Date: Sun, 1 Mar 1998 16:36:08 -0700
Subject: BHA: rts2-35


 The Logic of Scientific Discovery 199


5. OBJECTIONS TO THE ACCOUNT OF NATURAL NECESSITY
PROPOSED


Having outlined the principle advantages of my account of
natural necessity and natural kinds (on pp. 183-5 above),
I now want to consider some objections to it.  In Chapter
4  I will consider the conditions of the plausibility of
these objections.


     The chief Humean counter-arguments may be put in the
form of three theses :-


     (i) there can be no, or at least no knowledge of,
necessary connections between matters of fact;


     (ii) if there were necessary corrections between
matters of fact they would have to be known a priori; so
science could not be empirical;


     (iii) men are never directly aware of any causal
power or agency or necessary connections between matters
of fact, so these concepts cannot be justified by
experience (though they may be explained by it; or are,
for the neo-Kantian, imposed upon it).


     The argument for thesis (i) is typically constructed
as follows: there is nothing inconsistent about the
supposition that the cause of a phenomenon, say putting a
kettle of water on the stove and heating it, should not
be accompanied by the effect in question.  It is
conceivable that water might freeze instead of boil when
it is heated.  Now thesis (i) is, as stated, highly
ambiguous.  It is not clear whether it is an ontological
or an epistemological thesis (this ambiguity is of course
explicit in the way I have formulated it); whether the
`necessity' is logical or non-logical; and whether the
`matters of fact' are events and states of affairs or the
statements describing them.  Before returning to the
argument, we must see exactly what is at stake in it.


200 A Realist Theory of Science


     Now, it will be remembered, that for the
transcendental realist to say that a sequence E_a.E_b is
necessary is to say that there is a generative mechanism
at work such that when E_a occurs E_b tends to be
produced (is produced in the absence of interfering
causes).  If there is such a mechanism the sequence is
necessary; and its necessity is quite independent of any
knowledge of it.  To analyse the necessity of the
connection in terms of our knowledge of the necessity of
the connection would be to commit the epistemic fallacy
(see 1.4 above).  There is a real difference, quite
independent of men, between the fact that when I heat the
kettle of water it boils and the fact that it boils when
the time is half-past two or the colour of my socks is
blue.  The necessary connections that bind some but not
other events together (which are the enduring mechanisms
of nature) are quite independent of our knowledge of
them.


     Statements clearly belong to the epistemic not the
ontological order; and logical connections hold only
between statements, not between events and states of
affairs.  Hence the prima facie absurdity of those who,
in attempting to refute Hume, try to establish that nomic
necessity is, or may be, a species of logical
necessity.73  Natural necessity is not logical necessity.
Natural connections hold between things, events, states
of affairs and the like; logical connections between
propositions.  Moreover there could be a world without
propositions, in which the concept `logical connection'
had no application.  The laws of logic are not features
of the world, nor are they imposed upon it.  Rather, we
must say: the world is such that changes in it can be
consistently described.


     Neither natural necessity nor knowledge of natural
necessity can be identified with logical necessity.  But
our capacity to deduce the Wiedmann-Franz law from
Drude's theory of electrical conductivity may serve as a
*criterion* of our knowledge of the necessity the theory
describes.  I suggested in paragraph 3 above that three
levels of knowledge of the objective world order can be


 73 See e.g. A. C. Ewing, The Fundamental Questions of
Philosophy, pp. 159-81 and B. Blanshard, Reason and
Analysis, Chaps. 11-12.  Cf. also. N. Maxwell, `Can there be
necessary connections between successive events ?',
B.J.P.S. Vol. 19 (1967), pp. 1-25; and M. Fisk, `Are there
Necessary Connections in Nature?'  Philosophy of Science,
Vol. 37 (1969), pp. 385-404.


The Logic of Scientific Discovery 201


distinguished in the development of science; so that
statements can be classified as definitions, deductive
consequences of true theories and simple protolaws
according to the position they occupy (at any moment of
time) in the development of our knowledge.  Hence the
deducibility of a tendency from a nature may serve as a
criterion at the Lockean level for our knowledge of
natural necessity, just as a correct definition may serve
as a criterion at the Leibnizian level for our knowledge
of natural kinds.  But whether or not a sequence of
events is necessary is quite independent of the logical
status of the proposition used to express it; which is a
function of the way it is described in the context of our
knowledge; which in turn may be shown to have a certain
rationale in the development of science.


     Some causal statements expressing necessary
connections are logically necessary and some are
logically contingent.74  For the Humean, however, logical
and natural necessity are easily confused.  For given the
isomorphic relationship between knowledge and the world
assumed in empirical realism and restricting our
knowledge of nature to the protolegal phase of science
(see page 172 above) he naturally comes to regard
relationships between events as characterizable in the
same kind of way as the statements expressing their
relationships are at that phase typically, though not
invariably, characterized; namely as contingent.  But it
is into this very same trap that defenders of the
entailment view of natural necessity fall.


     I shall construe thesis (i) as an epistemological
claim to the effect that knowledge of necessary
connections between events is impossible.  And I will
attempt to refute it by arguing that unless there were
necessary connections between some (but not other)
events, science would be impossible; and that in science
the most stringent criteria for knowledge of natural
necessity may be satisfied.


     Unless there were necessary connections between
matters of fact neither confirmation nor falsification
would be possible.  For without them no confirmation
instance adds any probability whatever to any inductive
instance.75  On the other hand for it to


74 Cf. `Tania pushed the door open' logically implies
`the door opened'. As Davidson has put it: `the truth of
a causal statement depends upon what events are
described; its status as analytic or synthetic depends
upon how they are described', op. Cit., p. 90.

75. Cf. M. Fisk, op. cit., p. 390.


202 A Realist Theory of Science


be rational to reject what is falsified it must be
assumed that a hypothesis which has been false in the
past will not suddenly become true in the future.76
Whether the conclusions of inductive arguments are
weakened to probability judgements or it is denied that
science is inductive in nature there must be necessary
connections between matters of fact.  Such necessary
connections are provided by enduring mechanisms.
Moreover, if experimental science is to be possible,
there must be necessary connections between some but not
other events. This implies a dynamic principle of
indifference: to the effect that mechanisms not only
endure but are transfactually active.  Neither their
enduring nor their transfactual activity is in need of
explanation.


     Unless there were necessary connections between
matters of fact we could have no knowledge, even
particular knowledge (in as much as this depends upon
inferences beyond what is immediately observed), of the
world.  For science to be possible then the world must
consist of enduring and transfactually active mechanisms;
and there must be necessary connections between some but
not other matters of fact.


     Natural mechanisms are of course nothing other than
the powers or ways of acting of things.  Thus, if science
is to be possible, there must be a relationship of
natural necessity between what a thing is and what a
thing can do; and hence between what a thing is and what
it tends to do, in appropriate conditions.  The
deducibility of a tendency from a nature thus constitutes
a criterion for our knowledge of natural necessity.
Events are necessarily connected when natural tendencies
are realized.


     With this in mind, let us return to a detailed
examination of the argument for thesis (i).  Is it
conceivable that water should not boil when it is
heated?  Now it might be said straightaway that it is
inconceivable to suppose that water might not boil when
it is heated.  Since anything that did not boil when it
was heated could not properly be said to be `water' at
all.  That is that, in Lockean terminology, `boiling
when heated' specifies part of the nominal essence of
water; or we could say with Putnam that `water'
functions as a `law-cluster concept'.77  Now


76 Cf. R. Harre, Surrogates for Necessity', p. 380.


77 H. Putnarn, `The Analytic and the Synthetic',
Minnesota Studies in the Philosophy of Science, Vol. III,
eds. H. Feigl and G. Maxwell, p. 376.


The Logic of Scientific Discovery 203


the strength of this reply should not be under-rated.  I
have no doubt that we should ordinarily say something on
these lines.  Indeed, unless we have some criteria for
the correct application of the term `water' there is no
reason why we should use it to refer to substances which
as a matter of fact boil when heated rather than to say
desk lamps or Saturday afternoons (which do not boil when
heated).  And such criteria would be at least in part
dispositional; appearances, notoriously, can be
misleading.  Litmus paper that does not turn red when
dipped into acid, a metal that does not conduct
electricity, or petrol that does not explode when ignited
could not be said to be `litmus paper', `a metal' or
`petrol' respectively; since the point of referring to
the particulars concerned in those ways would be gone.78
A magnet that could not magnetize, a fire that cannot
burn or a pen that can never write would not be
`magnets', `fires' and `pens' at all.  Things must
satisfy certain criteria for them to be (correctly
identified as) the kinds of things they are.  By far the
most important of such criteria are those that depend
upon their powers to affect other bodies (a class which
may be extended, analytically, to include their powers to
affect observers under specified conditions in certain
standard ways).


     Such a reply will not however satisfy the Humean
(particularly if he believes that definitions are merely
matters of convention and cannot express empirically
ascertained truths about kinds of things).  More to the
point it will not satisfy the scientist: for, accepting
that `boiling when heated' specifies part of the nominal
essence of water, i.e. the criteria for the
identification of a substance as `water', he will want to
know what it is about water in virtue of which it boils
when it is heated.  That is, he will set out to construct
an explanation, in terms of the molecular and atomic
structure of water, from which he can deduce its tendency
to boil when it is heated.  Now it is clearly
inconsistent with this explanation to suppose that water
might freeze, blush shyly or do anything else rather than
boil when it is heated.  That is, if the explanation is
correct water must boil when it is heated.


     Suppose however we came across a stuff which in all
other respects looked and behaved like water but which
did not boil


78 Cf. E. H. Madden, `Hume and the Fiery Furnace',
Philosophy of Science 1971, p. 66.


204 A Realist Theory of Science


when it was heated.  Assuming standard conditions and a
closed system (so as to eliminate the possibility of
intervening causes) it would seem that we have the
following alternatives:


     (a) our explanation was false;


     (b) the fact that it was intended to explain,
viz. that water boils when heated, was false


     (c) the particular concerned had been wrongly
identified: it was not a sample of water after all;


     (d) the particular concerned had changed; so that it
had ceased to be water by the time it was heated.


Now the Humean asks us to imagine, and inductive
scepticism requires that it be possible, that the cause
event occurs and the effect event fails to materialize.
Let me call this the critical situation.  Now I want to
argue that, given only that possibility (a) is ruled out,
so that we have a correct explanation, the critical
situation is impossible; that is, it is not possible that
the cause event occurs and the effect event fails to
materialize - in our example, that water is heated and
does not boil.


     Let me show this.  If the explanation is correct
water must boil rather than freeze when it is heated
(though of course the converse is not the case); so
possibility (b) is ruled out.  Consider (c), the
misidentification of the particular concerned.  Now in
this case it is not true to say that water did not boil
when it was heated.  For what did not boil was not water
but only something which looked, and perhaps otherwise
behaved, like it, say `nwater'.  Finally consider (d), a
change in the particular concerned: what was water when
it was put into the kettle at time t_1, ceased to be
water by the time it froze at t_2.  Here again it is not
true to say that water did not boil when it was
heated.  For by the time it froze it had become something
else, say `retaw'.  Hence given only the possibility of
a realist interpretation of the entities postulated in
the explanation, the conditions for inductive scepticism
cannot be satisfied.  If there is a real reason, located
in the nature of the stuff, independent of the
disposition concerned, water must tend to boil when it
is heated (though in an open world any particular
prediction may be defeated).  The stratification of
nature thus provides each science with its own internal
inductive warrant


     Now it might be objected that I have omitted from my
list of


 The Logic of Scientific Discovery 205


alternatives the possibility of the explanation, though
correct up to time t_l, subsequently breaking down.  But
this possibility equally does not satisfy the
requirements of the critical case.  For, now at Stratum
II (defining the Leibnizian level of the particular
movement of science with which we are here concerned),
nothing which did not possess the molecular and atomic
structure that water has been discovered to possess could
be said to be `water'.  So, here again, it would not be
water that was freezing.  A stuff remains water only so
long as its nature (or real essence) remains unchanged.
(Of course scientists could make a taxonomic change, but
this does bear upon the argument.


     It is of course possible that the nature of some
particular will be transformed: in which event,
scientists will search both for an underlying substance
or quasi-substance which preserves material continuity
through change (e.g. a gene pool through species change,
an atom in chemical reactions, energy in microphysics)
and for the agent or mechanism which brought about the
change.  The principles of substance and causality are
interdependent and complementary.  Things persist (and
continue to act in their normal way) unless acted upon;
and their changes are explained in terms of the action
of persisting (and transfactually active) things.  If
science is to be possible changes must be
transformations, not replacements; and transformations
must be effected by the actions of causes (causal
agents).  Things cannot pass clean out of existence or
events happen for no reason at all.  These are ideals of
reason.  But if science is to be possible our world must
be such that they hold.  This entails that it must be a
world of enduring and continually acting things.  It is
o course true that it is impossible to prove that cases
of ex nihilo production and miracles cannot ever happen.
All we can say is that they cannot be known to happen.
For it always remains possible for the scientist that
what appears to be a case of an ex nihilo production or
a miracle at time t_1 can come eventually at t_2 to be
explained in terms of the transformation of real things
and the action of real causes upon them.


     I have argued that provided we have a correct
explanation the critical situation cannot occur; that,
for example, as long as the particular stuff remains
water it must tend to boil when heated.  But it might be
urged if, as I have acknowledged, the nature of


 206 A Realist Theory of Science


some particular may be changed does this not open the
floodgates of inductive scepticism once more?  The answer
is no: for there is a big difference between wondering
whether some particular will be so acted upon by real
causes in its environment that its nature (in this case,
molecular structure) will be transformed, so that it
ceases to be an individual of that kind, and wondering
whether, while remaining an individual of that kind, it
will cease to behave in the way that it has tended to
behave in tr the past.  The point is even clearer if we
generalise it, so raising the questions of the boundaries
of kinds and of the scope of application of laws.  The
difference is between wondering whether water will cease
to exist; and wondering whether, while continuing to
exist, it will stop boiling (in exactly the same
circumstances) when it is heated.


     It might be objected that while what I have said
clearly covers case (d), viz. that of a particular
changing, I have not taken the possibility of case (c),
viz. that of a particular being misidentified, of nwater
being mistaken for water, seriously enough.  What is to
prevent us continuously misidentifying particulars in
just this way ?  Now just as particulars may be
transformed, so they may be misidentified.  But the
situation the inductive sceptic asks us to imagine only
gets of the ground if we assume that the relevant
particulars have been correctly identified.  The problem
of induction is the problem of what guarantee we have
that the unobserved will resemble the observed, or the
future the past; it is not the problem of what guarantee
we have that we have correctly observed the observed or
correctly described the past.  The suggestion that what I
have here may in fact be a piece of lead piping is
irrelevant to the question of what warrant I have for
assuming that water will continue to boil when heated or
for supposing that there is a necessary connection
between water boiling and its being heated.


     Nevertheless despite this irrelevance to our present
concern scepticism about particular knowledge can and
should be met.  It might be met in the following way: Any
argument in which the case for the general
misidentification of particulars is stated itself
presupposes the capacity to identify certain particulars,
namely words as tokens of a type and hence possessing a
certain standard meaning in a given context.  Hence no
argument for the general misidentification of particulars
can be consistently


The Logic of Scientific Discovery 207


stated.  If this argument does not carry conviction try
to imagine a world in which we (a) systematically (b) at
random misidentified (alpha) some particulars (beta) all
particulars (a') all the time (i) some of the time.  A
world in which we systematically misidentified some given
class of particulars (such as books as saucers and vice
versa) would just be a world in which objects had
different names.  But a world in which our
misidentifications were haphazard or universal is not
coherently conceivable.  It makes no sense to say that a
particular has been misidentified unless one is prepared
to say in what respect it has been misidentified.  This
itself presupposes the capacity to identify the
particular as of a certain type.  Of course our capacity
to identify particulars presupposes the extended or
dynamic principle of substance enunciated above, namely
that things persist and continue to act unless acted
upon, and hence in this way it presupposes the existence
of necessary connections between matters of fact.  It is
up to the criteriology of empirical science to determine
whether a particular has been misidentified or a
perceptual report is nonveridical.  The point is,
however, that if science it to be an ongoing concern it
cannot persistently demand and persistently return
negative verdicts.


     It might be objected to my refutation of thesis (i)
that I have not considered the possibility that the
explanation, which gives each science at any moment of
time its own inductive warrant, is incorrect.  Now it is
of course always possible that we are mistaken in our
explanation of why water must boil when heated; that our
description of the mechanism in virtue of which it does
so is wrong.  But this is a general condition of all
knowledge; it does not bear on the argument of thesis
(i), which concerns the special difficulty of knowledge
of necessary connections between matters of fact.  I have
already argued against the idea that all knowledge is
conjectural on the grounds that refutations presuppose
acceptances (progress requires a material cause).  But
whether or not my account of the transitive dimension of
the philosophy of science is accepted, refutations
presuppose necessary connections between matters of fact.


     I have argued that scepticism about change, about
our capacity to identify particulars and about the
possibility of non-conjectural knowledge as such are all
distinct from the special kind of scepticism involved in
thesis (i), which is


 208 A Realist Theory of Science


scepticism about the possibility of knowledge of
necessary connections between matters of fact.  I have
shown how the second and third forms of scepticism,
though irrelevant to thesis (i), may be averted.  But how
can Heraclitean scepticism be countered?  Changes in
things, I have argued, are explained in terms of
unchanging things.  The world is stratified.  We need only
worry about whether atoms will cease to exist when tables
and chairs do; we need only worry about whether electrons
will cease to exist when atoms do.  It is contingent that
the world is such that science is possible.  But given
that it is the dynamic principles of substance and
causality that I have formulated must be true of it.


     Three further forms of Heraclitean scepticism are
possible in which we could be invited to imagine that our
world is replaced (a) by a totally different one; (b) by
one in which the principles of substance and causality no
longer held; and (c) by one in which science ceased to be
possible.  I shall argue that the replacements envisaged
in (a) and (b) are impossible, but that I am precluded by
my own premises from saying anything about (c).


     In (a) it is supposed that our world could be
replaced by a totally different one; but to which, once
it had come into being, inductive techniques could be
reapplied.  Now this is not an intelligible supposition,
not only because scientific continuity would be lost
during the replacement (so it would make little sense to
talk of reapplying inductive techniques), but because
there is no possible way in which such a replacement
could be affected save by the action of real causes.79  In
(b) it is supposed that our world might be replaced by
one to which the principles of substance and causality do
not apply.  Now although the existence of our world is
contingent, given that it exists the supposition that it
might be replaced in this way is not an intelligible one.
Transcendental realism demands that we reason from the
effect, science, to the condition of its possibility,
viz. a world of enduring and transfactually active
mechanisms.  So we can rest assured that long after
mankind has perished things will persist and continue to
interact in the world that we once lived in.  This leaves
us with (c), about which I have said


 79 It is of course inconceivable that a fundamental
entity or entities should act inconsistently with its
(their) nature.  Hence in the last (non-Laplacean)
instance everything is as it must be.


 The Logic of Scientific Discovery 209


my premises preclude me from speaking.  But a moment's
reflection shows that (c) is devoid of interest for us.
It is an empty counterfactual.  For we know as a matter
of fact that our world is one in which science is
possible.  Hence to assert the possibility of a world
without science is merely to reassert the contingency of
the circumstance that makes a study of the conditions of
the possibility of science possible.


     I have established that we can have (and that
science actually possesses) knowledge of necessary
connections between matters of fact.  And I have shown
how inductive scepticism proper, namely that arising from
the assumption of the possibility of what I have called
the critical situation, viz. the occurrence of the cause
event and the non-occurrence of the effect, can be
allayed, viz. by the provison of an adequate explanation;
and how the other forms of scepticism often confused with
inductive scepticism can be countered.  I now turn to
theses (ii) and (iii) which the Humean uses to bolster
his central contention.


     Thesis (ii) alleges that if there were necessary
connections between matters of fact they would have to be
known a priori, so that science could not be empirical.
It is clear that this argument trades on a tacit
conflation of logical and natural necessity and the
identification of the resultant concept with that of the
a priori.  To refute it, I will have to show how
knowledge of the natures or real essences of things,
which I have argued ground our ascriptions of natural
necessity, can come to be attained empirically; that is,
how a posteriori knowledge of natural necessity is
possible.


     As there is some misunderstanding about the role of
the concept of essence (and, as we shall see, the nature
of definition) in science, some preliminary
terminological clarification is necessary.  The nominal
essence of a thing or substance consists of those
properties the manifestation of which are necessary for
the thing to be correctly identified as one of a certain
type.  The real essences of things and substances are
those structures or constitutions in virtue of which the
thing or substance tends to behave the way it does,
including manifest the properties that constitute its
nominal essence.  Science, I have argued, seeks to
explain the properties of things identified at any one
level of reality by reference to their intrinsic
structures, or the structures of which they are an
intrinsic part (defining the next level of


 210 A Realist Theory of Science


inquiry).  Thus the dispositional properties of say
nickel, e.g. that it is magnetic, malleable, resistant
to rust, melts at 1445 degrees C and boils at 2900
degrees C are explained, in the context of
post-Daltonian atomic theory, by reference to such facts
about its intrinsic structure as that its atomic number
is 28, its atomic weight is 58.71 and its density is
8.90.  The atomic constitution of nickel is its real
essence.  But it was discovered a posteriori, in the
transitive process of science.  And it itself
constituted an explanandum of the next phase of
scientific inquiry.


     In general to classify a group of things together in
science, to call them by the same name, presupposes that
they possess a real essence or nature in common, though
it does not presuppose that the real essence or nature is
known.  Thus we are justified in classifying alsations,
terriers and spaniels together as different varieties of
the same species dog because we believe that they possess
a common genetic constitution which, despite their
manifest sensible differences, serves to differentiate
them from the members of the species cat.  A chemist will
classify diamonds, graphite and black carbon together
because he believes that they possess a real essence in
common, which may be identified as the atomic (or
electronic) structure of carbon, of which these are
allotropic forms.  To classify a thing in a particular
way in science is to commit oneself to a certain line of
inquiry.  Ex ante there will be as many possible lines of
inquiry as manifest properties of a thing, but not all
will be equally promising.  Thus if one's concern is to
account for the manifest properties of cucumbers it is
clearly preferable to classify a 12 in. long green
cucumber under the sortal universal `cucumber' rather
than under the universals `green' or `12 in. long'.  Not
all general terms stand for natural kinds or taxa;
because not all general features of the world have a
common explanation.  Carbon and dogs constitute natural
kinds; but tables and chairs, red things and blue, chunks
of graphite and fuzzy dogs do not.  The justification of
our systems of taxonomy, of the ways we classify things,
of the nominal essences of things in science thus lies in
our belief in their fruitfulness in leading us to
explanations in terms of the generative mechanisms
contained in their real essences.  Not all ways of
classifying things are equally promising because not all
sets of properties individuate just one and only one kind
of thing.


 The Logic of Scientific Discovery 211


     The distinction between real and nominal essences
should not be confused with that between real and nominal
definitions.  Real definitions are definitions of things,
substances and concepts; nominal definitions are
definitions of words.  (Nominal essences are the
properties that serve to identify things).  Real
definitions, in science, are fallible attempts to capture
in words the real essences of things which have already
been identified (and are known under their nominal
essence) at any one stratum of reality.  As so conceived,
they may be true or false (not just or even - more or
less useful).  The atomic weight of copper is 63.5.  It
would be wrong to claim that it was 53.4 or alternatively
that 63.5 was the atomic weight of tin.  Of course this
fact was discovered a posteriori; but it may now be said
to constitute part of the real definition of copper.  If
the real essence of copper consists in its atomic (or
electronic) structure, its nominal essence might consist
in its being a red sonorous metal, malleable and a good
conductor of electricity etc.  Something that did not
satisfy these properties could not properly be said to be
`copper'.  But conversely just because the word `copper'
in science has a history, and at any moment of time a
use, the nominal essence of copper cannot suddenly be
designated by the use of `reppoc' or `tin'.  Nominal
definitions in science cannot therefore be conceived as
stipulative, arbitrary or matters of convention.
Although there is a sense in which any other symbol could
have been used to refer to copper; given this usage and
that history `copper' cannot be replaced by `bronze' or
`female' for no reason at all.80  Changes in the
definitions of words in ongoing social activities require
justification.


     On the view advanced here science consists in a
continuing dialectic between taxonomic and explanatory
knowledge; knowledge of how the things there are
behave.  It aims at real definitions of the things and
structures of the world as well as statements of their
normic behaviour.  The source of the failure to see this
is the ontology of empirical realism which reduces things
to qualities, taxa to classes, enduring and active
mechanisms to constant conjunctions of independent and
atomistic


 80 For a discussion of the history of `copper' see
M. Crosland, Historical Studies in the Language of
Chemistry and R. Harre and E. H. Madden, Causal Powers,
Chap. 1.


 212 A Realist Theory of Science


qualia.81  Now if the world consists only of qualia and
qualia are independent of one another then the particular
names that we give to qualia cannot matter and all qualia
will appear on a par.  On this conception, predicates
must be independent of one another and classification is
ultimately arbitrary.


     Now just as it is a mistake to assume that science
is concerned with any and all behaviour it is a mistake
to assume that it is concerned with any and all things.
Scientists do not seek to describe the behaviour of or to
classify common objects like tables and chairs, though
the laws of physics and the principles of scientific
taxonomy (e.g. the identification of a table as an oak
one) may be brought to bear on them.  Now from the fact
that tables have no real essence it does not follow that
carbon has none.  Electrons are not related in the same
way as games.  A resemblance theory of universals works
best for the complex Strawsonian individuals of ordinary
life.  But the universals of interest to science are
real: they are the generative mechanisms of nature which
account, in their complex determination, for the
phenomena of the world, including (upon analysis) the
genesis and behaviour of ordinary things.  The dialectic
of explanatory and taxonomic knowledge must thus be
formulated as follows: science is concerned with the
behaviour of things only in as much as it casts light
upon their reasons for acting and hence upon what kinds
of things there are; and science is only concerned with
things of a particular kind, in as much as they
constitute the reason for some pattern of normic
behaviour and thus themselves become an appropriate
object of inquiry.


     The importance of taxa in science may be expressed
by saying that what is non-accidentally true of a thing
is true of a thing in virtue of its essential nature.  A
thing acts, or at least tends to act the way it is.  It
should be stressed that the difference between a thing
which has the power or tends to behave in a certain way
and one which does not is not a difference between what
they will do, since it is contingent upon the flux of
conditions whether the power is ever manifested or
tendency exercised.  Rather, it is a difference in what
they themselves are; i.e. in their intrinsic natures A
copper vase remains malleable even if it is never


 81 To use Goodman's very useful term.  See N. Goodman,
The Structure of Appearance, p. 130 and passim.  Goodman
himself attributes the term to C. I. Lewis, Mind and the
World Order.


 The Logic of Scientific Discovery 213


pressed out of shape.  It is contingent whether an
electric current is ever passed through a copper
wire. But it is necessary, given its electronic
structure, that it be a good conductor of electricity.
We know how things will behave, if certain conditions
materialize, if we know what the things are. But we can
only know what things are a posteriori, via the
empirical process of science.


     This view may be contrasted with the idea that
scientists are not concerned with questions such as `what
is energy?' or `what is an atom?' but only with questions
of the kind `how can the energy of the sun be made
useful?' or `under what conditions does an atom radiate
light?'82  Popper's `methodological nominalism' seems to
be based on the idea that to suppose that things have
essences is to suppose that it is possible to give
explanations which are `ultimate' in the sense that they
are insusceptible in principle of further explanation
(which is what he calls `essentialism').83  Although Locke
may have held this view, is it certainly no more a
necessary feature of the concept of real essence than it
is a necessary feature of the concept of behaviour to
suppose that because a thing can be described as behaving
in a certain way the behaviour itself cannot be subject
to further explanation.  It is clear that to suppose that
things have real essences is not to suppose that the real
essences of those things cannot be explained in terms of
more fundamental structures and things.


     Two other arguments sometimes invoked against the
concept of real essences should be mentioned.  The first
depends upon the assumption that differences in nature
are continuous, not discrete; that `God makes the
spectrum, man makes the pigeonholes';84 so that `genera,
species, essences, classes and so on are human
creations'.85  I can find no possible warrant for such an
assumption.  Taken literally, it would imply that a
chromosome count is irrelevant in determining the
biological sex of an individual, that the class of the
living is only conventionally divided from the class of
the dead, that the chemical elements reveal a continuous
gradation in their properties, that tulips merge into
rhododendron bushes and solid objects fade gaseously away
into empty space.  The second involves the belief


 82 K. R. Popper, The Open Society and Its Enemies,
Vol. I, p. 32.


 83 K. R. Popper, Conjectures and Refutations, Chap. 3,
esp. p. 102.


 84 A Flew, op. cit., p. 450.


 85 Ibid., p. 449.


214 A Realist Theory of Science


that to suppose that there are natural kinds is to
suppose that these kinds are fixed, and is in particular
to rule out the possibility of a mechanism of
evolution.86  Again, this is completely unwarranted.  For
natures may change; and whether, and if so the ways in
which they do, are matters for substantive scientific
investigation.  No spectrum exists between men and apes
but that does not preclude the possibility of a mechanism
of evolution (involving a whole sequence of `missing
links').  What happens in such cases is that biologists
posit a novel entity, a gene pool, as the underlying
continuant through the species' change.  The objection is
only valid at the level of ultimate physical entities
since necessarily if such entities exist they must be
enduring.


     Scientists attempt to discover the real essences of
things a posteriori, and to express their discoveries in
real definitions of the natural kinds.  From a
description of the nature of a thing its behavioural
tendencies can be deduced.  When such tendencies are
realized the events describing the stimulus or releasing
conditions for the exercise of the tendency and its
realization may be said to be necessarily connected.  Thus
scientists can come to possess knowledge of necessary
connections between events as a result of an a posteriori
process of discovery.  Scientists are not content to
collect conjunctions of events.  Rather they try to
discover the natures of things.  Given this, no problem
of induction can arise.  Since it is not possible for a
thing to act inconsistently with its own nature and
remain the kind of thing it is.  That is, a thing must
tend to act the way it does if it is to be the kind of
thing it is.  If a thing is a stick of gelignite it must
explode if certain conditions materialize.  Since
anything that did not explode in those circumstances
would not be a stick of gelignite but some other
substance.  Now given the satisfaction of the criteria
for the identification of a substance, say water, and the
recording, preferably under experimentally closed
conditions, of its most significant and suggestive
behavioural properties, scientists move immediately to
the construction and testing of possible explanations for
the protolaws identified.  But if there is an
explanation, located m the nature of the stuff or the
system of which the stuff is a part, whether or not it is
known by men, water must tend to boil when it is heated.
It is the real


 86 S. Toulmin, op. cit., pp. 135-6.


 The Logic of Scientific Discovery 215


stratification of nature that justifies induction in
science.  It is not we that impose uniformities upon the
world, but nature that makes induction (properly
circumscribed) a rational activity for men.


     The third Humean counter-argument is that we are
never directly aware of any necessary connection between
matters of fact or causal power or agency so that these
concepts cannot be justified by experience.  Thesis (iii)
thus completes a triangle, whose other sides are theses
(ii) and (i).  It could be argued that we are sometimes
directly aware of necessarily connected sequences (see
2.3 above), and that we are sometimes directly aware of
the exercise of causal powers (though the powers
themselves can only be known, not shown, to exist;
i.e. we are never directly aware of causal powers as
such).87  It seems clear that we are aware of ourselves as
causal agents in a world of other causal agents; and that
unless we were so aware we could not act intentionally,
or come to know ourselves as causal agents at all.
(Projective explanations of our idea of necessary
connection are clearly anthropocentric.)  However for the
transcendental realist this is incidental.  For, for him,
the status of the concept of necessary connection is
clear: it has been established, by philosophical
argument, as applicable to some but not other sequences
of events as a necessary condition of the social activity
of science.  (It should be stressed that this does not
mean that any particular science has correctly
identified, let alone adequately described, the necessary
sequences: it is a condition of the possibility of
science.)  Thus the concept of natural necessity does not
have to be justified in terms of or traced back to its
source in sense-experience; though there must be a
scientific explanation of how we come to possess the
concept.


     That science has a posteriori knowledge of necessary
connections between matters of fact is a proposition that
can be given no further justification.



 87 E. H. Madden and P. Hare do not clearly distinguish
powers from their exercise in their criticism of this
Humean argument in `The Powers That Be', Dialogue 1971,
pp. 12-31.



     --- from list bhaskar-AT-lists.village.virginia.edu ---