The "Roster Device": J. S. Mill and Contemporary Elitism

• DOI: 10.1177/106591296802100102
• Author: Willmoore Kendall,George W. Carey
• Date: 01 March 1968
• Published date: 01 March 1968
• Subject Matter: Articles

THE "ROSTER DEVICE": J. S. MILL AND
CONTEMPORARY ELITISM
WILLMOORE KENDALL, University of Dallas
and
GEORGE W. CAREY, Georgetown University
OLITICAL
THEORISTS, when faced with certain types of problems, often
p
move in their thinking from the tacit assumption that one is entitled, having
JL
fixed attention on any human quality or characteristic, to take for granted
the following premises: (a) That the individual members of any political society
possess that quality or characteristic in different degrees. (b) That every political
society may, therefore, for certain purposes be treated as a huge roster, on which the
names of the society’s individual members have been duly entered in &dquo;correct&dquo;
order -
that is, with that man who possesses most of the quality in question at the
top and that man who possesses least of it at the bottom and everybody else just
where he belongs in between the top and the bottom. (c) That, finally, one may
draw horizontal lines through this roster and so divide the membership of society
into significant groupings, especially for the purpose of allocating the individual
members of society to the various functions that must be performed within it. The
quality in question may be, as in the early pages of The Laws, the capacity to carry
one’s liquor well, and the roster may then be used, as Plato uses it, as a basis for
the proposal that masters of ceremonies at banquets be drawn from the top of the
relevant &dquo;roster.&dquo; Or the quality may be wealth, as in Aristotle’s Politics, and the
roster may then be used for a variety of purposes such as classifying types of gov-
ernment, or assuring stability, or preventing revolution. We could name any num-
ber of examples.
Now there are, we shall contend, many unnoticed difficulties involved in the
use of this &dquo;roster&dquo; technique. It can, let us notice first, be brought into play in
two quite different contexts. Plato, in the illustration cited above, is &dquo;model build-
ing&dquo; : he is speaking of a hypothetical situation, that is, no situation in particular,
and his roster is merely a theoretically satisfactory solution to a purely speculative
problem. He is, on the showing of the present article, operating within the &dquo;rules
of the game.&dquo; But if, say, a capacity-to-govern-well roster is put forth, as it usually
is, with the assumptions that those at the top of the roster can actually be identified
in a particular place and time, great difficulties arise. It now becomes incumbent
upon the theorist to offer proof either that the roster he appeals to actually exists
(like that which lists West Point graduates according to their &dquo;standing&dquo; ) , or that
it can be more or less readily produced, if and when needed, without expenditure
of time, effort and money out of all proportion to the benefits it is likely to confer.
To the extent that the theorist attempts to incorporate existential reality in his
asumptions, that is to say, he must assume responsibility for proving that his roster
is not impractical, and must recognize, to go no further, that the various rosters one
may posit on the level of abstract theory differ enormously as regards the resources
of time, energy, money, etc., it would take to construct them in an actual political
20

---

21
situation. This for several reasons: The &dquo;quality&dquo; in question, for one thing, may
or may not be &dquo;measureable,&dquo; as, for example, tallness is measureable. If measure-
able, it may be more easy to measure, or less easy. If not measureable, it may still
have quantitative dimensions, like the capacity to compose great music, of which
Beethoven presumably had more than, say, Rossini.’ Again: the quality in question
might be, e.g., &dquo;kindness,&dquo; but the relevant roster could be constructed, for practi-
cal purposes, only after agreement has been reached as to how to decide who has
more of it and who
has less of it. And there are further difficulties, which arise even
in connection with qualities we are accustomed to think of as readily measureable.
A &dquo;richness&dquo; roster, for instance, reaching from the richest to the poorest member
of a society, would, at first blush, seem easy to construct because we think of wealth
as measureable. But in some societies, including the United States, people do not
necessarily know how wealthy they are, and so might, if &dquo;polled,&dquo; misinform the
roster-builders. Or even if they do know how wealthy they are, they might attempt
to deceive the roster-builders, and describe themselves as richer or poorer than they
in fact are, thus getting themselves on the wrong line. Finally, even if we assume
an accurately constructed roster, baffling problems arise when it comes to drawing
significant lines across it (i.e., dividing it into significant sections). For rosters differ
enormously as regards the extent to which they yield up groupings that are other
than purely arbitrary.2
Another and less obvious problem is this: Political theorists sometimes talk as
if one and the same roster could simultaneously list the individual members of a
society in their right order not merely with respect to a single quality, but two or
even three qualities. John Adams comes quickly to mind here as an example, since
much of the controversy surrounding his &dquo;natural aristocracy&dquo; may be seen to turn
on just this point. At one juncture Adams asserts that his natural aristocrats (that
is, those at or near the top of his aristocracy roster) are those who &dquo;rate high,&dquo; as
we would put it today, in &dquo;birth, fortune, and fame.&dquo; Push him further, and the
relevant &dquo;qualities&dquo; turn out to be, among others, &dquo;merit,&dquo; &dquo;talents,&dquo; &dquo;wisdom,&dquo;
&dquo;learning,&dquo; &dquo;eloquence in council,&dquo; &dquo;confidence and affection of the citizens,&dquo; &dquo;sense
of duty,&dquo; &dquo;virtue,&dquo; &dquo;advantages of education,&dquo; &dquo;respect of the public,&dquo; &dquo;skill,&dquo;
&dquo;industry,&dquo; &dquo;great knowledge of public affairs,&dquo; and &dquo;honour.&dquo; 3 Now: Adams’
roster is intelligible, we see at once, only if (a) the qualities in question are insepa-
rable and, in addition, come in constant proportions, or (b) he can show us how
to &dquo;trade&dquo; x units of one quality for y units of the second, z units of the third, and
so on.4 (One of Adams’ natural aristocrats might, on the basis of such trading,
1
Cf. R. G. Collingwood, An Essay on Philosophical Method (Oxford: Clarendon Press, 1933),
pp. 69-72. Collingwood distinguishes between that which is "measurable in principle"
and "measurable in practice." He believes that philosophical concepts are, typically,
unmeasurable in practice.
2
Many of the difficulties noted here also plague our social scientists, particularly our sociolo-
gists, when they try to deal with the problem of "social stratification." In our terms,
they are engaged in constant controversy as to how many horizontal lines must be drawn
on such and such rosters in order to delineate significant "classes."
3
See his A Defense of the Constitutions of Government of the United States of America, par-
ticularly Volume III.
4
Leadership and power studies frequently illustrate this difficulty. Alvin W. Gouldner notes,
"Those proposing trait units [of leadership] usually do not suggest which of the traits are

---

22
turn out to be an uneducated and illiterate but supremely virtuous man). Adams
may well have supposed that the selfsame men could be counted on to rank very
high with respect to each and every one of the qualities in question; but once such
an assumption is hauled out into the open we see at once that it is unwarranted.’
Given the number and disparate nature of the qualities he compressed into his
aristocracy roster, we should expect them, in the absence of evidence to the con-
trary, to turn up in all manner of combinations. A &dquo;multi-quality&dquo; roster, then,
places upon the theorist using it very special demands; and any failure on his part
to meet these demands, as we have just noted them, renders his roster virtually
meaningless.
Yet another difficulty associated with the roster technique leaps to the eye when
we
find a theorist drawing a single line across a putative roster, and claiming for it
that, so to speak, it divides the sheep from the goats. Consider, for instance, the
following age-old argument against majority rule: Assuming a roster that ranks
individuals on the basis of virtue, we see at once (so we are told) that those at the
top of the roster are as a matter of course a minority of the members of society;
and, that being the case, that majority rule will result in the rule of the less vir-
tuous.6 But if we look a second time we see that the argument cuts both ways (i.e.,
can easily be turned against the man using it) , as follows: those who possess little
virtue, that is, the wicked, are a minority in society; and, that being the case, major-
ity rule will evidently give us rule by the more virtuous. The truth is, of course, that
neither argument proves anything at all, because each involves what we may now
identify as an improper use of the roster device -
improper because there is no
justification for treating individuals above and below a certain point on the roster
as constituting a homogeneous group, and no justification for assuming, as the argu-
more important and which least. Not uncommonly, lists of more than ten traits are
presented....