On Frustration of the Majority by Fulfilment of the Majority's Will

Author(s): G. E. M. Anscombe
Source: Analysis, Vol. 36, No. 4 (Jun., 1976), pp. 161-168
Published by: Oxford University Press on behalf of The Analysis Committee
Stable URL: http://www.jstor.org/stable/3327179 .
Accessed: 15/09/2014 21:34
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Oxford University Press and The Analysis Committee are collaborating with JSTOR to digitize, preserve and
extend access to Analysis.

http://www.jstor.org

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

ANALYSIS 36:4

JUNE 1976

ON FRUSTRATION OF THE MAJORITY BY

FULFILMENT OF THE MAJORITY'S WILL

G. E. M. ANSCOMBE
HERE matters are decided, yes or no, by a majorityvote, the
decision on each matteris the will of the majority. But it is also
possible that the majorityshould vote in the minority on a majorityof
the questions. This may be shewn in a table.
F
G
B
E
A
C D
I
H
J K

I
2
3

I
O

0o

I
O

I
I

C0
I

0O

I
O

I
I

I
I

I
I

I
I

O
I

I
I

o
I

o
O

o
O

o
O

I
O

IO

I
I

II

Here we have eleven voters, A-K, voting on eleven questions. Seven

of them, A-G, vote in the minority in a majorityof the decisions: A-F
in seven out of the eleven cases, G in six. The majorityis always 6-5.
These figures can of course be varied.
If we imagine an ideal democracywith a whole population voting
directlyon all questions,therewill obviously be room for much variation
in results over a long period, all of which however conform to the
description: the majorityvotes in the minority in a majority of cases.
This fact, I thought when I stumbledon it, must be familiarto voting
experts. But I have not found it remarkedupon. It sometimes startles
people, eliciting the reaction: 'But doesn't this make nonsense of
democracy?'
In the West, and perhapsin the whole world where Westernforms of
education prevail, men are brought up in a conviction of the unique
fairness of democracy. It is even conceived to be as it were the sole
legitimateform of government. 'It's not democratic'is a condemnation.
Pope Pius XII once spoke in a Christmasallocution of the right of
democraciesto defend themselves by whatever means they might think
necessary. One could hardly have better proof of the pervasivenessof
IGI

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

162

ANALYSIS

the attitude,if of all people the Pope of Rome should speakas if democracies had some special entitlement. Outside governments, we notice
that methods of proceeding and alterationsin arrangementsare often
proposed on no other ground than that they will be "more democratic".
I know a university,for example,where on these grounds representation
in the authoritativeassemblywas made proportionalto the numbersin
the different departments. Large departments proceeded to forbid
expansion to or even to reduce small ones, increasing themselves as
every opportunityoffered.
In the Liberaltradition of the West, democracyhas certainlybeen
favoured, and belief in it instilled as a fundamentaldogma. But there
are other strandsbelonging to that tradition. One of them is a concern
for minorities. Upholders of this, it seems, were conscious that a
majoritymay be a cruel and oppressive bully to a minority. In consequence "protectionof minorities"or "of the rights of minorities"also
had a certainsacredflavour. This it perhapssometimesretainsmore in
the United Statesthan in England, where a common denigratoryphrase
nowadaysis 'merelysectionalinterests'. However this may be, I notice
in my own country at least a hardening of sentiment exclusively in
favour of democracy, which is sometimes also called 'majority rule'.
to hear of
Such is the sentimentin favour of this, that it is not astonishing
an apparentlyreasonableand well-disposed man who says: someone
who is not preparedto accept a majoritydecision, or a law enactedby
democraticprocesses, ought to leave the society. And this, without
any limitation'sbeing put on what the matterof the law or decision may
be. We see that from such a mind the Liberaltraditionof concern for
the "rights of minorities"has been nearlyexpunged.
of the idea that decision accordingto the
Let us considerthe rationale
will of the majorityis superiorto any other. That a particulardecision
according to the will of a majority may be inferior, in the sense of
stupider,is evident. I takeit thata decisionis stupidif its implementation
is undesirablefrom the point of view of the very people who wanted it.
They would have it: but when they get it, it is evident that they would
rathernot have it; if they had only realizedwhat it would be like for their
decision to be implemented,it would not have been made. There was
anotherdecision available,perhaps:if the power of making the decision
had not been taken away from some particularperson and vested in the
majority. And it is now recognizedthat thatdecision would have been
better.
The superiorityof the majoritydecision cannot reside in its necessarilysuperiorwisdom. A superiority,even when the decisionis stupid,
is rather thought to be in this: participation in the decision-making is
itself valuable; or, perhaps one should say, is itself a value. Such a value
is mystical of course. I don't mean a criticism in saying so. But

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

ON FRUSTRATION OF THE MAJORITY

163

consider a man who always or most often finds himself voting in the
minority, and who judges that most of the decisions takenare extremely
harmful. Why may he neverthelessextol this method of arrivingat decisions? He may think that alternativesavailablein the situationwould
lead to even worse ones, but that is not to put a value on participationas
such. He may feel some satisfactionwith the method in that there was
alwaysa contraryvoice, and that his madepartof it-but why? Honour
satisfied? That is mystical. A nucleus of possible opposition and
difference of direction? There may be none such: the minorities in
which he votes may be inconstant sets of people. Does he simply have
a "we" feeling about the decisions, bad as he thinks they are? 'We
decide for ourselves'?--even though his vote was always contrary to
the decision? There is indeed such a sentiment, but, once again, it is
mystical. He has reason to feel as oppressed by the authority of the
majorityas he might by the authorityof an autocrator superior body,
who "hand down" decisions that he has to accept.
Now there is a quite different belief to yield the superiority of
majority decision except, indeed, where it is stupid in the sense I have
explained. Let us suppose a set of people who have to determine,or
for whom it must be determined,what they shalldo or what shallhappen
to them collectively. Suppose that they are all going to go somewhere
in a vehicle; the questionis, where? Some would like to go to one place,
some to another. They can't go to both. But they must all go together
where they go. Doesn't it then seem reasonablethat they should go
where the majority of them want to go? All cannot be satisfied; in
this way, perhaps,most of them will be. Complicationsarise, of course,
where the choices are multiple; but let me assume just a simple choice,
made by everyone, between a pair of alternatives. Then the majority,if
there is one, will be absolute. Now it does seem most reasonablethat,
where all cannot have their way, the greater number should not be
frustrated.
Note that this defence of majority decision in conveniently simple
cases does not assume that there is actuallya value in people's sharing
the job of makingthe decision. For the argumentonly concernspeople's
getting what they want.
The reasonablenessis meant to consist in a sort of fairness. If the
desires of the majorityare not to prevail, then a few perhapsget what
they want at the expense of a greaternumber who do not. This seems
to put the desires of those few in a privileged position; which is intrinsically unfair. When the majorityget what they want, they do so at the
expense of the minority'sfrustration,but at least nobody is privileged;
for the upshot results from making the desire of each count equally.
Ex bypothesino upshot could satisfy all equally; a method which derives
the upshot from an equal weight attached to each person's desire must

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

ANALYSIS
164
achievethe satisfactionof the desiresof the greatestpossible numberand
so be both the best and fairest.
This belief about the best and fairest upshot is independent of the
method adopted to secure that upshot. If the method adopted can be
that of majorityvote on simple alternativesthis, it is arguable, is the
best method becausepresumablypeople are themselvesin the best position to know what they want. Even though they may sometimesmake
mistakes,that is worth putting up with, becausethey will surely,at least
for the most part, vote for what they reallydo want. But the method of
majoritydecision is here espoused, not becauseof any intrinsicvalue in
people's sharing in making decisions, but on the assumption that it is
the best availablemethod for securing the upshot. The argumentthat
this is the best and fairest upshot does not depend on this being the
method for securing it; but, stupidity apart,this method would secure
it more surely than any other.
Note that that way of looking at the merits of majority decision
startsoff with an assumption:Namely, that a decisionhas to be madefor
a decision has to be made which determineswhat
people as a collection;
has
or
in some matter. This was why I supposed the
does
everyone
case where it was given that a party of people were in for going somewhere in a vehicle. We should distinguish between this set of people
and a set of people in a house who happen to have a practiceof determining daily by majoritydecision what all of them are going to do that
day. Here one might say: 'Why not let each choose his own activity?'
and the argumentI gave purportingto show that the will of the majority
should prevail will have no force.
At least, it will have no force until we take into account that some
people's desired activities include making other people act in various
ways. The philosopherthat I referredto, who thought that one had no
right to remainin a society if one was not preparedto acceptits majority
decisions, did not stipulate that the decisions should concern only
actions and fates where a decision for a whole collection of people is
necessary. This is perhaps realistic for the following reason: some
people desire to control others, to dictate to them, as also to get them
(in various ways) to do things they would not otherwise do. Now if
our initialassumptionis: each person's desireshallhave as much weight
as anyone else's in calculatingwhat to do-then these types of desire
can't be excludedfrom consideration. As soon as there is a proposal to
interfere with X's activity, someone presumablywants to do so. If
interferingcan come up for majoritydecision at all, and it is proposed,
some want and some perhaps do not want a decision that that set of
people collectively interferewith X. Even if those prevail who do not
want to interferewith X and do not want it to be a collective decision
to interferewith X, it has come about that therenecessarilyis a collective

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

ON FRUSTRATION OF THE MAJORITY

I65

decision whetherto interfere with X.1 What reason could be offered

against the paramountcyof the will of the majority? The "given", that
they are "in for" some decision (as it was a "given" that my set of
peoplewere going somewherein theirvehicle),has been contrivedmerely
by the question's coming up for majoritydecision at all.
Contrast the situation of an autocrat; he is the source of all the
decisions about collections of people. He is also so powerful that he
can interferewith any arbitraryX if he likes. If he does so (or abstains)
his action remainsan action just relatingto X; it is not transformedinto
a decision about people collectively.
I do not want to steer further out into these deep waters. For my
quarry is something quite different. We had an argument purporting
to show that where there are a lot of people collectively in for one of a
pair of alternatives, the best and fairest decision was the alternative
preferredby the majority. It seemed bestbecause we are envisaging no
criterion of what is good except that it is wanted, or of evil except
frustrationof desire; it seemedfairestbecauseeveryone's desire is given
equal weight. The tradition of concern for minorities is nourished by
far differentconsiderationsthan these. What those may be is not at all
my concern. I am impressedby the argument,even if I would like to
draw from some other source some principleof limitation of its applicability.
But now-the tablepresentedright at the beginning seemsto damage
this argument, which at first sight seems so impressive. For what do
we have here? Each decision is made according to a majority vote.
Each decision, then, (folly apart)would seem to be the best and fairest.
But when we look back and forth through many decisions, a different
picture emerges. No need to worry about minorities yet!--The
majority is frustratedmore often than not! Doesn't this destroy the
original argument? For what was appealing about the satisfaction of
the preferenceof the majoritywas preciselythis: grantedthat all cannot
be satisfied,the best is that most should be; the principleof letting the
desire of each have equal weight will both secure this and in itself be
fair, because none is given privileged considerationin the decisions as
they come up.
But the table shows that the best and the fairest (by that criterion)

may part companywhen we look througha multitudeof decisions.

However paradoxicalit may be, it seems that the principleof the
majoritygettingwhat they wantallowsthe majorityto get what they
do notwant in a majorityof cases. The air of paradoxis only partly
producedby its lookingas if some set of people,called'the majority',
werethe samein both partsof the observation-asof course,it is not.
1
May there not be some liberties which only stand so long as they are not brought into
consideration?-Bringing them into consideration is alreadymaking them subject to decision.

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

166

ANALYSIS

That it is not only shows that what seemed paradoxicalis not viciously
so at all. It does not show that there is any error in the judgment:
'So in this way the majoritymay be frustrated!'
Note that this is not a point just about the method of decision by
majorityvote. I have been at painsto distinguishthat from the principle
of the decision'sbeing what the majoritywants. The method of majority
decision may be regardedas a method for securing the latter. But the
paradoxto which I have pointed appliesto the principleitself as well as
to the method of majoritydecision.
It will now be plain why I have assumedthe simplest possible kind
of issues as the matterof decision, and assumeda perfect democracyin
which all concernedvote quite directly. It is a familiarpoint thatdifferent
voting procedures,multiplechoices, what is motion and what is amendment, and other such matters,make such a differenceto results that the
will of the majoritycan seem to be uncapturable,or even non-existent.
Also, we don't have pure, but rather delegated, democracies. One is
inclined to say: 'If only everything were simple, the total population
concernedvoting on straightYes-No alternativeswithout any question

of rigging(bynatureor design)fromtheorderin whichtheyweretaken

-if only thingswerelike that,thenwe'd havethe idealsatisfied(to the
extentthat peopleweren'tstupid)of the greatestpossiblenumberof
peoplegettingwhattheywant.' Whatthe tableshowsis thatthis is an
illusion.Weoughtthereforeto giveup anyideaof achievingas good an
as possibleto that ideallysimplekind of result,which
approximation
canitselfbe achievedonlyin verysimplesituations.Wherethe idealis
actuallyachievedthe achievementof the furtheridealis by no means
guaranteed.
It maybe, indeed,that thatcriterionof fairnessappealsto us: the
criterionby whichthe desireof eachhas equalweightin eachdecision.
But when we look backandforththroughmanydecisions,a different
ideaof fairnessin satisfactionobtrudesitself. Thusa parentof a large
familymightsay:'Eachchildshalltakeitsturnto decidesuch-and-suchthatwill be fair.'
Someonemaysay: 'Thepossibilitywhichyou havepointedto does
not matterat all, even if it is oftenactualized.(And,by the way, you
don'tknow whetherit is often,never,or very seldomactualized.')It
is simplyevidentthatmajoritydecisionis a soundway for a groupof
reasonable,
equal,autonomousbeingsto determinewhatis best to do.
Thatbeing so, it is of no importif, over manydecisions,the method
of eachquestion
leadsto this result. Whatmattersis the determination
in the bestway.'
To see both the force and the weaknessof this comment,let us
1 It would be of some
interest, perhaps, to scrutinize the voting in the English Parliament of the early eighteenth century with just this question in mind.

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

ON FRUSTRATION OF THE MAJORITY

167

imaginea methodof majorityjudgmentinsteadof majoritydecision.

Ourcasewill be ratherartificial,but no matter.We will supposesome
matterof fleetingobservationof whichconstantrecordis needed,and
where the observationalwaysdecidesbetweena pair of alternatives.
The matterbeingdifficult,a teamof observersis alwaysemployed,and
what is put down in the recordeach time is the observationof the
majority.This gives rise to just the samepossibilityas we have been
discussingfor decisions. When you look througha series of such
results,you may find that the majorityof the team have made the
minorityobservationin a majorityof the cases. But therewould be
whatis to
nothingdisturbingaboutthis. If themethodfor determining
in
the
that
will
record
deserves
not
of this
be
because
criticism,
go
if
is
that
is
a
the
actualized,
possibility;
possibility
just "giddyparergon",
a merelyincidentalcuriosity.It has no tendencyto invalidatea "principleof majorityobservation".Forall thatmattersis the determination
of each resultat the time. The same might hold if the principleof
majorityjudgmentwereusedin trialswitha constantpanelof judges.'
Suchno doubtis the thoughtof our objector. He is saying:'All
thatmattersis the determination
of eachcauseat the timein the best
But
that
can
be
way'.
hardly right,excepton the assumptionthatthere
is an objectivelybest or good decisionto make,the best way of determiningwhichis to takea majorityvote. By 'anobjectivelybestor good
decision'I meana decisionwhichis bestor good independently
of what
want.
if
can
an
such
Well, anyone give
account,morepowerto
people
him! Butas we havebeenconsideringthe matter,all thatis in question
is what, given theircircumstances,
peopleactuallywant,whichin the
individualcasea man gets by his choiceor actionif it is not stupid.2
But in the casewherepeoplemustdo or suffersomethingcollectively,
the idea was that the majorityshould get what they want. So the
frustrationof the majoritywhenwe considera whole seriesof majority
decisions-the failureof the majorityto havegot whattheywantmore
oftenthannot-cannot be regardedjustas a mildlycuriousby-product.
Indeed,we mightsee in ourproblema wayof givinga realmeaning
to the rhetoricallyattractivebut actuallysenselessexpression'the
greatesthappinessof the greatestnumber'.Thisis of the sameformas
'thegreatestnumberof wordsin the shortesttime',3andwhenwe have
two thingsactuallycountableandmeasurable
like wordsandtimeswe
see that it makesno sense. But we might say that what the famous
phraseintendscanbe givena sense:at leastit canif 'minimalfrustration
1 My colleague J. E. J. Altham has remarked to me that the smallest number for which
this result is obtainable is five.
2
This is not in fact true, because 'the race is not to the swift etc.' But the falsity of it
cannot play a part in our considerations.
3 The comparison was passed on to me by Gareth Evans, but I think he said he did not
invent it.

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions

168

ANALYSIS

of the majority' has one. How this could be planned for is indeed
difficultto see. It might have to involve departingfrom the method of
majoritydecision by the population.
That, of course, we do depart from anyway; we have delegated
democracy,and use majority decision in committees and Parliaments.
The aim of their decisions is not supposed to be satisfying the desires
of the majorityof their own membership. Here,then, the objectionthat
I considered has a point; there is not necessarily anything to worry
about, if the majorityof a committee should turn out to have voted in
the minorityin the majorityof issues it has decided.
Let us turn, then, to the question of satisfying the majoritydesire,
on any given issue, of the whole of a community or population. As
remarked,my table need not be a voting table: it may representa set of
questions being decided (no matter how) in accordancewith, or contrary to, the will of people (or groups) A-K. Here then we see the
naked point: the majoritymay be satisfiedon every issue, while nevertheless the majorityis frustratedover a majorityof issues. There is thus
the possibility of a certain technique of tyrannywhose every measure
has the support and is truly in accord with the desire of the majority,
those whom any given measurehurtsbeing in the minority;or again,one
by one "merely sectional interests" are damaged. Since everyone not
wretchedlyisolated belongs to several"sections",it will be possible for
the tyrantto damagethe interestsof anyone or any group (that does not
support him, say) while truthfully claiming "democratic"support for
his measures. Or again, the process of damage to sectional intereststhat is, to a majorityof the population-may occur in a democracyin a
haphazardfashion and without design, always in accordancewith the
will of the majority.
University
of Cambridge

? G. E. M. ANSCOMBE 1976

EGOISTIC HEDONISM

By LAWRENCE CROCKER
ESIRES differ from each other along several dimensions. They
differin intensity and probably in a hundred distinct ways with
respect to their phenomenologicalfelt character. Some desires are in
part linguistic in character-as when we desire something under a
certaindescription. Other desiresseem to have no linguisticcomponent.
For the purposes of this paper I will abstractfrom most of these ways

This content downloaded from 131.94.16.10 on Mon, 15 Sep 2014 21:34:59 PM

All use subject to JSTOR Terms and Conditions