Anda di halaman 1dari 28


,,-' --.-... :. ~-. ",:, . ,'~ _.

"'_ c

1) _ i I " ~ Ii! J' "

, ",," '~ .' - - " ,"

) , (.; J I 1,;"'; 1 . i ,"- ~
- -----

t ., ' :.. ' --:: ',' .: ':;: .' ~

. . ......
I . (~ I I ,. , ' .'"

(.\ ',".-) , I :):'lr~i~


.I e. 1 ) ' ;: '; -.;; :~~

,, ' , . " ,-

Copy Editor
Institute of Mathematics
Po l ish Academy of Sciences

Gdansk Teachers' p~
~':" ~:-:'
Acknow ledg7lLents. The author is gra teful to Prof\:~sor Zo ria DCllkow-
ska. Professor ZdzisJaw Skupicll a nd Dalliel Davie:) fur th eir llclpful Part A: Phrases Used in Mathematica l Texts
cr iti cism . Tlul!1ks are also due to Adam ~rysior an d ~Ll!"cill .\daw:,ki
for suggesting several il1l I' rm'clllcnt s, and to Hcmyb Walas for her Ahstract and introciuction ... .. . . . .... . . . . ...... . . . . . .. ... .... . .. ... .. . 4
painstaking job of typesettin g the continllo usly varying manu script. Definition ........... . ....... .. ..... . ... . . . ...... . .. . .. . . . .. . .. . .. .. . . , G
Notatioll . . . . . . .. ... .. ... . . .... ..... ... . ... ...... .... .. .. ...... . . . ..... 7
P ro perty ... . . . . ................. ..... . . . ..... . . .. . ........ . . . ..... .. . . 8
Assumptio n , cond iti o n, convention . ... . . . ... ... ..... ... . . ..... . ..... . 10
Thcorem: general rem arks .. ... . . ... . .. . . .. .... . ............ .. .. . .. ... 1~
Th eo rcm : illtrociuctofy plfrase . .-:-: . : .. ~-:-......-.. ... .. ..... . .. ....... 1:1
Theorem : formulation . .. . . ..... . ... . ..... .. . ..... . . ... . ......... ..... 1:1
Proof: beginning . . .. . . .. .... .. . . .. ...... . ... .. . ... . . ... . . . ..... . . .. . . 11
Proof: argu m e nts ..... . ... . .. . ....... . .. . . . ... . ... . ... . ... . .. . .. . . . . . 15
Proof: consecutive steps ... . . . . ..... . . ... . .. . ...... ... . ... . . .. . . .. ... . J (j
Proo f: "it is sufTic ient to .. ... " .......... ... .. ... . .. ............. . . . .. . . 17
Proof: "it is eas ily seen t ha t ..... " .. . . . . ... ........ .. . . . ..... . .. ... .... 18
Proof: conc lusion and r emarks .... .. .. ... ..................... . ... . . . . 18
n.efcrenccs to the literatu'r e . ... .. ... . . . .. . . . . ... ... . . . . .. . .... .. . .. . . . 19
Publ ished by Gdallskie W ydawn ictwo Owiatowe Acknowledgments ................. .... . . . ... ... . . .. . ... . ... ..... .. ... 20
How t o s horten the paper . . ....... . .... ..... . . ..... .. . . . ... . . .. . ... .. 20
(Gcl a llsk Teachers' Press ) Editorial correspondence .... . . . ..... ....... .. . . ................... . .. 21
P.O . Box 59, 80-876 Gdallsk 52, Pol an d R eferee's report .. ... ... .. ... .. . . . .. . . . . ..... ..... .. .. ... . . .. .... . ..... 21

Cover design by Agnieszka Polak

Part B: Selected Problems of En glish Grammar
Typ eset by Henryka vValas
I ndefini te a rt ic le (a, an, - ) .. .. . .... . ... . . ... . . . . .. . ....... ......... :':1
P;illt.ccl ill Poland by Zaklady Graficzne w Gdallsku Dcfinit e artic le (t he) .. . . . .. . . ... . . . .. . . . . . . . .. . .... . .. ......... . .. :' I
Article ornissio ll .. . . . .. .. . . . .. . . _..... .. . . . .. ..... . .... ........ . ' ..... :>;)
Infinitive .... . . . .. .... .. ...... . ... . . . .... . . . . . . . .... . ... .. .. . . ...... . . 27
l ng-form .. .......... . . ... ..... . .. .... . . . ........ . . .. ... . ... . . . ... ... . 29
Passive vo ice ...... .. .. .. ......... . ... .. .... . ..... . .. .... . ....... .. .. . 31
Quantifiers . ..... . . .. . .. . ... .... ... . .. ....... . ... . .. .. . . .... . . .... .. .. 32
Numher , quantity, size .... .. . .. . .. ...... .. ... . . . .. . .... ... . . . .... . ... 34
How to ;cvoid repetition . . . .... ....... .. ... .... . .. .. . .. . . . . ... . . . ... . . 38
Copyright by GdaI1skie Wydawnictwo Oswiatowe, 1993 \ Vord order .. .... . . ........ ..... .. ... . .. ...... . . ... .. .. . ......... . . . . 40
All rig hts reserved . No part of this publication may be reproduced \Vhere to insert a comma ... .. . . .. ....... .. . . . .. . . . .. ... . . ... .. . . . . . . 44
in any form \vithout the prior p ermission of the publisher. Some typic;).] e rrors ........ .. .. ....... . .... . .. ... . .. . . . . . ... .. ..... . . 46
Ind ex . .. . .......... . . . ...... .. .... ...... .. . ...... . . .... . ...... . .. ... . 48
ISBN 83-85G94-02-1
PART A: PHRASES USED IN MATHEM AT ICA L TEXTS It is Hot o llr pll r pose to s tudy .....
No atte mpt h:\.S be 'n made h ere to develop .. .. .
ABSTRACT AND INTRODUCTION It is poss ible th a t ..... but we will not develop this point here.
A more comp le te th eory lllay be obtained by .....
\\"' IHove th at in so m e fami li es of compa cta thcre are no u niversa l e lelllen ts.
this t op ic exceeds th e scope of this paper.
It is a lso s hown that .....
SO lll e r levant coun terexamples are indicatcd .
I ..
, we w tll not lise tlus fa ct III
any esse ntIal way.

It is o f inte rest to know whethcr ... .. We wis u to invest igate .....

.) I idea is to. apply
T lIe b aS .lc ( malIl . d.... .. .
geometnc mgre !Cut IS .. .. .
We are inter ested in finding .. ... Our purpose is to .. ...
Th e crucial fact is tbat the norm sa tisfies .... .
It is natural to try to relate .. .. . to ... .. Our proof invo lves looking at .....
T his wo rk was intended as an attcmp t t o motivate (at moti vat ing) b:\.Sed on the concept of .. ...
The aim of lIlis pap er is to bring together two areas in wh ich .. .. . Th e proof is s imilar in spirit to .....
review some of the st.andard I
adapted from .....
This idea goes back at least a5 far as [7)
facts on .....
have compiled some basic facts .. We em phas ize that .. ...
summari ze without proofs the Il is worth p oint ing out that .... .
rele vant material OIl . . .. . The important point to note here is the forlll of .. .. .
g ive a bri ef expositi on of ... . .
The adva ntage of using ..... lies in tlw fact that .... .
bri efl y sketch .....
Th e est imate we obtain in the course of proof sce IHS to be of ind epenrient
se t up nota tion and terminolo;;y.
inte rest.
d isc uss (study/trea t /e' amillc)
til e C; I S (~ ..... Our theorem provides a natural and intrinsic cil;u acter iza t.i on of.
Our proof m akes no appeal to .. ...
I I !i"1 l ilill :1 il ll.rocl\lcc the notion of .... .
Our view point sheds some new ligbt on .....
II III" III i I d ''' 'I I ill II II " dl'\,r lop tlt e theo ry of ... ..
I NI I/ ,' 11.1 1111', 1 ' '1i11 wi ll look 1Il 0re closely at .. .. . Our example demonstrates rather strikingly that ... ..
/ ",, 1111 11 1 lI'i li 1)(' c() Ilce m ed with .. .. . The choice of ... .. seems to be the bes t adapt ed to our theory.
JlI (lcel'li wit h the s tudy of .... . The probl em is t hat .... .
illlli (":1.lC' Ir w tilC'se techniq ues The m a in difficulty in carrying out this construction is th a t .....
III ;\Y b llsed to ... . ' In this case the method of ..... brea ks down .
ex te lId til e res ults of ..... to ... .. This class is not well adapted to .... .
de rive a lI int eres tin g formub for
Poi ntw ise converge nce presen ts a more d elica te p robleI1l .
it is s how n th at .. .. .
some of the recent resu lts arc Tbe res ults of this paper were announced witbout proofs in [8).
rev iewed in a more genera l sett ing. The detai led proofs will appear ill [8) (elsewhere/in a forth coming
. some applications arc indica ted . publication) .
our m a in resu lts arc s tated and proved. For the proofs we refer the reader to [G] .
con tai ns a brief su mmary (a discuss ion ) of .... . It is to be expected that .. .. .
deals witlL (discusscs) _the casc ~~ . .. . ____ __One may conj ec ture that .... .
is intended to m oti\ate our in vestigation of ..... One m ay ask wh etber t his is st ill true if
Srct ion 4 is d c\otcd to the study o f .... . One ques tion still unanswered is whether ... . .
provides a d ctailed exposit ion of .. .. . The affirmat ive solution would allow one to .... .
es tablis hes the relation between ... .. It would be d es ira ble to .... . but we haw not bee n able 1. 0 do
presen ts sOllie prelim inari es . thi s.
\\' . 11 1t o uch only a few aspects of the t heory. These results are far from being conclusive .
e WI r estrict our atten tion (the discussion/oursehes) to ..... This question is at present fa r from bcill~olved.
,.( .:;Y

Om method h<ls the disadv<l:ltagc of not I)f'ill;!; i:!trillsir . It is immateri,,-I which J\J we choose to denne F as 101lg as 1\1 contains x.
The solutio!l bll.; short of pr,.'\idil:::; :1:; l~xph !t fdrlllilh . This product is indepelldent of which Ill ell lber of !J we choose to define it.
\ rhat is still he!.;i :1::; i.-; :ell ex plici 1 ,j, ' ~nirl i,ll! of . It is Proposit ion 8 t h<lt makes this denllit.ion ,,- lIowable.
:\ s for pr(,I"I'fju isiks, the rC':lIler is l'Xr<'c'kd 1\1 11t ~ fal~lili:\i' '-':it;; .....
The first t\\"o ch<lpters or" ... .. CCll1stiwtp. ~ll:;:cil"lit pr,. p;'i<ltio!l .
Iwllh th e cla.sslcal one for .. ... is ... ..
Our definition agrees w~th the olle givell in [7] if II

:\0 prelilllill:ery kllo',\"led~,~ of ..... is r<' <i'.lirl'l;. this coincides with our previously introduced
To facilitate ;tcccss to tile' ilJd i':idlliti t')pic~. the ch;lprcl., iUI.: rClldel"t!d as Note that terminology if J{ is convex.
5elf-( 1)Ilt:liiled as jl<ls., i\'[(". Ithis is in agreement with [7] for .....
For the (tl;:\'e lli ellce of tLe r,'ad,' r we l"l'pe<lt the relc\':l!lt ll1aterial from [7]
witl 'O \lt proofs , th\ls lll:tking our cxpo:,iJiull self-colltil.incd . NOTATION

DEFINITION _ We \,'ill denote by Z I

Let us denote by Z - tJleset ..... Write (Let/Set) 1 = .....
.-\ ~e t 5 is dcn.,, >if ..... Let Z denote [Not: "Denote 1= ..... "]
A set 5 is c;1JJ,~d (said to be) ric71se if .. ... The closure of A will be denoted by clA.
\\'e c;t ll it set dense if ..... We wi ll use the symbol (Jetter) k to dellote .....
We call In tire product mcasU1C. [Note: The term defined appears last.] We write H for t.he value of .. ...
The function 1 is given (defined) by 1= ..... We will write the negation of]J as -'p.
Let 1 be given (defined) by 1 = .. ... The notation aRb means that .....
We define T to be .4B + CD. Such cycles are called hOlllologous (written e ~ e').
requiring 1 to be const;1llt on .. .. . Here
"[hi;; lIlap is defined by
the requirement that 1 be constant on .. ...
[Note the infinitive.]
Here and subsefj uelJ tly,
Throughout the proof, h
denotes I I
for t Ie map .....
imposing the following condition: ..... In the sefjuel,
From now on,
The k7!t;lh uf <l sequ ence is, by definition, the 11l1ll1ber of .. ...
The length of T, de110ted by I(T), is defined to be .... . 'vVe [ollow the not;ttioll of [8] (used in [8]).
D\' tl~e lcngth of T we mean ..... Our notation differs (is slightly different) from th at of [8].
Let uS introdu ce the temporary notation F J for 919 .
Defi11e (Let/Set) E = L1 , W
is ... ..
1Jere I , ,t f -
we la\e se
I - ... ..
, f being the sol u lion of ... ..
Wit h tIl(! 11ot;1tio11 J = ..... ,
With this notation,
Iwe have .....
with f satisfying ..... In the notation of [8, Ch . 7]
If f is real, it is cus!ol1l:try to write ..... rather thil.1l .....
\\'(' wi il 'ol,ider I the behaviour of the family 9 defined as follows .
. ..,.~ the height of!J (to be defined later) and ..... For simplicity of llot:ttion, write f instead of .....
To 1l1(~ ;1~llr " the growth of!J we make the following definition.
To (simplify/shorten) notation, we use the same letter f for .... .
I3y ,,-buse of not.ation, continue to write f for .. .. .
we shall call
In Ihis \\';\:; we ol)tain what :vill be referred to as the P-s?Jstcm.
I For abbreviation, let 1 stand for .....
We abbreviate Fallb to b'.
15 known as
We denote it briefly by F . [Not: "shortly"]
We write it F [or short (for brevity).
S I rl:t~ norm of f is well defined. The Radon - Nikodym property (RNP for short) implies that .....
Illce ..... , the definitiOll of the llorm is unambiguous (makes sense).
\Ve will write it simply x when llO confusion can arise.
t-..(" 7
II \~ ill (,, \II Se no confusion if we us e th e sam e lett er to designate we have jus t d efined
we wish to study (we used in Chapter 7)
" "" '"Ille r o f A and its rest.riction to [\' ,
\ \'1' ', 11.111 WI i II' the above exp ressio n as
II" , "I I/IVII I'x press io ll m<l.)i be written as
It = ..... tbe (an ) e lemen l to b e defined later [= which wi ll be defined]
in questio n
under study (consideration)
\\ '1' 1 ,1 11 IV I iii' (,1) ill tbe form

' I II" (:II 'I'k illl li l'l'S !<tilel co mponents of sections of E ,

..... , tbe constant C being iudependent of .... . [= where C is .. .. ,]
..... , th e s upr emum b eing taken over a ll cubes .....
1' lllIt il'IIIIIII OloIl,Y' ... .. , the limit being taken in L,
'1'1 ", 1',X PII ':,:, io ll ill it a li cs (ill italic type), in large type, in bo ld print; is so chosen th a t .. .. ,
III 11 ,'" '111.11" '01':1 ( ) ( rOlillcl brackets) , . is to be chosen later.
iii 1/1.\111 ' 1:1 \ I ( hCjU I\ r e bra ckets), is a su itable constant.
111111 ,11 '1':, ( ( r lilly I>I'al' kets ), in angular brackets (); .... " where C is a conveniently chosen element of .....
WII hili L111' IllIllIl ~i; ', II :; invoh'es tbe d e rivatives of ,.. ..
( ', qlil.iI 11'1.11'1 '1 11I1pI ' I r ; 'iI' Idt ers i= s mail let. tcrs = lower G\SI! It: t.ters ;
ranges over all subsets of .... ,
may be made arbitrari ly small by .... ,
(;oLlli r (;1'"11;111) lei LI' I ~ ; sc ripl (ca ll i~ra rhic) let.ters F;
s l' cri :ti I{ Ulil l lIl l<'t lle l S If'!
have (share) many of the properties of .....
Dot " prime " as t. eris k = st. ;ll' ., tilde - , bar - [over a symbo l], ha t - , have still better smoothness prop:-ties.
ve rti cal s troke (vertica l bar) I, s l;Ls h (diagona l s troke/slalll)/, lack (fail to have) the smoothness properties of .....
da.5h - , s harp # still have norm 1.
Dotted line . , , , .. , das hed lin e _____ , wavy line _ _~ not m erely symmetric but actually se lf-adjoint.
not necessarily monotone.
both symmetric an d positive- definite.
PROPERTY not cont inuons, nor do they satisfy (2).
The operators Ai
[Not e the inverse word order after "nor" ,]
such that (w it h the proper ty that) are ne ither symmetric nor positive-d efinite.
[Not : "such an element that " ] only nonnegative rather than strictly
with the follow ing properties: ..... positive, a.s on e may have expec ted,
sat isfy ing LJ = .... , allY self-adjoi nt operators, possibly even
wit h N J = 1 (with coordin ates J:, 1j , ;::) nnbounded .
of norm 1 (of the form .... ,) still (no longer) self-adjoint..
whose norm is ..... not too far from being se lf-adjoint .
all of whose subsets are .....
by means of whic h 9 call Iw COIIlPIlt.I~ d preceding theorem
for whi ch this is true indicated se t [B ut adjectival clauses with
tbe (an) element at wh ich 9 bas a local ma.ximum Th e above-mentioned group prepositions come after a nou n ,
descr iber! by th e eq uations .. ... r es ulting region e ,g. "llle group de fin ed in Sect ion 1" .]
g;ivclI by LJ = ... .. req uired (des ired) element
d epe ndin g olIly on ..... (ind ependent of .. ... )
13otli--X- allCi Yarc finite :- - - - - - - - -
1I0 t in A
so small that (sm a ll eno ugh that) ..... Neit her X nor Y is fi n ite.
as above (as in the prev io us theorelll) X and Yare countable, but neithe r is finite,
occurring in the cone condition Neither of the m is finit e. [No te: "Neither" refers to two a lt ernat ives,]
[Note the do ub le "r" .] None of the fUllctions Fi is finite,
gua ranteed by the assumption .... ,
so obtained X is not finit e ; nor (n ei th er) is Y.

8 9
x is not finite, nor is Y countable. [Note the inversioll .] This involves no loss of generality.

X is empty I;, ,()iJuti~ . ~" IS: !lot.

siuce otherwise .....
for .. ... [= because]
\Ve can certainly assume thilt .. ...
, for if not, we replace .. ...
X belongs to )' I : ~() cl(~l~S. Z .
I . uut Z aoes not. \ Indeed, .... . .

i\'eit he r the hypothesis nor the conclusion is affected if we replace .. ...

ASSUMPTlmJ, CONDITION, CONVENTION By cllOosing b = a we may actually assume tuat .... .
If f = 1, which we may assume, tben .. ...
\Ve will make (Ileed) the follo\\'ing assumptions: .....
For simplicity (convenience) we ignore the ciepeudence of F on g .
From IlOW on we make the assum p tion : .. ...
[Eg . ill IlotZltion]
The following assumptioll will be needed throughout the paper.
_ Our baSiC assllm[Jtioll i.'i..S1L~Jul lowillg-,- It is convenient to choose .....
Unless otherw ise stZltecl iUntil further ;;-otice) we ;L'is\l~le that ... -.. --~ - ~~ ~ - - \ \ 'e C;tLl assunre; by decreasing k if necessary, that .....
In the [,(;Illainder of tbis section we assume (require) g to be .... .
F meets 5 transversally, say ilt F(O) .
In order to get asymp t otic results , it is necess;try to put some restrictiolls There exists a minimal element , S;t)' 71, of F.
on f.
\Ve sh;tll m;tke two st;tDding assumptions on the maps under consider;ttioll.
G acts on H as a multiple (say n) of V .
For definiteness (To be spec ific ) , cOll sid er .....
It is required (;tssumed) th;tt .. .. .
The requirement on g is that .... . is not particularly rest r ictive.
is surprisingly mild .
is subject to the condition Lg = O. admits (rules out/excludes) elements of .. .. .
.. ... , where g satisfies the cond ition Lg = O. This coLld ition
I is merely required to be positive.
is essential to the proof.
cannot be weZlkened (relaxed/improved/om ittedj
the requirement that g be positive. d ropped).
[Note the infinitive.]
Let us orient M by
requiring g to be .. ... Tbe tbeorem is true if "open" is deleted from tbe hypotheses.
imposing the condition : .. ... The assumption .... . is superfluous (redunclant/unnecessar ily restrictive).
\Vc will no\v show how to dispense with the assumption o n .....
for (provided/whenever/only in case) P i= 1.
unless p = l. Our lemma does not involve allY assumptions about CUrva llir ' .
(4) bolds the condition (hypothesis) that .... . We bilve been working under the ass umption that .. ...
under the more general ;:;ss:1nlptioll that .... . Now suppose that this is no longer so.
some further restnctlOns on .. ...
To study the general case , take .. .. .
additional (weaker) assumptions.
For the general case, set .... .
satisfies (fails to satisfy) the assumptions of .....
ha.'i the desired (asserted) properties.
()io:ides the desired diffeomorphism .
The map f Irca
w ill be viewed (reg;trded/thought of) as a fUlll clO r .... .
lZll1g .... .
F still satisfies (need not satisfy) the requirement that ..... think of L as being constant.
meets this condition. From now on we reg~rd f as a map from .. ...
do~s not necess;trily have this property.
satIsfies all the other conditions for membersh ip of X.
I tZlC\t/y assume that .....

There is 110 loss of generality in assuming ..... It is und erstood that r i= l.

vVithout loss (restriction) of generality we can assume .... . We adopt (a dhere to ) the convention that 0/0=0.
f' 11
. an extension (" birly straightforward rephr ase Theorem 8 as follows.
gf'neralizatioll/a sharpened \'Crsion/ \ Ve have thus p rove d ... ..
\Ve can llOW state the analogue of .....
Summarizin g , we have ... ..
a rcfillement) of ..... 1 formulate our maill results.
IS a reforlllulation (restatement) of .....
We are thu s led to the following strengtheuillg of Theorem G: .... .
in terms of .... .
all aualogue of .... . The re mainder of this section will be (b'oted to the proof of .... .
analogous to .... . The coutinuity of A is established by our next theorem.
' I'hi:: tll l'on' lIl a partial converse of .... . The following result Illay be proved in much the same way as Theo rem G.
a ll ans wcr to a question raised by .....
Here arc some ele mentary properties of these concepts.
d ea ls with .....
(,IISllr s th e existence of ..... Let us mention two important consequences of the theorem.
('X P I' sscs th e equivalence of ..... \Ve begin with a general result on such operators.
p l'ov' icl cs a c rit e rion for .. ..
[Note : Sentences of tIle type "We now have the following lemnw.",
y i(: ld: ill[ I'mali o n about .. ".
carrying no inforlllation , can ill gelleral be cancelled.]
Illakcs it leg it imate to apply .....
Th e tit 'o rCIII s ta tes ( ; L"iS ' r l ,/s hows ) th a t .....
l to ug ld y (Loosel y ) spe:tkillg, th e fUl'lIllila says t hat ..... THEOREM: FORMULATION

\\,1 f IS
. opell, (3.7)'.Ill S t tl to saying that ... .. If ..... and if .. ... , then .....
. ' len ,\lll OUIl S to th e fact that ... .. provided III i= l.
Suppose that .. "' 1Then unless m = l.
Here is another way of stating (c): .... . Let M be ..... Assume that ..... ..... ,
with g a constant
Ano th e r way of stating (c) is to say: .... . 1Write .. ... satisfying .... .
II 1\ ('q uiva le nt formu\atiou of (c) is: .... . Furthermore (Moreover), .....
T I\( :o rc Ill S 2 and 3 may be sUlllmarized by saying that ..... In fact, ..... [= To be more precise]
}\ ssc r n (ii) is nothing but the statement that .... . Accordingly, ..... [= Thus]
(;('() lli elrically speaking, the hypothesis is that ..... : part. of t.he conriusion
is th a t ..... Gi\'en allY f i= 1 suppose that .. ... Theu .... .

T he int.e res t he hypotheses of .... .

. t I " fi
T IIe pnUClpa
f I I . I'111 the assertIOn
. ..... Let P satisfy the above assumptious.
Slgru canee 0 t. Ie emma IS tl t't II t 1 Then ... ..
'Ih . t la I a ows olle 0
e POIll 1 N(P) = 1.

The theorelll gaius in interest if we realize that ..... Let assumptions 1-5 hold. Then .... .
Under the above assumptions, .... .
TI Ie tl leo re m is 's1111
till true I'rl we drop the assu;nptioll ...
1 I . Under the sallle hypotheses, .... .
I s t l 10 (S It .IS onlY
I assumecIt Ila t .... .
Under the conditions stated above, .....
I [ wc t;l!<C
. f = ..... I
I we recover th
_ e standard clemma .....
Under the assumptions of Theorem 2 with "convergellt"
1(l:p lar lll ;; f by - J, [I, Theorem Cl]. . replaced by "weakly conv ergent" :~~ .. . - ---
Under the hypotheses of Theorem 5, if moreover .....
'l'l li :; s pec ializes to th e result of [7] if f = g.
Equality holds iu (8) if and only if .....
be Ileed ed in 1 The following conditions are equivalent: .....
' ,'Ii i:: I (, ~; lI l l will pro ve extremely useflll in Section 8.
I not be needed until [Note: Expressions like "the following inequality holds" can in
general be dropped.]
rpi .' .
~,~ ,
12 13

pro\'c (sh()\,,/rrcall/obscr\'e) tklt .... . definition, ... .. , which follows from ... ..
We I
fi rs t pro\'(~ " rC l !ll(,l~ci form of the th core:lI. Iassumptioll,
the defini tion of .... .
..... But I =g as was described
Let us outli!lC C;i \'t: t!:r: mai n ideas of) U!\, p roof. (shown / ment ioned/
examill e BI . t he compactness of .... . noted) in ... ..
By Taylor's formula, ... ..
I To ~cc (pro':e) this. Id f = .....
a similar ::l.rgument, ... .. shows that .. ...
Out. A = n. \:I\!l:':)\'e th is i\s,f()!:~\:',~~_
I 1,1,., b proved b~ ,\. )l ld o g - ... ..
the above, .. ...
the lemma below, .....
Theorem 4 now
yields (gi ves /
implies) I = .....
To his end. consider .. ...
t co nti nuity, .... . leads to 1 = ... ..
\V f t t If I [= For tbis Dllrposc; not: "To this aill l"]
\! Irs COlllpl1 e '1 To do this. tah; ..... L1 = o. [Not: "Since ..... , then .... .-']
- fl)r this Jllirpose,we se t .... . --- Since .hs co mpact~ w~ ha.'ieLI_= O.
it fo llows"that L1 = 0
To deduce (3) froD! (2) , LIke .... . we see (co llclude) th at LI = o.
\ \ 'e claim that .. ... Indeed , .. . .. B ut L1 = 0 since I is compact.
We begin by proving ..... (by recalling the notion of .. ... ) 'I'Ve bave If = 0, beca use ..... [+ a longer ex planati o n]
Our proof starts with the observation that ... .. \Ve must have L1 = 0, fo r o th erw ise we can rep lace .....
1')1e procedure is to find .. ... As 1 is co mpac t we L1 = O.
The proof consists in the construction of ..... Therefore LI = 0 by Theorem 6.
That L1 = 0 follows from Theorem G.
straightforward (quite involved).
The proof is by induction on n. we conclude (deduce/see ) th at .. ...
(5) we have (obtain) AI = N.
left to the reader.
based on the followillg observatioll. From this [Note: without "that"]
what has already
been proved, it follows that .....
The main (bas ic) idea of the proof is to take ..... it may be concluded that .....
The [lroof falls l1at~r ~lly i:ltO three parts .
wdl be clJvlded mto 3 steps.
According to (On account of) the above remark, we have AI = N .

We have diviued the proof into a sequence of lemmas . It-follows that

Hence (TllUs/Consequently,/Th erefore)
M= N
the assertion of the lemm a is false.
S uppr)se [hence = from th is ; thus = in this way; therefore = for t hi.-
I , contrary to our c l'aJm , t h a t
it follows that = from the above it fo llows th a t]
I (';(!,{) I I ;

() !l t IJe contrary,
Conversely (To obtain a contradiction), suppose tlla t .... .
and so AI = N.
Sllppose the IClllIlIa were false . Th e n we could fiIlt! .. ...
This gives M = N . and co nseq ue ntl y M N . =
We thus get AI=N . and, ill consequ ence , Ai = N.
t here existed n
an x ..... ,
If x were no t I l l ,
Iwe wou ld have ..... The result is M =N. F is compact, and hence bounded.
tI Id b (3) now becomes M=N. which gives (im pli es/
it were true that ..... , .lere wou e ..... Th is clearly forces M = N. yields) 111 = N.
Assume the formula holds for the degree k; we will prove it for k + 1. [Not: "what gives"]
Assullling (5) to hold for k, we will prove it fo r k + l. F =G=H I
the last equality being a consequence of Theorem 7.
, which is due to the fact that .....
We only give tbe main ideas of the proof. ~

We give the proof only for the case n = 3; th e other cases are left to the Si nce .. ... , (2) shows that ... .. , by (4).
reader. We conclude from (5) that ... .. , hence that .. ... , and finally that .. ...
t;:, ,:~~ -
14 15
Th!! 'quality f =g. which is part of the conclusioll of Tlteorem 7, implies Repealed app li c:tI.ioll of Lemma 6 enables us to write .....
thal ..... . \Ve now J r r' 'd I Y illduction.
\Ve can 1I0 W Pl(l 'ced annlogonsly to the proof of .... .
As in the proof of Theorem 8, equation (-I) gives .....
Analysis similar to that in the proof of Theorem 5 sI10\\'5 \Vc nex t I claim (show/prove that) .. ...
that ..... [Not: "similar as in"] sharpen these results and prove that .....
A passage to the limit similar to the abo\'e illlplil's that .....
Similarly (Likewise), ..... claim is that .....

Similar arguments applyI' It 0 tl Ie case ..... Our next goal is to determine the number of .....
obj ective is to evaluate the integral I.
} 'I Ie same reasoulllg
. app les concern will be the behaviour of .....
Tlte sa me couclusion can be drawn for .... . \Ve now turn to the case f =1= l.
This foll olV by th e s;].me method as iu .... . \Ve are now in a position to show..... [= We are able to]
T ill! telill T f can be haudled in much the same wav, the only difIereIlce \Vc proceed to show that .... .
Iwillt'. ill lli e ;]'Dalys is of ..... - The task is llOW to find .... .
II I llll! saJllI! mann er we can see that .....
Ha\'ing disposed of this preliminary step, we can now return to .....
T lt l' rc:; t of lh e proof runs as before.
W nolV apply this argument again, with I replaced hy J, to obtain ..... \Ve wish to arrallge that J be as smooth as possible.
[Note the illfinitive.]
PROOF: CONSECUTIVE STEPS We are thus looking for the family .....
We have to construct .....
evaluate .... .
Consider ..... Define
Choose.... . Let
If = ..... Let. us
compute .... .
apply the forl1lllia to .....
In order to get this inequality, it :vill be ll~cessary to .... .
IS convelllent to .... .
Fi x ..... Sel . suppose for the III0nWIIt.
regard s ;L~ fixed and ..... To deal with I J,
To estimate the other terIll.
Iwe note that .....
iN(}l e: The imperative mood is used wheu you 0l'del' the reader to do For the general C<L5e,
sO l1l ethi ng, so you should not write e.g. "Give an exalllple of .... ." if
yo u mean "~Ve give an example of ..... "]
Adding 9 to the left-hand side
Subtracting (3) from (5) ffi I Ishow (prove) that .....
Writing (Taking) h = H f
Substituting (4) into (6)
yields (gives) It = ..... I
It ~u 1 cflies. t to make the following observation.
IS SUI! clen
use ( 4) together WIth
. the observatIOn
. that .....
we obtain (get/have) J =.rJ
Combining (3) with (6) We need only consider 3 cases: .....
[Note: without "that"]
Combining these We only need to show that .....
[E.g. th ese in equnli ties] we conclude (deduce/sec) that .. ...
we can assert that .... . It remains to prove that .... . (to exclude the case when ..... )
Replacin" (2) by (3)
we can rewrite (5) as ... .. What is left is to show that .....
Letting n co
Appl ying (5) ~ ~ We are reduceclto proving (4) for ..... ---
11I 1,!' I,(," ;1111~ i l l g J ,wd 9

IN,, /, '1'111' ill :~ fOrlll is eit.he r th e subject of a sentence ("Adding ..... gi-
V(':I" ), (II' Il'ljll ires th e subject "we" (HAdding ..... we obtain"): so do
. We are left with the task of determilling .... .
The only poiut remaining concerns the behaviour of .....
The proof is completed by showing that .....
/11)/ IV I ill; e.!;;. "Adding ..... the proof is complete."] We shnll have established the !emnw. if we prove the following:
If we prove that ..... , the assertion follows.
1,\ ',. rO lll.illlle in this fashion obtaining (to obtain) J= ..... The statement O(g) = 1 will be proved once we prove the lemma below.
W, 1I1 ;\Y now integrat e J.: times to conclude that .... .
~'-~::-~ 17
PROOF: "IT IS EASILY SEEN THAT ..... " Note tbat we have actually p roved that .... .
[= We have proved more, namely that ... '
clear (e\'i<ieli! /ir;1llIu::lte/oi;vio'lS) that .....
It IS casiI-. ~CC!l t!;.l~ .... . W I ll oll ly the fact that ... ..
easy "to cLx!': t ilat .... . e 1;"\"(, user the existence of only the ri ght-han d derivati\c.
a sinlp le !!1~:.:ter to .... .
for f = 11 it is no longer true that .. ...
\Ve SCI: (check) at O!lC" ::, : ,; . " .. ..... , wi:!r! 1 is cl,~ar from (3). the argument breaks down .
F is ca.;ilv seen (Cilcckc<, to be sm oo th. . .... , itS i~ ot<J.~y to check. The proof strongly dep ended on the assumption that
It folluw,; cas i'" (inllll<~dl:l'c1y) t i!:lt , .... 00tc that we did . not really ha\'e to usc .. .. . ; we could bave applied .....
Uf cour~l' (CI l'a riy/Olwiol\siy) ...... For more details we refer the reader to [7].
The rroo[ is "traigLr:,,[\\'ard (stallJardi i!1Jl1l,~diate) .
The details are left to the reader.
All easy camp':' ~!: in" {A trivi:11 '{erii:C!l.tion) sho',\'s that .. ... We !cal'e it to the reader to n:rify that... . [Nole: "it" necessary]
- !:':2) ,Hakes -it ob '.<oUS 1hQt. ~ r=
By (21 it is obviolls that] Thi s fillishes the proof, the d'.! wiled verification of (4) being left to tbe
Tile fac:.tO f Gj I'oses ilU prQblel:1 lkC,-lii.-;r. Gis .. ...
(see [or instance [7 , Th. 1]) (sec [7] and the references giVl'll tl ll 'I" )
proves the theore m.
completes the proof. more details)
I:stablishes the form ub . (sec [Kal] for the d efinition of ..... )
..... , wlliciJ
: .Vot: .. '.\' h,l.T.. j is the desired conclusion. the complete bi bliograpby)
is our claim (assertion). [Not : "thesis"]
gives (-1) when substituted in (5) (combined with (5)). TIH~ bes t general reference here
TIlP. sUllldard wo rk Oll .. ...
IS .. ...
wa~ proved hy L lX [t>].
This can be found ill
the proof is complete.
this is precisely the assertion of the lemma . The cl:Lssical wo rk Iwre Lax [7, eh. 2] .
..... , and the lemma follows.
(3) is proved. . is due to Strang [8].
f = 9 as claimed (required). I
goes back to the work of .... .
as far as [8].
ThIS CO l!t[,',J ids our assumption (the fact that ..... ). was motivated by [7].
.. ' .. . conrra ry to (3). This construction generalizes that of [7] .
, C,1
. ..... '.VU1 1 15 .IlllpOSSJ'ble. [N a t : 'w
., It''']
la IS follows [7].
...... '.vLie!! con t r:\dicts the maximality of ..... is adapted from [7] (appears ill [71) '
.... .. :l cOill.r'ldiction. has previously been used by Lax 7] .
The proof for G is similar. a rece nt account of the theory
C may be handled in much the same way. a treatment of a more genera l case
Silllilar considerations app ly to C. a fu ll er (thorough) treatmell t
Ie Silme proo
f works (remains valid) for .... .
still goes (fails) when we drop theassumptioll .....
For a dee rer discussion of .... .
direct constructions along more
we refer the reader to [7].

cl assical lines
The l1letbod of rroof carries over to dom aills .... . yet anothe r method
The rroof a/)o\'e gives more, namely f is ..... We introduce the notion of .. ... , following Kato [7].
A s li ght change ill the proof actually shows that ..... We follow [Ka] in assuming tilat .....

0.. 7
' I'hl! lll aill results of this paper were announced in [7]. Tak illJ', in l.o I\ITO I III~ (tl) --> By (4)
Silll il:tr results havebecn obtained independently by LiLX and arc to be Oyv illIl P of( ,I) ,13y (4)
pub lished in [7] . Oy 11!lalio ll (d) , lIy (4)
III tit illll'l vii i [0 , II ~. In [0,1]
ACKNOWLEDGMENTS There (' X i S I.~l :t f ili i tion f E C(X) --> There exists f E C(X)
T he a uthor :vishes to ~xpress his thauks (gratitude) to .....
IS greatly Illdebted to .. ...
For every po ill t)J !II -v-+ For every p E AI
F is defill (:d I>y tit , for mula F( x) = ..... -v-+ F is defined by F( x) = .....
his <1.ctive interest iu the publication of tllis paper. Tb eo rem 2 a ll cl Th eorem 5 "" Theorems 2 <1.ud 5
suggestiug the problem and for many stimulating conversations. This follow s frolI l (1), (5), (6) and (7) --> This follows from (4)- (7)
for sever<1.l helpful comments concerning ... .. For deta ils!i ' [:31. [4J a ud [5] ..... For details see [3]-[5]
drawing the author's attention to .. ...
poi nting ou t it mist< in ..... The derivaLi v wi Lit respect to l ..... The l-deriv<1.tive
hi s coll abor:tt ion in proving Lemma 4. A fun ct ion [ cI <1.SS C 2 "" A C 2 function

' I'lt l' :t IlL hor g ra tefull y ac knowledges the many helpful suggestions of .. .. . For a rbitra ry 1; ~. For all x (For every x)
li m ill g Lit pr para tion of tile paper. In the case n = 5 ..... For n = 5
Thi s is p<1.r t of th e a uth or's Ph .D. thesis, written uuder the supervision This leads t o a contradiction with the maximality of f
of .... . a t th e Uni versity of ..... ........ .. , contrary to the maximality of f
Applying Lemma 1 we conclude th<1.t "" Lemm:t 1 shows tll<1.t
The author wishes to tlt a nk the University of .. ... , where the paper WiLS ... .. , which completes the proof --> . . . . .
written, for financial s upport (for the invitation and. hospitality).
, lI era l rul es: I would like to submit Ithe enclosed manuscript " ..... "
I am submitting for publication in Studia Mathematica.
J. n.eme mber: you are writing for an expert. Cross out a ll th at is trivial or routine.
~. Avoid repeti ti on: do not repeat the assumptions of a theorem at th e beginning I have also included a reprint of my article .. ... for the convenience of the
of its proof, or a complicated conclusion at the end of the proof. Do not repeat referee.
th e ~s uillptions of a previous theorem in the statement of;\ next on e (instead, I wish to withdraw my paper ..... <1.S I intend to make a m<1.jor
writ e e. g. "Und er t.he hy potheses of Theorem 1 with f replaced hy g, ..... ") . Do - revision of it.
not re peat the same formula-use a label instead . I regret any inconvenience this may have caused you.
3. C heck all formul as : is e<lch of them necess<1rY?
I am very pleased that the paper will appear in Fundamenta.
Phrases you can cross out:
Thank you very much for accepting my paper for publication in .... .
We denote by IR the set of all re:tl numbers.
lYe have the following lemma. Please find enclosed two copies of the revised version.
T he followin g lem ma will be useful. As the referee suggested, I inserted a reference to the theorem
.. .. . the follow illg inequ a lity is s<1.tisfied: of .... .
We have followed the referee's suggestions.
P h r <l~ ~ yo u ca n short en (sec <1.lso p. 38):
I have complied with almost all suggestiolls of the referee.
I.e'/. Ill' <1.11 a rb iLrary but fi xed positive numb er""" Fix c; >0
I,e'/. II:; fi. Htb ilra ril y x E X """ Fix x E X REFEREE'S REPORT
1,I'i. 11:; firs t obs(rvc th :tt """ First observe th<1.t
IV" wi ll lirs t co mpute """ \Ye fi rs t compute The author proves the interesting result that .....
II I' II{"(; we 11(1Xe :r =1 """' Hence x = 1 The proof is short and simple, and the article well written.
II (, ll cc it [o llolVs tha t x=l""" Hence x=l The res ults prese nt ed are original.

20 21
The paprer is a good piece of work Oll a subject that attracts PART B: SELECTED PROBLEMS OF ENGLIS H GRAMMAR
considerable attention.
I ;lm plca..c,cd to I j' r bl' " INDEFINITE ARTICLE (a, an, - )
T. I'. ~pI .. , ... ,rt'commcllC It. lor pc lcatlOll III
1l " " e'L'loLl:'OI S .. ~1 I '
I strongly tuella. at..:ematlca.
Note: You use "an or "an" depending on pronunciation and not
The only remark I wish to make is that condition B should be formuhted spelling, e.g. a uuit, an x.
more carefully.
:\ fe\\' minor typographi cal errors arc listed below. 1. Instead of the number "one":
I !J,lxe indicated \'arious corrections 011 the manuscript. The four centres lie in a plane.
The results obtained arc not particularly surprising and will be A chapter will be devoted to the study of expanding map ~ .
of limited interest. For this, we introduce an aux il iary variable z .
'Tile re<-l'lts- '
, ., ~ I
I correct
"re I ,
but ouly,. moder<1.tely
c f kintrrcstilig.
__ _ -~ 2:- Me3ning "member of- a class of objects" ,- "some" ,-"one of":
i r<'.: :-, cr cas:; mOlll 11cations 0 :;own w.cts .
The cx:ullple is worthwlli:e but not of sufficient interest for a research Then D becomes a locally convex space witll dual space D'.
article. Tbe right-hand side of (4) is then a bound ed functi on.
The Eng lish of the paper needs a tborough revision. This is easily seen to be an equivale nce relation.
T he paper does !lot meet the standards of your journal. Theorem 7 bas been extended to a class of bound a ry val\l e jll'Oh ll'III ';
The transitivity is a conseque nce of the fac t th at .... .
T ueorClll '). I
- IS fit Ise.III
as stated .
t h'IS genera j'Ity. Let us now state a corollary of Lebesgue's theo rem for .....
After a change of variable in the integral we get .... .
Lel1lma 2 is known (see .. ... )
Accordingly, I recommend that the paper be rejected. \Ve thus obtain the estimate ..... with a constant C .
in the plural:
The existence of partitions of unity may be proved by .. .. .
Tbe definition of distributions implies that .. ...
..... , with suitable constants. .
..... , where G and F are differential operators.
3. In definitions of classes of objects
(i.e. when there are many objects with the given property) :
A fundamental solution is a fu nction satisfying ... ..
We call C a module of ellipticity.
A classical example of a constant C such th a t .. ...
We wish to find a solution of (6) which is of th e form .. ...
in the plural:
Tbe elements of D are often called test functions.
points with distance 1 from K
tl 1e se t 0f
Ia II functIOns
. . . I compact support

The integral may be approximated by sums of the form .... .

Taking in (4) functions v whicb vanish in U we obtain .... .
Let f and g be functions such that .. ...

22 23
4. In the plural-when you are referring to each element of a class: DILl : Ev el Y (\()Il Clllpty open set in IRk is a union of disjoint boxes.
Direct sums exist in the category of abelian g roups . [I ( YO II wi s h to stress that it is some union .of not too well
~ ; p( !c i(i ,t! objects.]
I II particular, closed sets are Dorel sets.

Borel measurable functions are oftell called Borel mappings. 4. In front of;1 ::lrdinil l number if it embraces all objeds considered:
Th is makes it possible to apply fh-results to fUllctions in any Hp. The two !',rOll[1S have been shown to have the same number of
If you are referring to all elements of a class, you use "the": gel) mtors. [Two groups only were mentioned.]
The real measures form a Sll bclass of the complex ones . Each of th e th ree products on the right of (4) satisfies .... .
[The re arc exactly 3 products there.]
5. In front of an adiedive which is intended to mean
"having t his particular quality": 5. In front of an ord in a l number:
Tllis map extencls to all of M in an obviolls hshion. The first Poisson integral in (4) converges to g.
A remarkable feature of the solution should be stressed . The seco nd sta tement follows immediately from the first.
S ' ct' 0
C 1 n
1 gives a condensed exposition of .....
describes ill u unified manner the recent results .....
6. In front of surn a mes used attributively:

A s imple computation gives .....

the Dirichlet problem
th e Taylor expansion Bul:
ITaylor's formu la
[without "the"]
Combining (2) and (3) we obtain, with a llew constant C, ..... th e Gauss theorem a Banach space
A more general theory must be sought to account for these
irregularities. 7. In front of a noun in the plural if you are referrin g to a class of objects
The equatioll (3) has a unique solution g for every f. as a whole, and not to particular members of the class:
But: (3) has the unique solutiou 9 = ABf. The real measures form a subclass of the complex ones.
This class includes the Helsol1 sets.
MC',liling "mentioned earlier", "that": A e X. If aB = 0 for every B intersecting the set A, thell ... .. 1. In front of nouns referring to activities:
' ex p x = L.Xi Ii!. The series can easily be shown to converge.
I ) <'Ii II Application of Definition 5.9 gives (45).
2. III rront of a noun (possibly preceded by an acijedive) referring Repeated application (use) of (4.8) shows that .....
to a single, uniqu ely determined object (e.g. in definitions): The'last formula can be derived by direct consideration of .....
A is the smallest possible extension in whic h differen t iation
Le t f be the linear form I~efi'lli~' [/(2). [If there is only one.] is always possible.
Using integration by parts we obtain .. .. .
u = 1 in the compact set J{ of a ll. points at distance 1 from L. If we apply induction to (4), we get .. .. .
We denote by B(X) the Banach space of all linear operators in X. Addition of (3) and (4) gives .....
.. ... , under the usual boundary conditions.
This reduces the solution to division by Px .
.. .. . , with the nat ural defini tio ns of addition and multiplication .
Comparison of (5) and (6) shows that ... ..
Us ing the standard inn er product we may identify ~..~..~.~ _
-~ [Nol e : In constrilctions with "of" you can also use "the" .]
In th e co nst ru d ion: th e + property (or another charaderistic) + of +
o bj ect: 2. In front of nouns referring to properties if you mention no
'1'1' 0 co ntinuity of f follows from ..... particular object:
'1'1", (' x i ~ t c nc e of tes t functions is not evident. In question of uniqueness one usually has to cOllsider ... ..
'1'11('1'<: is fi xed compact set containing thE; supports of all the fj.
it By continuity, (2) also holds when f = l.
'1'1 1(': 11 X is the c entre of an open ball U. By duality we easily obtain the foHowing th eorem .
Th e int e rsection of a decreasing family of such sets is convex. Here we do not require translation invariance.
q 25
3. After certain expressions with "of": a zero of order at least 2
rauk 2.
a type of cOlv:crgcllce the hypothesis of positivity y a t t he origin .
F h;LS cabrdilnalit Ie. But : F has a density g.
a problem of U!:i'llll'll"~S the method of proof a so ute va ue l.
the condition of ellipticity the point of i:;cre;LSe determinant zero. [U.nless 9 h;LS app eared
I earlier; then: F has density g.]
4. In front of numbered obj"cts: 9. In expressions with "as" :
It follows from Thcort~m 7 that ..... Any rauciom variable can be taken as coordinate variable on Y.
Section <1 give.; a cOllcise presentation of .... . Here t is interp reted as area or volume.
Property (iii) is "calkd the tr iangJe inequality.
This h;LS been proved in pnrt (a) of the proof.
But: tht~set of sol utions of the form (4.7)
We show that G is a group with composition as group operation .
But: G is well defined as the integral of f over U.

To pro,'e the estim ate (5.3) we first extend ..... I

~ I
10. In front of the name of a mathematical discipli ne:
We thus obtain the inequality (3). [01': inequality (3)] Tliis idea cQ."rnes fWlll game theory (homologica l algebra).
The ;LSymptotic form ula (3.G) follows from ..... ! But: in the theory of distributions
Since the reg ion (2.9) is in U, we have ..... 11. Other examples:
5. To avoid rEpetition: I \Ve can ass ume that G is in diagon a l for 111.
t !:e order and symbol of a distribution Then A is deformed in t o B bv pushiu<T it at cons t nnt ~rH ... d ,illIl I1',
the ;:::'soci;ltivity and commutativity of A the integral curves of X . 0

the direct sum and direct product G is uow viewed ;LS a set, without group structure.
tLl: i!lilCr and outer factors of f [Note the plural.]
S;d: a d.?ficit or an excess INFINITIVE

6. In front of surnames in the possessive: 1. Indicating aim or int ention:

~lillkowsk i 's
inequality, but: the Minkowski inequality To prove the theorem, we first let .....
Fefferman and Stein's famous theorem, to study the group of .....
We now apply (5) to deri~e the foll?willg theo re m .
more usual: the famous Fefferm?,n-Stein theorem
7. In some expressions describing a noun, especially after "with" and "of":
_ I
to obtam an x WIth uorm not exceediug l. an! some examples to show how .....
an algebra with unit c; an opemtor with domain 1l 2 ; a. solution with 2. In constructions with "too" and "enou gh":
vanislliug Cauchy data; a cube with sides para llel to the axes; a
JOIlIilin with smooth boundary; an equation with constant coef- This method is too complicated to be used here.
ticient s; a function with compact support;ranclom variables witli This is import ant enough to be stated s par;ltc ly
zrro expectation 3. Indicating that one action leads to another one:
the equ a tion of motiou; the velocity of propagation;
We uow apply Theorem 7 to get N f = O. [= ..... alld w ' !,e t N J IlJ
an element of finite order; a soJu tiou of polynomial growth; Insert (2) iuto (3) to find that ..... '
a. ball of radius 1; a function of norm p
4. In constructions like "we may assume lIf to be .. ... ":
Bl1t: elements of the form f = ... \\'e may assume Jt! to be compact.
Let B be a Banach space wit h n weak symplectic form w .
Two random variables with a commou distribution. We define ]{ to be the section of Hover S.
If we take tbe contour G to lie in U, then .... .
8. After "to have": We extend f to be homogeneous of degree 1.
Icompact support. But: F has Inn COfillite
F h;LS finite no r111. UOf111 not exceedi~g 1..
IIIP ;lcL support contallled III I.
The .class A is defined by requiring all the functiOllS f to satisfy .....
Partlally order P by declaring X < Y to 111e,.aI1 th at .. ...
f.:T ..
26 27
5. In const ructions like "M is assumed to be ... ..": At fil !l ll,. lll ll (,(! tl'I appears to differ from N. in two major ways: .. ...
is ass umed (ex pectcd/found! considercd/ t akc ll / A III O I ( ~ ~1() Jllii s ti ca ted argument enables one to prove that .....
claimed) to be opell. [Noll' : "cr lab le" requires "one", "us" etc.]
will be chosen to contain O. H prop osed to study that problem. [Or: He proposed studyillg .. ... J
NI can be t ake n to be a constant. We 111!lku ; net trivially on V .
can easily be shown to have ..... [Note: "cas ily" after "ca n" ] Let f sn ti ;; fy (2). [Not: "satisfies"J
is also found to b e of class S. \Ve n eed t o conside r the following three cases.
T his investigation is likely to produce gooe! results. \Ye n eed !lot co nsider this case separately.
[= It is very probable it willJ ["nee I to" in affirmat ive clauses, without "to" in negative
The close agreemen t of the six elemcnts is unlikely to be claus s; a lso no te: "we only need to consider", but: "we
a co incidence. [= is probably not] need only consider" J
(I . III tll C st ru cture "for this to happen":
For liti s to ha pp e n, F must be compact. lNG-FORM
[= In order th a t this happens] 1. As the subject of a se ntence (llote the absence of "the"):
F r th e las t estimate to hold, it is enough to assume .....
The u for s u ch a map to exist, we Illust have ..... R epeating the previous argument and using (3) leads to .... .
7. As the subject of a sentence:
Since taking symbols commutes with lifting, A is .... .
Combinin g Proposit ion 5 anu Theorem 7 gives .... .
To see that this is not a symbol is fairly easy.
[Or: It is fairly easy to see that .. ... J 2. After prepositions:
To choose a point at random in the interval [0, 1J is a concep tual After making a linear transformation, we may assume that .....
experiment with an obvious intuitive meaning. In passing from (2) to (3) we have ignored the factor n.
'1'0 S ;1Y that u is m<L'<imal means simply that .....
In deriving (4) we have made use of .... .
Il/ier expressions with "it": On substituting (2) into (3) we obtain ... ..
II. i'j I\ ( ' ( ' (" l!i nry (usefu l/very important) to consider .... . Before making some other estimates, we prove .....
I i. /I ".I( n il Sf' Il !:l t o s p eak of ... .. Z enters X without meeting x = O.
I i. jj l.llC ' j('for of ill t erest to look at ..... Instead of using the Fourier method we can multiply .... .
n Afl er "b," : In a"ddition to illustrating how our formulas work, it provides .... .
Our goal (Ill cthou/ a pproach/ proccdurc/ objective/aim) is t o filld .... . Besides being very involved, this proof givcs no info rlll a tion Oll .... .
The problem (difficulty) here is to construct ..... This set is obtained by letting n -> 00.
It is important to pay attention to domains of d efin ition
9. With nouns and with superlatives. in the place of a relative clau se: when trying to .. .. .
The theorem to be proved is the followillg. [= which will be proved] The following theorem is the key to constructing .... .
This will be proved by the method to be described in Section 6. The rcason for preferring (1) to (2) is simply that .... .
For other reasons, to be d iscussed in Chapter 4, we have to .. .. .
He was th e first to propose a complete theory of .. ... 3. In certain expressions with "of" :
Th ey appear to be the first to have suggesCed the now-accepted - - - - - - - - - - T he id ea ~f combining (2) anel (3fc~-tnleI'ionl---:~ --
interpretation of ..... The problem cOl!sidered th ere was that of determining WF( u) for .....
10. After certa in ve rbs: \Ye use the t echnique of extending .....
Th ese proper ties led him to suggest that .... . being very involved .
Th y believe to have discovered .. .. . This method has the disadvantage of requiring that f be positive.
L::lx claims to h ave ob tained a formula for .....
This map turns out to satisp.f .. ... Actually,S has the much stronger propert~f being COllvex.
[Note the iu fin it ive.]

t".;:? .
28 29
4. After certain verbs, especially with preposi tions:
We begin by analyzin g (3).
The ideal is defined by m = . . . , it be ing u nd erstood that ... ..
F ~on ti.nuous,.we ca n assu me th;1 t ..... [= Since F is .....]
(It b~lI1g .m:p~ss lbl e ~o make A an d B intersect)
\Ve succeeded (were successful) in proving (4).
[Not: "succeeded to proye"] I [= slllce It IS 1m POSS Ible]
\Ve next tur n t o estimating ..... [Do not write "a func tion being an element of X" if you mean
They persisted in investigating the case .... . "a fu nction which is a n element of X". ]
\Ve are interested in finding it solution of .... . 8. In expressions which can be re phrased as "the fact that X is .... ... :
\Ve were surprised at finding out that .... .
[Or: surprised to find out] Note th a t M being cyclic implies F is cyclic.
Their study resulted in proving the conjecture for ..... The probab ility of X being rational equals 1/2.
__ The suscess of our method will depend on proving t hat ..... In addition to f b eing convex , we requi re that .....
To compute the j;Z;:m of ~amo"i:Jnts to findirlg---: ...~ --
We should avoid using (2) here, since .. ... PASSIVE VOrCE
[Not: "avoid to usc"]
We put off discussing this problem to Section 5. 1. Usual passive voice:
It is wo~th noting th at ..... [Not: "worth to no t e" ] This theorem was proved by l\Iilnor in 1976.
It is worth whi le discussing here this phenomenon.
[Or: worth while to discuss; "worth while" with ing-forms is In ite ms 2-(j, passive voice str uc tures re place sente nces wit h s ub ject .. IV. .. <J I

best avoided as it often leads to errors.] imperso nal constructions of otber languat:;es.
It is au idea worth carrying out. 2. Replacin g the structu re "we do something":
[Not : "worth while carrying out", nor: "worth to carry out"]
This identity is establis~ed ,by observing that .':...
After having finished proving (2), we will turn to .....
This difficulty is avord~d_' above. '-- - - . .
[Not: "finis hed t o prove" ]
(2) needs handling with greater carc. When this is sub~tituted in (3), au analogous description of J(
is obtained. . .
One more case merits mentioning here.
In [7] he mentions having proved this for I not in S. Nothing is assumed concerning the expectati~n of X.
5. Present Participle in a separate clause (note that the subjects 3. Replacing the structure "we prove that X is":
of the main clanse and the subordinate clause must be the same):
We show that I satisfies (2), thus completing the analogy with .... .
M Ii5mayeasbeily said
shown to have ... ..
to be regular if ... ..
Restricting this to R, we can define .. ...
[Not: "Restricting ..... , the lemma follows". The lemma does This equation is known to hold for .. ...
not restrict !] 4. Replacing the construction "we give an object X a stru cture Y" :
The set A, being the union of two continua, is connected.
Note that E can be given a complex structure by .... .
6. Present Participle describing a noun: The let ter A is here given a bar to indicate t hat .... .
\Ve need only consider paths starting at O. 5. Repla cing the structure "we act on something":
We interpret f as a function with image having support in .....
We regard f as beillg defined on ..... Thi s ord er behaves well when 9 is act ed upon by an opera tor.
F can be thought of as .. ...
7. In expressions which can be rephrased using "where" or "since" :
So a ll the terms of (5) a re accounted for.
J is defined to cqual AI, the function I being as in (3). This case is met with in diffraction problems.
[= where f is ..... J III the phys ical context already referred to, K is .. .. _
This is a speci al case of (4), the space X here being B(K) .
Tile preceding observation, whcn looked at from a more D'eneral
We construct 3 maps of the form (5), each of them satisfying (8).
point of view, leads to .... . Co
..... , the limit being assumed to exi st for every x.
V'. 31
II Mr. !l ing "which will be (proved etc.),,: If ]( is ll OW lilly (fll ll p:,rL Sllbset of H, there exists .... .
[Any W il ld . I'V I' ! you li ke ; write "for all x", "for every x" if you
1klore st a ti ng the result to be provcd, we give .....
Th is i: a special case of convolutions to be introduced in Chapter 8. just J! II" ; , II I q ililiit ifi r r.]
Every m ';1." 11 1'(: ca ll \)u comp leted, so whenever it is convenient, we may
\V, co ncl ude with two simple lemmas to be lIsed mainly in .....
assum th i,l /Ill y give ll meas ure is complete.
QUANTIFIERS T hc r fJ i ~l a !i ubsequence such that .....
Tlt m'o cx i !ll!l an x with .... .
" l' hAt'
T l11 5 Imp les t at
all open subsets of U .
con alllS a 11 y wily
'tI G = 1. . . [O/: tll ere ex ists x, but: there is an x]

ll r I
f all transforms F of the form .....
I~e t B b e tI1e co ec Ion 0 all A such that .....
Th cr e a r c sets satisfying (2) but not (3).
T il r is ;t 1iu ique function f such that .....
Ea 11 J Ii s in zA for some A (at least one A/
/.' j.] cil' li!l ('d at a ll points of X . exac tl y O ll C A/ at most one A).
1111 Idl II I
0; fo r a ll m which have ..... ; for all other m; Not th at some of the Xu may be repeated.
fnl d l h uL it !ill ite ulllll ber of indices i F has 11 0 fi xed vecto r (no pole) in U. [Or: no pol es]
X cOIlt<1.ins a ll th e boundary except the origin. F has no limit poi nt in U (hence none in J() .
T he ill tegml is t a ken over all of X. Call a se t dense if its closure contains no nonempty open subset.
If no two members of A have an element in common, then .... ,
all extend to a neighbourhood of U.
all have their supports in U. No two of the spaces X, Y, and Z are isomorphic.
E , F <1.ud G are all zero <1.t x. It ca n be see n that no X has more than one inverse.
are all equal. III other words, for no real x does lim F,J x) exist.
[Note the inversion after the negat ive clause.]
' 1'1, 1' , f: ex is t fu ncti ons R, all of whose poles are in U, with .. .. . If there is no bounded functional such that .... .
1';,11 Ii ()f i.Ill! fo llowing 9 conditions implies all the others.
:;,11 1, .111 .,. ,'x is l.s iff a ll th e intervals Ar. have ..... .. .. . provided none of the SUillS is of the form .... .
Let Au be a sequence of positive integers none of which is one less than
1\ " I'VI'''Y 9 ill X (not iu X) there exists an N .... . a power of two.
1 11 11/ ' for all J <1.1ll1 g, for a ny two maps J ancIg; "every" If there is an f such that .... . , we put .. ... If there are (is) none, we
i:; fo llowed by a singular noun .]
define .....
'1'0 t: V'~I'Y J th re co rresponds <1. unique 9 such that .... .
N one ofthese are (is) possible.
Fvcr y illvari<l nt su bsp<lce of X is of the form .....
[Do lIot write : "Every s ubspace is llot of the form ..... " Both f and 9 are obtained by .... .
if you mean: "No subspace is of the form .. .. ."; [Or: f and 9 are both obtained]
"every" must be followed by an affirmative For both Gee and analytical categories, .....
statement.] C behaves covariantly with respect to maps of both X and G.
Thus f of. 0 at almost every point of X. We now apply (3) to both sides of (4).
Both (these/the) conditions are restrictions only au .. ...
Since A " = 0 for each TI, .... . [Each = every, considered separately]
[Note: "the" aft er "both"]
Each ter m ill this series is either itor~l ,--~ ~ C lies on no segment both of whose endpoints are in J( .
F is ho unded on each bouuded set.
Two consecutive elements do not belong both to A
j:.IC" h of t il ese fo ur integrals is finite .
01' both to B .
T hese curves arise from ..... , ancI each consists bf .. ... Both its sides are convex . [01': Its sides are both co nvex.]
Tilere remain four intervals of length 1/16 each. Bane! C are positive numbers, not both O.
X ass umes values 0,1, ... ,9, each with probability 1/10. Choose points x in /11 and y in N, both close to z, and .....
P I , .. . ,Fa me each definecI in the interval [0, 1]. We now show how this method works ill 2 cases .
Tilose n d isjoint boxes are translates of each other . In both, C is .. ...

In either c;ese, it is clear that ..... [= In both cases] The gain up to and in -' lll dill h( til , n th trial is .....
Each f can be ex pres~('d ill either of the forms (1) and (~). The elements of th e third a lld fOllrth rows are in I .
[= in any of the two fOi"!;lS] [Note the pluraL]
TL(~ den~ity of X + Y is gi\"Cll by either of the two illte~r:ll ~. F has a zero of at leas t t.hird order a t x.
The t\\o da.sscs coiuciJe if X is cOlllpact. In tbat case we write C(X) for 3. Fractions:
either of tu em. Two-thirds of its diameter is covered by .... .
Either f or !l must be bo u nded . Bu t: Two-thirds of the gamblers are ruined _
Ld 11 and v be two distributions neither of which is ..... G is half the Sllm of the positive roots.
[Use "neither" WhCll there arc iwo alternatives.] [Note: Only "half" can be used with or without "of".]
This is true for n ei th er of the two functi ons. On the average, about half the list will be tested.
Neithel' statement is true. J contains an interval of half its length in which .... .
In neither case CUI f be smooth. F is greater by a half (a third).
[:\ote- tllp. inversion after a negative cbuse. ] - ' - -The- other- pl a);er is- haW(one third tas- fast:-
lIe proposes two coutiitions, but neither is satisfactory. We divide J in half.
All sides were increa.sed by the same proportion.
NUMBER, QUANTITY, SIZE About 4 0 percent of the energy is dissipated .
A positive percentage of summands occurs in all the k
1. Cardinal numbers:
parti tions.
A aud B are also F-funct. ions, any two of A , B, and C being
independent. 4. Smaller (greater) than :

t1 I .. j 'tl I all entries zero except the kth which is one great er (less) than k.
18 lllll tl-Illl ex WI 1 the last k entries zero much (substantially) greater than k .
n is no greater (smaller) than k.
This shows that there are no two points a and b such that .... . greater (less) than or equal to k.
There are three that the reader must remember. [= three of them] [Not : "greater or equal to"]
\\'e have defined A, B, and C, and the three sets satisfy .... . strictly less than k.
For the two maps defined in Section 3, .....
[;'The" if only two maps are defined there.] All points at a distance less than K from A satisfy (2) .
R is conce ntrated at the n points Xl, . .. , Xu defined above. We thus obtain a graph of no more than kedges.
for at least (at most)' one k; with norm at least equal to 2 TI11S I
. se t Ila.s fewer elements than f( has .
no fewer than twenty elem ents.
Tbere are at most 2 such Tin (0,1) .
There is a unique map satisfying (4). F can have no jum ps exc eeding 1/4.
(~) !la.s a unique solution g for each f The degree of P exceeds that of Q.
But: (4) has the unique solution 9 = ABf find the density of the smaller of X and Y .
(-n has one and only one solution. The smaller of the two satisfies ... ..
Precisely T of the intervals are closed. F is dominated (bounded/estimated/majorized) by .....
In Example 3 only one of the Xj is positive. 5. How much smaller (greater):
If p = 0 then there are an ad itional m arcs. 25 is 3 greater than 22. 22 is 3 less th an 25.
2. Ordinal numbers: Let an be a sequence of positive integers none of which IS one less
The first two are simpler than the thil-d . than a power of two.
Let Si be the first of the remaining Sj . The degree of P exceeds that of Q by at least 2.
The nth trial is the last. f is grea ter by a half (a third).
X 1 appears at the (k + 1 )th place. C is less t h a n a third of the distance between .....
t. _(:~
Within J, the functiail f varies (ascillates) 9. Man y, few, a nUl11 be r of:
by less t hali l. a large number of illustrations.
The upper anel lawer limits af f differ by at most l. T h re are only a finite number of f with Lf = l.
\Ve thus have iu A one element too many. [N te the a small number of exceptians.
O n applying this argument k more times, we abtain ... .. plural.] an infinite number of sets .....
T his met had is rece ntly less and less used. a negligible number of points with .....
A s uccess ian af more and more refined discrete mod els. Ind c is th e numb e r of times that c winds around O.
6. How many tim es as great: \Ve give a numb cl' of results concerning ..... [= same]
twi ce (t en timeslone third) as long as; half as big as This may happeu in a number of cases.
T he langes t edge is at mast 10 times as lang as the shartest ane. They carrespond to the values of a countable number of invariants.
A has twic e as m a ny elements as B has. .. ... far all n except a finite number (for all but finitely many n).
.J mllt.;tiIl S a su binterval of half its leng th ill which ..... Q cantains all but a countable number of the t.
1\ 11 :\.'; f')IIl' lillies the rad ius of B. There are only countably many elements fJ af Q with dam fJ = S.
Till' d iall ll'lcr of L is 11k tim es (twice) that of M.
The thearem is fairly genera!. Th ere are, however, numerous
r. MIJi Lip l : exceptions.
T in; k-fold illtegratia ll by parts shows that .. ... A variety of other characteristic functians can be constructed iu this
F cavers Ai twofold. way.
AI is oouncbl by a Illultiple of t (;c constant tillles t). There arc few p.xceptians t.o this rlll(, . [= not lllany]
This distance is less than a constant mt~l tiple of d. Few of varia liS existing proofs are coust.rllctivc.
G acts 011 H as a multiple, say n, of V. He accounts for all the majar achievements in topology
over the last few years.
n Mos t, I as t, grea test, smallest: The generally accepted point of view in tuis domain af
" It ;l ~; tit mos t (the fewest) points when .... . science seems to be changing every few years .
I II 1I11 "d, r :l.'c. it turns out that .. ... There are a few exceptions to this rule. [= some]
"'I wit, Ill' t.1t!! t.heorems presellted here are original. MallY interesting examples are knawn. vVe now describ e
T it " fI' Ol)r:: for tit m os t part , o.nly sketched.
a l l!, a few of these.
Mwd. pl' (J !) lI bl y,Lt is lIlet had will prove useful in .. ... Only a few of those results have been published befare.
\Vhitt Ill os t ill terest us is whet her ..... Quite a few of them are naw widely used.
[= A considerable number]
T Lte leas t such constant is called the norm of f
This is the least useful of the faur theorems. 10. Equ ality, difference:
The method described abave seems to be the least camplex. A equals B 01' A is equal to B [Not: "il. is equal B"]
T hat is the least one can expect. The Laplacian af 9 is 4r > O. Then T is about kn.
Tue elements of A are comparatively big, but least in nu mber. The inverse af FG is GF. The norms af f and 9 coincide.
Nane of those proofs is easy, and John 's least of all.
T he best estimator is a- linear -combinatian U- such that-
__Y' ll a:'~1C s~n-.:e number of z~ros al151_poles ~n U.

var U is smallest possible. F and G differ by a linea r term (by a scale factar) .
T he expected waiting time is smallest if ..... The differen tial af J is different from O.
L is the smallest number such that ..... Each memher of G other than the identity mapping
F has the smallest norm amal1g all f such that ..... is .....
[( is th e largest of th e functians which occur in (3). F is nat id entically O.
T here exists a smallest algebra with this praperty. Let a, band c be distinct complex numbers.
Find the second largest clement in the list L. Each w is pz far precisely Tn distinct values of ::.
36 37
functiolls which are eq,I;11 :l.e. ,HC indistinguish:Lble as felr a.s We may replace A and B by whicheve r is the larger of the two.
illtegra tion is concerned. [Not : "the two ones"]
This inequality applies to cond itio nal expectations as well as [0
11. Numbering: ordinary ones.
Exercises :2 to 5 furuish l,tbcr applic;1tions of this tt>cllli:'lIW.
One has to examine the equ at ions (4) . If these ha\'e no solutions
lAma.: Exercises :2 th;'ough 5]
then ..... '
ill the third and fourth ro\\'s the deri\-ati,,'cs \lp to order k
D yields operators D+ and D-. These are formal adjoillts of each
from lin e 16 onwards the odd-numbered terms
other. .
in Jines Hi-10 This gives rise to the maps F i . All the other maps are suspensions of
the next-to-bst CO!Ulllll
the par;1graph but one of the pre\'ious proof
F is the sum of A, B, C and D. The last two of these are zero.
~t . ~-- . 1-11 in the (i,j) entry <lnd .zero (lsewhcrc
TI Ie 111<l flX W I L I:1! I entlws Both f a nd' g-;:'lfe connected, but the latteris- in addition compact.
. '
zero ex cept t'ur,\ ' - ). at (\'
' .)')
[The latter = the second of two objects]
Tl' .
liS IS
hinted in Sectio!!s 1 ;1nd :2.
on page 3G of [.1].
Doth AF and BF were first co ns idered by 13(ln<lch, but onlv the for-
mer is referred to as the Banach map, the latter bei ng "called th e
Hausdorff map. '
HOW TO AVOID REPETITI ON \\'e have thus proved Theorem s 1 and 2, the latt e r wit ho n t using .. ...

1. Repetition of nouns : Si nce the vectors C i are orthogonal to th is last space, .....
As a consequencc of this last result, .....
Note that the continuity of f implies that of g.
Let us consider sets of the type (1), (2), (3) and (4) .
The passage from Riemann's theory to that of Lebesgue is .. .. .
These last two are called .... .
The diameter of F is about twice that of C ,
His method is similar to that used in our previous pap er. We shall now describe a general situation in which the last-mentioned
The nature of this singularity is the same as that which f has at functionals occur naturally.
x = O. 2, Repetition of adjectives. adverbs or phrases like "x is .... ,.. :
Our results do 'Ilot follow from those obtained by L<LX ,
ff f and g are mcasurable functions, then so are f + 9 and f . g.
Olle can check that the metric on T is the one wc have jus t described.
The union of mea.surable sets is a measurable set; so is thc complement
It follo\\'s that 5 is the union of two disks. Lct D be the one that
of every measurab le set.
contains .... . Tue group C is compact and so is its image under f.
The cases p = 1 and p = 2 will be the ones of interes t to us.
It is of the same fundamental importance in analysis as is the
We prow a uniC[ueness result, similar to those of the preceding section .
construction of .....
Each of the functions on the right of (2) is one
F is bounded but is not necessarily so after division by C.
to wuich .....
F has many points of continu ity. Suppose x is one , Show that there are many such Y,
In addition to a contribution to WI, there may be one There is only one such series for each y .
to W 2 . Such an h is obtained by .....
\.Ve !lOW IHOV<! t.hat the constant pq cannot be replaced by a smaller 3. Repe tition of verbs:
Consider the differences between these int egrals anel the A geodesic which meets blvI does so either transversally o r .....
corresponding ones with f in place of g, Th is wi ll ho ld fo r x > 0 if it does for x = O.
On account of the estimate (2) a nd s imilar ones which call be ..... T he iu t gritl migh t not converge, but it does so after .... .
The geodesics (4) are the only ones that realize the distance between No te th a t w have not required that ..... , and we shall not do so ex 'c pt
their endpoints . wli '11 exp licitly stated.
\Ve will show below that the wave equation call be put in thi s form, There has since been little systematic work Oil .... .
as can many other systems of cqu<1.tions. " It has recently been pointed out by Fix that .... .
The elements of L are not ill 5, as they are in the proof of ..... It is sometim es difficult to .....
'I . I ~ petition of whole sentences: This usually implies further conclusions about f.
It often does not matter whether .. ...
The same is true for j in place of g.
The same being true for j, we can ..... [= Since the same ..... ] Adverbs like "also", "therefore", "thus":
The same holds for (applies to) the adjoint map . Our presentatiou is ther efo re organized ill such a way that .... .
\Ve shall assume that this is the case. The sum in (2), though form ally infinite, is therefore actually finite .
Such was the case in (2) . Oue must therefore also introduce the class of ... ..
The L2 theory has more symmetry than is the cas e C is connected and is therefore not the union of ... ..
in L1. These properties , wit h the exception of (1), also hold
Then ei ther ..... or ..... In the latter (former) case , ... .. for t.
1.'(1 1 k this is no longer bue. \Ve will also leave to the reader th e verification that ... ..
'f 1,iB is Hot tru e of (2) . It will thus be sufficient to prove that .... .
This is not so ill other queuing processes. (2) implies (3), since one would otherwise obtain J,; = O.
If this is so, we may add .... . The order of several topics has accordingly been changed .
If fi ELand if F= h + ... + fn then FE H, and eve ry
Emphatic adverbs (clearly, obviously, etc.):
F is so obtained.
It would clearly have been sufficient to assume that .....
We would like to ... .. If U is open, this can be done.
F is clearly not au I-set.
Ou 5 , this gives the ordinary topology of the plan e.
Its restriction to N is obviously just f.
N()I.(~ t hat this is not equivalent to ... ..
This case must of course be excluded.
INote t he differeuce between " this:' and "it" : you. say "it i.s :lOt
<'<I \I iva.I nt to" if you are refernng to some olJject exphcl tly
The theorem evidently also holds if x = O.
(1 1i' lI l, io Ii CcI ill lhe preceding seutence.] The crucial assumption is that the past history in no
way infl uences .....
I" I",. t.Il1 ' s t. a t. ed (rl'sired/cl a imed) properties.
\Ve did not really have to use the existence of T.
The problem is to decide whether (2) rea lly follows
from (1).
General r emarks: The normal order is: subject + verb + d irect ohject + ,,, h 'erbs ill The proof is 110W easily completed .
the order manner-plac e-t illle. The max imum is actually attained at some point of AI.
Adverbial cla uses Cnn also be pbceJ at the beginning of a sentence, and SOllie adverbs
\Ve then actually have ..... [= \Ve have even more]
a iways come betwee n subject anJ verb. Subject almost always precedes ve"rb, except
At present we will merely show that .....
in questio ns and some negat ive clauses.
A stronger result is in fact true.
1. ADVERBS - - -
Throughout integration theOl")" one inevitably encounters 00 .
1,1. B lween subject and verb. but after "be"; in compound tenses after
--nut H ltself call equallY well be a menlDer- or- S. - - - - - -
firs l :w xiliary lb. After verb-most adverbs of manner:
I I 'qu ell yadverbs: \Ve conclude similarly that .... .
Thi s has a lr eady been proved in Sectio n 8. Oue sees immediately that .... .
Th is res ult will now be deri ved computationally. M uch relevant information can be obtained directly from (3) .
Every measurable subset of X is again a measure space. This difficulty disappears entirely if .....
Vie first prove a reduced form of the theorem . This method was used implici t ly ill random walks.
40 41
Ie. After an object if it is sh ort: ?b. At the end (normal position):
\ \<.: will pnwe \.lIe t ilt:(lrl' ll! din~ctly widll'llt \!.,ing the lclllllla. The n.verages of Fll become small ill small neighbourhoods of I.
iJul: \\'e will prm'(' din'ctly a tlieorcm st:1:ing t!\:1t .....
2e. Between subject and verb, but afte r fi rst auxiliary-only sh ort claus es:
This is trl\e for evcry ~Cql!i"llC t ~ thllt ~!Jril!k..:; to J: lIicely.
Denne F!J an:llo::;ollsly as lLt~ lil:lit of .. ... The observed values of X will on the a verage cluster around .. ...
(2) deline's g 1II1aIllbig\!Oll~ly [or e\Try g'. This could in principle imply an n.dvantage .
f or simplicity, we will for the time being a ccept as F oniy C? milps .
ld . At the beginning-adverbs referrin& to the whole sentence:
Accordingly we are in effec t dealing with .. ...
Incidentally, we h:1\'c 110\\' construCtcd . ... .
Th e k nowledge of f is at best equivalent to ... ..
Actuallv. Theorem :3 "i\'CS lllore. llamely .... .
Finally,' en
shows (h a'~ j =!l. [:\'at: "At bst"]
Ncv-erl he kss, it tur ns t)ut that -:-.. .. -~----~ - - - - - - - - - - - - - - - -- -
The stronger res ult is in fact true.
It is ill all respects similar to matrix mu ltiplicatioll .
Nex t, let V Lc the \'('ct()[ Sp:1CC of .... . 2d. Between verb and object if the latter is long:
T'.Iore precisely, Q consists of .... . It suffices for our purposes to assume .....
Explicitly (Intuitively), this means that .... . To n. given density on the line there corresponds on the circl e the
Needl ess to say, the bounded ness of f was assumed only density given by .. ...
for simplicity.
Accord ingly, either f is asymptotically dense or ..... 3. INVERSION AND OT HER PECULIARITIES
Ie. In fro nt of adjectives-adverbs describing them: 3a. Adject ive or past partici ple after a noun:
a slowly varying function Lct Y be the complex X witb the origin r emoved.
prohabilistically significant problems T heorems I and 2 combined give a theorem .....
a method better suit ed for deal ing with ..... We uow show thn.t G is in the symbol indicated.
F and G arc similarly obtained from H. We conclude by the pn.rt of the theorem already proved that .....
F has a rectangularly shaped graph. The bilinear form so defined extends to ... ..
Three-quarters of this aren. is covered by subsequently Then for A sufficiently small we .... .
chosen cubes. [Note the singular. ] 13y queue length we mean the number of customers present including
the,customer being served.
If. "on ly"
The description is the same with the roles of A and [J reverse d .
\Vc need t lj(~ ()p(~nness only to provc the foll ow ing.
It reduccs to the statement that only for tbe distributio n F do th e 3b. Direct object or adjectiva l clause placed fal'ther than usual- wh en
m;lp~ Fi satisfy (2). [Note the inversion.] they are long:
III thi:, cbapter we will be concerned only with ..... We must add to the right side of (3) the probability that .. .. .
In (3) the Xj assume the values 0 aile! 1 on ly. This is equivalent to defining in tbe z-plane a density with ... ..
If (iii) is required for finit e unions ollly, the n ..... Denote for the moment by f the element sat isfying .. .. .
F is the r es triction to D of the unique linear Dlap .... .
We Ilced oilly require (5) to hold fo r b \Il1decl sels.
Tbe probabi li ty at birtb of a. iifetime exceeding t is n.t most ... ..
The prouf of (2 ) is simibr , n.ud will o lll y be illdicated briefly.
To pro\'c (3). it only remains to ver ify ..... 3e. In version in so me negat ive cl auses:
\Ve do not assu me that ... .. , nor do we n.ssnme n. priori that .... .
N e ithel' is the problem simplified by assuming f = g.
2a . At the beginning: The "if" part now fo llows from (3), since at no point can S exceed
In testing the character of .. ... , it is sO II It'lilllt':1 difli ' ult to ..... tIte large r o f X n.nd Y.
For n = 1,2, . . . , consider a famil y o f .... Tile fa c t th a t for no X does Fx coutain y impli~s that .... .

42 43
F(x) = G(x) for all x E X.
III case does the absence of a reference imply any cbim to
11 0
Let F be a nontrivial continuous lillear operator in V.
originali ty 011 my part.
2. Com ma required:
3d . In ve rsion-other examples:
The proof of (3) depends on the notion of !If-space, which has already
F is compact a nd so is G. been used in [4].
If f, 9 arc measurable. then so are f + 9 and f . g . We will use the map H, which bas all the properties required .
o 1
n y or
f - 1
Ican one expect to obtain .....
does that limit exist.
There is only one such f, and (4) defines a map from .....
In fact, we can do even better.
3e. Adj ective in front of "be" -for emphasis: In this sect ion, however, we will not use it explicitly.
By far the most important is the case where ..... Moreover, F is countably additive.
wlu ch more subtle are the following results of John. Fiualiy, (d) and ( e) are consequences of ( 4).
l';sse ll t ia l to th e proof are certain topological properti es of III. Nevertheless, he succeeded in proving that .... .
Conversely, suppose that .....
f .)l liJj 'c t o n lin g soo ner than in some other languages: Consequently, (2) takes the form ... ..
/':q lt nlity occ urs ill (1) iff f is cons tant. In particular, f also satisfies (1) .
Til n a tural ques tion arises whet llCr it is possible to ..... Guidance is also given, whenever necessary or helpful on further
III the foll owing app li cations use will be made of .... . reading. '
Recently proofs have been cOllstructed which use .... . This observation, when looked at from a more general point of view
3g. Incomplete clause at the beginning or end of a sentence: leads to ... .. '
Put differently, the moments of arrival of the lucky customers con- It follows that f, being COllvex, cannot satisfy (3) .
st itute a rellcwal process. If e = 1, which we may assume, then .....
Ttnthet than discuss tllis in full generality, le t us look at ..... \Ve can assume, by decreasing k if necessary, that .....
I L is impo rtant that the tails of F and G are of comparable magnitude, Then (5) shows, by Fubini's theorem , that .....
" Ht:lt lII e nt 1l1 ade more precise by the following inequalities . Put this way, the question is not precise enough.
Being open, V is a union of disjoint boxes.
WHERE TO INSERT A COMMA This is a special case of (4), the space X here being IJ(K) .
C:C1Lc,.,,1 r uks: Do no t over-use com mas-English usage requires th em less often than - In [2], X is assumed to be compact.
in many other languages. Do not use commas around a clause th~t defines (li- for all x, G(x) is convex .
mits, makes more precise ) some part of a sentence. Put commas before and after [Comma between two symbols.]
non-d efining clauses (i .e. on es which can be left out without damage to the sense) . In the context already referred to, K is the complex field.
Put a comma where its lack Illay lead to ambiguity, e.g. betwee n two sy mbo ls. [Comma to avoid ambiguity.]
l. Comma not required : 3. Comma optional:
We s llall now prove that f is proper. By Theorem 2, there exists an h such that .....
The fact that f has radi2..1 limits was pro"ecl in [4J.
For z near 0, \~e have .....
It is reasona ble to ask wllether this Ilolds for 9 = 1.- - ---
AI is til e se t of all maps which take values in V. If /; is sm~oth,the-Il M is compact.
Th ere is a polynomial P such that P f = g. Since h is smo-oth, A1 is compact.
T he clement give n by (3) is of the form (5). It is possible to u~e (4) here, but it seems preferable to .....
Let 1\1 be tile manifold to whose boundary f maps IC This gives (3)! because (sinc-e) we may assume .....
Tak an element all of whose powers are in S.
Int egrating by parts, we obtain .....
F is called proper if G is dense .
To do this! put ..... -
There exists a D such that DxyH wuellever HxyG.

X, Y, nlJd L nrc compact. Sillce f = 0 thCll M is closcd ---; Si nce f = 0, M is closed
X = :1;'G. \Ilien! F is defilled by ... .. [Or: Since /=0, wc conclude t hat M is cl osed ]
.. ... as it is shown in Sec. 2...,... ..... , as is shown ill Sec. 2
Thus (ll~!lr e /Thl!rdor e }-, we itn\"(' .... .
J:: ', u)' function being an e leme nt of X is co nvex
--+ [,'cry fUllction which is an element of X is convex
SOME TYPI CAL ERROR S S.-,ttillg 11 =]/, the equation can be .. .. .
--+ Se tting n = p, we can ... .. [Because we seL]
L Spelling errors:
3. Wrong word used:
Spelling; should be cOIJ.ji~tellt. eiilIl!r Dritisu or .-\lllcriciln t hrougliouL:
\Ve now gi ve few examples [= not many]
[Jr.: colom, llci,~h1Jo11r. celltre, fibre. labelled, I'1()ddlillg ....... vVe now give a few examples [= some]
Amer'.: co lor , llei:.,.;llboL ccnter , fiber, labe led, tllodelillg SUlllming (2) and (3) by sides...,... Summi ng (2) and (3)
;lll lluir.l'd appronc!l ~ a unified npproilch In the first paragraph...,... In the first section
a ;\ f s \:ch thnt ~- ;lll .\I such that _ ... .. , \Wich_Rro\,CS gur_thesi!? _______
[U~,! "a" or " <lU" a ccording to rrollllnciation .] ...,... ..... , \\!Jich proves our assertion (concl usion/sta te me nt}
[Thesis = dissertation]
2. Grammatical errors: For n big enough...,... For 11 la rge enough
Let f denotes ....... Let f denote To this ;"tim ....... To this end
Mos t. of them is ~ i\ Jos t of them are At first, note that -v-+ Firs t, note that
There is ;t finite numbe r of...,... There are a finite number of At last, wc obtain -v-+ Finally, we obtain
In 196-1 La...>: has slImvll ...,... III 19G4 La...x showed For every two elements...,... For any two elements
[Usc the past tense if a date is givell.] .. ... , what comp letes the proof...,... ..... , wbich completes the proof
The Taylor 's fo rmuh ...,... Taylor's formula [Or: the Taylor formu la] ...... what is impossible -v-+ ..... , which is impossible
The sertion 1 ---; Sect ion 1 4 . Wrong wo rd order:
Such lllilp exists ....... Such a map exists [But: for every such mar] The described above condition ---; The condition described above
In the C;loSe ,\1 is compact ...,... In case AI is compact
[Or: In the case where AI is compact] The both conditions...,... Botb conditions, Both the conditions
Its both sides...,... Both its sides
In case of smooth norms""'" III the case of smooth llorms
We are ill the position to prove""'" \Ve are in a position to prove The three first rows...,... The first three rows
'. The two following sets...,... The followin g two se Ls
F i ~ ,'q ual G ....... F is equal to G [Or: F cquals G]
F is grl';ller or eqllal to G ...,... F is greater thall or equal to G This map we denote by f --+ We denote tbis m a p by f
Continuous in t.he point x ...,... Continuous at x 5. Other exa m ples :
Disjoint with B ...,... D isjoin t from B W(! have (obtain} that [J is .... .
Eqllivalent with B ...,... Equivalent to B ....... We sec (conclllcie/cleullce/fimi/ infer} that [J is .... .
Indcpendent on B -v-+ Independent of B \Ve arc done --+ The proof is completc,
[But: depending on B]
Similar as B ...,... Similar to B
Simi larly to Sec. 2
As (J ust as} in Sec. 2
Similarly as in Sec. 2 -v-+ As is the case in Sec. 2
In m uch the same way as
in Sec. 2
In the end of Sec. 2 ...,... At the end of Sec. 2
On Fig . 3 -v-+ In Fig. 3

It, I,l l, 2:) , 1\6 generality, 10 print, 8
1\(cHdi nr, ly , 1:1 greater, :.15 same, IG, 40
,\f \ 1I .d ly , I V, 4 1 say, 11
half, :.15
adject i v, I clauses 9 have , 2G second largest, :.I(j
Ildverb i. 1 cla us es : 42 "have that", 15. 16, .\7 section, 4, 47
ad ve rbs , ,to hence, 15, 4G short-cu (s, 20
a fe w, :.1 7, 47 shortly, 7
a ll, 32 if necessary, 11 similar, 16, 46
a lso , 41 imper:ttive, Hi similarly, 46
a nu m be r of, 37, 46 in a position, 17, 45
since, 15, 47
a ny , ::13 independent of, 8, 4G
sm:tller, 35
as , 15 , 18 , 2 7 induct ion , loI smallest possible, :l(j
.s ic., :JO , H in fact, 13 so is, 10, 39
III [i,,;l, 4 7 infinitive, 10, 17, '27, 20
SOlne, 33
III I. LS ( , <1 2, 4 7 introduction, 4 succeed, 30
:tv id , :.10 inversion, 9, 10, :.1:.1, 42, 43
such , :.19 , 46
it, Itl, 19, 28, 40 such that, 8
b eca use , 15 it follows that, 15
being, g, 30, 47 th<1.t, 38
both, :13, 47 largest, :.16 th e, 24
brackets, 8 last but one, :.18 the one, :.Itl
lat ter, :.19, 40 th e reforc~ , 1;;, 11, '1<i
cardin:tl numbers, :J.1 lC <lSt , :H, :3(j
case , 40, 46 th e re is, :\:l
less, :.IS t hese , :j<J
co nl ra d ic ti o n, 11, Hl let, 46 thesis, 18, 20, ,17
d 'n o te , 7 likely, 28 the two, :H , :J9
d e pe nd in g o n , tl matrices, :l8 this, 40
d ill('l' , :Hi , 'J7 more , :.I(j th is 1<1.5t, :lV
c1i njo in t fro lll , 4G mos t, :.14, :.Ib, 46 those, :li'!
d is t in c t , :)7 multiple, :.16 thus, 15,41,46
each, ::\2 t o be defi ned, 9, 28
need, 17, 29
eithe r , 34 too, 27
neither, 9, :.14, 4:.1
en<1.b le, 29 t o thi s end, H, 47
nex t-to-last, :.18
en o ug h, tl, 27 twice as long iIS , 36
no, :.1:.1
eq ua l, 37, 46 t wo-thirds , 35
no greater, 35
e rro rs , 46 typefaces, 8
none, :.13
every , 'J2 , 4 7 nor, 9, 10, 43 ullion, 25
f,'w , :l7, 4 7 number in g, 26 , 38 unique, 24, 34
f,\V,;r, :,5 unlike ly , 28
[i ,, :t lly. '12 , 47
"obtain that" ~15 16, 4,
of, 25, 26, 29' .
u pTo , :35 , :>8 - - - - - -
fin is h , :50 one, 23, 38 what, 15, 18, 47
k- fo ld , :J6 only, 29, 42 w hi ch, 15, Hi , 47
fo llowin g, 13, 19, 20, 47 ordinal numb e rs, 25, 34, 38 with,2 6
fo r, 11, 2tl wOTth,30
forn1e r , 39, 40 p aragraph, 4, 47
worth while, :10
for s h o rt, 7 participles, 30
fr a ctions , :.15 percent, 35


The _booklet is intended to provide practical. help for

authors - ofrnatIierriatic8]- papers. It will -be - useful
both as a guide for beginners and as a reference book
for experienced Writers.

The first pa...r-t of the booklet provides a useful collection

of ready-made sentences and expressions occurring .
in mathematical papers. ' The examples are divided
into sections according to their use (in introductions,
definitions, theorems, proofs, comments, references
to the lilerature, ackrlOwledgments, editorial correspon-
dence and referee's reports). Typical errors are also
pointed out.
The second part concerns selected problems of English
grammar and usage, most often encountered by
mathematical writers. Just as in the first part,
an abundance of examples are presented. all of them
taken from the actual mathematical texts.

The index enables the reader to find many particular

pieces of information scattered throughout the text.

Jerzy Trzeciak, formerly of Polish Scientific Publishers,

is now the senior copy editor at the Institute
of M?-thematics, Polish . Academy of Sciences. He is
responsible for journals including Studia Mathematica;
Fundarri.enta Mathematicae. Acta Arithmetica and others.

ISBN 83-85694-02-1