Professional Documents
Culture Documents
Schrödinger Logics
Newton C. A. da Costa
Department of Philosophy
University of São Paulo
Décio Krause
Department of Mathematics
Federal University of Paraná
Abstract
Schrödinger logics are logical systems in which the principle of identity
is not true in general. The intuitive motivations for these logics are both
Erwin Schrödinger’s thesis (which has been advanced by other authors)
that identity lacks sense for elementary particles of modern physics, and
the way which physicists deal with this concept; normally, they under-
stand identity as meaning indistinguishability (agreemment with respect
to attributes). Observing that these concepts are equivalent in classi-
cal logic and mathematics, which underly the usual physical theories, we
present a higher-order logical system in which these concepts are system-
atically separated. A ‘classical’ semantics for the system is presented and
some philosophical related questions are mentioned. One of the main
characteristics of our system is that Leibniz’ Principle of the Identity of
Indiscernibles cannot be derived. This fact is in accordance with some
authors who maintain that quantum mechanics violates this principle.
Furthermore, our system may be viewed as a way of making sense some
of Schrödinger’s logical intuitions about the nature of elementary parti-
cles.
∗ The authors very much would like to thank the anonymous referee for his quite interesting
1
Schrödinger Logics 2
Contents
1 Introduction 2
2 First-Order Systems 6
3 Higher-Order Systems 7
4 The semantics of Sω 9
1 Introduction
It is well known that there are fundamental differences between the classical
and the quantal descriptions of the world, as for instance in the problem of
describing a system of ‘identical’ particles. Of course, if we admit a system
composed by two electrons then, according to classical mechanics, it is possible
to follow the trajectories of the electrons and to distinguish them in every instant
of time. But, according to quantum mechanics, if there is a superposition of the
wave functions associated to each particle, it becomes impossible to say which
electron is associated with what wave function (see for instance Eisberg 1961,
Chap. 12 and French 1989).
This situation, which occurs for instance in a helium atom in the fundamental
state, where the wave functions of the electrons are completely superposed, has
motivated some discussions not only about the meaning of concepts like identity
and indistinguishability in physics (see French 1989), but also about the nature
of semantics and the manner in which we ‘talk’ about this kind of entities
(Toraldo di Francia 1986, Dalla Chiara and Toraldo di Francia 1985). In fact,
entities like the electrons of an helium atom in the fundamental state suggest
that ‘classical’ concepts like identity apparently do not apply as they do in the
case of macroscopic bodies. Physicists, of course, use this concept (identity) in
the sense of indistinguishability; when they say that two electrons are ‘identical’,
they intend that the electrons are ‘indistinguishable’ in the sense that they have
all their ‘intrinsic’ properties in common (cf. Jauch 1968, 275. See below).
Of course, there is some confusion regarding the terms identity and indis-
tinguishability in the physical literature. The use of identity as agreement with
respect to attributes (that is, as indistinguishability) demands some qualifica-
tions. In fact, let us observe that since the usual underlying logic of the physical
theories is classical logic and mathematics, it would be natural to suppose that
the concept of identity is the one ascribed by classical logic and mathematics.
Then, to say that ‘a is identical with b’ (in symbols, ‘a = b’), simply means
Schrödinger Logics 3
that there are not two distinct items, but only one, which can be referred to as
either a or b (see Krause 1990a).
From the technical point of view, in a second-order language with identity,
the definition of this concept (Leibniz’ Law) may be written as
x = y := ∀F (F (x) ↔ F (y))
where x and y are individual terms and F is a variable ranging over the set of
attributes of these individuals. Thus, the ‘traditional’ theory of identity may be
said to be the theory of absolute identity, since according to it, if two objects are
identical, they are absolutely identical , that is, they are the very same object.
The above definition does not permit treating identity and indistinguisha-
bility (agreement with respect to attributes) as distinct concepts: if two entities
are indistinguishable, that is, if they share all their attributes, then they must
be the same entity, and that is all. As it is well known, from the philosophical
point of view this characterization is attributed to Leibniz, who said that there
are no two objects which differ solo numero (see the references).
But, apparently, in the field of quantum mechanics the concepts of identity
and of indistinguishability are not equivalent in this sense. Physicists use the
term indistinguishability to mean that entites agree with respect to attributes,
and not to express the idea that they are the same object. This point will be
analysed below. But there are more radical opinions which we want to consider.
For instance, according to Erwin Schrödinger, we can talk about the indistin-
guishability of elementary particles, but the concept of identity simply has no
meaning with respect to them. In fact, in his book Science and Humanism
(1952, 16-8), he said that
When you observe a particle of a certain type, say an electron, now
and here, this is to be regarded in principle as an isolated event.
Even if you observe a similar particle a very short time at a spot
very near to the first, and even if you have every reason to assume
a causal connection between the first and the second observation,
there is no true, unambiguous meaning in the assertion that it is the
same particle you have observed in the two cases. The circumstances
may be such that they render it highly convenient and desirable to
express oneself so, but it is only an abbreviation of speech; for there
are other cases where the ‘sameness’ becomes entirely meaningless;
and there is no sharp boundary, no clear-cut distinction between
them, there is a gradual transition over intermediate cases. And
I beg to emphasize this and I beg you to believe it: It is not a
question of being able to ascertain the identity in some instances
and not being able to do so in others. It is beyond doubt that the
question of ‘sameness’, of identity, really and truly has no meaning.
The most plausible interpretation of the above quotation is that Schrödinger
referred to the impossibility one has of recognizing a particle, once observed in
Schrödinger Logics 4
a neighborhood of another one of similar species, in a time later than the first
observation. According to this view, we should be faced with the problem of
re-identifying the electrons, that is to say, with the problem of transtemporal
identity, a concept which has been widely discussed in the literature (see Quin-
ton 1973, Chap.3). But we think that the last part of the quotation suggests
that Schrödinger’s view is not only related to the case where physicists are in-
capable of individuating an electron once observed. Schrödinger suggested that
there is a circumstance in which it has no meaning to affirm that an electron
is identical or different from another. This view is in a certain sense supported
by Toraldo di Francia, who emphasized that our use of common languages to
talk about the entities of microphysics lacks, in general, adequate foundation.
In particular, he said (1986, 192-3) that
2 First-Order Systems
The system we will call S has the following categories of primitive symbols: (a)
connectives: ¬ (not) and ∨ (or) (the other ones are defined as usual); (b) the
universal quantifier: ∀ (for all) (the existential quantifier is defined as usual);
(c) parentheses and comma; (d) a denumerably infinite set of individual vari-
ables of first species x1 , x2 , . . . and a set of individual constants of that species:
a1 , a2 , . . .; (e) a denumerably infinite set of individual variables of second species:
X1 , X2 , . . . and a set of individual constants of second species: A1 , A2 , . . .; (f )
the symbol of equality: =; (g) for each natural number n > 0, a non-empty col-
lection of n-ary predicate symbols. We utilize x, y, etc. as syntactical variables
for any terms.
By a term, we understand a variable or a constant; hence, we have terms
of first species (which we call m-terms) and of second species (called M -terms).
Intuitively, the m-terms may be thought of as denoting the microscopic objects
of our environment, while the M -terms denote the macroscopic ones. To this last
category, we suppose that classical logic is valid in all its aspects. The formulas
of S are defined in the standard way (see Wang 1952), but it is convenient to
observe that x = y will be considered as a formula if and only if x and y are
terms of second species.
Since the expression x = y is not a formula if x or y are not terms of
second species, the language of S does not permit us to talk about either the
identity or the diversity of the m-objects. This concept (of identity) simply
does not apply to them, as Schrödinger’s above quotation suggests. The other
syntactical concepts are the usual ones. The propositional postulates (schemes
of axioms and inference rules) for S are also standard, based on the primitive
connectives we selected (for instance, the system of Hilbert and Ackermann
1950); concerning quantification and identity, we use the following postulates:
1. ∀xA(x) → A(t), where x is a variable and t a term of the same species of
x which is free for x in A(x).
2. C → A(x)/C → ∀xA(x) where x does not occur free in C.
3. ∀x(x = x) where x is a variable of second species
4. t = u → (A(t) ↔ A(u)) with the usual resctrictions, provide that t and u
are both terms of second species.
Schrödinger Logics 7
3 Higher-Order Systems
The system we will construct here, denoted Sω , is a natural extension of the first
order logic S to the simple theory of types (STT); therefore, we will not describe
all the details concerning its language, which contains the same connectives, the
equality symbol and auxiliary symbols as in S. With respect to variables and
constants, we observe firstly that the set Π of types is defined as the smalest
set such that: (a) i ∈ Π (i is the type of the individuals) and (b) if k1 , . . . , kn
belong to Π, then hk1 , . . . , kn i ∈ Π.
Then, for each type k ∈ Π and k 6= i, there exists a denumerable infinite
set of variables X1k , X2k , . . . of type k and a set of constants Ak1 , Ak2 , . . . of that
type. When k = i, there are two sorts of terms: the M -terms and the m-terms.
The former are the variables X1i , X2i , . . . and the constants Ai1 , Ai2 , . . ., called M -
variables and M -constants respectively, whereas the latter are the m-variables
xi1 , xi2 , . . . and the m-constants ai1 , ai2 , . . .. In other words, we have a two sorted
language at the level of the individuals. We use U k , V k , . . . and u, v, . . . as
syntactic variables for terms of type k (including the M -terms of type i) and
Schrödinger Logics 8
for m-terms respectively; U, V, . . . are used as syntactic variables for any terms
in general.
The definition terms of type k and of atomic formulas are the standard ones
(see Hilbert and Ackermann 1950, Chap. 4) but, with respect to the predicate
of identity, we require that only expressions of the form U k = V k are atomic
formulas; hence, expressions like U k = u, or like u = v etc., are meaningless.
Consequently, we cannot talk either about the identity or about the diversity
of the objects denoted by the m–terms. The postulates of this logic are similar
to those of the standard higher-order systems (Ibid.), including the axioms of
Extensionality, Separation and Infinity. The case of the Axiom of Choice will
be discussed below. With respect to equality, we introduce the following axiom:
U k = V k ↔ ∀F (F (U k ) ↔ F (V k ))
where U k and V k are terms of type k, but not m-terms, and F is a variable of
type hki.
U ≡ V := ∀F (F (U ) ↔ F (V ))
Theorem 3.1 Let A be a formula of STT and Aω its translation in the language
of Sω . Then, if A is a theorem of STT, Aω is a theorem of Sω . In particular,
all the mathematics developed in STT can be developed in Sω .
We will stop here with the development of Sω . We hope that what was said
is sufficient to provide an idea of the nature of Schrödinger logics.
Schrödinger Logics 10
4 The semantics of Sω
In this section, we outline a proof of a ‘weak’ completeness theorem for Sω in
the sense of Henkin (1950) (cf. Church 1956, Section 54). In what follows, the
language of Sω will be denoted by L.
If D is an infinite set such that D = m ∪ M and m ∩ M = ∅, then a frame for
L based on D is a function M whose domain in the set Π of types such that M(i)
= D and, for each type k = hk1 , . . . , kn i ∈ Π, M(k) ⊆ P(Mk1 , × . . . × Mkn ).
If the inclusion in this last expression is replaced by the equality symbol, than
the frame is standard .
If we write Mk instead of M(k), then the frame can be viewed as a family
(Mα )α∈Π of sets satisfying the above conditions. In what follows, we will refer
to both this family and the set F = {X : ∃k ∈ Π ∧ X = M(k)} indiferently as
the frame for L based on D.
A denotation for L based on D is a function φ whose domain is the set of
constants of L, defined as follows:
1. φ(a) ∈ m
2. φ(Ai ) ∈ M
3. φ(Ak ) ∈ Mk for every k 6= i
With these results at hand, we can prove the following theorem (see Church
op. cit., 314; for the proof of this particular case, see Krause 1990):
Schrödinger Logics 12
charge 4.8 × 10−10 e.s.u.’ and ‘to have spin 1/2’ –see Toraldo di Francia 1981),
while L would stands for an appropriate label (for instance, spatio-temporal
location).
Even in the case when the particles have the same E, as in the case of
the electrons in the above considerations, they might be distinguished by their
labels. That is, if hE, Li i, i = 1, . . . , n, with Li 6= Lj , classical mathematics
applies to such ordered pairs, and there is a sense in which the particles can be
counted. We observe that in this case we are not dealing with the ‘particles’
directly, but with a ‘classical’ representation of them (the ordered pairs –see
below).
But, when the labels are the same, that is, when Li = Lj , the tools of
classical mathematics cannot be applied to count the pairs as representing dif-
ferent particles, since hE, Li i = hE, Lj i and the particles denoted by them must
be identified. But in quantum mechanics, at least in the case of bosons (van
Fraassen argues that in certain circumstances Leibniz’ Principle of the Identity
of Indiscernibles may be ‘saved’ –see van Fraassen 1991, 422–3), this is the case
when even the labels are the same (technically, the wave functions representing
the particles are completelly superposed) but the particles are not counted as
distinct. In this case, physicists cannot count them, but they use other kinds
of procedures to say that there are a certain number (say n) of particles in a
particular state. The quantity of them may be ascribed (in some instances) by
the measurement of some observables, the occupation number operators (it is
important to note that some cautionary remaks are needed in using the occu-
pation numbers to count the quantity of particles in a state –see van Fraassen
1991, 441). From the mathematical point of view, the aggregates of such parti-
cles have a cardinal number, but not an ordinal (this fact was observed firstly
by Dalla Chiara and Toraldo di Francia 1985). But there remains the problem:
employing our above terminology, we have n entities instanciated in just one
ordered pair.
Hermann Weyl also pointed out the importance, in various fields of knowl-
edge, of considering such collections of indistinguishable objects. But his ap-
proach to the subject used classical mathematics. In the symbolism introduced
above, Weyl’s ideas can be adapted as follows (for more details concerning Weyl
on this topic, see Weyl 1949, App. B, Krause 1990, 1991): let us call E the
collection of the pairs hE, Li i, i = 1, . . . n. Suppose now that there is an equiv-
alence relation R defined on E such that the elements of each equivalence class
of the quotient set E/R are the pairs having the same labels (in reality, there
would be just one pair in each class, but we will proceed with the argument).
If we call Ci , i = 1, . . . n these equivalence classes, and call them states, as did
Weyl, then if the elements of each class are absolutely indistinguishable, the
most information we may have is the decomposition n1 + . . . + nk = n, where
n is the number of elements of E and the ni are the quantity of elements in the
respective class. Weyl called this the effective state of an aggregate of individ-
uals. But, as is easy to see, this procedure translates the problem to another
Schrödinger Logics 14
References
[1] Ben-Menahem, Y. (1992) “Struggling with realism: Schrödinger’s case” ,
in Bitbol, M. and Darrigol, O. (eds.) (1992) Erwin Schrödinger: philosophy
and the birth of quantum mechanics, Editions Frontières, Gif-sur-Yvette :
25–40.
[2] Blizard, W. O. (1989) “Multiset theory” Notre Dame Journal of Formal
Logic 30 (1) : 36–66.
[3] Barnette, R. L. (1978) “Does quantum mechanics disprove the Principle of
the Identity of Indiscernibles ?”. Philosophy of Science 45 : 466–470.
[4] Church, A. (1956) Introduction to mathematical logic, Princeton Un. Press,
Princeton, Vol. 1.
[5] Cortes, A. (1976) “Leibniz’s Principle of the Identity of Indiscernibles: a
false principle”. Philosophy of Science 43 : 491–505.
[6] da Costa, N. C. A. (1980) Ensaio sobre os fundamentos da lógica, Hucitec–
EdUSP, São Paulo.
Schrödinger Logics 16