
[-Perfect] n [+Perfect] = A.
(vi) PT = [-Plural] U [+Plural);
[-Plural] n [+Plural] = A;
theY4n+3' we s [+Plural];
all members of BT other than theY4n+3' we,
are members of [-Plural].
Quantification rule
2.2. TRANSLATION
«s,«<s,t'>,t'>,t'»,«s,t'>,t'»;
(7) for all a £ BTA , c(a) = <i' ,t'>;
(8) for all at £ BRFA , c(at) = «i' ,t'>,«i' ,t'>,t'»:
(9) for all a € BrQ U Bt/t , C(a) = «s,t'>,t'>:
A FRAGMENT FOR FREE ADJUNCTS AND ABSOLUTES 379
Pk , Qk k'
pi, Qi in'
pi in"
p, q, r, s <s,t'>
t, tl' t 2 , i'
I, I1 ' 1 2 , <i' , t' >
pt , Qt <s,<i',t'»
pt <s,«s,<i' ,t'»,t'»
M, Ml' M2 , <i',<i',t'»
L, L1 ' L2 , «s,t'>,«s,t'>,t'»
t~t,]j
tomorrow BTA ) translates as
(€
AtVt 1 [[daY'(t 1) & At 2 [todaY'(t 2 ) ~ [t 2 ( t,
& At 3 [[t 2 ( t3 & t3 < t,] -+ todaY'Ct 3)]]]] &
t ~ t,];
during the paat aummer BTA ) translates as
(~
R(xs,x k )];
[be] (e PIVi/PREDs) translates as
APSAxiVxS[R(xS,xi) & pS{x s }];
[be] (e PIVi/PREDi) translates as Api[vpi];
(viii) beCore translates as
AptAtVt,[[t < t, & M(t,t,)] & pt{t,}];
after translates as
AptAtVt,[[t, < t & M(t,t,)] & pt{t,}];
when translates as AptAtVt,[t = t, & pt{t,}];
while translates as
AptAtVt,[[t ~ t, & ~ment(t,)] & pt{t,}];
(ix) at-that-timen translates as
AptAt[t = tn & pt{t}];
since noon translates as
AptAt([xn(t) & Vt, [[noon'(t,) & Vt 2 [day'(t 2 )
& [t, ~ t2 & nov~ t 2 ]]] & t, < t]] & pt{t}];
(x) [have] (c BTAB//TAB) translates as
AptAt[perC(t) & vt,[t, ~ t & pt{t,}]];
(xi) John, Mary, Bill, Jane, and Fido translate as
APopo{John'}, Apopo{Mary'}, APopo{Bill'l,
Apopo{Jane'}, and APopO{Fido'}, respectively;
A FRAGMENT FOR FREE ADJUNCTS AND ABSOLUTES 383
we translates as Apopo{we'};
hen' she n , itn translate as APOpO{X~};
that and this translate as APopo{that'} and
APopo{this'}, respectively;
theYn translates as Apipi{x~};
(xii) his n , hern' its n translate as
APOAQO[Qo{1Xo[POSS(x~,xo) & pO{xo}]}]:
theirn translates as
APOAQO[Qo{ 1Xo[Poss(x~,xo) & pO{xo}]}]:
our translates as
APOAQO[Qo{ 1Xo[Poss(we',x o ) & pO{x o }]}]:
(xiii) where g is a biunique function whose domain
is the set S of members of UA e: Cat BA not
mentioned in (i)-(xi) above, g(a) e: Conc(a)
whenever a e: S.
,
then F 3 n (a, B) translates as
.
a,(AAt[AT(t, Vxs[R(x s , x~) & B'(x s )])]).
T 17. I f a e: PM 0 D' cP E: P t' and a, cP t ran s 1 ate a sa' ,CP " the n
F 17 (a,CP) translates as a'(C(cb»)(~cp').
T18. If a E: P MOD ' cP E: Pt " 1jJ e: Pt' and a, cP, 1jJ translate as
a', cp', 1jJ', the n F 18 ( a, cP ,1jJ) t ran s 1 ate s as
a'(D(cb) (~cp'»)(~1jJ').
130. I f a E
PIVi/Meas' 13 E PMeas ' and a,S translate as
a',S', then F 30 (a.S) translates as a,(A S ')'.
T36. If a E PIVs, 13
Ppps, and a,B translate as a',B', then
E
T61. If ex E: PMTA , e £:
PTAB , and Ot,B translate as Ot',B', then
f 61a (a,6}, F 61b (ex,a) translate as ex'C .... B').
A FRAGMENT FOR FREE ADJUNCTS AND ABSOLUTES 389
T69. If a E: PTSC ' S £ PTAB ' and a,S translate as a',S', then
F 69(a,S) translates as at(AAt[NONPAST(t) & s'(t)]).
390
REFERENCES 391