Aspect and Aktionsart: Fighting or Cooperating? 
Allan Ramsay 
Centre for (?omputal;iolml Linguistics 
UMIST, PO Box 88, Manchestel' M60 1QD, England 
Abstract 
It is widely accepted that semanti(: ttle- 
ories should, as far as possible, be com- 
positional. The claim that a dmory is 
(:omt)ositional, however, lacks bite if lex- 
ical and pre-lexical items are allowed to 
mean dilrerent things ill difl'erent con- 
texts. The aim of the current paper is 
to show how to deal with a well-known 
t)henolnenon by relying on (:ombinato- 
rim effects to infer difforent consequences 
from the same items in difl'erent contexts 
without altering the contributions that 
these items make individually. 
1 Compositionality vs. Coercion 
Consi(ler the thllowing sentences: 
1 Henrietta was crossing the wad. 
2 Harry was hiccupping. 
in (1) it seems as though the present participle 
marker is being used to indicate, that some event 
with a well-defined end point was in progress at 
some time in the past, and that it is reasonable 
to suppose that this end point was eventually 
reached Ihmrietta did cross the road. Cases 
like (1) are generally taken to be prototypical: the 
present parti(:it)le marker indicates the progressive 
aspect, which says that sonto extended event with 
a recognisable end t)oint is in progress and will 
probably reach its conclusion. 
(Asher, 1992) considers the circumstances un- 
der which (1) will lend you to conclude that Hen- 
rietta did indeed reach the far side of the road, 
arguing that this conclusion can only be reached 
by using a default inference rule which would be 
cancelled in cases like: 
3 Ilcnrictta was crossing th, c wad, when sh, e was 
hit by a bus. 
I have. no argument with his analysis of (1) and 
(3). What concerns me here is the apparent 
change in the contribution of the present partici- 
ple marker in (2). In (2) we have a (conceptually) 
instantaneous (',vent, namely a hiccup. Since hic- 
cups are generally thought of as taking no time, 
it; is not possible~ to be in the middle of a single 
hiccup and hence we are solnehow driven to con- 
elude that Harry was in the middle of a series of 
hiccut)s. A similar problem arises with: 
4 Allan is living in Bray. 
Here we have a homogeneous state where there is 
no result to be achieved no interesting state of 
affairs that arises as a consequence of reaching the 
end point. As such the present participle cannot 
be taken as an indication that the cuhnination 
of my living in Bray has not been reached, since 
there is no such cuhnination to reach. In this ease 
the present l)artieiple somehow transforms itself 
to an indicator of temporariness, so that you {:an 
gel; exchanges like the following: 
Allan's living in Bray 
I thought he lived in Buxton 
.... Yes, but he's on a visit to Ireland at 
the inoInent 
(Smith, 19911) deals with this phenomenon by ap- 
pealing to a notion of "derived interpretations", 
though with very little discussion of how the 
derivations take place. (MoSns and Steedman, 
1988) deal with it by invoking a process of co- 
ercion which changes the meaning of the ast)ect 
markers as required by the properties of the verb 
to which they are attaehe(t. Much of what I want 
to say below follows their anMysis, with one m~> 
jor difference.. Coercion changes tile meaning of 
the aspectual marker in response to the semmltic 
properties of the marked verb. But if items are al- 
lowed to change their meanings as a consequence 
of the semantic properties of other items then the 
principle of compositionality that 
"the mealfing of the whole is made out of simple l 
(:onlbinations of the meanings of the parts" 
t If we alh)w arbitrary rules of combination then 
we (:an include rules which make arbitrary changes 
889 
becomes rather ineffectual. We are, alter all, led 
to describe a word as being a homonym in exactly 
those cases where the meaning of what appears to 
be a single lexical item depends on the semantic 
properties of the words it is being combined with. 
In 
5 IIe keeps his money tied up in the bank. 
6 He keeps his boat tied up by the bank. 
the fact; that; the interpretation of bank depends 
on the semantic properties of money and boat is 
what; persuades us that the form bank is being used 
to realise two diffe.rent lexical items. We do not, 
however, want to describe the present participle 
marker as being ambiguous, with different inter- 
pretations which depend on the semantic context 
in which it occurs, unless we are absolutely forced 
to. The analysis in this paper attempts to show 
that the effects described by Mogns and Steedman 
(:an be achieved without any meaning-changing op- 
erations or unwanted ambiguities. 
The basic tool that I will use is the observation 
that A U F ~ A can hold when neither A ~ A nor 
\]7' ~ A does, and ill particular that if P and F' are 
different then A U F ~ A and A U P' ~ A' can 
hold where A and A' are different, or even incom- 
patible. If we back ut) the labels representing lex- 
ical and predexical items with appropriate sets of 
meaning postulates then we may well find that dif- 
ferent things can be inferred from a single itein in 
different semantic contexts without being forced 
to conclude that those items themselves mean dif- 
ferent things. In this way the meanings of words 
will cooperate to convey more complex inessages 
than each can carry alone. 
2 Meaning Postulates 
Consider tile following analysis of 
7 Harry was hiccupping 
,.A::{ (subset(A, .\[ B, name(B, Harry)\]) 
A IAI = i)} 
~C :: {past( C) } prog( C, 
• IV, (vati ,,,t(D, A) 
A event(D) 
A tvp (D, 
This 2 a is all very well as far as it goes, but unless 
the consequences of saying that something is an 
to the meanings of the parts. If this happens then 
the principle has no force. "Simple" combinations are 
usually taken to be things like function application 
and set union or intersection. 
~The analyses in this paper require a combination 
of truth functional operators and l-abstraction. I 
use the notation .\[x, P\] rather than txP to empha- 
sise that I am relying (Turner, 1987)'s treatment of 
abstraction, where you can safely combine the two, 
rather than classical l-calculus where you run the risk 
of paradoxes if you combine them. 
a~X::{P}(Q) says that Q is true of the X which 
satisfies P. As such it performs much the same role 
event of type "hiccup", or that someone is the 
patient of an event, and so on, are spelt, out ill 
detail then it is not possible to perform any non- 
trivial inDrences on tile basis of this interpretation 
(and hence not; possible to argue about whether or 
not it is right, so that the clainl that a fragment 
of natural language should be paraphrased ill a 
parl;icular way becomes vacuous). You might, for 
instance, disagree with my decision to label the 
sleeper as the patient of the event. Unless I spell 
out what this label commits me to, there is no 
way for me to defend it; or for you to attack it,. 
Simply appealing to our everyday intert)retation 
of tile term will not do. 
We therefore need to develop a collection of 
meanin.q postulates (MPs) to specify tile conne(> 
dons between the terlns that will appear ill our 
interpre.tations. This is perhaps an obvious point, 
but apart from a few honourable exceptions (the 
attempt in (Dowty, 1988) to specify the conse~ 
quences of assigning all item to a thematic role 
is a notable case) it is too often neglected. The. 
central claim of the current paper is that; tile inter- 
actions between Ineaning postulates can produce 
sut)tle effects which you may miss if you simply 
label items as belonging to (:lasses or as being in 
relationships with one another and leave it at that 
if you simply say, for instance, that some event 
is progressive, without spelling out the MPs for 
progressive. 
3 Aktionsart and Aspect Revisited 
I will now look in some detail at aspect and aktion- 
sart. For the remainder of this section I will say 
that; the relationship specified by an aspect marker 
holds between a tilne and all event type, where an 
event type is nothing more. t, hall all abstraction 
over a proposition about ewmts. What we need 
are tile meaning postulates that spell out the con- 
sequences of saying that a time and an event; type 
are in the relationship specified by some aspect 
marker. I make the following assumptions: 
• The aspect; of the core verb specifies a re- 
lationship between all instant and all event 
type. The details of these relationships are 
spelt; out via MPs. 
• The tense of the core verb, together with 
ally auxiliaries, specify a relationship between 
the present time now, all anaphoric reference 
time re\]', and the time mentioneA in this rela- 
tionship. Nothing much in tile analysis below 
depends on the particular properties of the. 
time lille. The only assumption that i will 
make any use of is t;hat there are intervals 
and instants. 
• The MPs for the core verb specify the tempo- 
ral properties of the event type. If the verb 
as (Barwise and Perry, 1983)'s notion of anchoring. 
890 
shares temporal 1)roI)erties with a range of 
other verbs, then dwse are gathered together 
as MPs for the (:lass as a whol(;, which is r('- 
ferred to as an aktionsart. 
The first MP we, will consid('r deals with the pro- 
gressive asi)e(:t , as folh)ws: 
MP I)rog: 
Vtgl'(prog( t, 1") 
A -\]el (lnc"m.bc','(cl, E) 
A \]t,l(t,l ( t 
A ,~t,.,./4..(.,, ~,))) 
A ~e.2 (m,e.m.r(e2, E) 
A - ~-fll,,e (l:~ < t 
A e~,,dpl,(e.~, t2))))) 
This says that the relationshi I) prog hol(ls t)(!tween 
an instant /; and an event l;yi)(; P if there is a sel, 
}\]\] of events of t;h(; al)t)roprial;e tyl)c , at; least one 
of whi(:h starts })efore the insbmt and at least, one 
of which does not end 1)efor(~ it;. 
Meaning 1)osl;ulates are 1101; ileCe.ssary alia Slltli- 
('.lent; (:on(tit;ions. They are ('.onstellations of \['acts 
which, serve to sl;rllCtlll'('. \[;he (;on(:et)imal st)a(:c. 
They do not exhaust thai; sl)a(:e , an(l they do not 
ne,(-essarily botl;om ()lit in sense-dal;a l)ased prim- 
itives (Carnap, \]936; Quine,, 1.960). The mosl;, 
and least, you ca.ii say about them is that Lhey 
help d(;lin(',t~te, a set of (:oncep\[;s a.n(l relatiolls \[)(> 
t;ween con(:el)ts which can 1)e use(\] to l)oint out I;he 
relations that \]mid among words, and at eerta.in 
l)oinl;s between wor(ls an(1 exl)(',ri(m(:es. In ((buse, 
\] 986)'s t)hrase, l;hey e.xl)r('ss s EMANTI(; 'I'I{,A ITS 
state.merits about so'me of the things that tyt)Jcally 
follow from asserting tltat some r(;lationshi I) hohts. 
There is th(;reIore n() irres()lwd)le (:lash 1)etween 
MP prog and Asher's MPs whi(:h des(:ril)e t;he 
conditions under whi(;h you wouhl expect a relic 
event des(:ribed using the progressive Lo t)roceed 
to its cuhnination, and i would expect, to supple- 
menl; what I have to say in this 1)ap(;r wid~ his 
defaull; 1;real;l\[tellt of t, his ol;her issue. In t)articu- 
lar, it should be. note(l l;\[lat; MI)-prog entails l;he 
existence of a starl; l)oint for the rei)orte(l a(;l;ioil 
t)llt 1lOt that of an end 1)oint. 
The de(:ision to talk in terms of sets of ev(;nl;s 
provides exl;ra tl(;xil)ilil;y, ill the same way that the 
de.(:ision l;o deal with NPs in terms of sets of indi- 
vi(hmls supt)orts th;xible tr('.atmenI;s of plurals and 
of otherwise awkward l)henoinena such as generics 
and bare plurals(l{amsw, 1992). We (:an always 
(:onstrain a set of eve.nts to t)e a singleton if we 
need to, so certainly nothing is, lost by talking 
about sets rather than individuals. 
SUl)l)ose w(; have the following MPs for aktion- 
sarts and thematic roles: 
mP ,;vent: Ve(event(e) -+ ~t(,(startpt(c, to)) 
A ~t. (crulpt(e, 1,1))) 
MP telic_event: 
v,:(.,a:t.,:o..~,,,r.t ( ,:, tet,:c_.ve,,,t) 
:J,.'(','e.~,,dt (., ,J 
Avt~ (.',,Or(< t,~) 
+ at(t1, .~) 
A Vl,'(I,' < t, 1 
MP action: Ve(o,M.ion.sa,rl.(o., action) 
MP state: 
Ve( aktion.~arff e, .star.c) 
At,<i 
A pal'ie',d,(c., :,:)) 
-÷ ..z(t, z'.:,:))) 
MP o.xt;end('.d ,;vent: 
V e ( ak l.ion.sar t ( e., e.x te'ndo.d .eve.n/.) 
~ w.,,w:, ((.~t..,,'tvt(e, t.,) A .,,.@t(,,., /.,)) 
-~ ~t(~.0 < l. < t, )) 
MP inst_event: 
V~@d~:tion, sa'rt(c, i'n.sl,_cvc'n,t) 
~ ~.,,:\]t, (,..t..,.tW.(., 1.,,) a .,,.@t(., t, ) 
A -,~t(t0 < tAt < till) 
MI" cxt;ended_telic_aetion: 
Ve( ,'.:,:t~',.1,,,.d_tcl'ic .,.tio..(o.) 
> .,.t',:o,,~ (.) 
A c:rl, c.n, dcd e'vc'n.t(c) 
MP agent: V:,:Vc(o, gc'n,l,(G a:) 
..,u.~c(:,:, o,) 
A Vs('r~s',,.U.(.., s) 
MP-in|;end: V:,:V.s('intend(:,:, s) + ,.ni'm,,,l.c(:,:)) 
mP patient: Va:Ve(1,al.ic,tt(c ,:,:) -~ animal.e(x) ) 
fl'ht;st, arc' all straighl;forward enough. I,\]vents have 
sta.r\[; and end points. Telic events have results, 
which are charat:terised by propositions which be- 
come true at the (;ii(t point of the. ev(;nl;. A stal;e 
is charaeterised 1)y a prot)erty P l;hat hol(ls of the 
state's patient a: throughout some interwd i. A('.- 
tions are ev(;nl;s with agents, where an agent is 
a being that inl;entionally causes the resull; of the 
eVellt \[;o \])CCOllle ti'ile,~ aild oIlly aliiHlaLe t)eiltgS ca.Ii 
intend to bring things about, l?atients are.just; an- 
imal, e, beings. 10,xtended events take time (th(;re ix 
some instant 1)etween their start an(l end points), 
instantaneous events do llOt (tiote that this may 
or not lneall that the. stm'l; anti end points of an 
instantaneous (;veltt are identi(:al, depending on 
whether we regar(l the time line as d(mse. As far 
as th(; (:urrent t)at)er is concerned this is a free 
ehoic('.). Exl;(;n(h;d leJi(: actions are. just extended 
evellts with results whi(:h become true at; their end 
t)oints and agents who inten(1 I;hose results 1;o })('~- 
(:ore(; true. 
All we need to know at)out eat and hicc'u.p for 
the moment is that; cat denotes &ii extended relic 
action and hiccup denotes an instantaneous evenl;: 
891 
MP eat: Ve(type(e, eat) 
--+ extended_relic_action(e)) 
MP hiccup: Ve(type(e, hiccup) --> inst_event(e) 
MP eat says that eating events take time, and 
MP hiccup says that hiccuping events doift (or 
rather that we don't think about the time they 
take). There are many other MPs dealing with 
these verbs, since there is a great deal more to be 
said about them, but we do not need this extra 
detail here and hence we will omit it. 
Compare now the following analysis of 
8 Ite is eating a peach. 
3A ::{A C .\[B,peach(B)\] A IAI =1} 
,,C::{O c_ .\[D, male(D)\] A ICl =1}, 
prvg(now, 
• \[E, object(E, A) A agent(E, C) 
A event(E) A type(E, eat)\]) 
with the interpretation of (7) given earlier. MP 
prog says in each case that there must be an 
event whose start point is before now and an event 
which does not have an end point, before now. In 
the case of (8) this is compatible with the possibil- 
ity of there being exactly one such event. Indeed, 
since only one peach is involved, the remainder of 
the MP for eat '(which would include the informa- 
tion that you can only eat something once) would 
presumably force this conclusion. It is further- 
more compatible with the requirement that there 
should be an event whose start is before t and an 
event, whose, end is not before t, since eating events 
are extended - if they have end points then these 
are after their start points. In the case of (7) it 
is not, possible for there to be a single event, since 
the start and end points of a single hiccup are 
taken to occur with no intervening instant. We 
therefore find that (7) must denote a set of hic- 
cups, simply by inspecting the MPs and without 
resorting to a process which turns hiccupping from 
an instantaneous act to a homogeneous sequence 
of acts. In both cases, the sentence reports a se- 
quence of events. But in (8) there is nothing to 
say that this sequence has more than one mem- 
ber, and the fact that only on(; peach is involved 
suggests that it has exactly one member; whereas 
in (7) the temporal properties of the conceptually 
instantaneous act of hiccupping mean that there 
must be more than one such event. 
Returning to 
4 Allan is living in Bray. 
we get the following interpretation: 
~A::{A C_ -\[B, name(B, Allan)\] A IA t -1}, 
prog(now, 
• \[C, agent(C, A) 
A event(C) A tyve(c, live) 
A ~D::{D C .\[E, name(E, Bray)\] 
A iDI--1}, 
in(O,D))\]) 
Why does this carry an overtone of "temporari- 
ness"? Assuming that live denotes a state, we 
need to look at the interactions between the MP 
for the progressive aspect and the MP for the ak- 
tionsart state. MP state says that the charac- 
teristic property of the state is true of its patient 
throughout some interval, but unlike MP relic- 
event it says nothing about the start and end 
points of that state, not even whether or not they 
exist. Of course in general we know that most 
states do have start and end points, but in many 
cases that is all we know about them. A speaker 
who is committed to the existence of a state, then, 
may not be concerned about the existence of the 
start or end point of that state they may not 
know when it started, they may not care whether 
it has ended, as far as they are concerned it may 
have been going on since the beginning of time and 
it; may continue to the emt of time. If, however, 
their report of tiffs state invokes the progressive 
aspect then they do become committed to know- 
ing something about the start and end dates. If, 
for instance, we were considering Allan was living 
in Bray rather than Allan is living in Bray then 
we would assume that the speaker knew enough 
about the end of this state to place it before the 
reference point marked by the past tense of the 
auxiliary. Thus the use of the progressive aspect 
here commits the speaker to the existence of an 
end date for the state in a way in which commit- 
ment to the existence of tile state does not: it 
is this that gives (4) its feeling of being about a 
temporary state of atfairs 4. 
We now turn to the simple aspect. Consider the 
following pair of sentences: 
9 Allan lives in Bray. 
10 Mary eats a peach for her" lunch. 
(9) describes a simple homogeneous state of af- 
fairs. The properties of the verb live and t, he sin> 
ple aspect seem to collude in this case, and there is 
no need for anything like coercion. In (1.0), on the 
other hand, the.re does seem to be a problem. Eat- 
ing denotes an activity with a definite final state, 
where what was eaten ends up inside the eater's 
stomach. Somehow (10) conveys the message that 
Mary habitually eats a peach for her lunch: note 
in particular that it is not the same peach or the 
same lunch every day! 
We therefore need a single MP for the simple as- 
pect which enables us to conclude different things 
tbr the two cases. For (9), where the verb denotes 
a homogeneous state of affairs, the simple aspect 
supports tile conclusion that such a state of af- 
fairs does indeed hold. For (10), where the verb 
denotes an activity, the simple aspect supports the 
4of. (Smith, 1991)'s observation that aspect pro- 
vides a spotlight on some portion of the event. 
892 
conclusion that such an activity hal)pens on a reg- 
ular basis. The following meaning postulate says 
that the relationshit) simple holds between an in- 
stant t and all event type P if there is an interval I 
which contains t, and for any instant t' in I there is 
some event e of the apt)ropriate tyI)e which starts 
before t' and finishes after it. 
MP simple: 
VWP(.simple(t, P) 
-÷ _~ I ( interval (1) 
AtCI 
A Vt'(t' C 1 
-* )\]c(l'.e 
A ~to(to < t A 1,\[) ~_ I 
A startpt(c., to)) 
A ~tl(tt < tl Atl C .\[,tl 
A c~'tdpt(c, tt))))) 
Consider tilt', interactions between this MP and 
the fonowing analyses of (9) and (lO). 
simple (now, 
• \[A, m::{.'_~ c .It, ....,,e(C, All..,O\] 
A IBI =:l}, 
a.qent ( a, B) 
A event(A) A type(A, live) 
A ,,D:: {D C .\[E, name(E, l/ray)\] 
A IDI =~}, i,,(A, D)))\]) 
simple (now, 
• \[A, ~B ::{v c- .\[C, pe,,eh(C)\] 
A II~l =I} 
tD::{D C .l E, ,name(E, Mary)\] 
A IDI =1}, ot:#et ( A, I5 A (,qent ( A, D) 
A event(A) A type(A, eat) 
A VF ::{for(A, F)}luneh(F))\]) 
RemeInber that the MP for live sws nothing 
about the start and end points of the specilied 
state. Then there is nothing in MP sinlple to 
lead us to infe, r the existence of more than one 
such state, of affairs. There is also n(/dting to en- 
able us to infer that there is no more ttmn one: I 
will return to this below. 
If, on the other hand, tile Mt ) for eat says that 
the start and end points of the action must be 
quite close together, then MP simple entails that 
there must be several such actions in the specified 
interval. Whi(:h is, after all, as much as you can 
infer fi'om the simple aspect itself. Note that the 
wide scope of the aspect oi)erator si'mple ineans 
that for (f0) we are conside, ring ewmt types in 
which there is a pea(:h, and a hmch, for every in- 
stance of the type. So unlike (8), where there was 
one peach and the event type we were consider- 
ing dealt with eating that one peach, here there is 
nothing driving us to (:onclude that there is only 
one peach and hence that the set of events nmst 
be a single, ton. 
The combination of have and a t)ast-participle 
(I will call this the pmuq.~(;q'wt,; different people 
use different terminology \['or this) prese.nts similar 
problents. We can obtain the same kind of inter- 
pretation for such sent;en(:es, paraphrasing 
11 He had slept. 
as 
~A ::{past(A)} 
reI =J 
A m::{u c_ .\[c,.,../e(c)\] A I~1 =1} ped(,'ei, .\[m, ,*ge,"(D, ~) 
A event(D) 
A type(I), sleep)\])) 
"All" we need now is a suitable MP for the relation 
per5 
Part of the diffcre, nce bet;ween this construction 
and the simple past arises from the explicit men- 
tion here of the R.l<lq,;R.l:;N(llr; TIME (l{eiehenltach, 
1956). Sentence, s like (1l) make reh~renee to s()nle 
anaphorically determined instant, and this gives 
them a slightly difli;rent flawmr t¥om sinqfle past 
sentences. But there is more to it than tha~. 
Consider the following examples: 
12 He lived in Bray for five years. 
la lie has lived in Bray for five years. 
14 He had lived in Bray for five years. 
The striking thing about these is that itt each of 
(13) and (:1.4) the obvious interl,retadon is tsha/; 
his period of living in Bray continued after the 
reference time, so that he probatfly live.d there for 
more than five years in total; whereas it is atl but 
impossilfle to read (112) as saying anything other 
than that his residence in Bray took no more or 
less than five years. This distinction becomes even 
clearer when we consider 
15 In 1919 It(; had lived in Bray for five', years. 
16 In 1919 he lived in B~zty. 
17 * In 1919 h,e lived in BTny for five years. 
It; seems that whereas you can have both a date 
and a duradon with the perfective, you Call have 
either but not both with the simple past. One 
way to account for this is to argue that the simple 
past deals with the end point of the event whereas 
the perfective deals with the end of some related 
interval. We have to be carefifl here. The MI' for 
the simph', asliect given above is designed'to be 
open to readings where some single past event is 
being report;ed and to the possibility of a "hat)if 
ual" reading. The perfective is also open to the 
same ol)tions: 
18 I had read the Times for years, but had grad- 
ually come to reeognise it as a capitalist rag. 
We have further to acknowledge the correct intu- 
ition that for telic events the perfcctive fbcuses on 
the end point of the event where the sinq)le aspect; 
views it as a whole. My current approach takes 
the MP given above for simple as a basis for t)oth, 
893 
but adds an extra clause saying that for the simple 
case all the events in the specified set, end within 
the interval: 
MP simple': 
VtV P( simple( t, P) 
-+ ~I(interval(I) 
AtEI 
A Vt'(t' C I 
--+ ~e(I'.c 
A 3to(to < t'Ato E I 
A startpt(e, to)) 
A 3tl(t' < tt Atl E I 
A cn@t(e, t~)))) 
Ve'((P.c' 
A 9t2(,startpt(e', t2) A t2 C i) 
--+ Vta(er~,dpt(c', ta)) 
Omitting this extra clause from the MP for perf 
means that the set, of ewmts in question could in- 
elude one that is not yet complete, so that 
19 I have also read the Guardian for years, but 
I am now becoming dissatisfied with it as well. 
has a past habitual reading which is ()pen to con- 
timmtion in a way that the habitual reading of 
the simple past cannot be. The ramifications of 
this require further exploration, perhaps in con- 
junction with a treatment of implieature, like thai; 
given in (Gazdar, 1979) to explain why examples 
like (19) generally give rise to the feeling that the 
event sequence in question is not yet over and done 
with. 
4 Conclusions 
The analysis above of the interaction between the 
simple and progressive aspects and various kinds 
of verb shows that at least some of the phenomena 
dealt with by (Mo~ins and Steedman, 1988) can 
be explained without appealing to actions which 
change the meanings of the lexical and pre-lexical 
items involved. In the approach outlined here, 
every sentence reports a set of events. Aspect, ak- 
tionsart and temporal modeifiers then provide in- 
formation which can be used to determine the car- 
dinality of this set and to draw other conclusions 
about its temporal characteristics. Each compo- 
nent of the report is allowed to make a very weak 
contribution, and then the interactions between 
these contributions construct a larger, and more 
subtle, set, of conclusions. The fact that most 
sentences report singleton sets of events arises, in 
the absence of information to the contrary, by a 
process of implicature, though the adverb once is 
available to reinforce this conclusion if necessary. 
i have only dealt with a small subset of the rele- 
vant phenomena here. It seemed better to use the 
space available to explore a small number of cases 
in some detail than to cover a wider range without 
being convincing about any particular case. Simi- 
lar analyses of other aspects and other aktionsarts 
are also easy to devise. Inventing analyses that 
cover specific phenomena is fairly easy. The diffi- 
cult part is ensuring that all your analyses work 
at the same time and without introducing large 
nuinbers of spurious readings. 
It is important for my claim to have preserved 
compositionality that all the analyses in this paper 
have been obtained on the basis of the interpreta- 
tions of the lexical items that appear in them and 
the semantics of the rules of combination, using a 
version of the system described in (Ramsay, 1992; 
Ramsay and Sch/~ler, 1995). 
References 
Asher, N. (1992). A default, truth conditional 
semantics for the progressive. Ling"aistics arm 
Philosophy, 15:463 508. 
Barwise, .l. and Perry, J. (11983). Situatiorts and 
Attitudes. I\]radford Books, Cambridge, MA. 
Carnap, R. (1936). Testability and meaning. Phi- 
losophy of Science, 3:401 467. 
Cruse, D. A. (1986). Lcxieal Semantics. Cam- 
bridge University Press, Cambridge. 
Dowry, D. R. (1988). Type raising, flmctional 
composition and non-constituent conjunction. 
in R.T. ()ehrle, E. B. and Wheeler, D., ed- 
itors, Catcgorial Gram'mars and Natural Lan- 
guage Structures, pages 153 198, l)ordrecht. 
Kluwer Aca(lemic Press. 
Gazdar, G. (1979). PTugrnaties: l'mplicature, Pre- 
s'apposition and Logical Form. Academic Press, 
New York. 
Mo~ns, M. and Steedman, M. (1988). Tempo- 
ral ontology and temt)oral reDrence. Compu- 
tational Linguistics, 14(2):15 28. 
Quine, W. V. O. (1960). Word and Object. MIT 
Press, Cambridge, MA. 
Ramsay, A. M. (1992). Bare plural NPs and ha- 
bitual VPs. In COLING-92, pages 226 231, 
Nantes. 
Ramsay, A. M. and Schgler, R. (11995). Case and 
word order in English and German. In Re- 
cent Advances in Natural Language Processing, 
Tzigov Chark. 
Reichenbach, H. (1956). The Direction of Time. 
University of California Press, Berkeley. 
Smith, C. S. (1991). The Parameter of Aspect. 
Kluwer Academic Publishers, Dordreeht. 
Turner, R. (1987). A theory of properties. Jour- 
nal of Symbolic Logic, 52(2):455 472. 
894 
