MARITAIN, INSTRUMENTALISM, AND THE PHILOSOPHY OF

EXPERIMENTAL SCIENCE

MATIHEW S. PUGH
Is jacques Maritain's philosophy of the experimental sciences, the
sciences of observation, fundamentally at odds with Maritain's causal
realist understanding of the philosophy of nature as one of the
genuinely deductive sciences of explanation? This is an important
question in Maritain studies, for on it hinges the broader question of
the coherence of Maritain's philosophy of science as a whole. Yet how
we answer it very much depends on how we characterize Maritain's
philosophy of the experimental sciences. If Maritain's conception of
experimental science is a conventionally instrumentalist one, as it at
first glance appears to be, one could well make the case that his
philosophy of the sciences of observation is truly incompatible with a
causal realist view of nature. If, on the other hand, Maritain is not an
instrumentalist, or at least not an instrumentalist in the conventional
sense, then his philosophy of the experimental sciences may turn out
to be compatible with his causal realist philosophy of nature, for not
every version of instrumentalism is incompatible with causal realism.
The purpose of this paper, then, is to find out if Maritain is an
instrumentalist, and if so, just what kind of instrumentalist he might
be. In order to do this, I must first define what instrumentalism is, and
then determine if Maritain's position coheres with that definition.
Thus, I define instrumentalism in section I, and then in section II
elaborate Maritain's philosophy of experimental science, showing
exactly how Maritain's position is instrumentalist and in what ways it is
not. I conclude in Section III that, although Maritain was indeed an
instrumentalist, his version of instrumentalist philosophy of the
experimental sciences leaves room for completion by a causal realist
philosophy of nature.

I. INSTRUMENTALISM

In order to see how Maritain's philosophy of experimental science is

a qualified instrumentalism, we must first address the basic question:
What is instrumentalism?
32
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 33

In the broadest sense instrumentalism 1 in the philosophy of science

can be defined simply as the view that scientific theories are not true
pictures of reality, but merely instruments or tools useful for making
calculations and predictions. More specifically, instrumentalists claim
that the categories of truth and falsehood cannot be used to evaluate
scientific theories. Rather, what counts in instrumentalism is that
theories "work," and theories work when they successfully provide the
scientist with guiding principles or "inference tickets" which enable
him to organize and/ or relate observation statements which, when
used in conjunction with other observation statements, entail
subsequently confirmed predictions. Questions about the truth of a
theory play no part in this process for the instrumentalist, for theories
contain T-terms (theory terms) that are observationally unverifiable
and hence meaningless. Theory terms are simply functional notions
that enable the scientist to make predictions based on the theory in
question. Hence, for the instrumentalist, theories do not and cannot
explain in any conventional, or causal realist sense. In fact, the basic
goal of science according to the instrumentalist is not explanation, but
prediction.
Now this view of science rests on a number of important
distinctions, such as the distinction between law and theory. For the
instrumentalist, this distinction is one of kind, not degree. The key to
this distinction turns on the notion of observation. Experimental laws
pertain to observable things. In fact, as one prominent instrumentalist
has said, "Experimental laws are statements that formulate an
observable relation between things, and can be validated by controlled
observation." 2 Examples are: the pressure of an ideal gas at constant
temperature varies inversely with the volume; when water in an open
container is heated, it eventually evaporates, etc. Experimental laws, in
other words, contain 0-terms (observation terms) like" ... is sharp";" ...
is cold"; " ... expands"; " ... points to 20"; etc. The meanings of such terms
are observationally fixed, or are at least operationally definable, like
"length," "mass," "pressure," "rate of acceleration." In effect, 0-terms

1 It should be noted that instrumentalism belongs to a broad family of typically

anti-realist positions in the philosophy of science, including but· not res-
tricted to: logical positivism, phenomenalism, conventionalism, fictionalism,
and constructive empiricism.
2 Ernest Nagel, The Structure of Science: Problems in the Logic of Scientific
Explanation (New York: Harcourt, Brace and World, 1961), 80.
34 MATIHEW S. PUGH

of experimental laws are linked to a procedure for predicating the

terms of some observational traits, and this is what gives experimental
laws their empirical content. Consequently, an experimental law is an
inductive generalization of constant relations holding between
observed data. Furthermore, experimental laws are independent of
theories, and in fact often outlast the theories used to explain them.3
Because the meanings of 0-terms can be determined (at least in part)
independently of the theory used to explain them, experimental laws
take on a life of their own. But precisely because experimental laws
have determinate empirical content, and are thus verifiable or (at least)
falsifiable-provided one carries out the operations specified by the
conditions of the law-experimental laws are either true or false. So
once an experimental law is established, that is, is empirically verified,
its truth-value is independent of any theories used to explain the law. It
is this independence of experimental law from theory that enables the
scientist to use experimental law to determine-along with other
features such as coherence, simplicity, fecundity, etc.-if a new theory
is plausible. Experimental laws draw their support strictly from
observation.
Theories, on the other hand, contain T-terms like" ... is an electron";
" ... IS . an e1ect romagne t'IC f'Ie ld"; " ... Is
. a quark" : " ... IS . a grav1'ty wave ";
etc. For the instrumentalist, these terms are unobservable and some-
times even operationally indefinable, since there is no specifiable, that
is, empirical way to identify their referents. T-sentences (sentences
containing T-terms) are hence meaningless. As a consequence,
instrumentalists either claim that T-sentences are simply linguistic
tools for calculating and predicting, or that there are no inductive
grounds for believing that the referents ofT-terms exist.
More specifically, instrumentalists maintain that the meanings ofT-
terms are not determined by experimental procedures. Rather, the
meanings of T-terms are only defined implicitly by the theories of
which they are a part, and only indirectly by the experimental uses to
which they are put.
There are two reasons for this. First, all theories employ an abstract
calculus that is the logical skeleton of the theory, and gives to it its

3 Ibid., 86. Nagel states that Wien's displacement law remained in place even
after the theory that explained it (classical electrodynamics) was replaced by
Planck's quantum mechanics.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 35

"explanatory power." This structure is arrived at by ignoring the non-

logical or descriptive terms of the theory, like "quark," "wavicle,"
"mass," etc., and focusing exclusively on the logical relations in which
the terms stand to each other. This abstract set of postulates thus
determines the meanings of the non-logical terms in virtue of their
place as variables within the postulates. Because these postulates are in
fact statement-forms 4 rather than statements, they assert nothing. But
other statement-forms can be derived from them using the standard
rules of logic. Second, theories must be linked in some way to a-
statements, and this linkage usually takes the form of correspondence
rules. 5 Without that linkage, no theory could be used to "explain" a-
statements, let alone make predictions. Yet rules of correspondence do
not provide explicit definitions of the T-terms used in theories, and for
the following reason. An explicitly defined term can always be replaced
by its equivalent defining expression. But theoretical expression in
scientific theories cannot be so replaced because the replacing ex-

4 A statement-form, as opposed to a statement, is simply a sentence that has

the grammatical form of a statement, but which is in fact not a real
statement. For example, the sentence "For any x if x is an elementary
particle and x is P, then x is a neutron," is a statement-form because P is an
unspecified predicate variable, whereas the sentence "For any x, if x is an
elementary particle and x had no charge and a rest mass of 1.6749542 x 10-27
kg, then x is a neutron," is a genuine statement. Following Nagel, "the
expression w(x,r) is employed in the Schrodinger equation in quantum
mechanics for characterizing the state of an electron. There is in effect a
correspondence rule for the expression \jf(x,r) \jf*(x,r) (where w* is the
complex conjugate of \jf), but no such rule for \jf(x,r) itself.... [T]heories
containing such terms are statement-forms and cannot be said to be either
true or false" (Ibid., 132-33).
5 An example of the problem here arose regarding the Bohr theory of the atom,

which accounts for a number of experimental laws of spectroscopy. Since the

electron, their orbits, their jumps, etc., are concepts that do not apply to
anything that is observable, how are connections to be found which will link
these notions to what is experimentally observed? How was it done? A line in
the spectrum of an element was associated with an electromagnetic wave
whose length can be calculated, in accordance with the theory, from
experimental data on the position of the line. But the Bohr model associates
the wavelength of a light ray emitted with the jump of an electron orbit. This
jump is thus correlated with the experimental notion of a spectral line. Cf.
Ibid., 94-95.
36 MAITHEW S. PUGH

pressions do not explicitly define the replaced expression. 6 In short, the

inability of correspondence rules to "tack down" or link every T-term
in a theory to 0-terms enables theories to have great flexibility and to
range over many experimental concepts. Theories must, therefore,
never refer to any one set of experimental concepts, lest they be
restricted to that one set. "A theory seeks to formulate a highly general
structure of relations that is invariant in a wide variety of
experimentally different situations." 7 In order to achieve the generality
required for theories, the scientist very frequently uses the symbols of
logic and mathematics to effect transformations not linked to
experimental concepts. Because there are no procedures for directly
applying T-terms to experimental instances of the term, a theory
cannot be directly put to experimental test. T-terms cannot be
understood apart from the particular theory that implicitly defines
them. Unlike experimental laws, then, a theory is not an empirical
generalization.
Of course, instrumentalists recognize that theories must be linked to
experimental laws, and that this takes place via correspondence rules.
But they reject the notion that the abstract calculus of a theory
functions as a way to draw conclusions from experimental postulates
that are held to be premises. Rather, the function of a theory is to
provide a rule or principle for analyzing observational data, a tool so to
speak, for inferring observation statements from other observation
statements, or for "making logical transitions from a set of
experimental data to another set." 8 In effect, a theory is a leading
principle "in accordance with which conclusions about observable facts

6 Ibid., 98. The point here is that correspondence rules do not supply explicit
definitions of theoretical notions used in theories, but at best implicit
definitions. For example, the definition for the expression "X is a mammal,"
could be replaced by the equivalent expression "X is a member of the
vertebrates that nourish their young with milk and bear live offspring," but
the expression "X is the wavelength of the radiation emitted when an
electron jumps from the next-to-smallest to smallest orbit of the hydrogen
atom," is not equivalently replaced by "Y is the line occurring at a certain
position in the spectrum of hydrogen."
7 Ibid., 10(

8 Ibid., 129.'
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 37

may be drawn given factual premises; not as a premise from which such
conclusions are obtained." 9
Now this view of the role which theory plays in scientific enquiry
has a number of important consequences for how the instrumentalist
understands: (i) explanation; (ii) causality; and (iii) the cognitive status
of theories.

1. Explanation
As far as explanation is concerned, instrumentalists typically
embrace some form of the famous deductive-nomological {D-N) or
covering-law model of scientific explanation developed by Hempel and
Oppenheim. This model is useful and satisfies the instrumentalist's
conception of the goal of science because "it formulates the conditions
under which events of various sorts occur, the statements of such
determining conditions being the explanations of the corresponding
happenings." 10 In its most basic form, the explicandum E, or event to be
explained, is explained when certain laws are adduced which, when
joined with antecedent conditions, make E empirically necessary. The
explanans (the conditions C1,C27C31 ... Ck1 together with the law or laws
L 1,L21L31 ... Ln) answer the "why" question fundamental to all
explanation. 11
It is important to point out, however, that the explanans (at least
insofar as it contains or uses experimental laws) must in all cases be
empirical and true. The explicandum, of course, can be either an
individual event or a law, but if it is a law, or even a set of laws, the
principles used to explain them must be general rules specifying
repeatable patterns of dependence among the observed properties of
the subject matter studied. Relations are thus established between
apparently unrelated or disparate phenomena. The explicandum is
explained if it is subsumed under a general representation or principle.
Thus laws which are generalized patterns of relations are in the D-N
model subsumed under laws which identify patterns of relations which
have greater range than the laws which they subsume. But these

9 Ibid., 129-30.
10 Ibid., 4.
11 For example, if we want to know why a rod expanded when heated, we could

say, "All metals expand when heated; this rod is metallic and it was heated:
therefore this rod expanded."
38 MATTHEW S. PUGH

repeatable patterns of relations, or patterns of dependence, are best

expressed in mathematical form, wherein relations of dependence can
be expressed in formulae whose variables are related to other variables
by some mathematical function. Examples from physics include the
superposition principle, various laws of constancy-such as Galileo's
law of freely falling bodies- and Kepler's third law of planetary
motion, etc.

2. Causality
Nevertheless, though the principles in question operate in a purely
functional way, they typically give the appearance of identifying
causes, and of thus providing us with causal explanations. Yet they do
not. Thus in a functional law of the type X = YZ, the variables can be
rearranged mathematically in such a way as to make it impossible to
know which of the variables in question is a cause, and which an effect;
that is, X = YZ can become y = ..!_
or z = X 12 As physicist Gerald
z y
Holton says, in such cases "it is on the whole more fruitful to think of
an interaction rather than simple causation, and to ask 'to what factors
is X related,' instead of 'what causes X."' 13 Indeed, the D-N model of
scientific explanation and its variants (the Inductive-Statistical, or IS
model, etc.) 14 do not embrace a realist notion of causation, but rather
some kind of regularity or qualified Humean account of causation. The
regularity model of causation simply states that events that are held to
be causally related are so related if and only if the events in question
instantiate a general regularity between like kinds of events. In other
words, A causes B if A is of a kind of event X, and B is of a kind of event
Y, such that events of kind X are regularly followed by events of the
kind Y. In that case, the regularity or law pertaining to kinds X and Y
covers or subsumes events A and B. The regularity is in turn written in
terms of a formula whose variables are related to each other as a
mathematical function. Thus it is the functional regularity that law
encapsulates, and that, when joined with relevant conditions, explains

12 Gerald Holton, Introduction to Concepts and Theories in Physical Science

(Princeton, New Jersey: Princeton University Press, 1985), 182.
13 Ibid.
14 For a good general overall discussion of the different kinds of explanation
see "Scientific Explanation," in the Stanford Encyclopedia ofPhilosophy Online.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 39

the explicandum. Given the explanans, the explicandum "is to be

expected."
The realist or non-Humean view of causation, on the other hand,
maintains that laws express a "necessitation relation" between the
properties of relata. Accidents, however, do not express relationships
of necessity between relata. Accidents involve constant conjunction,
but laws involve something more. Thus, when under standard
conditions one billiard ball strikes another billiard ball, the second
must move. Yet taken by itself, this approach fails to address the real
question: What is the nature of this necessity or "must"? The regularity
theory overcomes this problem by treating necessity hypothetically or
conditionally. That is, effects or consequences are to be expected given
the laws of nature, which are themselves contingent. If the laws of
nature were different, then when A strikes B, B might not move, though
B will move given the current laws of nature. In addition, the regularity
theory can accommodate non-deterministic or probabilistic accounts of
nature. (We are reminded here, of course, of the whole of quantum
mechanics.)
For the instrumentalist, then, not only are laws that assert
functional dependencies between variables non-causal in the narrower
realist sense, they are non-causal even in a more broadly defined sense.
Typically, laws are held to be minimally causal when they satisfy four
conditions: (1) the relations between the relata are uniform; (2) the
relations between the relata are directly or indirectly spatially
contiguous; (3) the relations between the relata reveal a temporal
sequence; and (4) the relations between the relata are asymmetrical.
Yet laws expressing functional dependencies fail to satisfy at least one
of these four conditions. An example frequently used to demonstrate
this point is Boyle's law of ideal gases, formulated as pV = aT. In this
formula, no claim is being made that a change in temperature is
followed by a change in pressure or volume. In other words, nothing
about sequential order is asserted by the formula. The law captures a
functional dependency, not a sequential order, even though the latter
might figure prominently in experimental verification. (The same
observation applies to laws that only assert invariable statistical
relations between events.)
But even when the laws which function as rules or guidelines in the
explanans of the D-N model do occasionally satisfy the above-
mentioned conditions, these laws are not the kinds of laws favored by
40 MATTHEW S. PUGH

causal realists, for realists presuppose that that type of causal law
centers on a relation of necessitation holding between the properties of
relata. However, the relation of necessitation in turn rests upon the
above-mentioned four conditions for causal law being satisfied. In fact,
however, and again, at least one of these conditions is violated by
scientific law, for the causal realist assumes that there are substances
with determinable properties which take numerous forms. And herein
lies a problem, for a substance is identified by the type of determinate
properties that it has, which in turn differentiate the substance in
question from other substances (as long as it differs in at least one form
of a set of determinable properties). Thus to assert that X is a rock salt
is to assert that it has the following forms of a set of determinable
properties: crystalline structure = cubical crystals; color = colorless;
density = 2.163 g/cm 3; melting point = 804°C; hardness = 2 on Moh's
scale; etc. 15 Using induction, one could say that, because all observed
rock salts are X, Y, and Z, therefore all rock salts are X, Y, and Z. And
this claim will acquire nomic necessity if backed up by a theory (itself a
set of laws) that shows why rock salt must have the properties it has.
Yet this law is non-causal in that it makes no claim about any
properties preceding or, in effect, causing other properties in
sequential order. For example, rock salt's color is not caused by its
hardness, nor is its hardness caused by its melting point. As a result, a
crucially important support for the necessitation relation in the causal
realist conception of law is thereby undermined.

3. The Cognitive Status ofTheories

Finally, if the instrumentalist view of science is correct, then it
would appear that factual truth or falsity cannot be predicated of
scientific theories. Since laws operate functionally in the
instrumentalist conception of explanation, that is, as guides or leading
principles for organizing data and making logical transitions from one
set of observable data to another set of observable data; and since these
laws are constructed using T-terms only implicitly defined by their
place in an abstract calculus and by borrowing ideal concepts from
mathematics that are not descriptive of anything real, theories cannot
be said to be factually either true or false. Factual statements are only
true if they formulate relations between existing things or events, or
between the properties of existing things. If those observable pro-

15 Nagel, The Structure ofScience, 75.

 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 41

perties reference unobservable theoretical entities, then the

statements containing terms referencing them cannot be factually true
or false.
As we have seen, the instrumentalist makes a hard distinction
between 0-terms and T-terms. 0-sentences (sentences containing 0-
terms) and generalizations of 0-sentences can be true or false (since
they are empirically verifiable). ButT-sentences (sentences containing
T-terms) are for the instrumentalist empirically unverifiable, and
hence meaningless. Instrumentalists view T-sentences with their T-
terms as instruments for calculating and predicting. As a result, we
have no good reason to believe that the referents of T-terms actually
exist. Rather, T-terms have the status of variables, and as such
constitute statement-forms, instead of genuine statements. But
statement-forms, unlike statements, are neither true nor false. 16
The concern of the instrumentalist is thus not whether theories are
true or false, but whether they work. Theories work if they can be
successfully used to organize and/or relate observation statements,
which can in turn be used (in conjunction with other observation
statements) to make subsequently confirmed predictions. The question
of a theory's truth is irrelevant in this regard. The task of a theory is
not to provide a true description of what takes place in the world, but
to provide a way to analyze and symbolize certain properties of the
subject matter studied. As these properties reveal themselves in
experimental situations, good theory makes it possible to infer
additional information (subsequently confirmed) about other
properties of said subject matter. Since theories are neither true nor
false, there is no problem in instrumentalism with conflicting or
incompatible theories. 17 Theories can only conflict or be incompatible if

16 Statement forms such as "For any x, if X is an animal and X is P, then X is a

vertebrate," are neither true nor false. However, a statement like the
following (which does not contain variables like P) "For any x. if X is an
animal and xis a mammal, then X is a vertebrate," is factually true or false.
There are examples of statement forms in quantum mechanics theory, and
the molecular theory of gases. Cf. Nagel, The Structure ofScience, 132.
17 " ... it is not a source of embarrassment to the instrumentalist position that in

inquiries into the thermal properties of a gas we use a theory which analyses
a gas as an aggregation of discrete particles, although when we study
acoustic phenomena in connection with gases we employ a theory that
42 MATTHEW S. PUGH

they are held to be true or false. Yet instrumentalism does not hold that
theories or theoretical entities are mere fictions, but rather that they
are not true to the facts. One might simply call such theories
"instrument theories." Indeed, some theories are superior to other
theories; for example, if they serve as effective guides for subsuming a
greater range of experimental data or laws, or if they make it possible
to infer more observational data than other theories, or if their
inferred conclusions agree with further observation.
Of course, if theories are leading principles in accordance with which
conclusions are drawn, rather than premises from which they are drawn,
the derivations that the instrumentalist must use rely on an intuitionist
logic in which the classical law of the excluded middle no longer holds.
(Derivations in classical logic are based on the assumption that the
statements used as premises are true.) Clearly, one cannot assert "A,"
or "not-A," if one cannot decide whether "A" is true or "not-A" is true.
And since T-terms and T-sentences are meaningless because their
referents are unobservable and hence unverifiable, T-sentences
containing T-terms cannot be true or false. Hence, given that "A" is aT-
sentence and "not-A" is a T-sentence, one cannot assert either "A" or
"not-A." And if one cannot assert "A" or "not-A," then neither can one
assert "A or not-A." This is important given that the instrumentalist
must derive his predictions from theories, and these derivations must
be valid. If the instrumentalist uses the standard rules of inference of
classical logic, he is applying these rules to derivations the premises of
which are incapable of being either true or false. An intuitionist logic
which rejects the law of the excluded middle enables the
instrumentalist to escape from what otherwise might be a fatal
inconsistency.18

represents the gas as a continuous medium. Construed as statements that are

either true or false, the two theories are on the face of it mutually
incompatible. But construed as techniques or leading principles of inference,
the theories are simply different though complimentary instruments" (Ibid.,
133).
18 W. H Newton-Smith, The Rationality of Science (New. York: Routledge and
Kegan Paul, 1981), 33-34. For a possible way aroundthis problem without
having to abandon classical logic, see Nagel, The Structure of Science, 138-40.
For the .sake of clarification, it should be pointed out here that the
instrumentalist does not reject classical logic altogether. When it comes to
0-sentences, for example, and the derivations derived from them, the
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 43

II. MARITAIN'S POSITION

Having said all this, we are now in a position to answer the question,
Is Maritain an instrumentalist? The Degrees of Knowledge and The
Philosophy of Nature make it clear that if Maritain does embrace
instrumentalism, it is not instrumentalism in the commonly accepted
sense/9 yet it is similar in enough features to allow us to tag it with that
label. This is so, for example, when it comes to the 0-term/0-
statement, T-term/T-statement distinction. Like all instrumentalists,
Maritain clearly separates fact from theory, and maintains that this
distinction is a distinction that is based on a difference of kind, not
degree. In keeping with Aristotle and St. Thomas, Maritain believes that
all knowledge begins with facts, which, as he says, are well established
existential truths given to a mind which receives them. The mind, in
turn, discerns in the objects of its concepts (which correspond to what
is given) certain connections pertaining thereto, connections existing
in the real. 20

instrumentalist is just as willing as the realist to embrace classical logic with

its traditional rules of inference. The problem arises when the
instrumentalist attempts to derive observable predictions combined with 0-
sentences stating initial conditions from theoretical sentences. Since T-
statements are neither true nor false, that is, since they have no truth status,
they cannot be treated as real premises from which conclusions are drawn,
but rather must be likened to rules in accordance with which predictions are
made. Nevertheless, whether T-statements are viewed as premises or rules
or guiding principles, the derivations must be valid, and herein lies the
logical challenge for the instrumentalist. Since T-statements are not real
premises, one cannot use the truth preserving rules of inference of classical
logic to check derivations made from such statements. In what sense, then,
does the concept of validity still apply in instrumentalism? The
instrumentalist's best recourse is to turn to a logic that rejects, at the very
least, the classical law of the excluded middle. For more on intuitionism, cf.
Arend Heyting, Intuitionism: An Introduction (Amsterdam: North-Holland
Publishing, 1956); Michael Dummett, Elements of Intuitionism (Oxford:
Clarendon Press, 1977).
19 As we shall see, Maritain's instrumentalism is really a version of ontological

instrumentalism, as opposed to semantical instrumentalism. Newton-Smith

uses the term "epistemological instrumentalism" to refer to roughly the
same position as Maritain's. See Newton-Smith, The Rationality ofScience, 30.
20 jacques Maritain, The Degrees of Knowledge, trans. 4th French ed. under

direction of Gerald B. Phelan (New York: Charles Scribner's Sons, 1959), 57.
44 MATIHEW S. PUGH

Nevertheless, the facts that are given to a mind are not mere copies
of the external world, for facts are judged by the mind that receives
them, indeed, even by the senses which also judge what they receive.
Every fact, says Maritain, is discriminated/ 1 that is, judged, either by
the intellect or the senses. But judging does not distort or deform what
is judged. Rather, judging is a matter of the senses/intellect becoming
assimilated or conformed to what is judged. And in the process of
judging, senses and intellect work together. Maritain's critical realist
epistemology will not allow him to embrace a simplistic view which
sees all facts as arising from sense, and all theory from intellect. Hence,
in the determination of facts, the intellect often intervenes with
already formulated theories, but it does so to "discern and formulate
that which is furnished by sense intuition." 22 In the determination of
theories, the intellect works with sense intuition to uncover "essences
or laws and explanatory reasons.'t23 The intellect thus picks out what is
of interest to it from the scientific standpoint, but it does so using
certain principles which give to certain facts a value and reference that
they might not assume under different guiding lights. However, this
activity on the part of the intellect (highlighting and selecting certain
facts) does not constitute creation and distortion. Certainly, basic a-
statements presuppose a number of theoretical propositions, but these
propositions are propositions having to do with what is to be measured
and how to measure it; that is, they are propositions whose terms are
operationally defined. In more complex cases, facts mediately rather
than immediately disclosed to the intellect are taken from data
conceived in reference to already formed explanatory theories, or
which are derived from these explanatory theories. 24 Consequently,
scientific facts make up a hierarchy of value wherein those facts which
bear on real physical causes assume a higher value than those which
are the result of the physical being reduced to an instrument used for
discriminating between mathematical reconstructions or theoretical
entities understood to be beings of reason (entia rationis). According to
Maritain, such mediately established "facts" belong to explanatory
theory, rather than genuine fact.

21 Ibid.
22 Ibid., 52.

23 Ibid.

24 Ibid., 58.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 45

For Maritain, then, though fact and theory are truly distinct, they
nevertheless operate as dual partners in the scientific endeavor. Facts
are not the creations or distortions of the senses/intellect, but they are
the result of the senses/intellect exercising their judging capacity.
Following closely upon the distinction between fact and theory is
another distinction of great importance for Maritain's philosophy of
the experimental sciences: the distinction between what is observable,
and what is unobservable. As we have seen, in conventional inst-
rumentalism the 0-term/0-statement, T-term/T-statement distinction
turns on the fact that 0-terms reference what can be directly or
indirectly observed, while T-terms reference entities that cannot be
observed either directly or indirectly. This distinction in particular is
the basis for the semantical instrumentalists' claim that theories are
neither true nor false, and that consequently no incompatibility
between theories can arise. (The entities referenced by T-terms used in
T-statements are unobservable and hence unverifiable. As a result, T-
terms and the T-statements that use them are meaningless.)
Maritain, on the other hand, maintains that the best that the
experimental sciences can do is to construct theories that are true
(even though at times apparently incompatible) as long as they "save
the appearances," and are fruitful for making predictions and
determining/formulating new experimental laws.
Maritain arrives at this position precisely because the sciences of
observation necessarily employ a kind of intellection that cannot
capture essences, or lead to a direct apprehension of causes. Indeed, he
has even coined a term for this special kind of intellection, which he
calls "perinoetic" or circumferential knowing, which is to be
distinguished from what he calls "dianoetic" intellection, or
intellection which can grasp essences (though only indirectly). In
general, the intellect in its abstractive activity attempts to give us
knowledge of a thing's essence, but in the beginning it succeeds only in
revealing to us the commonest and poorest notes of intelligible being.
Maritain illustrates his point using the example of fire. 25 From the
outset, in forming the idea of fire, the intellect knows only that fire
represents some determined thing which produces certain sensible
effects like burning and glowing. But the essential characteristics of fire

25 Ibid., 30.
46 MATTHEWS. PUGH

that explain these sensible effects or properties elude the initial grasp
of the intellect. In fact, even if the intellect were finally to succeed in
grasping the essence of fire, it would grasp it only in and through its
essential properties-it would never grasp the essence directly in itself.
Perinoetic intellection fails to grasp the essences in their specificity
precisely because, below man, these specificities belong to the purely
sensible world. For man, whose specificity is purely intelligible, it is
possible to deduce essential properties from the difference "rational."
This is not the case with species below man. Taking an example from
Garrigou-Lagrange, 26 Maritain says that we are able to identify many of
the common sensible (descriptive) characteristics, of say mercury-it's a
liquid at ordinary temperatures; silvery in color; solidifies at 40°C; boils
at 360°C; is very heavy; is very toxic; etc. The specific difference that
explains why mercury has these properties remains hidden to us. Not
even the philosophy of nature can capture the specific differences of
beings lower than man on the scale of nature.
Thus, the type of intellection employed by the experiment-
tal/inductive/ observational sciences is not able to uncover the
intelligible constituents of the beings they study. For these sciences,
essences and the intelligible constituents of essences are always
hidden; essences, causes, or reasons for being only reveal themselves in
effects-never in themselves-and even their effects must be grasped in
signs and symbols that are substituted for them. But effects do reveal
experimental constancies, which are in turn signs of necessities or
essential connections hidden in constancies. And these constancies are
formulated in science as experimental laws. As Maritain says,
experimental laws enfold essences but without revealing them.
The sciences of explanation, on the other hand, are so-called
because the type of intellection that they employ genuinely penetrates
to the essence, the cause, or the reason for being of the phenomenon or
thing that it studies, though again, never directly or in itself, but only
in and through its proper accidents. These sciences, like mathematics
and philosophy, are strictly deductive in nature. Both mathematics and
philosophy use the dianoetic mode of intellection in which "the
intelligible constitutive is objectivized in itself (if not by itself at least
by a sign which manifests it, by a property in the strict sense of the

26 Ibid., 176, n. 2.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 47

word.)" 27 Yet, once again, even here the essence is only indirectly or
mediately known. In other words, in dianoetic intellection, essences are
known indirectly through the accidents which manifest them. More
exactly, essences reveal themselves to the intellect in and through
their proper accidents. For example, human nature is known by
rationality, animal nature by sensitivity. These are proper accidents
because they have generic fecundity. From rationality, one can deduce
docility, risibility, etc.
In perinoetic intellection, however, even the proper accidents
remain unknown. Rather, the intellect grasps sensible or common
accidents (as observable or measurable) which mask genuine
properties. Though the intellect succeeds in perinoetic intellection in
circumscribing the intelligible in the sensible, the intelligible "core"
escapes its grasp, and so fails to uncover the essence. Instead, the
substantial nature is known by signs which hide rather than manifest
the essence. While dianoetic intellection enables us to know substances
by proper accidents (which in turn are known by other accidents that
are their operations), perinoetic intellection gives us knowledge of
substances and properties by signs and in signs. These signs reference
descriptive properties such as density, atomic weight, melting point,
spectrum of high frequency, etc., which, though indispensible (for
example, in chemistry), nevertheless mask real ontological properties. 28
Hence, for Maritain, the essences or causes which the intellect
naturally seeks cannot become known in the inductive sciences, the
sciences of observation. And of course in this respect, then,
causes/essences are also unobservable. But though they are not know-
able or observable, this does not, as we shall see, make the theoretical
terms which reference theoretical entities meaningless. Maritain was

27 Ibid., 203, n. 1.
28 For those Thomists who may be wondering whether this distinction is

grounded in the texts of St. Thomas, see: In II Analyt., cap. XII, lect. 13; In spir.
creat., a. 11, ad 3; ST I, q. 29, a. 1, ad 3; Dever. I, q. 4, a. 1, ad 8; In VII Meta., lect.
12; In I De an., lect. 1; De ver., q. 10, a. 1, ad 6. These are just a few of those
places in the texts where this distinction is supported. I intend, in a future
article, to examine this issue much more closely.
48 MATIHEW S. PUGH

no positivist. 29 But our inability to penetrate beyond the sign

substitutes for essences in perinoetic intellection creates a distinction
that is absolutely fundamental for Maritain.
Nevertheless, because essences are the connatural object of the
intellect, the intellect naturally tries to close the gap between the
sciences of observation with their perinoetic form of intellection, and
the sciences of explanation with their dianoetic form of intellection.
The result has been, as all readers of Maritain know, the creation of
mathematical physics, a scientia media30 which has the physical for its

29 For Maritain's negative assessment of positivism, cf. Philosophy of Nature,

translated by Imelda Choquette Byrne (New York: Philosophical Library,
1951), 45-73.
30 On the notion of the scientia media, see Ibid., 102-18. This is a crucially
important notion for Maritain, and one that is quite central to his whole
concept of science. The experimental sciences naturally seek completion in
the sciences of explanation. The sciences of explanation are constituted by
mathematics and the various forms of philosophy, including, of course, the
philosophy of nature. As noted, these sciences are genuinely deductive in
form, and hence truly explanatory in nature. The experimental sciences
need them because, as Maritain says, "the resolution of concepts into the
observable and measurable as such is not sufficient" (Ibid., 102).
Consequently, the experimental sciences must be subalternated to either
mathematics, or philosophy. Now one science is subalternated to another
science when the subalternate science borrows its principles from the
subalternant science. The subalternant science resolves the conclusions of
that science into first, self-evident principles, but the subalternate science
does not by itself do so. Take a science such as geometrical optics. This
subalternate science is subalternated to the subalternant science of
geometry both as regards principles and subject. Optics borrows its
explanatory principles from geometry, since it understands light rays in
terms of geometry. And its subject is taken from geometry as well, for optics
studies visual line. But it adds something to line, namely, an accidental
difference grounded in the sensible (visual), in respect of one of the proper
objects of the subject of geometry, which in this case is line. Here the visual
becomes drawn into the mathematical sphere of intelligibility. The visual, in
other words, becomes conceivable entirely in terms of mathematics. Hence,
optics sits astride both the sensible and the mathematical. It is materially
physical, but formally mathematical. But, because it is subalternated both
with regard to principles and subject, it belongs to "the physical degree of
abstraction materially, and to the mathematical formally" (Ibid., 105). Thus,
geometrical optics is truly an intermediary science. The problem, of course,
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 49

subject matter, which it then joins to mathematics, a deductive science

which gives to physics its principles and mode of explanation. In other
words, mathematical physics is a science that is materially physical and
formally mathematical; a science of the physical real, but which knows
the physical real by transposing the physical real into entia rationis.
These mathematical beings of reason then take the place of real causes.
Thus, mathematical physics attempts to satisfy the intellect's demand
for causal explanations, but it does so by necessarily replacing real
causes with entia rationis.
However, because mathematical physics is materially physical, the
mathematics in question is applied to sensible nature as this is given to
the intellect through the senses. That is, the sensible effects and
experimental constancies that reveal essential/necessary connections
in and through sign substitutes are reconceived in terms of the formal
connections of mathematical relations. These mathematical relations
then become the substitutes for real causes. As a result, mathematical
physics, "has given up the direct search for real causes in themselves,
and aims to translate ... its measurements of things into a coherent
system of equations." 31 Mathematical physics builds a hierarchy of at
least formal causes, which in this science are simply the "conformity of
phenomena to mathematicallaw." 32 Also, "Here the substitute for the
ontological quid est is not an inductively established law, but a
mathematics quid est, an algorithm of the physical real."33
Thus, in mathematical physics, causes are reduced to the
measurable, because by its very nature it conceives the real causes in

is that the system of explanatory reasons and causes that mathematics

constructs to explain the sensible, taken as these are from the second degree
of formal abstraction, or abstractio formalis, is made up of entia rationis, rather
than entia realia: beings of reason rather than beings of real ontological
causes and principles. Nevertheless, even entia rationis never become entirely
severed from the real, for no matter how far removed or indirect from the
real, entia rationis are ultimately grounded in observed and measured real
beings. Yet, at some point, the sensible is so drawn into the mathematical
that unreal/ideal mathematical entities become the means by-'which real
beings are deduced.
31 Ibid., 45.

32 Ibid., 47.

33 Ibid., 55.
50 MATIHEW S. PUGH

terms of the quantitative, which mathematics abstracts at the second

degree of formal abstraction.34 In mathematics, only the quantitative
has meaning. Measurements, then, are organized into formulae that
capture the relations among them, and it is these relations captured by
formulae that become substituted for real causes, and that function as
explanations in mathematical physics. As long as these explanations
cohere with the initial measurements taken from instrument readings,
the theories are true. So, although real physical causes are
mathematically reconceived in mathematical physics, they are
nevertheless grounded in the physical real through measurement.
In this type of science, then, "a physico-mathematical theory will be
called true when a coherent and fullest possible system of math-
ematical symbols and the explanatory entities it organizes coincides
throughout all its numerical conclusions with measurements we have
made upon the real."35 Thus, any theory that satisfies this definition
"saves the appearances" and is thereby true. Maritain is quick to add in
a footnote that this criterion of saving the appearances does not mean
that mathematical physics rejects causal research as the search for
causal explanations. 36 Rather, these mathematical formulae are themselves
causal explanations, and the theoretical entities they reference are true
(that is, they exist) insofar as they save the appearances without
"making any claim to penetrate the nature of things themselves."37
The example Maritain uses here is that of the electron, which for
some has only a mathematical existence, since it is a center of vibration
in a wave system, which is itself taken to be real. For others, only the
waves have mathematical existence, having been substituted for a
surrounding but nevertheless real discontinuous field. Over time,
however, mathematical physics does at least point to the existence of

34 For those unfamiliar with Maritain's philosophy of the three degrees of

formal abstraction, cf. The Degrees of Knowledge, 35-46; Philosophy of Nature, 12-
31; Existence and the Existent, translated by Lewis Galantierre and Gerald
Phelan (Lanham, Maryland: University Press of America, 1987), 10-46. See
also Matthew S. Pugh, "Maritain, the Intuition of Being, and the Problem of
the Proper Starting Point for Thomistic Metaphysics," The Thomist 61, no. 3
(1997): 405-24.
35 The Degrees of Knowledge, 62.
36 Ibid., 63,·n. 1.
37 Ibid.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 51

certain entities, even if it cannot agree about their nature. This seems
to be true in the case of atoms.
Maritain's instrumentalism is therefore best described as a form of
ontological instrumentalism. Theoretical entities (here reconfigured as
entia rationis) point to the existence of something, and so entia rationis
have meaning-the meaning that all mathematical entities have insofar
as they are sign and symbol substitutes for what is ultimately grounded
in quantity, i.e. the measurable real-but their true nature entirely
eludes the grasp of the intellect. Therefore, whether such entities exist
is in the end of no real consequence for science, as long as they
function explicatively in the equations of the physical theory in
question, that is, as long as the entities in question are defined by at
least theoretically realizable operations of measurement.38 Indeed,
according to Maritain, mathematical physics is full of various kinds of
entia rationis, which together form a hierarchy of such beings. For
example, there are beings of reason that simply correspond in a more
or less direct way to experimental observations and causes. Other
beings of reason are genuinely theoretical entities like the electron or
quark, which appear to be real, but which can only be grasped through
symbol substitutes. These are mathematically reconstructed beings.
Finally, there are beings of reason in mathematical physics that,
although they are founded on the real in that they are taken from
measurements of the real, are absolutely incapable of existing as such.
In this instance, Maritain mentions "Einsteinian times." 39
Of course, these beings of reason are the direct creation of the mind
insofar as the mind is able to view the quantitative dimension of the
real quantitatively, rather than ontologically. That is, the mind is able
to formally abstract quantity from the real and see in it, "the very
relations of order and measurement which the objects of thought
discernible in it, as forms or essences proper to it, maintain among
themselves."40 When so abstracted, quantity is reconfigured by an
intuition of the internal sense, namely, imaginative intuition, which is

38 Ibid., 140.
39 Ibid., 141. Maritain is simply referring to Einstein's "space-time" as opposed

to either "space" or "time" alone.

40 Ibid., 143.
52 MATIHEW S. PUGH

linked to the real only in that it presupposes external perception.41 As

mathematics has progressed, however, it has discovered beings of
reason not directly figurable in imaginative intuition, such as irrational
number, imaginary number, and transfinite number. Hence, the
universe that mathematical physics creates based on the use of these
beings ofreason is also unfigurable in the imagination.42 Nevertheless,
no matter how abstract or unfigurable to the imagination physics
becomes, all of its entia rationis are ultimately grounded in real beings,
because they are ultimately grounded in real quantity.
Furthermore, at the first degree of abstraction, the mind attempts to
move in either of two directions; either upward toward the ontological
(this leads to the philosophy of nature) or downward toward the
sensible (this leads to the experimental sciences). In the latter case, the
observability of the object is crucially important. Its terms therefore,
must reference, either directly or indirectly, what can be observed, or
at least be reconfigured in imaginative intuition. But it never succeeds
in disengaging the ontological for itself; it never rises, in other words,
above perinoetic intellection. This also holds for mathematical physics,
for even when the experimental sciences try to rise above perinoetic
intellection by joining mathematics to physics and thus do embrace a
dianoetic mode of intellection, they have to be content with quasi-real
entia rationis, rather than real causes or reasons for being.

41 Maritain's explanation of the imaginative intuition employed in math-

ematical physics constitutes one of the most impressive parts of The Degrees
of Knowledge. Quantity precedes quality in the priority of accidents, yet
quantity can only be made known to us through the sensible qualities of
things. Thus, the imagination, which presupposes perception but is free from
it, is able to penetrate to quantity formally abstracted from matter. In this
way, pure quantity becomes known to us in sensible symbol substitutes of
the object of pure quantity that are free (because imaginatively recon-
figured) of every sensible or experimental condition. Yet these beings of
reason (the quantitative) are not purely intelligible, and so must be
reconfigurable in imaginative intuition in order to assure us that they are
grounded in genuine essences, that is, in the real. Those mathematical beings
that cannot be reconfigured in imaginative intuition must, by analogy, fall
indirectly into the imaginable. (Cf. Ibid., 144. See also Philosophy of Nature, a
genuine Classic.)
42 Maritain, The Degrees ofKnowledge, 146.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 53

The consequences of such a view are dramatic for a Thomistic

philosophy of science, for, according to Maritain, the sciences of
observation must give up-which, as a matter of historical fact, they
have already done-the search for essences, for what things are in
themselves. Empirical science shifts to a consideration of what is
measurable, and to how these measurements can be linked together to
form mathematically expressed laws. As Maritain says, "Every
definition should be given not now by means of the proximate genus
and specific difference, but by well-determined observable and
measurable properties, with the means of rediscovery and practical
verification being related in each case."43 And, most importantly, "the
possibility of observation and measurement replaces the essence or
quiddity which philosophy seeks in things." 44
These same considerations apply to the notion of causality. Under
the mathematization of physics, with its ever-increasing reliance on
beings of reason, science moves historically from (i) an initial
ontological notion of cause as that which is productive of being, to (ii)
an empiriological-ontological notion of cause as a phenomenon
productive of another phenomenon, to (iii) a mechanistic notion of
cause as a phenomenon to which another phenomenon is linked by a
universal necessary connection expressible as a "law" of nature, to (iv)
a pure empiriological notion of cause "as the spatia-temporal con-
ditions of a phenomenon ... the observable and measurable deter-
minations to which a phenomenon is linked."45 The latter finds ex-
pression in mathematical formulae capturing functional relations-as
in differential or tensorial calculus-but even more prominently in
wave mechanics, where waves are ultimately viewed as mathematical
or statistical constructs expressed as a mathematical symbol, and
which appear to eliminate strict mechanistic causality or deter-
mination from the subatomic world. Here, waves are transformed into
a series of probabilities. Heisenberg's uncertainty principle, for
example, enables the physicist to specify the measurable deter-
minations to which a phenomenon is linked, to specify the spatia-
temporal conditions of the phenomenon in question, but not to
determine simultaneously both the location of a particle and its

43 Ibid., 149.
44 Ibid.

45 Ibid., 150.
54 MATTHEWS. PUGH

momentum. That a mechanistic notion of causality would give way to

indeterminism was inevitable, given physics' ever-increasing
dependence on mathematical beings of reason. Strictly speaking,
causes for Maritain are either proximately observable, or are
theoretical causations. Proximately observable causes are causes that
are indirectly observable through experimental constancies, as when
molecules are broken down into ions in electrolysis, or when the height
of liquids in barometric tubes changes with atmospheric pressure.
Theoretical causations, on the other hand, are truly unobservable and
have to be mediated through mathematical beings of reason as used in
complex physico-mathematical theories, which, in turn, have to use
correspondence rules to "tack down" the theory to experimental
observations. Examples abound from quantum mechanics, subatomic
physics, photon theory, string theory, etc.
In effect, the mathematization of physics has freed physics from
ontology.46 Simultaneity, time, space, etc. are freed from essentialist
conceptions and instead are understood in purely empiriological ways.
Indeed, for Maritain, mathematical physics frees science not only from
ontology, but from philosophy and even common sense as well. 47 The
New Physics does, and must, give up the search for essences. Instead,
the contemporary physicist is confronted with symbols, rather than
materials understood as forces that are, as Eddington says, "familiar in
the workshop." 48
Maritain even goes so far as to call these symbols "myths."49 The real
is known via the mathematical preater-real, he says, which transforms
the world of qualities into a world of quantities expressible as
mathematical functions. Hence the "myths" in question are only tied to
the real insofar as they agree with the measurable. As we have seen, to
that extent theories are true; true that is, as long as they save the
appearances. 50

46 Ibid., 158.
47 Ibid., 159.

48 Ibid.

49 Ibid.

50 Ibid., 162.
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 55

However, symbolism (myth) and realism are united in mathematical

physics and "constitute the warp and woof of the same cloth."51 But-
and this is often overlooked in Maritain's philosophy of science-the
realism that he speaks of is the result of (i) our pre-philosophical
understanding of nature, and (ii) our ontological knowledge of nature,
first in the philosophy of nature proper, and second in metaphysics.
(This is why in Maritain's, at times confusing, diagram of the three
degrees of formal abstraction in The Degrees of Knowledge, the second
degree resulting in the creation of mathematical physics lies off the
line that leads from the philosophy of nature to metaphysics.) Science,
which includes: (a) infra-scientific experience; (b) empirical science
(not yet mathematized); (c) physico-mathematics, and; (d) mathe-
matics, lies below (epistemologically speaking) philosophy, which
includes (1) the philosophy of nature, and (2) metaphysics. By itself,
mathematical physics cannot give us knowledge of essences or causes.
Yet, because it rests upon infra-scientific experience, and because it can
avail itself of the philosophy of nature, mathematical physics can never
stray so far from the real that it becomes purely positivist or
pragmatist in its outlook. But, in its day-to-day operations, the question
of the reality of the referents of its theoretical terms, whether they are
mere beings of reason or actualities, is utterly irrelevant to it.
In light of our brief summary of instrumentalism as well as our brief
summary of Maritain's philosophy of the experimental sciences, we are
now in a position to address the question posed at the beginning of our
discussion: Is Maritain an instrumentalist? The answer must be a
qualified yes, though only slightly qualified, for it would appear that
Maritain would agree with most of the elements of an instrumentalist
philosophy of experimental science. With only minor exceptions,
Maritain, I believe, would agree with the following statements:
1. The primary function of experimental science is not to explain in
any conventional or causal realist sense, that is, is not to search for and
identify real essences, causes, or reasons for being, but rather to predict
and control.
2. Theories are successful when they work, that is, when they save
the appearances.

51 Ibid., 163.
56 MATTHEW S. PUGH

3. a-statements and T-statements, fact and theory, represent a

difference in kind, not degree. The referents of T-terms are
unobservable, being hidden by symbol and sign substitutes, which are
conceived in mathematical physics as entia rationis, though (and this
represents a difference between Maritain and semantical instru-
mentalism, though not ontological instrumentalism) this does not
make T-statements meaningless, since the symbol/sign substitutes (as
entia rationis) are ultimately grounded in the real. Hence, for Maritain,
both a-statements and T-statements are true or false, though T-
statements are only true insofar as they are based on, or taken from
accurate measurements and pointer readings, and draw conclusions
that cohere with all the relevant measurements. But unlike a-
statements, which can be directly verified in sensory experience, T-
statements along with T-terms must be mediated via mathematical
beings of reason that necessarily misrepresent what they purport to
reference. T-statements, then, are not true in the way that a-
statements are factually true. Rather T-statements are only true in the
sense that they save the appearances. Consequently, two apparently
incompatible theories in mathematical physics can be true as long as
they save the appearances.
4. a-statements are operationally definable. a-statements or
experimental laws are inductive generalizations of constant relations
holding between observed data.
5. Theoretical entities, as the referents ofT -terms, are operationally
indefinable. Hence they become useful tools for calculating and
predicting, rather than causes or reasons for being.
6. T-terms are defined by the theory that uses them, and so cannot
be put to direct empirical test. Thus, theories are not inductive
generalizations of experimental laws.
7. As mathematized, T-statements or theories and the theory terms
they use are functional notions that capture regularities or constancies
evident among the relata of phenomena.
8. T-statements are linked to a-statements via correspondence
rules, but they do not function as premises from which conclusions are
drawn. Instead, they are guidelines or rules for making the transition
from one set of a-statements to another set of a-statements. T-
statements constitute an abstract calculus in accordance with which a-
statements are inferred. Here, the nature of, or even the reality of, T-
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 57

referents is irrelevant. The logical relations in which the T-terms stand

to each other, however, are crucially important.
9. The deductive-nomological (D-N) model of scientific explanation
must be understood in the above way. The explanans provides a set of
logical rules, or statement forms mathematically expressed as func-
tions or equivalencies between phenomenal relata, (variables) that
provide guidelines for inferring the explicandum, which "was to be
expected." These mathematical functions and equivalencies are them-
selves the explanations in experimental science, which take the place of real
causes.
10. T-statements and T-terms only appear to reference real causes
necessarily connected to real effects. In fact, what theories identify are
regularities, not necessary connections, which again are expressed as
mathematical functions subsuming a whole host of lesser experimental
regularities or laws. The regularities in question are thus functional
regularities. Necessity is treated hypothetically or conditionally. Since
this regularity model of causality does away with necessary connection,
experimental sciences can even embrace something as startling as
Heisenberg's uncertainty principle.
11. Because T-terms are implicitly defined by their place in theories,
and because their referents are unobservable and thus use ideal
mathematical sign/symbol substitutes in their place, theories are not
true in any conventional sense. (T-statements are in fact statement
forms, rather than real propositions.) Since entia rationis do not
represent things, T-statements cannot in the strictest sense, therefore,
be factually either true or false. But T-statements are not for Maritain
meaningless. So theories are only true in an unconventional sense,
insofar as they save the appearances, and enable the scientist to make
predictions that cohere with measurements and pointer readings.
Thus, incompatible theories, as in electron theory or photon theory,
can all be true under this broad conception of truth.
12. Entia rationis, although symbol substitutes for the real, are
nevertheless grounded in the real, that is, in quantity, even when they
become unfigurable in imaginative intuition. They reference something
that exists (for the most part), though they necessarily misrepresent it.
But as long as entia rationis are true to the facts, that is, cohere at some
point with 0-statements, the theories to which they belong are true. In
other words, entia rationis are double-valenced when it comes to truth
and meaning; on the one hand, they are true insofar as they are taken
58 MATTHEW S. PUGH

from the real, either directly or indirectly, and so have the meanings
that measurements taken from the real give them; but, on the other
hand, they are convenient fictions whose meanings are determined
strictly by the logical place which they hold as variables in math-
ematical formulae. At that level, they have merely the logical meaning
that statement-forms have, which are not true propositions. At that
level, their truth lies purely in their workability in terms of saving the
appearances.
So conceived, the experimental sciences may well have to make way
for a non-classical intuitionist logic which rejects the law of the
excluded middle, which certainly would be more compatible, for
example, with quantum mechanics. In other words, entia rationis are
double-valenced-insofar as they are grounded in the real through
quantity.

III. CONCLUSION
It seems then, that Maritain is an instrumentalist of sorts when it
comes to the experimental sciences. Nevertheless, as any reader of
Maritain knows, his philosophy of the sciences of observation is
incomplete without a complementary causal realist philosophy of
nature. As Maritain never tires of pointing out, empiriological analysis
must be completed by ontological analysis. The first analyzes and
resolves its concepts in the sensible, the observable; the other analyzes
and resolves its concepts in intelligible being. In the first, definitions
are sought through observation, measurement, and physical oper-
ations; in the second, definitions are sought by means of ontological
properties constituting essences. It is the task of the philosophy of
nature, operating at the first degree of abstractive visualization, to
provide this completion. This is crucially important, for it is only in this
way that science as a whole for Maritain is able to overcome positivism,
and only in this way that Maritain's instrumentalism/constructive
empiricism can be separated from the verificationist-based instru-
mentalism of the logical positivists. Indeed, it is the very possibility of
ontological analysis which is the gllarantee of a genuine (truly
explanatory) philosophy of science. ·In other words, it is Maritain's
critical realism that creates the foundation necessary for the
completion of empiriological analysis in the ontological analysis of the
philosophy of nature. Infra-scientific experience, along with the
ontological analysis of dianoetic. intellection operative in the philo-
sophy of nature, embeds or enfolds the perinoetic intellection
 THE PHILOSOPHY OF EXPERIMENTAL SCIENCE 59

operative in the experimental sciences. For Maritain, these two parts of

the philosophy of science really need to be thought together. Hence, it
would a gross injustice to Maritain to accuse him of simply holding to a
qualified instrumentalist philosophy of science. Nevertheless, whether
these two halves of the picture really fit together is a question for
another study altogether.