F()C, I )N "()NI_X" ANI) "N()T" 
Allan Ibmlsa.y 
I)el)l. o1! (2onll)Ul;(+r Sci('nc(:, \[)uivcrsit;v (<,ollcgc I)ul)tin, 
al lan(~monkey, ucd. ie 
lldficld, l)ubliu 4, Ir(',hmd 
I(rif'l,;.;~, \[1993j Ires suggesl,exl t, ha,t focus 
should be seen a.s a m~,a,nn of provid- 
iitg tna.t;eria.\] I+(w a. rang,:! of senm.nti,:; 
a.nd pra.gtna.l, ic \['uuwtit)Jts to v,:ork on+ 
ra.t, her than a,s a. Slmcili(: senla, nl, ic: or 
F,r;t,gtna,tic. finncl.i<m it, sdl'. 'l'he ctnrrent 
i):-~F, er describes .an iulplemeul;ationl o\[' 
t, his genera.I idea,, :-t.ud a,l)l>lies it 1.o t.he 
inl,,rF, ret.a.t.ic, n ()\[' onlff a.ud ?m:. 
1 Background 
(\]onsidev the \['ollowing sentences: 
(1) / only borrow(:d yo+tr ca?" 
(2) / o,d!l bor~v,'cd 9otn' c.r 
(4) I only bovr:owcd flour ,"av 
All of t,h<'m ent.a.il l, he sa.me h,a.sic m(,ssa.ge, na.mely 
thai: \[ b(~rr,r+wc.d yottr Ca.l'. It+ ad(litiou Ic, the l>a- 
sic, message, however, 1.hey a.lso ca.rry iuroHua.tion 
a.l)oul, what \[ didn't do. (1) sa,ys tha,t l(li(\]nh, })or 
row any of your el, her I)OSSessi(ms, (2) says thM; I 
didn'l; borrow a.nyoim else's car, and (3) says tit;d. 
I didn't do :.;llyt;hing else to your car+ II seems 
;*IS Ihollg)l t;)le wor(I oTzl.q a.rld the focus tnarl.:er 
(indi,:'.u.l.(,d here ILy t~rt,d(,rliuliag the str.ess,:u:l word) 
coral)the Lo a.dd a,n ext.ra IllessaK(~ a.I)oul, what 1 
did n't: do. 
A simila, r l)henolnetum appea.rs Co be takil~g 
place in t, he next. sol; or selllellC(!S: 
(,l) I did?//, .','Zeal f/o+:r car 
(5) / di&*'l sl, cal //otzr (:.r 
(6) / didn'l s/ca/ yo::: + car 
I!;ach of lhese sa.ys t, ha,t l didn't steal your c;u:, 
but, again they ea,ch carry some extra, t'llO, SS~,l,~,r(% 
('1) says 1.ha.t 1 did st.,...al sonlething which I)el,.:,ngs 
t:o you. (5) says tha.I; I sl.,:>le s(.n,:fl:,ody's cat:, I)ul. 
ltot yours, ~+ll(I (6) sa,ys that; I did do soJnethilJg 
to your ca.r (l l)rol>a.bly t)orrowed it, though that; 
is not, m+t,a.iled l)y (6)). 
i(,'ifka, \[1993\] a,t'g,ies I, hat, in (I 3)and (4 6), 
and inl a. number of' el.her situations a.s we\]l, the 
I'c, cus marker I is tts,:;,:l t.o exl;ras;t pa.rt; or the iu 
I, erprct;;-t,l:ion. ()pera.txJrs like +mly gum so-called 
+'\['ocu ssed tm~\[a.l;ion" 1,hen cc, m biule l;\]l('., ex I.ra.cl.ed 
element, rff l;he inll:(!rprcd,:-l.l.ion with wha, I, was left: 
Imhind to ca.try <:omplex nnessage+s of t.he kind dis+ 
Ctlsse(\] ;%l)ove, 
The ctlrrellt pa.per shows how to impletm;nl. 
this geuera.l notion, wil, hout following l'(ri\[k+Cs 
a, na,lysis in ,:let;a.Jl. The crucial l>oitll; is 1,he F,r(> 
vision or some wa.y of st;orinp; the extra:::l.ed p;trt 
<~\[' the interpreta.tion arm makinp( it ~wa.ila.ble wheu 
required. The i td:erprcl.a.l.ioll of ,'ndff and focussed 
nee;a,I;ion is Ia.irly sl.ra.ight;forward, so long a.s the 
l.real, ll/ellt; o\[' I, he \['~)clissed item itself is coherettl.. 
2 Abstraction and Focus 
The g,~mera.I aim of ISis l)a.l~er is to show h()w t,o 
use locus I;o decotnpose the ii~l, erpreL:.~t, ioJl of a. 
phra, se iul;o two) pa..rl, s, where, oue pa, rt is I, he iit- 
tez'prel;.;:d, ion ,.ff I, he \['o,::t+ssed item a,n(l the or;her is 
sotn,:; o1:,\],::,,::1; with which this ca.n comlJine. Sup-- 
pose, \['or exa.ml>le, we t.h(+ughl, l.ha, t; the VP ./., a 
pea.oh shoutd Im iul:;rF, reted ~ts: 
A47v(.+;,,,z(Y) A z:,a,.(Y, (:.~) 
Ap,":,s/(Y) A affe?zl.(Y, +4) 
AJX(v,~.,,/,( X ) n .b.#:::l(Y, x))) 
In other words, this VP is au M>str~tction over 
(Welll, S whel'() SOlllel)of\[y ,Fbl, e a \[)ea,(;h. :l'hen we 
w,mhl w~-/,tl \[ the two objed, s corresponding to 1,he 
in terpl'eLal, iolt o\[ a/,c a l)Cac/z to be somet h ing \]ike: 
AZi:ach (Z) 
l'l'h(! Lcrl\[l loO(',US h~l,,'4 b(!(!ii used J:tt a, w\[(le v~u'i(,.ly (;\[ 
WayS. In tJl(: i)rescnt, p;tpel: I simply use. it, t,o d¢m+)l,e the 
\[)a,l:l~(s) Of a.II II|.I.CI'II,11C(~ \[C) which attention is drawn I)3' stress 
Ill ~l,l:kcl'S. 
881 
a, nd 
~p~A_-\]Y(,+,,,~,,,z(Y) A tsm<,(Y, cot) 
Apa.vt (Y) A ogc,,l( Y, A) 
A~X P.X A ob.~ct(Y, N))) 
llere w0 ha, ve extra,cto(I th(: (hmotatiotl of peach a,s 
tim i)rol)erty of heing a I)each, and (:onv(u't(~d tho 
int(u'l)reta,tion o\[ t.h(> VI ) to an al)sl,:ra,(:l;iOli which 
will (:onil)hm a, pl)rol)ria,tely with this t)rol)el:ty to 
r(;l)roduce tile orJgit:ia,1 ilito.riJr(%i~tioll 2 
~;llere 0.Is(; do we see it. p\[lenon~enon o\[ this 
kind? CoJisi(ler l, he \[ol\[owing 17)hra,ses: 
(7) lhe man who stole your bike 
(8) the man a,,h,o l wanlM ;~jo'ts to meal 
l:n (7) Irh(} i)roperl;y o\[' being it illa,ll combhlos wil, h 
l,h(~ l)IX)liorl;y o\[' boing SOllIOOtle who sto\[o .yoII 1' })i ke 
to (:onst ru(% ~ context ually .fi n i m a\] u niq uo (: har- 
a,(:lerisai;ion of i;hc' I'Ol(wa;iil; individ iial, an(t silni- 
\]a, rly in (S). 'Po a, chi(;v(~ this> wc need (o hit;orpl:ot 
tl:to rc\]~tive pl;O:llOUllS in the two rcla.tivo clauses a,s 
leaving a holo ill th(; int(u'iireta,tlon ()\[' claltso and 
ilion ++bstrax:l;hlg with rosl)ocl0 to that hoh~. This 
is clea, r rot (8), b.t it, also holds for (7) ir we w+.., 
to int(;rp:r('A; a, S(~l\]l,oli('(~ lik(; a man ,</,olc o bide a,s 
A\]Z(,nan( Z) A age'nl( Y, Z )) 
A~X(bike(X) A obj,sct,(Y, X))) 
wh(we 1,ho qua,ntifier introduced I)y tho su\[)ject 
does uol; in fa,(:l; ha.v(', maximal SCOl)e (an axlaly- 
sis l ha,re argued for e\]sewhere \[l{~uilsa,y 1992a,\]). 
The t rc, i~t men t o1' (8) cleady req ui resm u ch t he 
samo me(:hanism as we will require i\[' w(:, want to 
d(>a,I with f()(:us ~'ts outline(I above, a.ud tills ma.y 
or may not a, lso hold rot (7). Any serious NI,\] ) 
systom will inelu(l(; some w~y of (h~a, lii~g with tlio 
iut(u't)retation oi' ca,sos like (8), a,nd almost a, ny 
such :mech<~nism should I)e Ol)Cn 1o a,da, t)tation to 
dea,\] with focus a,long l;he suggesl, e(l lines. ()no 
811c\[i a,l)l)roa,(:\]i is outline(t bolow. 
2you eali/lot \[reciy mix A-(:a,l(:ulus and the trui;h \[un(> 
tionM conne(:tivcs o\[ pr(xlical,e cal(:uhls as wc li~vc }icr(~ 
wil, hout rlll/llillg inl,o I,\[ic pa,radoxes o\[sclf-rcl'or(uicc \[{/is 
sc,.l/'s p;n'adox, tim \],i;tr, and so on. Tim nolal.ion used in 
this paper looks, for lJie s;tkc of \[a.miliarity, llke a. comb\[ 
n,~tion of X-calculus ;tnd pl:edlcate c~dcuhls, but is hi \['act 
t lurncr s\[1987\] grounded in the l:evlsion4)~sed semantics o .... 
Pll.O Pl",ll.ri'Y '\['H\['\]()I\[Y. 
3 Quantlfica ran, Presupposi- 
tion, Abstraction and Focus 
W(t (;x\])(;cl; to inl;(wpret rcla,tJve (l~u,qes (un(:onlro- 
versbdly) a,,(I phra,ses with f'ocu,<~,sed (:onstituonts 
(more co:nt rov(',rsia,lly) a,s a, bstra,ctions over the in- 
1;erprel;a,tions o\[' simple senl,(;n(:es. \]n ord(',r to cot> 
struct intc'rl)rotations or the ldn(Is of ol)je(t,~ wo 
~{ire int(;resl,(xl \]ll~ then, we l\]a,v(~ l,o sl;a, rl; I)y look- 
ing; a,~ sire pie sen fence's. The analyses l)l:ese:l:ll;ed in 
this paper sta,rl; \[rom th(; \['ollowing obs(wvath)ns, 
most of whMi a,re \[;-t,irly orthodox: 
Indefinite N1)s shouhl t)e viewed a,s a, way 
o1" introdu('ing itoms (or prefc>d)ly sets o1' 
items) inl, o t Ne disco,, rse. U niv('.rsa Ily (1" a,n- 
tifi(;d Nl)s sa,y theft a,Jl it(.ms o\[ the Sl)(~('ifh;(l 
type sa, Lis\[~y soul(; I)rOl)(wty. 
VPs should be viewed a,s a wa,y of introduc 
iug ove.ts or possibly sets of (;vents into tim 
(I isco. rse. 
If you construct i,~t(~ri)r(;l~tions by i)ara,- 
i)hrasil:lg N\], s(,ntenco, s Lnto a, \[ornia,\[ hu~ 
gun,g(, which e×tell(I,q i)redicat;(, c~dculu,~, you 
haw~ to r(~aliso tha,t the s(:OD(; o1' qua, ntifi(~rs 
iN you r para, i)h rases may not l)e (h'l,(;rmi.cd 
by shnl)l(; stru(;tu:ral prop(+rt;ies o\[' the sol, r(:(; 
text. 
\])efinite N1)s and other i)resuppositio.a,1 con 
structions pl~w.c (:onstra, ints on the discourso, 
so tlitLt a. s(mtelice conta,iHing the 1)hrtuse the 
man will I)e unh,torpretable i:N ('ontexts not 
conta,iNing a. imi(lu(,, illa.li (a. vorsion or this 
i)oint h~s I)(,.on maxle \[% a.motig otliers, liar 
wise ,~ p(,.,:y \[Jgsa\], l(~t,,p \[l~s.l\], (:;,.,),~- 
nendijl~ ,t~ Stokhof \[:1987\]). 
Thor(; a, rc intera, ctio.s of scope I)etweoll (leF 
inite NI)s a,n(I other t,yl)oS o\[ exl)ression: i~l 
I:,'ach ma~, kill,~ t/u: U~,ing hc love,s, the pre 
supposition:d ('o,struct I, he thing h~ /o~c,~ re- 
(luires I;hc (~xiston(t(~ o\[ a single ta, l:gd, oF air 
Fe(:tion per man. 
The' stand~u'd way to deal wil~h tho poto, ntial 
discropa, ncy botwoen where a, phraso al)pear,~ and 
the width o\[its scopo is by storinK quaxltiliors ON a, 
qumltific'r ~tack u.dl the o.tiro sontence ha,s bccn 
interl)r(',ted , ;t,n(\] 1,h(',n using (',xpJicil, in\[ornlal.ion 
a,I)out the priority o\[ va, rious quantitiers I,o sort 
thhlgs out \[Cooper 1983, \/cstr(~ 199\]\]. The work 
882 
r(,polfl;e,:l I\]Ol'(} tbllows l;his Lr(m,t;lllen{, I)llI, (,xi;e:n(ls 
il, I)y inl, rod il(;illg (luantifier-lik(~ onl;il;ies \[or dealin<n; 
w\]l, ii \[)resUl)l)osii;i()na, l il;olns ,<4tlCii zl,,s (l(:\[init,(~ Nl>s 
(,so(~ I{a.l\]i,sa..y \[19921>, 19!),l\] \[or a. \[(>rln;-i.\[ ;-i.('(;Olliil 
O\[ Sll(:\]l CONSTRAIN'?S 011 wh(~tll(H' a H(UIL(HI(;(~ iS 
Inea.liin'~\[)ll wit. h I'eSl)e(:l, 1,o :.1. ,sil, uat.i,:m). /\s a,n 
oxa.r11\[)Io, the S(Hit.('.i1('.(~ I,\]Z(: tv(mzalt .+~O/C a \[)71C(! iS 
iill;(,rl)rel.(,d as 
L\]A ,.'1 < now 
A ~t~ : (V(,' m~mbcr(C,/~) ~ woman(f:) 
A Inl :- I) 
i\]19 Vlq ~m'mb,Ye( l',', I)) -~ bik~( I;') 
A Inl =,, l 
A si~nple(/1, Al" rmnl(l") 
A Z!IP<( I", sZcal) 
A CU, l~l(l", B) 
A obj, cl(l", I))) 
This sa,ys thai: t, lm relal, ioushil) ,d'mpb' holds I>~, 
lw('en som('~ I:.a,,sl; in,sl, a, iil. A a, nd I;h(~ i)r,,\]l:.erl,y ~fl 
I)ein K a, ('erLaiN sort o\[ eveld,. \VhaJ; sort, d' ewulI,? 
One where a I)i\],:e is stolen ILy somenl~e (or ralher, 
where a, siuglel, on set o1' bikes is si:olen). Writ- 
in K som(qhing lib, ~,11 : (VC 'mcmb~ r( (.', H) -. 
.,o,,,,,.,.(c) A Inl - ~)~¢'. ',vhcre M," lll;/,y (t()l,|;,~/,il, 
occurrences o\[ /\], sa,.ys 1;ha, i, I'V holds I'~r' I, he cou 
texl;ua.lly unique individuM /, ~ which sa,l, isfies the 
i:esl, r'i(:l,i(m l.ha.t, l\] is a w(mlan (is :-/. singletou set, 
or \V(}I\]\[Cll). \[\[' 1,his resl,rhtioli fails lx) pick out, ~/ 
unique in(livi(ll:la.I the whole (,Xllressioli is III0~I,11 
il\]gless in the c, onlexl,. 
Mosl, of Lhis a.na,ly,qs is rairly orl;\[Iodox. The 
l;wo ma.in l)oini, s l;hal, might i'(,quir(~ S()ll10 (\]('\['OAI('(' 
axe llle a, na,lysis of asp(,d, in LerI11B oi";-I rda,tinrmllip 
I)el;ween t(,nl l)Ol'a,I ol)j('ct:s a n(I (v,,el\[t; I yllOS, W h ich 
is discussed ill \[R.a.msay 1,093\], mid l;l,e l;r(41.t, lu(mt, 
of (le\[inile I'(;\['01'(}\[\]C(} ill L(?I'\[IIR or ('(>nstrn,int, s on 
mea,ningrulnoss. Neii, h(;r or these is crucial t;o tim 
renm,ind(,r of i, he pa, i)er, bul if you (hm'l, like t, hem 
you will have I:() rel)la,(:e them wiI, h som('l, hiuK I>el, 
t;er, and w)n a,:re unlikely Lu find SOLn¢'lhi~i K whi('h 
is both I)elter ml(I simpler. 
The a, na,lysi,s a,I)ove wa,s obl;a,ined in a, fra,nle- 
worl,: where qua.nl:ifier SCOl)(t is (hfl;ermillo(t on the 
I)a,sis o\[ informa,Lion exl)licitly associa,l;(,(I wil, h a. 
\['()rill 0\[' COOPER. STOII,AGF, (\[(.',O0\[)Cl: \],()S;I\]), us- 
ing al)stra,cl, ioti Ol)era, I;ors of the \[orm AIV~JXI'V, 
,\I'I"V.\" 14/" ~I" ,\i.'V~X : IMV whic\]l can he ~q>l>NOd 
1\]() ~r fl)rmula, to hind its fr(~(~ varia,t)te:L Within lifts 
\['ranmwurk, it; is perR~(;I;ly o;+sy lo deal w;I, tl ca,ses 
like (S) hy allowing t\],~ r01aJ;iw! pronuull Lo ~uld 
I.h(' exl)rcssion ,\IV,\X W 1.o th(' qua.ui.ilier stor('., 
a,n nol;a,t, ed i;o speci(y 1,ha,i, t, his expression \]la,s 1\[la.x-- 
iula,\] sc(>pe. If Lhis expression is ~-Lp\[)lied to a, for 
mula, ,:;(ml;:.l,il\]in g a, \['rn<, occiirl'e\[ic(~ o\[' \],~ \]1; will re- 
t;lll'\]/ a,l/ a,I)sLra,cl;ion with respocI, to X exacl;ly 
what we wa, rll,. Tile re<luir(;metll, Lh:d; l;h\]s should 
ha, w', ma,xima.I SCOlm will ensure 1,ha, t, X is l;he la.si; 
I!l:ee va, riMfle in W. 
But, if we ca,n use I;his mechanism 1;o conslru,::l: 
4,1\] aJ)slA'a,(:l~ioll a,s |,}le itil;el;J:,l'(!l, aLiO\]l 0\[' a rela,tive 
Cla, LlSe~ WC' (;all also use il; Lo ('O\[iSLI:LI(',L ~t, tl a, bsl;l:a.c- 
Lion ,a~s the inl;ei:l)r(;t, aA, ion o\[' a phra,se con La,hfing 
a, fo<:ussed iI,enl. The only exi;ra, work we ha,ve I,u 
i)er\[brnl is I, lla,I; we ha,v(, 1;(, \[iHd som(,wh(,re to iml; 
l, he int,(,rl)r++l;a.l, init (ff Lh(' foclmsed itx, m Jtsel\['. 'l!(i 
<1(> this. all I,\[HI,I, is nee(led is aJl (!xi, ra, \['c,:M, ur(! focus 
ill the descriptions oflinguisl, ic hxmu~. The wilue 
o\[ focus is lhe \['ucussexl il,eln itsd\[, focus hehaw:s 
Nko a, GI',~(; Foot FEATURE.~ ill I;\[la~t a,I, I11(),S\[ t)llC 
(l~~ughl;el: of an il;(:l\[l (:a, ll ha,re a, Not\[ VaCttOUS value 
for focus, ~m(l Lha.l; i\[' an il, en) do(,s llame exaci, ly 
ono daughl;or with a, n(m-w1.(wolts vallie for (his 
fea.ture i,h(;n (,he it(m) will share this vn, IIm with 
I;\[IM, (la, ughi;er, focus is thus very like l, he sl,a.n.- 
dard I'eat;ure z\]_ash which is us(,d l'()t' (h;alillg with 
left; ext, ra, i)osil, ion it, is a, fool, \['ea, lwre whose 
va,luo is s(>mc item which is somel\]ow "'ouL (*i' t>().. 
,sii;ion". 
4 Applications of Focus 
Once we ha,w; t;his \]neclm, llism, we ca, n /IHO j| J,o 
construcl, inbu'l)rei,a,l,i<m,s of'sentences like( ) (6). 
(lonsider, \['or insL;mco, the exa.nH)le: 
(9) I only bm'rolvcd a car 
,ml;q(A4 cc~r(A ), 
AI; -J(/ G' < 'now 
A ~11) V/',' ~n~mbcr(l'/, I)) ~ ILl,; 
A I Z)l = l 
A I,l" : (V(;~mm~b~!','((,', l") 
• > ,sp,:ak<v((i) 
A IN-~ i) 
.~rm, pld(;, l~))) 
where /q is Allcv(~d,(ll) A typc(ll,borrou:) A 
a,g(!'n.t(H, I,')A ob.jt~cl,(ll, I))(this has I)een ('.x- 
Ir'ax;ted from tim di,~pla,yed \['ornluh~. to get, it in- 
sido Lhe a, va,ilable M>a.ce il, is in \[a, ct l)a, rI, ¢~\[' t,h&e 
form ulay. 
This says Lha, L 1,he rela.ti~mshil) on.l 9 holds l)(;l,w(~en 
the l>\['~fl)erty o\[' l)ehb'; a, cam a.i\]d ,some oLlmr oh joel. 
This is Nile ;~s fax ~s il Koes, buL iL isH'l; worth 
883 
very much unless we spell out the conditions un 
der which this rel~tti(mship holds. The tbllowing 
mea.ning postub~te does just that: 
VPVQ(o,dy(l', (2) ~ Q.t' A (VP'(@I" -+ P' = P))) 
In other words, i\[' only(P, Q) holds then t' s~l;is:lies 
(2 and nothing else (It)ks. In ~he present ('a.se, the 
first of th(;se consequences lnea.ns that I did indeed 
bo\]H'()w & C&l': 
:JC C < ~ww 
A ~z) w,: ,,,,(~,.,,z)~,,(zs', p) -, .,,,.(zc) A IZ)l ~ 
A ~,F : (VG morn.bet((;, t") ~ speaker(C) 
A Iv'l = ,) 
simple(C, K ) ) ) 
where K = A Hc'v(nt(II) A Z'ype(ll,borrow) A 
,:~,,,~(//, F) A o~,.jec~( n, z)))) h,s ~,,~,~.i, I,ee~, e> 
tra.cted to sa.ve spa.ce. This was ol)ta.ined from the 
mea.ning :l)ostul~te I)y substituting kAea'r(A)) for 
, ~,.,,d ,,si~g (,X,~,'~,'(~4)).~': =_ ,:o,,'(lC). 
The second eonsequen('e of the MI ) for mdy s~.ys 
thaJ, there is no other (;~ttegory o\[ item which sat 
islies the ~bstra.(:tiou tha.t the only thing \] bor- 
I'owed w;-l,s a, Cai'. 
\[\[ we put the Ibcus somewhere else, we get 
anot her hlterl)retation: 
(10) I only borrowed a car 
o~d~(),A AB M/ event(C)A /.ypc(C, loom.owl 
A ILkl)(agent((;, I))) 
A A.k\[';(objec*(C, \[,:)), 
ki,' ~(;' G < now 
A qlY gI member(l, \[I) -~ car(l) 
A IHI -- :t 
A ~,1 :(VKmcmbcr(k-, ,1) 
-~ speaker(K) 
A IJl-~) .,im#4c;, K)) 
whe,'e *,': (*,:M~(~,.H)).AM(M.,J))) 
This s~ws theft only holds between ~ descrip- 
tion el'the type of event C where somebody h' bor- 
rows something A, ~tnd ~u-, a.bsiraction over situa.- 
tions in which I did something 1o some (-a.r. Then 
the first consequence of o/.dy sa.ys tha.t what I (lid 
to this c~u' was 1 l)orrowed it: substituting the de- 
scril)tion of the event type tbr the abstra.cted va.ri- 
~l)l(,. I,' produces ((,\A,\ll,\Cevent(C) A tvpe(C', 
borrow) A I:}.A l)(age'nJ((;, \]))) A A.~\ It(object((:', 
or ,~i'm, plc, a,nd this i:educes to ACe'venl(C) A 
type(C, borrow) A a, gcnt(C, J)) k object(C, II), 
which is wha.t we want. The second sa.ys tha.t I 
didn't do mtything else to it. 
Much the s~/.ltle will happen with 
(\] \]) I didn't ,s'teal it 
A lLAI)(agent(C, I))) 
A A,AJC(obi~ct((:', l,:)), 
A t" _~(,' (7 < now 
At, l\] :(VI m, ember(l, ll) ~ 7~euter(l) 
/, IHI = 1) i.,l :(V/i" w.cmbcr( K, J) 
--~ specl~:cr( k') 
A FI = I) 
,,,i,~/.e.( (;, l~') ) 
whe,'e t~" =/,'.,\/;(~./~).A~'VI(A4.a))) 
llere we ha.ve a, 2-pla.(e rel;-/.tiol\] /'tOt, wh'lch \]8 
ba.cked tip by the following M P: 
w'vo(,,o~(P, O) -, (~0./' A ~ ~"((,8. l,'))) 
This sa.ys t hat this Ibrm of negation holds between 
/ ) ~;LI/(I Q if Q does not hold of P, but does hold 
lbr so:me other entity P'. In the present case, this 
\[l:le.a.ns tha.t I did do seine.thing; to it (wha.tever "il" 
is), I)ut what I did wa.s not stea.ling. 
This contrasts with simple :negation, with :no 
fb(:ussed item, a.s ix,: 
(12) l didn't steal it 
~(3A A < now 
A td.¢ : (VC rnc~nber((7, B)-~ neuter((;') 
A Ilq : ,) 
tl) :(VH member(f;, 1)) 
-+ spe,/,:er( l';) 
A fl)l--t) 
where K -- ,\l,'cvent(l/) A type(la,.steol) A 
,,.g~,,,~( ,,; D) A obj~,.Z( /,; ,))) 
This shnply says Lh~tt it is uot the c~se tha.t 
there is a 1)a.st stea.ling event inw~lvh~g me a.nd 
it. The choice I)etween the two is forced by l, ho 
presence or ~bsence of" a foeusse(I item. 
As a finaJ exa.mpte, consider a. sentence which 
(;outa.ins ~ focussed item I)tlt lie oper~or for using 
it up: 
(13) A man ate i/ 
The a.nalysis of this is a.n a.bstraction over Idnds 
of indivM ua.ls who ate it 
884 
A ~(; VI) mo,~n6~r(I),C) . A.\]) 
A I(:1-- l 
A t, lU : (VI" ~m!mbcr(I",/d) 
~e',Zer(I") 
A I/':1-= :l) 
,~implc( H, h') 
with k' :: A(;cvc'n,t(i(; ) A I, ypc((;,~:at) A 
.,<s~,,,z((;,(;') A <,b.i~..<:#.((;,/,;)), ~,.n(i with the r,,,:,,,~ 
seL a,s l;he descril)l;ion (inchl(ting l,\[l('~ s(,nia, iiLic 
a, ua,l.ysis) of I, he \[()cussed i:llira,s(! ml'U~. This is ,i,st, 
the ki lid ofol)ject r(;( I Lli re(I \['or a, (Jiscou I;S(; Ol)(H'~¢t(ir 
,Sli('.h ;,I,s (tolll;l';-/,si; 01; (4a, i)ora,tiou exa,ci, ly whM/ 
Sli(-li ()l)0ra,l, or is a,i)t)rol)rial,e (lel)etids oii f;-i,(;LOl','4 
ii()l, visible iu (l'l) itsolf, bill, wha,1;ev(~r ii, i,:; ii; will 
require a, l)~ir of ii,i'~lllil(',iil,'40\[ i, his kind. 
5 Conclusions 
The discussion ~d)ove shows wlia~ ca.li I)(~ a(:hieved 
I)y 1,rea, l, ing focus a,s a ,~ynl, acl;i(; nia,rl,;er whi('}l 
llla, i,;es illFOrlll:g,l;\]Oll availa, l)le 1,o a variety of Ol)- 
er:q,l,Ol;,q. The 1\]/o(;ha, iliSJll \[I,)l' doing ttlis involves 
ilitrOdllCillg :<i, \['ooi, \['O~/,I, IIIX~ \[,O (u/,rry the \['O(:llss(xl 
il;eli/ a, rC~llnd, aild covlst;rucl, ing a,i)l)roliria, l,e ~tl) 
sl, ra,ctions l)y u,gin<K i, he si;a, nlJ;-l.r(I qu:aul, iiier stop. 
in<K nlecha, nisul which is required \[llr ol, h(;r t)lie- 
ilOlli()lia ;l l i .V w :46,. I)i\[\[hrenl, N I,I > ,s.7,sl,(!lihS will dea, i 
wil, h i;he S.ylil, a,x a, il(I ,Seliialil;ics or t)h0nonl(ui;-i, sucli 
a,s left; a, nd righl;+exLra>l)(isiLion in (\]ifl'or(uit wa,ys. 
\&/ha,l; \[ h;-wo :+~i'glmd is l, ha{ aln/o,sI; a, lly a,pl)ro~ch 
Lo l;hese phenonmna, (;;-/,li I)0 a,(h~jfl;od to deal wil, h 
\['ocus :-is w(~ll. '\['he exa,nil)les iu ,'-JectiOl\[ 4 siiowed 
hiiw VOll (t:rt, il COlllllill(~ tllese an:4,l,ys(~,~ o\[ \['oIHIS with 
~c va, riel;y o\[ ol)et'a,l;ors 1,o ('ouvey a, I';-i.lig(~ oF iill, ei'- 
\[)l'(H, al, ioll,S o\[ tim s{/\[l}(~ SO(lll(ql(:O ()\[ words. II, is 
ilill)()l'l,a,\]ll; 1,o reca,II ;-i,I, this I)oiuL |;hal: 1;he ilil,or. 
i)l;ela.iion la, nt~;u~<Ke lining used here i,', ,~i, hi<Khiy iu 
I;ensiona,I logic which imrinil, s qua,ntilica, t, lou oww 
a, rlfil;ra,ry kinds o\[ iudiv\]dua, I, including qua,uti\[i- 
calion ove:r I)roperl;i0s ,:l, lid proposil, ions. I h:+l, vo 
a,l'gu(xl elsowhere l;h:a,l; ,such a, f~-~ilp~lla,ge is r('quired 
I'()r a, wide va, rieLy of iihell(line~i;-i. 'l'he ilil, erl)l'el;:+/,- 
l, ion i,)l" locus is just :,'ln(~l;her exa,:Uil)ie. 
6 hnplementation 
All Lhe a,nalyses in this i);,i,l)(U ' wer(~ liro(llJ(X,d> and 
/\-i;0(lllC,(x\[ (a,n(I 1,Urlle(I iltlo I~'I'I,;XI), by a, v('r,sion 
of l, he sy,sl,eu:i (lescrilmd in \[l{,antsa,y i992~Q. 'l'liis 
con,sisi, s o\[' a, highly lexicM glu-i, tilllHtr wil, h a, ('(llil- 
I)osilional s(~lrlalil;\](',<-;~ I)a,i:~ed via a I)i-dir('ctionaJ 
hea.d-dr\]ven chm'l pa.rser. 1 b('fieve it is vh'tua\]ly 
iiupossil)le to do this Idnd of work without em- 
h¢)dyi\[Ig it; in a. win'king system. 5g)u simply can- 
not explore i;he consequences or doing something 
one wa.y ra.ther tha,n another, or of com l)iniag a.n 
a.nalysis of this wigh au a.na,lysis o\[' Lha.1,, unless 
a,ci:ivities such a,s (;oinl)Osil, ionaJ co:n,qtru(:tion and 
,subseClllCUll, )~ i'(Xiil(',t, iciii o\[ i:llt(3ri)reb:4{ioll,<-; is (Iol\](; 
\['or ,yOII by nlax;hine. 
\]{efei'ellces 
Ila, rwise ,1. <~.~, I)orry J. (I 983): ,S'it,ual, ion,v mzd it I,- 
~i/,udcs: II ra.xl liu'd \]looks, (Ja, iii b ridge, M A. 
(~,(Jol)0r IL. (1983): Qua'n,l'~ilicatio, and ,5'ynZaclic 
'l'h~ ory: LLeid el, I )o r(l rech t. 
(',rooneiidijk J. & ~l, okliof M. (\],()87): /)ym'lm/c 
IJrcdical, c I,ogic: 7'o'wa, rd,w a Composigional 
m~xt Non-rcprcscntational Discourse 73c<>ry, 
Te(;huica,l lix~l)orL , Universil,y of Amster(laJn. 
|{a, llli) I1. (1984): A '\['h0o:ry o\['Tl'u{\[i ~~nd ,r"Je\[i/:4,1i. 
1;\](: liepresenl, aJ, ion~ in Formal Mcihod.+ in th+~ 
5;V,'udy of L<m,9uage (eds. J. Groenendijl(, J. 
,la, n,ss(;n & M, S1;okho\[): IJ'oris l)ulilications, 
\])(u'(Irechl,: 27'7 372. 
t(ri\[l,:a, M. (I ,()93): I"ocil,% Pre,~il pposil;ion a,n(t l)y- 
na,iiliC Ititc~i'l)rel;~lJ, iOli, .Jo'al'l~(l,\[ of D'emanlic,s 
\]0. 
l~,a, uisa,y A.M. (1!)92:a,): lla, r(! \])lu:i:a,I Nl)s a,n(l Ila, 
l)itua.I Vl)s, ('OI, ING-92, Na,l\]tes: 226-231. 
Ranlsay A.M. (1992b): \])resuppositjolls a.ud 
W l I-clauses, ,/ourltal of ,~'~nzanlics 9:251 
186. 
ll.a.uisay A.M. (1{).();I): Asp+cl mul AkZim~,~arl 
I,ViU~,o~I Co(reims, slll)uiitt(,d I,o A C, \[, 9,1. 
l{;,i, iiisay A.M. (19!)4): Mea~ninKs a,s (~,oiisl, railll, S 
on Infl)rina, I;i(m SLa,lJes, in ()'ol~<~lraild,,% l,alz- 
(.lltCl.fJe alzd C'om, p'ul, ation (eds. C.,J. Rut)p, 
M.A. l{osiler & 11.1~. ,lohJison): A(;.~l,(lerni(; 
Press, \],Oil(ion: 249 2?(7. 
'l'urii(!r I~,. (1.987): A 'l:heory of I'roperl;ies, Jour- 
nal c U" ,6'ymbo/ic I,ogic 52(2): d,55 /172. 
Vesl, re I';. (1991): Ali Algorit, lml for Geuer- 
a,i, iug N()ii-r(!dunci,~uil, Qua,nl,iJier gco\])in~;s > 
I,'iJTh Co~@~',~n<:c of the Iqlu'opr'a'l~ (;'tm.pt~:r 
o/" l/~( Assoc. for Compulal, ioiz(d J, in.guisl, ics, 
Ilorlin: 25l 256. 
885 
