A (Leibnizian) Theory of Concepts

A (Leibnizian) Theory of Concepts

?

Edward N. Zalta

Center for the Study of Language and Information

Stanford University

?

Ifwehadit[a characteristica universalis], we shouldbeabletoreason

in metaphysics and morals in much the same way as in geometry

and analysis. ... If controversies were to arise, there would be no

more need of disputation between two philosophers than between

two accountants. For it would su?ce to take their pencils in their

hands, to sit down to their slates, and to say to each other ... : Let

us calculate.

1

G. W. Leibniz

I have therefore, in what follows, ...endeavored as far as possible

to exhibit the theory of monads as a rigid deduction from a small

number of premisses.

2

B. Russell

In the eyes of many philosophers, Leibniz established his credentials

as a clear and logically precise thinker by having invented the di?erential

and integral calculus. However, his philosophical and metaphysical views

were never expressed as precisely as the mathematics he developed. His

ideas about concept summation, concept inclusion, complete individual

?

This paper was published in Philosophiegeschichte und logische Analyse / Logical

Analysis and History of Philosophy, 3 (2000): 137–183.

?

I am indebted to John Perry and the Center for the Study of Language and Infor-

mation (CSLI) for continued support of my research at CSLI. I am also indebted to

Chris Swoyer, who read earlier drafts of this paper and made numerous valuable sug-

gestions; to Godehard Link, discussions with whom led to a significant improvement in

one of the definitions needed to represent Leibniz’s view; and to Andrew Irvine, David

Barker-Plummer, Eric Hammer, and Anthony Everett, for their critical observations

on this material.

1

This is the translation in Russell [1900], pp. 169-170. The source is G.vii 21 and

G.vii 200. (‘G’ refers to C. Gerhardt (ed.), Die philosophischen Schriften von Gottfried

Wilhelm Leibniz; see the entry for Leibniz in the Bibliography.)

2

See Russell [1900], p. viii.

Edward N. Zalta 2

concepts, possible worlds, and the containment theory of truth were never

developed within a single, overarchingsystem as preciseas the mathemat-

ical calculus. In this paper, we describe a system that may rectify this

omission.

Inwhatfollows,wearticulateatheoryofconceptsusingLeibniz’swork

on the logic and metaphysics of concepts as a guide. The development of

thetheoryconsistsprimarilyofprovingfactsaboutconceptsthatareoften

just stipulated to be true. The derivations of these facts take place within

the context of a precise metaphysical theory that has been developed in

previous work, namely, the axiomatic theory of abstract objects.

3

The

truth of the axioms of this theory will be preserved in the theorems of the

theoryofconcepts, butputting asidethequestionoftruth, weputforward

the following system as one way of reconstructing many of Leibniz’s ideas

about concepts.

Leibniz’s views on concepts are developed in two strands of his work,

which we shall refer to as his ‘logic of concepts’ and his ‘(modal) meta-

physics of individual concepts’, respectively. The first strand, the logic

of concepts, is outlined in the series of unpublished sketches in which he

formulated and reformulated algebras of concepts.

4

We shall focus on

what appears to be his most mature logic, namely, the one developed in

the 1690 fragment G.vii 236-247. We shall derive both the axioms and

theorems of this fragment in what follows.

The second strand of Leibniz’s work on concepts is the metaphysics

he develops in connection with the notion of an ‘individual concept’. This

notion played a significant role in the Discourse on Metaphysics,theCor-

respondence with Arnauld,theTheodicy,andtheMonadology.Inthese

works, it seems clear that Leibniz thought that the properties of an in-

dividual could be derived from its individual concept. To account for

contingent truths, he theorized about individual concepts in the context

of his metaphysics of possibility and possible worlds.

3

See Zalta [1983], [1988a], [1993], and [1999]. Certain Leibnizian features of this

metaphysics were first sketched in Zalta [1983], but this work is redeveloped and en-

hanced here (Section 8) in numerous ways. The present e?ort represents a more con-

sidered view.

4

See the fragments in G. W. Leibniz, Logical Papers, edited and translated by

G. H. R. Parkinson. Hereafter, we refer to this work as ‘LP’. (The full reference

is in the Bibliography.) Leibniz referred to these algebras as ‘calculi’ and Rescher

[1954] separates these calculi into 3 basic systems. He documents Leibniz’s intensional

and extensional interpretations of these systems, noting that the second system has a

propositional interpretation.

3 A (Leibnizian) Theory of Concepts

Although Leibniz didn’t integrate his modal metaphysics of individual

concepts and his logic of concepts, it seems clear that this is something

that should be done, since he applied the concept containment theory of

truth in both of these strands to his work. In what follows, we integrate

the logic and metaphysics of concepts and prove many of the claims that

Leibniz asserted in connection with individual concepts, including what

appears to be the fundamental theorem of his metaphysics of concepts,

namely, that if an object x has F but might not have had F,then(i)

the individual concept of x contains the concept F and (ii) there is a

individual concept y (a ‘counterpart’ of the concept of x) which doesn’t

contain the concept F and which appears at some other possible world.

I think that the formal derivations that we construct in developing the

logic and metaphysics of concepts goes some way towards understanding

Leibniz’s idea of a characteristica universalis and calculus ratiocinator.

Our principal goal in what follows, then, is to develop a correct theory

of concepts that links the two strands in Leibniz’s work and to do so by

provingmetaphysicalclaimsthatareoftenstipulated. In workingtowards

this goal, we shall use neither ‘possible world semantics’ nor set theory,

nor shall we build a semantical system which models the truth of logical

and metaphysical claims; indeed we hope to show that no set theory or

other mathematics is needed to develop a Leibnizian theory of concepts

with rigor,clarity, andlogicalprecision. Concepts will be systematized by

preciselyidentifying themwithin anaxiomatizedrealmofabstractobjects

which have well-defined, and antecedently-specified, existence and iden-

tity conditions. The interesting web of logical and metaphysical theorems

that result should establish that the existence and identity conditions for

abstract objects correctly apply to (Leibnizian) concepts. Moreover, the

containment theory of truth will be defined so that: (a) it applies both

to the logic and modal metaphysics of concepts, and (b) it is consistent

with the existence of contingent facts.

These accomplishments should justify the present contribution both

to the literature on concepts and to the secondary literature on Leibniz.

Of course, some theories of concepts in the literature do not look back

to Leibniz for inspiration.

5

Of the ones that do, some only treat the

logic of concepts and not the modal metaphysics of individual concepts.

6

5

See, for example, Peacocke [1991] and Bealer [1998].

6

See, for example, Rescher [1954], Kauppi [1960], [1967], Casta?neda [1976], [1990],

and Swoyer [1994], [1995].

Edward N. Zalta 4

Others treat only the modal metaphysics of individual concepts and not

the logic.

7

Although Lenzen [1990] attempts to treat both, it is not made

clearwhichofthe Leibnizianlawsthatheformalizesareaxiomsandwhich

are theorems.

8

Most of these works in the secondary literature assume

certain basic axioms (often stated in set-theoretic terms) which are then

used in the reconstruction of Leibniz’s ideas. In what follows, we plan to

derivetheseaxiomsastheorems.

§1: Concepts, Properties, and Concepts of Properties

Before we turn to the definitions and theorems that articulate the theory

of concepts, it will serve well if we spend some time discussing the dis-

tinction between concepts and properties in what follows. We shall not

identify concepts and properties. There are good reasons to distinguish

these two kinds of entity. Two of these reasons are ‘negative’ ones; they

tell us why we shouldn’t identify concepts and properties. There is also a

‘positive’ reason for distinguishing concepts and properties, which we will

get to in just a moment.

Thefirstreasonnottoidentify conceptsandpropertiesisthatitwould

get the Leibnizian logic of concepts wrong. The most important theorem

in Leibniz’s of logic concepts is:

(A) The concept F contains the concept G i? the concept F is identical

with the sum of the concept F and the concept G.

9

Swoyer [1995] and [1994] designates this principle ‘Leibniz’s Equivalence’.

We may represent this formally as follows:

(A) F followsequal G ≡ F=F ⊕G

Principle (A) becomes false when analyzedin terms of oneof the property

7

See Mates [1968], Mondadori [1973], and Fitch [1979].

8

See the various formalizations that Lenzen presents throughout [1990] and which

are listed on pp. 217-223.

9

Principle (A) is the ‘containment’ version of the conjunction of Principles 13 and

14 in LP 135:

Principle 13: If F ⊕G is identical to F,thenG is in included in F.

Principle 14: If G is included in F,thenF ⊕G is identical to F.

The source is G.vii 239.

5 A (Leibnizian) Theory of Concepts

theories that are now available.

10

Just consider the following property-

theoretic analysis of the notions involved:

The concept F =

df

F

FcontainsG(‘F followsequal G’) =

df

F?G (i.e., a50?x(Fx→ Gx))

The sum of concepts F and G (‘F ⊕G’) =

df

[λxFx&Gx]

11

These definitions do have the virtue of preserving the Leibnizian intu-

ition that the conceptpersoncontainstheconcept rational.

12

Butitis

now easy to see that (A) is false from the point of view of any property

theory that treats properties in a fine-grained way. For on the property-

theoretic analysisof containment andsummation just proposed, Principle

(A) would amount to the following:

F?G ≡ F=[λxFx&Gx]

This is clearly false as a principle governing intensionally conceived pro-

perties.

13

10

See Cocchiarella [1978], Bealer [1982], Zalta [1983], Chierchia and Turner [1985],

Menzel [1986], and Swoyer [1998] and [1996]. These theories allow for the possibility

of distinct, though necessarily equivalent properties.

11

The λ-predicate used in the definition of concept summation denotes a complex

property. We read the λ-predicate as follows: being an x such that both x exemplifies

F and x exemplifies G. Intuitively, the λ-predicate denotes the complex conjunctive

property: being F and G.

12

To see why, note that Leibniz regarded the concept person (or ‘man’) as the sum

of the concepts rational and animal. On the above analysis, once the concepts rational

and animal are identified with the properties of being rational and being an animal,

respectively, the sumof the concepts rational and animal is identified asthe conjunctive

property being rational and an animal ([λx Rx & Ax]). It therefore follows that the

concept person (i.e., the sum of the concepts rational and animal) contains the concept

rational, since the property of being rational is necessarily implied by the conjunctive

property being rational and an animal.

13

From the fact that a property F implies a property G,itdoesnot follow that

F just is identical to the conjunctive property being F and G.Intuitively,F may be

distinct from being F and G even if F implies G. Indeed, on a fine-grained theory of

properties, the right hand side of this biconditional is always false—it is always false

that F is identical with the property of [λxFx&Gx], no matter what G you pick.

Of course, one could try to reinterpret the identity sign in Principle (A) in terms

of some weaker notion, but what notion that might be remains a mystery. Casta?neda

[1976] and [1990] suggests that Leibniz’s relation of congruence is a weaker congruence

relation on concepts, but Ishiguro [1990] argues against this idea, in Chapter 2. Among

other things, it conflicts with Leibniz’s reading of the symbol ‘=’ (which, in his logical

papers is usually symbolized as ‘∞’) as ‘identity’ or ‘sameness’.

Edward N. Zalta 6

The second reason not to identify concepts and properties is that it

would get the Leibnizian metaphysics of concepts wrong. It is central to

Leibniz’s view of individual concepts that each individual x has a unique

individual concept. So suppose that Adam’s complete individual concept

is the concept (i.e., property) P. Then pick your favorite proposition,

say q, and consider the distinct property [λy Py &(q ∨?q)]. Call this

property ‘Q’. P and Q are exactly the kind of necessarily equivalent

but distinct properties that are the subject of property theory. But if

P is a complete individual concept of Adam, so is Q. If Adam is the

unique thing exemplifying P, he is the unique thing exemplifying Q.

14

So we have distinct complete individual concepts that are both complete

individual concepts of Adam, contrary to the assumption that there is a

unique individual concept of Adam.

Although, in what follows, we shall employ a fine-grained, intensional

property theory as a part of the theory of abstract objects, we shall not

identify (Leibnizian) concepts with such properties, but instead identify

them with abstract objects that encode such properties. But before we

turn to the developmentofthis idea, it is importantto discussthe positive

reason and textual support for distinguishing properties and concepts.

The positive reason for distinguishing properties and concepts is that

it allows us to separate individuals and properties, on the one hand, from

the concepts of individuals and the concepts of properties, on the other.

In what follows, we shall not simply distinguish the individual x from the

concept of that individual c

x

, but also distinguish the property F from

the concept of that property c

F

. This exploits a distinction that seems

to be latent in Leibniz’s views on truth and predication. Here is how.

Leibniz agrees with Aristotle that a substance is something of which

attributes can be predicated but which itself cannot be predicated of

14

In the formal terms of property theory, one defines a property F to be complete

i? for every property G,eitherF implies G or F implies the negation of G.Wherethe

negation of G (‘

ˉ

G’) is defined as [λx?Gx], this definition becomes:

Complete(F)=

df

?G(F ? G ∨ F ?

ˉ

G)

A property F is an individual property i? necessarily, at most one object exemplifies

F:

Individual(F)=

df

a50?x?y(Fx&Fy→ x=y)

(This is a weak notion of an individual property, and although stronger notions are

definable, the problem we are discussing applies already to the weaker notion of indi-

vidual property.) It is now straightforward to show that if P is a complete, individual

concept of Adam, then so is Q.

7 A (Leibnizian) Theory of Concepts

anything.

15

ButLeibniz o?ersatheoryaboutwhatitisforanattributeto

betrulypredicatedofasubstance,namely, whentheconceptoftheformer

is contained in the concept of the latter. This is his concept containment

theory of truth. Leibniz frequently states this theory in what appear to

be linguistic terms: a ‘proposition’ with subject-predicate form is true i?

the concept of the predicate is contained in the concept of the subject.

Thus, in the Discourse on Metaphysics (Article 8), we find:

It is necessary, therefore, to consider what it is to be truly at-

tributed to a subject. ... The subject-term must always include

the predicate-term, in such a way that a man who understood the

notion of the subject perfectly would also judge that the predicate

belongs to it.

16

In this passage, Leibniz is talking about a subject term that refers to an

individual substance. He says something similar about ‘universal propo-

sitions’ in a piece entitled ‘Elements of a Calculus’, written in April 1679:

...every true universal a?rmative categorical proposition simply

showssomeconnection betweenpredicateandsubject(a direct con-

nection, which is what is always meant here). This connection is,

that the predicate is said to be in the subject, or to be contained

in the subject; either absolutely and regarded in itself, or at any

rate, in some instance; i.e., that the subject is said to contain the

predicate in a stated fashion. This is to say that the concept of the

subject, either in itself or with some addition, involves the concept

of the predicate....

17

Noticethatinthelastlineofthispassage,Leibniztalksabout‘theconcept

of the predicate’. This seems to distinguish the predicate as a linguistic

entity from the ordinary concept it expresses. But consider what hap-

pens if we interpret Leibniz in the material mode, by understanding his

use of ‘predicate’ as referring to the property (or attribute) expressed.

15

In the Discourse on Metaphysics, Article 8, Leibniz says:

It is very true that when several predicates are attributed to one and the

same subject and this subject is not attributed to any other, one calls

the subject an individual substance.

This is the translation in PW 18. The source is G.iv 432.

16

This is the translation in PW 18. The source is G.iv 433.

17

This is the translation in LP 18-19. The source is C 51. Parkinson entitles the

piece from which this quote is taken ‘Elements of a Calculus’.

Edward N. Zalta 8

Then ‘the concept of the predicate’ would be some further entity, namely,

something like the concept of the property. But this is the distinction we

are suggesting is latent in Leibniz’s work, namely, between the property

F and the concept c

F

of the property.

18

To exploit this distinction further, let us return to the previous pas-

sage,in whichLeibniztalksnotonlyofthe ‘subject’but alsoofthe‘notion

of the subject’. Let us interpret this talk also in the material mode, so

that ‘subject’ refers to the individual substance being discussed in that

passage from the Discourse and ‘notion of the subject’ refers to the con-

ceptofthe subject. Puttingthese suggestionstogether, wecould interpret

Leibniz as having invoked a distinction between the individual substance

xandthe concept ofthe individual substancec

x

, on theone hand, and the

property or attribute F of the substance and the concept of that property

c

F

, on the other. In somewhat more formal terms, we might say that the

Leibnizian analysis of the ordinary singular statement ‘x is F’amounts

to: c

x

contains c

F

. Whereas most philosophers distinguish the individ-

ual x from its concept c

x

, it is rare to find the property F distinguished

from its concept c

F

. But it seems natural to group ordinary individuals

and properties together and to suppose that there is a kind of concept

appropriate to each.

Further evidence for this view occurs later on in Article 8 of the Dis-

course on Metaphysics, where Leibniz talks about the ‘notion’ of an ac-

cident:

...itisinthenatureofanindividualsubstance,orcompletebeing,

to have a notion so complete that it is su?cient to contain and

render deducible from itself, all the predicates of the subject to

which this notion is attributed. On the other hand, an accident is

a being whose notion does not include all that can be attributed to

18

The reader should note that our particular distinction between properties and the

concepts of properties is drawn here for the specific purpose of developing a Leibnizian

theory of concepts. In the context of Frege’s work, however, the distinction between a

property and the concept of that property would be drawn rather di?erently. Elsewhere

we have suggested that a predicate denotes an ordinary property and has an abstract

property as its sense. Abstract properties encode properties of properties and they are

axiomatized in the type-theoretic development of the theory of abstract objects. As

such, these abstract properties would serve as the ‘concept’ a person x might have of

the property F, for such a concept would encode the properties of the property F that

x takes to be characteristic of F. But this is a matter that does not a?ect the present

development. Interested readers should consult Zalta [1983] (Chapters V and VI), and

[1988] (Chapters 9 – 12).

9 A (Leibnizian) Theory of Concepts

the subject to which this notion is attributed. Take, for example,

the quality of being a king, which belongs to Alexander the Great.

This quality, when abstracted from its subject, is not su?ciently

determinate for an individual and does not contain the other qual-

ities of the same subject, nor everything that the notion of this

prince contains.

19

In this passage, Leibniz seems to distinguish an accident from the notion

ofthataccident(“anaccidentisabeingwhosenotion...”). Butthen,

in the sentences that follow, he reverts back to talk about the quality of

being a king, as if that were the entity that is contained in the concept of

Alexander. In what follows, however, we continue to distinguish between

properties and their concepts. We think that it provides the key to a

thorough reconstruction of Leibniz’s logic and metaphysics of concepts.

§2: Concepts As Abstract Objects

In what follows, we shall try, whenever possible, to match our results to

the propositions asserted or proved in the 1690 fragment.

20

Often, how-

ever, the propositions that Leibniz considers in the 1690 fragment appear

in somewhat di?erent notation in some of his earlier logical sketches.

We proposethat Leibnizian concepts (both the concepts ofindividuals

and the concepts of properties) are abstract objects that encode proper-

ties. Encoding is a mode of predication that has been formalized, axiom-

atized, and applied in Zalta [1983], [1988a], [1993], [1999], and elsewhere.

Those readers unfamiliar with this system should consult Appendix I,

which contains a reasonably thorough sketch of the theory. In what fol-

lows, we shallidentify the principles ofthe theoryaccordingto the scheme

in Appendix I.

Here is a one paragraph, intuitive description of the idea that we shall

formalizein what follows. The Leibnizian concept (of) F will be identified

as the abstract object that encodes all and only the properties necessar-

ily implied by F.Thus,the concept person will encode the property of

being rational, assuming that the property of being rational is necessarily

implied by the property of being a person. Moreover, we shall define:

x contains y i? x encodes every property y encodes. So, assuming that

19

This is the translation in PW 18-19. The source is G.iv 433.

20

Parkinson entitles this paper ‘A Study in the Calculus of Real Addition’. See LP

131-144 (= G.vii 236-247).

Edward N. Zalta 10

the property of being a person necessarily implies the property of being

rational, it will follow that the concept person contains the concept ra-

tional.

21

By way of contrast, the individual concept of Alexander will be

identified as the abstract object that encodes all and only the properties

the ordinary object Alexander exemplifies. It will then follow that the

concept Alexander contains the concept person if given the assumption

that Alexander exemplifies the property of being a person.

Given this sketch, we begin our formal analysis by identifying Leib-

nizian concepts in terms of the abstract objects of our background meta-

physics. Principle 3 (Appendix I) is our comprehension principle for

abstract objects and it asserts that for any condition φ (in which x isn’t

free), there is an abstract object x that encodes all and only the proper-

ties satisfying the condition. Since our system uses the predicate ‘A!x’to

denote the property of being abstract, we can turn our theory of abstract

objects into a theory of concepts by employing the following definition:

Concept(x)=

df

A!x

Instances of the comprehension principle for abstract objects now assert

the existence of Leibnizian concepts, and so we may think of the compre-

hension schema as providing the existence conditions for concepts. Simi-

larly, Principle 4 (Appendix I) o?ers well-defined identity conditions for

abstract objects and these conditions now tell us that concepts x and y

are identical whenever they necessarily encode the same properties. Note

that the identity conditions for concepts are defined in terms what we

take to be their distinctive feature, namely, their encoded properties.

It would serve well to look at a particular example of a concept, say

the concept rational. It should now be clear from our discussion that the

concept rational (‘c

R

’) is to be identified with the abstract object that

encodes all and only the properties necessarily implied by the property of

being rational (‘R’):

c

R

=

df

?x(Concept(x)&?F(xF ≡ R?F))

21

For a quick proof sketch, assume that the property of being a person necessarily

implies the property of being rational. To show that the concept person contains

the concept rational, assume that the concept rational encodes property P (to show

that the concept person encodes P). If the concept rational encodes P,thenby

definition, the property of being rational necessarily implies P. But since being a

person necessarily implies being rational, and being rational necessarily implies P,

being a person necessarily impliesP. So the concept person encodes P, since it encodes

just the properties necessarily implied by the property of being a person.

11 A (Leibnizian) Theory of Concepts

In this definition, ‘the concept rational’ is defined in terms of a definite

description. Principle 3

prime

and its Corollary (Appendix I) establishes that

every such description denotes, so our new term ‘c

R

’ is well-defined. To

take another example, we may identify the concept animal (‘c

A

’) as the

(abstract) object that encodes just the properties implied by the property

of being an animal (‘A’).

In general, we define the concept G (‘c

G

’), for any property G,as

follows:

c

G

=

df

?x(Concept(x)&?F(xF ≡ G?F))

That there is a unique such object for each property G follows from Prin-

ciple 3

prime

.

22

Clearly, then, we are distinguishing the property G from its

concept c

G

.

§3: Concept Identity

Now let x, y,andz be any Leibnizian concepts. Then the following

three facts concerning concept identity are immediate consequences of

Principles 4 and 5 (Appendix I):

23

Theorem 1: x=x

Theorem 2: x=y → y=x

Theorem 3: x=y &y=z → x=z

So we should represent Leibniz’s claim:

24

If A=B,thenB=A

as the claim:

c

A

=c

B

→ c

B

=c

A

This is an instance of Theorem 2. Leibniz proves other corollaries and

theorems with respect to conceptidentity.

25

However,these areallsimple

22

The following is derivable from Principle 3

prime

by generalizing on the variable G:

?G?!x(A!x&?F(xF ≡ G?F)).

23

See LP 131 (= G.vii 236), Propositions 1 and 3, where Leibniz proves the sym-

metricality and transitivity of concept identity, respectively.

24

See LP 131 (= G.vii 236), Proposition 1.

25

See LP 131-132 (= G.vii 236-237), Proposition 2, Corollary to Proposition 3, and

Proposition 4.

Edward N. Zalta 12

consequences of the fact that concept identity is an equivalence condition,

and we shall omit mention and derivation of these simple consequences

here.

26

§4: Concept Addition

If we continue to let x,y,z range over Leibnizian concepts, then we may

define the real sum of the concepts x and y (‘x⊕y’) as follows:

x⊕y =

df

?z(Concept(z)&?F(zF ≡ xF ∨ yF))

In other words, the real sum concept x⊕y is the concept that encodes a

property F i? either x encodes F or y encodes F. Principle 3

prime

(Appendix

I) guarantees that there is a unique concept that meets the definition of

x⊕y, for any x and y.

27

To take a particular example, the real sum of the

concept rational and the concept animal (‘c

R

⊕c

A

’) is now well-defined.

It is a simple consequence of these definitions that the sum of the concept

G and the concept H is identical to the (abstract) object that encodes

just the properties implied by G or implied by H:

Theorem 4: c

G

⊕c

H

= ?x(Concept(x)&?F(xF ≡ G?F ∨ H?F))

The proof of Theorem 4 is given in Appendix II. Note that we must prove

that x and y necessarily encode the same properties to show that they

are identical. Since the logic of encoding (Appendix I) guarantees that

encoded properties are rigidly encoded, it su?ces to show that x and y

encode the same properties to prove them identical. So, in the left-to-

right direction of the proof, we assume that c

G

⊕c

H

encodes an arbitrary

property, say P, and then show that the object described on the right

hand side of the identity sign encodes P, and similarly for the right-to-

left direction.

28

26

Note also that Theorems 1 – 3 remain true when the range of the variables are

extended to include ordinary individuals. This is a consequence of Principles 2 and 5

(Appendix I). Since the domain of objects contains only ordinary and abstract objects,

our theorems of identity can be thought of as completely general, covering any objects

whatsoever.

27

The following is a consequence of Principle 3

prime

, by generalizing on the variables x

and y in the relevant instance: ?x?y?!z(A!z &?F(zF ≡ xF ∨ yF)).

28

Intheproof,weusethenotation‘c

G

⊕c

H

P’ to assert that the sum concept c

G

⊕c

H

encodes property P. This is an encoding formula of the form ‘xG’ in which the variable

‘x’ has been replaced by a complex object term.

13 A (Leibnizian) Theory of Concepts

Now to confirm that ⊕ behaves in the manner that Leibniz prescribed,

note that it follows immediately from these definitions that the operation

⊕ is idempotent and commutative:

Theorem 5: x⊕x = x

Theorem 6: x⊕y = y⊕x

Leibniz takes these two principles as axioms of his calculus, whereas we

derive them as theorems.

29

Leibniz omits associativity from his list of ax-

ioms for ⊕, but as Swoyer ([1995], [1994]) points out, it must be included

for the proofs of certain theorems to go through. In our system, the asso-

ciativity of ⊕ is almost as immediate as idempotency and commutativity:

Theorem 7:(x⊕y)⊕z = x⊕(y⊕z)

The proof of Theorem 7 is also in Appendix II. Recall that to show that

(x⊕y)⊕z and x⊕(y⊕z) are identical, we take an arbitrary property P

and show that (x⊕y)⊕z encodes P if and only if x⊕(y⊕z)encodesP.

30

So ⊕ is associative. In virtue of this fact, we may leave o? the paren-

theses in ‘(x⊕y)⊕z’and‘x⊕(y⊕z)’. Indeed, this suggests that we may

generalizethe realsum operation. We may define the realsum ofconcepts

x

1

,...,x

n

as follows:

x

1

⊕...⊕x

n

=

df

?z(Concept(z)&?F(zF ≡ x

1

F ∨ ... ∨ x

n

F))

That there is a unique such object, for any concepts x

1

,...,x

n

, is guaran-

teed by Principle 3

prime

.

31

Since c

G

1

⊕...⊕c

G

n

is aninstanceof this definition,

‘the sum of the concept G

1

and ... and the concept G

n

’ is well-defined.

We can therefore generalize an earlier theorem:

Theorem 8: c

G

1

⊕...⊕c

G

n

=

?x(Concept(x)&?F(xF ≡ G

1

?F ∨ ... ∨ G

n

?F))

29

See LP 132 (= G.vii 237), Axioms 2 and 1, respectively. Other idempotency

assertions appear in LP 40 (= G.vii 222), LP 56 (= C 366), LP 85 (= C 396), LP 90

(= C 235), LP 93 (= C 421) and LP 124 (= G.vii 230). Swoyer [1995], in footnote 5,

also cites C 260 and C 262. Lenzen [1990] also cites GI 171 for idempotency. Other

commutativity assertions appear in LP 40 (= G.vii 222), LP 90 (= C 235), and LP 93

(= C 421).

30

The proof uses the notation ‘(x⊕ y)⊕zP’ to express the claim that (x⊕ y)⊕z

encodes P. Again, this is an encoding formula of the form ‘xP’ in which the variable

‘x’ has been replaced by a complex term.

31

By the following consequence of the comprehension and identity principles govern-

ing A-objects: ?x

1

...?x

n

?!z(A!z &?F(zF ≡ x

1

F ∨ ... ∨ x

n

F)).

Edward N. Zalta 14

The following two theorems also prove to be of interest:

32

Theorem 9: x=y → x⊕z=y⊕z

Theorem 10: x=y &z=u → x⊕z=y⊕u

Leibniz is careful to note that the converses of these two simple theorems

are not true, for there are counterexamples.

33

We can produce our own

counterexample to the converse of Theorem 9: let x be c

F

⊕c

G

⊕c

H

,let

y be c

F

⊕c

G

, and let z be c

H

.Thenx⊕z=y⊕z, but xnegationslash=y. Similarly,

as a counterexample to the converse of Theorem 10, let x be c

F

⊕c

G

,let

z be c

H

⊕c

I

,lety be c

F

, and let u be c

G

⊕c

H

⊕c

I

. Then, x⊕z=y⊕u,

but neither x = y nor z = u. Finally, Leibniz describes other theorems

concerning concept addition and identity, but these are less interesting

(since easily derived by substitution of identicals), and will be omitted

here.

34

§5: Concept Inclusion and Containment

Leibniz defined the notions of concept inclusion and concept containment

in terms of ‘sameness’ or ‘coincidence’. He says:

Definition 1. Those terms are ‘the same’ or ‘coincident’ of which

either can be substituted for the other wherever we please without

loss of truth—for example, ‘triangle’ and ‘trilateral’. ... ‘A = B’

means that A and B are the same.

35

Definition 3. That A ‘is in’ L, or, that L ‘contains’ A, is the same

as that L is assumed to be coincident with several terms taken

together, among which is A.

36

In other words, Leibniz defines: x is included in y just in case there is a

concept z such that the sum of x and z is identical to y.

However, in our reconstruction of Leibniz, we shall define concept

inclusion and containment in terms of our notion of encoding and then

prove Leibniz’s definition as a theorem (in Section 7). We define the

notions of concept inclusion (‘x precedesequal y’) and concept containment (‘y followsequal x’)

in our system as follows:

32

See LP 133-4 (= G.vii 238), Propositions 9 and 10.

33

See the Notes (Scholia) to Propositions 9 and 10 in LP 133-134 (= G.vii 238).

34

See LP 133 (= G.vii 237-238), Propositions 5, 6, and 8.

35

This is the translation in LP 131. The source is G.vii 236.

36

This is the translation in LP 132. The source is G.vii 237.

15 A (Leibnizian) Theory of Concepts

x precedesequal y =

df

?F(xF → yF)

y followsequal x =

df

x precedesequal y

To show that these definitions are good ones, we establish that the no-

tions defined behave the way Leibniz says they are supposed to behave.

In what follows, we identify the relevant theorems in pairs, a theorem gov-

erning concept inclusion and the counterpart theorem governing concept

containment. We prove the theorem only as it pertains to concept inclu-

sion. First, we note that concept inclusion and containment are reflexive,

anti-symmetric, and transitive:

37

Theorem 11i: x precedesequal x

Theorem 11c: x followsequal x

Theorem 12i: x precedesequal y → (xnegationslash=y → y negationslashprecedesequal x)

Theorem 12c: x followsequal y → (xnegationslash=y → y negationslashfollowsequal x)

Theorem 13i: x precedesequal y &y precedesequal z → x precedesequal z

Theorem 13c: x followsequal y &y followsequal z → x followsequal z

The proof of Theorem 12i is in Appendix II. Leibniz goes on to provethat

when concepts x and y are included or contained in each other, they are

identical:

38

Theorem 14i: x precedesequal y &y precedesequal x → x=y

Theorem 14c: x followsequal y &y followsequal x → x=y

Two other interesting consequences of concept inclusion and identity that

Leibniz doesn’t appear to have considered are that x and y are identical

concepts if (a) concept z is included in x i? z is included in y,orif(b)x

is included in z i? y is included in z:

Theorem 15i: ?z(z precedesequal x ≡ z precedesequal y) → x=y

Theorem 15c: ?z(x followsequal z ≡ y followsequal z) → x=y

Theorem 16i: ?z(x precedesequal z ≡ y precedesequal z) → x=y

Theorem 16c: ?z(z followsequal x ≡ z followsequal y) → x=y

37

See LP 133 (= G.vii 238), Proposition 7, where Leibniz proves the reflexivity

of inclusion. See LP 135 (= G.vii 240), Proposition 15, where Leibniz proves the

transitivity of inclusion. See also LP 33 (= G.vii 218) for the reflexivity of containment.

38

See LP 136 (= G.vii 240), Proposition 17.

Edward N. Zalta 16

I conclude this initial discussion of concept inclusion and containment

with two other consequences of the foregoing:

Theorem 17i: c

G

1

precedesequal c

G

1

⊕c

G

2

precedesequal ...precedesequal c

G

1

⊕...⊕c

G

n

Theorem 18i: c

G

i

1

⊕...⊕c

G

i

j

precedesequal c

G

1

⊕...⊕c

G

n

,

where 1 ≤ i

1

≤ ...≤ i

j

≤ n

Similar claims apply to concept containment.

§6: Concept Inclusion, Containment, and Addition

Leibniz’s theorems that govern just concept inclusion (containment) and

addition are also derivable. The concept x is included in the sum x⊕y,

and so is the concept y:

39

Theorem 19i: x precedesequal x⊕y

Theorem 19c: x⊕y followsequal x

Theorem 20i: y precedesequal x⊕y

Theorem 20c: x⊕y followsequal y

Moreover, if y is included in z,thenx⊕y is included in x⊕z:

40

Theorem 21i: y precedesequal z → x⊕y precedesequal x⊕z

Theorem 21c: y followsequal z → x⊕y followsequal x⊕z

It also follows that if x⊕y is included in z,thenbothx and y are included

in z:

41

Theorem 22i: x⊕y precedesequal z → x precedesequal z &y precedesequal z

Theorem 22c: z followsequal x⊕y → z followsequal x&z followsequal y

And it follows that if both x and y are included in z,thenx⊕y is included

in z:

42

Theorem 23i: x precedesequal z &y precedesequal z → x⊕y precedesequal z

Theorem 23c: z followsequal x&z followsequal y → z followsequal x⊕y

39

This fact appears not to have been mentioned in the fragment of 1690, but see LP

33 (= G.vii 218) for the corresponding fact about containment.

40

See LP 134 (= G.vii 239), Proposition 12. See also LP 41 (= G.vii 223), for the

version governing containment.

41

See LP 136 (= G.vii 240), Corollary to Proposition 15.

42

See LP 137 (= G.vii 241), Proposition 18.

17 A (Leibnizian) Theory of Concepts

Finally, we may prove that if x is included in y and z is included in u,

then x⊕z is included in y⊕u:

43

Theorem 24i: x precedesequal y &z precedesequal u → x⊕z precedesequal y⊕u

Theorem 24c: x followsequal y &z followsequal u → x⊕z followsequal y⊕u

§7: Concept Inclusion, Addition, and Identity

The most important theorems of our Leibnizian calculus are the ones

that relate the notions of concept inclusion (containment), addition, and

identity. RecallthatLeibnizdefines xis included in y i?thereisaconcept

z such that the sum of x and z is identical to y.

44

This definition, and

the corresponding definition of x contains y, fall out as theorems:

45

Theorem 25i: x precedesequal y ≡?z(x⊕z = y)

Theorem 25c: x followsequal y ≡?z(x = y⊕z)

(The proof of Theorem 25i is in Appendix II.) Second, our definition of

precedesequal also validates both forms of the principal theorem governing Leibniz’s

calculus: (i) x is included in y i? the sum of x and y is identical with y

and (c) x contains y i? x is identical with the sum of x and y.Informal

terms:

46

Theorem 26i: x precedesequal y ≡ x⊕y = y

Theorem 26c: x followsequal y ≡ x = x⊕y

(The proof is in Appendix II.) Though Leibniz proves Theorem 26i using

our Theorem 25i as a definition, on the present theory, no appeal to

Theorem 25i needs to be made.

We can now prove Leibniz’s Principle (A) discussed at the outset of

the paper as a simple instantiation of Theorem 26c. Recall that Principle

(A) is:

(A) The concept F contains the concept G i? the concept F is identical

to the sum of the concept F and the concept G.

43

See LP 137 (= G.vii 241), Proposition 20.

44

See LP 132 (= G.vii 237), Definition 3.

45

Compare Swoyer [1995], p. 99, who builds this condition into Leibnizian Relational

Structures.

46

See LP 135 (= G.vii 239), Propositions 13 and 14.

Edward N. Zalta 18

The formal representation of Principle (A) is a simple corollary to Theo-

rem 26c by instantiation:

Corollary: c

F

followsequal c

G

≡ c

F

=c

F

⊕c

G

So Leibniz’s main principles governing the relationship between con-

cept inclusion (precedesequal), concept containment (followsequal), concept identity, and con-

cept addition (⊕) are derivable. Moreover, since Theorem 26i establishes

the equivalence of x⊕y = y and x precedesequal y, and Theorem 25i establishes the

equivalence of x precedesequal y and ?z(x⊕z = y), it follows that:

Theorem 27i: x⊕y = y ≡?z(x⊕z = y)

Theorem 27c: x = x⊕y ≡?z(x = y⊕z)

These results show that our definition of precedesequal and followsequal preserve Leibniz’s logic

of concepts.

We conclude this derivation of the basic logic of Leibnizian concepts

with the following extended observation. Since we have derived Principle

(A), the following instance of Principle (A) is therefore provable: the

concept person contains the concept rational i? the concept person is

identical with the sum of the concept person and the concept rational. In

formal terms:

Instance of Principle (A): c

P

followsequal c

R

≡ c

P

= c

P

⊕c

R

Thereasonthisistrue(intheleft-right direction) is that if everyproperty

encoded in the conceptrationalis encoded in the conceptperson, then the

concept person encodes the same properties as the concept that encodes

all the properties encoded in either the concept person or the concept

rational(since c

R

contributes to c

P

⊕c

R

no properties not alreadyencoded

in c

P

).

47

However, recall that Leibniz believed that the concept person is

itself the sum of the concept rational and the concept animal. In formal

terms, Leibniz’s suggestion amounts to:

c

P

= c

R

⊕c

A

If this is an accurate representation of Leibniz’s view, then if we add this

asan hypothesis to oursystem, we may derivefrom it the followingfact:

48

47

Of course, we have assumed here that the property of being a person implies the

property of being rational; i.e., P ? R.

48

Given the identity of c

P

with c

R

⊕c

A

,thefactthatc

P

followsequal c

R

is a simpleconsequence

of Theorem 19c: x⊕y followsequal x.

19 A (Leibnizian) Theory of Concepts

c

P

followsequal c

R

So from Leibniz’s analysis of the concept person, we may derive the left

side of the above Instance of Principle (A). Such a derivation, together

with the instance itself, yields the consequence that the concept person is

identical to the sum of the concept person and the concept rational.

However,it seems reasonableto arguethat, strictly speaking, c

P

is not

the same conceptas c

R

⊕c

A

. By definition, c

R

⊕c

A

encodes all and only the

properties either implied by the property of being rational or implied by

the property of being an animal. It therefore fails to encode the property

[λx Rx & Ax], since this conjunctive property is neither implied by the

property of being rationalnor implied by the propertyof being an animal.

Butonemightarguethatthe concept person does encode the conjunctive

property. This suggests that it might be preferable to distinguish c

P

from c

R

⊕c

A

by identifying the former with the concept that encodes just

the properties implied by the conjunctive property of being a rational

animal:

49

c

P

= c

[λx Rx & Ax]

This makes good sense. For one might argue that, strictly speaking, it is

the property of being a person that is identical with the property of being

a rational animal (just as the property of being a brother is identical with

the property of being a male sibling, and the property of being a circle is

identical with the property of being a closed, plane figure every point of

which lies equidistant from some given point, etc.). So simply by adding

the hypothesis that P=[λxRx&Ax] to our system, we could derive that

c

P

is identical with the concept c

[λx Rx & Ax]

(since abstract objects that

encode the same properties are identical).

Noticethatevenifwewereto representthe conceptpersonin thisway,

it still follows that the concept person contains the concept rational, since

the conjunctive property of being a rational animal implies the property

of being rational. But, strictly speaking, this is a departure from the

letter of the Leibnizian corpus, in which a complex concept such as c

P

is analyzed in terms of the sum of its simpler concepts. Our theory of

concepts suggests that there is subtle and important di?erence between

c

R

⊕c

A

and c

[λx Rx & Ax]

. This di?erence may not have been observed in

Leibniz’s own theory.

49

Recall that c

[λx Rx & Ax]

is defined as the abstract object that encodes all the

properties necessarily implied by the conjunctive property [λxRx&Ax].

Edward N. Zalta 20

§8: The Modal Metaphysics of Individual Concepts

In this section, we define individual concepts and derive the basic modal

facts that govern them.

50

In the next section, we show how to com-

bine these facts with the containment theory of truth to establish the

fundamental theorem of Leibniz’s metaphysics of concepts. To define

the notion of an individual concept, we employ the notion of a possible

world. However, our use of possible worlds is based on an explicit defini-

tion of worlds that can be given in object theory. This definition and the

derivation of the theorems of world theory can be found in previous work

(Zalta [1993]). Readers unfamiliar with this work may wish to consult

Appendix III, where one will find a review of the theorems of world the-

ory. Of course, the fact that Leibniz appeals to the notion of a possible

world throughout his work is well known.

51

§8.1: Must We Use Counterpart Theory?

Inwhatfollows,weplantoo?eranalternativetothewidespreadviewthat

the best way to reconstruct a Leibnizian modal metaphysics of concepts

is to use some version of D. Lewis’s [1968] counterpart theory. This view

traces back to work of Mondadori ([1973] and [1975]), who notes that the

natural readingof certainpassagesin the Leibnizian corpusare suggestive

of counterpart theory.

52

Here is an example from the Theodicy which

Mondadori cites:

I will now show you some [worlds], wherein shall be found, not

absolutely the same Sextus as you have seen (that is not possible,

he carries with him always that which he shall be) but several

Sextuses resembling him, possessing all that you know already of

thetrueSextus, but not that is already in him imperceptibly, nor in

consequence all that shall yet happen to him. You will find in one

50

The ideas that follow were initiallydeveloped, in somewhat di?erent guise, inZalta

[1983] (pp. 84-90). The reader familiar with the earlier work will find that we have

replaced our [1983] definition of ‘monad’ by a definition of ‘individual concept’ and

replace the [1983] definition of ‘concept containment’ by the definition developed in

the present paper. It should be mentioned that the material in [1983] was indebted to

Parsons [1978] (pp. 146-148) and [1980] (pp. 219-224).

51

Two particularly clearcut examples are in the Theodicy (see T 128 = G.vi 107)

andintheMonadology §53 (see PW 187 = G.vi 615-616).

52

See also Ishiguro [1972], pp. 123-134), for a possible version of this view.

21 A (Leibnizian) Theory of Concepts

world a very happy and noble Sextus, in another a Sextus content

with a mediocre state, ...

53

Mondadori also cites the letter to Landgraf Ernst von Hessen-Rheinfels of

April 12, 1686, where Leibniz talks about the di?erent possible Adams,

all of which di?er from each other:

Forby theindividual notion of Adam I undoubtedlymean a perfect

representation of a particular Adam, with given individual condi-

tions and distinguished thereby from an infinity of other possible

persons very much like him, but yet di?erent from him... There is

one possible Adam whose posterity is such and such, and an infin-

ity of others whose posterity would be di?erent; is it not the case

that these possible Adams (if I may so speak of them) are di?erent

from one another, and that God has chosen only one of them, who

is exactly our Adam?

54

Many commentators have accepted Mondadori’s understanding of these

passages, though some have had certain reservations and qualifications

about understanding Leibniz’s work in terms of counterpart theory.

55

The case for reconstructing Leibniz’s work in terms of counterpart

theory is not perfectly straightforward, however. When Mondadori intro-

duces the suggestionofusing counterparttheorytomodel Leibniz’s views,

he notes that whereas for Lewis the counterpart relation is a relation on

individuals, “in Leibniz’s case, it is best regarded as being a relation be-

tween (complete) concepts” ([1973], 248). This is explicitly built into

the Leibnizian system described in Fitch [1979]. The idea is that in a

Leibnizian modal metaphysics, the possible worlds are not inhabited by

Lewis’s possibilia, but rather by complete individual concepts. Indeed,

these authors just model possible worlds as sets of compossible individual

concepts. So their reconstruction of the modal metaphysics of complete

conceptsinvolves: (1)equivalenceclassesofcompossiblecompleteindivid-

ual concepts to define the possible worlds, and (2) a counterpart relation

which connects each complete individual concept of a given world to var-

ious other complete individual concepts in other worlds. The Leibnizian

53

This is the translation in T 371. The source is G.vi 363.

54

This is the translation in PW 51. The source is G.ii 20.

55

See, for example, Fitch [1979], Wilson [1979], and Vailati [1986]. Lloyd [1978]

also accepts that Leibniz ‘resorts to counterparts’ (p. 379), though she discovers some

Leibnizian features in a Kripkean semantics of rigid designators, which supposes that

the same individual can appear in other possible worlds.

Edward N. Zalta 22

analysis of the contingent claim ‘Alexander might not have been a king’

becomes: there is a possible world w (other than the actual world) and

an individual concept c such that: (i) c is in w, (ii) c is a counterpart

of the concept c

a

of Alexander, and (iii) c doesn’t contain the concept of

being a king.

So when Leibniz talks about the ‘many possible Adams’ and ‘several

Sextuses’ which are all distinct from one another, these commentators

take him to be talking about di?erent individual concepts rather than

di?erent possible individuals. This is, strictly speaking, a departure from

the texts. The commentators suppose that it is legitimate to ‘interpret’

Leibniz’s talk of ‘several Sextuses’ and ‘the many possible Adams’ not as

referringtopossibiliabut rathertoindividual concepts. Similarly, inwhat

follows, we describe a modal metaphysics in which we interpret Leibniz’s

talk of ‘the many possible Adams’ and ‘the several Sextuses’ as referring

to various distinct complete individual concepts. However, we shall not

employ counterpart theory to link the counterparts of the concept Adam

which appear at various other possible worlds. Instead, we shall utilize

the variety of complete individual concepts which become defined when

we consider the properties that Adam exemplifies at each possible world.

In the modal metaphysics we present in what follows, it is part of the

logic that the single ordinary individual Adam exemplifies properties at

every possible world. Of course, the properties that Adam exemplifies at

one world di?er from the properties he exemplifies at other worlds. For

example, althoughAdam is concrete(spatiotemporal)atour worldandat

certain other possible worlds, there are possible worlds where he fails to

be concrete. (Some philosophers have been tempted to call those worlds

where Adam isn’t concreteworldswhere ‘Adam doesn’t exist’, but we will

not follow this usage.)

56

There will be numerous possible worlds where

Adam is concrete and his ‘posterity is di?erent’. (At worlds where Adam

is not concrete, he has no posterity.)

57

56

Given our fixed domain quantified modal logic governed by the Barcan formulas,

every object exists necessarily in the sense that ?xa50?y(y=x). The contingency of an

ordinary object such as Adam is preserved by the fact that he is not concrete at every

world. While other philosophers would say that these latter worlds are ones where

Adam fails to exist, it would be a mistake to say this in the present version of modal

logic, which was defended in Linsky and Zalta [1994] and Williamson [1998].

57

It might help to point out here that, in the case of ordinary objects such as Adam,

the claim ‘Adam is F essentially’ is to be understood as the claim that Adam is F

in every world in which he is concrete. For example, the claim ‘Adam is essentially a

23 A (Leibnizian) Theory of Concepts

Now although the same ordinary individual Adam exemplifies proper-

tiesatotherworlds,aLeibnizianmetaphysicsof‘world-bound’individuals

emerges once we consider, for each world, the ‘individual concept’ (i.e.,

abstract object) that encodes exactly the properties that Adam exempli-

fies at that world. At each world, Adam realizes a di?erent individual

concept, since individual concepts will di?er whenever they encode dis-

tinct properties. The individual concept that encodes all and only the

properties Adam exemplifies at one world is distinct from the individual

concept that encodes all and only the properties Adam exemplifies at a

di?erent world. Of course, we will define only one of these individual con-

cepts to be ‘the concept of Adam’, namely, the concept that encodes just

what Adam exemplifies at the actualworld. Strictly speaking, the various

individual ‘concepts of Adam’ are ‘concepts-of-Adam-at-w’(forsomew)

and so the concept of Adam will be identified with the concept of Adam

at the actual world. When Leibniz talks about ‘possible Adams’, we may

take him to be talking about di?erent individual concepts, namely dif-

ferent ‘Adam-at-w’ concepts. We’ll explain this in more detail once the

definitions and theorems have been presented.

What is more interesting is the fact that these individual concepts

have certain other Leibnizian features. By way of contrast to other com-

mentators on Leibniz, we shall not just stipulate that individual concepts

are partitioned into compossibility (equivalence) classes but rather define

compossibility and prove that it is a condition which partitions the in-

dividual concepts. Moreover, we shall not define possible worlds as sets

of compossible individual concepts, but rather prove that there is a one-

to-one correlation between each group of compossible concepts and the

possible worlds. We shall also prove that any arbitrarily chosen member

of a groupof compossible individual concepts is a perfect mirrorof its cor-

responding world. And, finally, we shall prove that individual concepts

are complete concepts that are realized by at most one object if realized

at all.

So, although we do not use counterpart theory in our reconstruction,

and we suppose, in departure from the strict Leibnizian text, that there is

exactly one Adam who exemplifies properties at di?erent worlds, we nev-

ertheless recover a Leibnizian modal metaphysics in which the complete

individual concepts are the ‘world-bound’ individuals, not the ordinary

man’ is to be understood as: Adam is a man in every world in which he is concrete.

See Linsky and Zalta [1994], §4, for further discussion.

Edward N. Zalta 24

objects themselves. These individual concepts will provide an interpre-

tation for much of what Leibniz says about necessity, contingency, com-

pleteness, mirroring, etc., as we shall see.

§8.2: The Definitions and Theorems

We begin by defining what it is for an ordinary object to realize a concept

at a world. We shall say that an object y realizes a concept x at world

w just in case y is an ordinary object which exemplifies at w all and

only the properties x encodes.

58

This definition can be formalized with

the help of notions of object theory defined elsewhere. First, the notion

of an ordinary object (‘O!x’) is defined as any object that might have

been concrete (Appendix I). Second, the claim ‘proposition p is true at

world w’(‘|=

w

p’) is defined and properly systematized (Appendices I

and III) by the claim that w encodes the property being such that p (i.e.,

w[λy p]). With these notions, we can formalize our definition of ‘realizes

at’ as follows:

59

RealizesAt(y,x,w)=

df

O!y & ?F(|=

w

Fy ≡ xF)

Or, letting ‘u’ be a restricted variable ranging only over ordinary objects,

we may define this notion more simply as:

RealizesAt(u,x,w)=

df

?F(|=

w

Fu ≡ xF)

Important notions ofLeibniz’s modal metaphysicscan be defined in terms

of realization:

Appears(x,w)=

df

?uRealizesAt(u,x,w)

IndividualConcept(x)=

df

?wAppears(x,w)

The first definition tells us that a concept appears at a world just in case

some ordinary objectrealizes that concept at that world. The second tells

us that a concept is an individual concept just in case it appears at some

world.

58

This represents another minor change of terminology from our previous work. In

Zalta [1983], p. 85, we said that the ordinary object y ‘is the correlate of’ the abstract

object x when this condition obtains.

59

The symbol ‘|=

w

’ should be read with the smallest possible scope. For example,

|=

w

p ≡ q is to be parsed as (|=

w

p) ≡ q. To claim that the biconditional p ≡ q is true

at w,wewrite:|=

w

(p ≡ q).

25 A (Leibnizian) Theory of Concepts

To get some practice with these notions, let us now introduce concepts

ofparticularindividualsandshowthattheyareindividualconcepts. (This

is not a triviality!) In general, whenever u is an ordinary individual (i.e.,

whenever O!u), we define the concept of u (‘c

u

’) as the concept that

encodes just the properties u exemplifies:

c

u

=

df

?x(Concept(x)&?F(xF ≡ Fu))

It follows from the Corollary to Principle 3

prime

(Appendix I) that for any

ordinary object u, the concept of u exists; i.e., ?u?z(z=c

u

).

We can now prove that the concept of u is an individual concept. The

proof depends on a simple lemma, namely, that c

u

encodes a property G

i? u exemplifies G:

Lemma: c

u

G ≡ Gu

This Lemma is a consequence of the logic of definite descriptions.

60

(It

is important to remember here that our definite descriptions are rigid

designators. The logic of these descriptions is governedby a logicalaxiom

thatisacontingentlogicaltruth. SeetheaxiomDescriptions inAppendix

I. In the next section, we’ll see how this fact about Descriptions becomes

important.)

Now with the help of this Lemma, it is straightforward to show that

the concept of u is an individual concept:

Theorem 28: IndividualConcept(c

u

)

(The proof is in Appendix II.) A fortiori, for every ordinary object u,

there is exactly one individual concept that serves as c

u

.

To see how these theoretical definitions work in a concrete case, let

us assume the contingent fact that Alexander (‘a’) is a concrete object

(E!a). It then follows by the T schema of modal logic and the defini-

tion of ordinary object (Appendix I) that Alexander is an ordinary object

(O!a) and it follows from the foregoing that the concept of Alexander (c

a

)

exists and is an individual concept. So, thus far, we have used the contin-

gent premise that Alexander is a concrete object to show that individual

concepts exist.

However, it is possible to prove that individual concepts exist without

appealing to any contingent premises. Notice that our theory doesn’t as-

sert the contingent claim that Alexander is a concrete object as an axiom,

60

The Lemma follows immediately from A-Descriptions (Appendix I), which in turn

is provable from the Descriptions axiom.

Edward N. Zalta 26

nor does it assert either the existence of any other particular concrete ob-

ject or the existence of concrete objects in general. However, we can and

ought to extend our theory with the aprioriaxiomthatitispossible that

there are concrete objects; i.e., a51?xE!x.Suchanaxiom,inthecontext

of an S5 modal logic, is provably a necessary truth. Moreover, by the

Barcan formula, it follows that ?xa51E!x. In other words, it follows from a

priori assumptions alone that there are ordinary objects (?xO!x). It then

immediately follows from Theorem 28 and the Corollary to Principle 3

prime

:

Theorem 29: ?x[IndividualConcept(x)]

Thus we have an aprioriproof that individual concepts exist.

Now one of Leibniz’s most interesting metaphysical ideas is that of

mirroring. We can show that individual concepts mirror any world where

they appear. We define the notion of mirroring as follows:

Mirrors(x,w)=

df

?p(x[λy p] ≡ w[λy p])

In other words, a concept x will mirror a world w just in case x encodes

all and only those propositions true in w. It now follows that:

Theorem 30: Appears(x,w) → Mirrors(x,w)

(The proof of Theorem 30 is in Appendix II.) From this theorem it now

follows that every individual concept appears at a unique world:

Theorem 31: IndividualConcept(x) →?!wAppears(x,w)

(This too is proved in Appendix II.) This last theorem demonstrates that

individual concepts are, in an important sense, ‘world-bound’. Since the

properties these world-bound individual concepts encode are necessarily

encoded, we have recovered a kind of Leibnizian ‘super-essentialism’ at

the level of individual concepts.

61

(In the next section, we will see just

how this super-essentialism is consistent with the existence of contingent

truths.)

Since it is now established that every individual concept appears at a

unique world, we may legitimately talk about the world w

c

where indi-

vidual concept c appears. Letting ‘c’ be a restricted variable ranging over

individual concepts, we may define:

w

c

=

df

?wAppears(c,w)

61

See Mondadori [1973].

27 A (Leibnizian) Theory of Concepts

Theprevioustwotheorems,thatanyconceptappearingataworldmirrors

that world and that every individual concept appears at a unique world,

allow us to assert that every individual concept c mirrors its world and

indeed, even contains its world:

Corollary: Mirrors(c,w

c

)&c followsequal w

c

The proof of the left conjunct is immediate from the preceding theorems

and the proof of the right conjunct is immediate from the definition of

containment and the fact that propositional properties of the form [λyp]

are the only kind of properties that worlds encode.

Another important notion of Leibnizian metaphysics is compossibility.

Two individual concepts are compossible just in case they appear at the

same world. Compossibility should partition the individual concepts into

equivalence classes that correspond with the possible worlds. We define

compossibility formally as follows:

Compossible(c

1

,c

2

)=

df

?w(Appears(c

1

,w)&Appears(c

2

,w))

Given the previous theorem, it is an immediate consequence of this defi-

nition that:

Theorem 32: Compossible(c

1

,c

2

) ≡ w

c

1

= w

c

2

With the help of this theorem, we get the following immediate results:

Theorem 33: Compossible(c

1

,c

1

)

Theorem 34: Compossible(c

1

,c

2

) → Compossible(c

2

,c

1

)

Theorem 35: Compossible(c

1

,c

2

)&Compossible(c

2

,c

3

) →

Compossible(c

1

,c

3

)

Since compossibility is a condition that is reflexive, symmetrical, and

transitive, we know the individual concepts are partitioned. For each

group of compossible individual concepts, there is a unique world where

all of the individual concepts in that group appear.

62

Each member of

62

Compare Mates [1968], Mondadori [1973], and Fitch [1979], who stipulate most

of these claims. Note also the di?erence between our definition of compossibility and

that in Lenzen [1990], p. 186. Lenzen defines compossibility as follows: concepts x and

y are compossible just in case it is possible that both x and y contain the concept of

existence. On this definition, the transitivity of compossibility does not follow. But

if individual concepts mirror their worlds in a genuine sense, transitivity has to be a

property of compossibility.

Edward N. Zalta 28

the group not only mirrors the world corresponding to that group but

also contains that world.

We look next at what kind of concepts individual concepts are. Indi-

vidual concepts are suppose to be ‘complete’ and ‘singular’. Intuitively, a

complete concept is any concept x such that, for every property F, either

x encodes F or x encodes

ˉ

F:

63

Complete(x)=

df

?F(xF ∨ x

ˉ

F)

The theory now predicts that every individual concept is complete:

64

Theorem 36: IndividualConcept(x) → Complete(x)

(The proof of this theorem is in Appendix II.)

65

Finally, aconceptxissingularjustincase,foranyworldw, ifordinary

objects u and v realize x at world w,thenu=v. In formal terms:

63

Recall that

ˉ

F is defined as: [λx?Fx].

64

However, see Mondadori [1973] (p. 239) and [1975] (p. 257), who argues that

Leibniz doesn’t regard the complete concept of x as involving all of the properties

that x exemplifies. Mondadori cites CA 48 (= G.ii 44). However, the reader should

consider the passages cited at the end of this section, where Leibniz seems to say that

individual concepts ‘involve’ even extrinsic, relational properties.

Nevertheless, if Mondadori is right, there could be a way to capture this view, if

there is some clearcut notion of ‘primitive’, ‘simple’, or ‘basic’ property. Both Leibniz

(CA 48 = G.ii 44) and Mates [1968] appeal to some such notion. If such a notion can

be well defined, we could stipulate:

Thecoreconceptofu =

df

?x(Concept(x)&?F(xF ≡ Fu& Basic(F)))

Asanexample, the coreconcept of Adamwouldencode alland onlythe basicproperties

exemplified by Adam. However, now we have two candidates for the concept of Adam:

c

a

(as originally defined) and the core concept of Adam. These are distinct concepts.

We will leave it to others to determine which one is best matched to Leibniz’s entire

corpus. Of course, before one can utilize ‘the core concept of Adam’, one will have

to somehow demonstrate that all of Adam’s other properties can be derived from this

core concept.

65

It might have been more in the spirit of Leibniz’s framework to have defined a

complete concept to be one which, for any property F, contains either the concept F

or the concept

ˉ

F. In other words, we might have defined the notion of completeness

as follows:

Complete(x)=

df

?F(x followsequal c

F

∨ x followsequal cˉ

F

)

This, it turns out, is equivalent to the above definition, at least as far as individual

concepts go. For by Theorem 39b (which is proved in the next section), it follows

by simple disjunctive syllogism that if individual concepts are complete in the sense

proved in Theorem 36, then they are complete in the sense defined in this footnote,

and vice versa.

29 A (Leibnizian) Theory of Concepts

Singular(x)=

df

?u,v,w(RealizesAt(u,x,w)&RealizesAt(v,x,w) → u=v)

It now follows that every individual concept is singular:

Theorem 37: IndividualConcept(x) → Singular(x)

(The proof of this theorem is in Appendix II.) So every individual concept

is complete and singular.

§8.3: Some Observations About the Metaphysics

The results just outlined o?er a precise picture which seems to underlie

such claims as the following (Discourse on Metaphysics):

...it is in the nature of an individual substance, or complete

being, to have a notion so complete that it is su?cient to contain,

and render deducible from itself, all the predicates of the subject

to which this notion is attributed. (Article 8)

66

...every substance is like an entire world and like a mirror of

God, or of the whole universe, which each one expresses in its own

way. ... Thus the universe is in a way multiplied as many times

as there are substances. (Article 9)

67

This very last claim is made true by the fact that every substance (i.e.,

non-concept) has an individual concept which mirrors, and indeed, con-

tains, the actual world (by the Corollary to Theorem 31). A world w is

‘multiplied’ since each individual concept in the corresponding group of

compossible concepts has w as a part.

There is no doubt that Leibniz scholars would question these results

by pointing out that they are grounded in a logic which accepts that

ordinary individuals have properties at every possible world. But there

is also no doubt that the fact that these results are provable constitute

a reason to take them seriously that Leibniz himself would have found

compelling. (The results proved in the next section may be even more

compelling.) The subsystem of individual concepts and worlds (i.e., the

group of theorems that pertain to individual concepts and worlds) still

preserves Leibniz’s conception of the work that God had to do to ‘create’

66

This is the translation in PW 18. The source is G.iv 433.

67

This is the translation in PW 19. The source is G.iv 434.

Edward N. Zalta 30

the actual world. If we focus narrowly on that subsystem, it becomes ev-

ident that, in order to evaluate all the possible worlds, God simply has to

inspect an arbitrarily chosen individual concept from each group of com-

possible individual concepts. His inspection will reveal to Him the entire

corresponding world, since every individual concept of the group mirrors

that world. Note that our metaphysics doesn’t tell us which world is the

actual world other than by describing the actual world as the one that

encodes all and only the true propositions. So God could ‘create’ (i.e., ac-

tualize)worldw (after deciding it was the best) bymaking it the casethat

every proposition p encoded in w is true. Alternatively, when inspecting

the possible worldsby examining a representativeindividual concept from

each group of compossible concepts, he could ‘create’ (i.e., actualize) the

individual concept c which was representative of the possible world which

turned out to be the best. To do so, He would make it the case that

there is in fact an ordinary object which exemplifies all the properties c

encodes. In the process of doing this, he would have to create an entire

world, since c encodes all the propositionsencoded in the worldit mirrors.

We nowadopt the strategyofthe otherLeibnizian commentators, who

interpret Leibniz’s talk of ‘the several Sextuses’ and ‘the many possible

Adams’ as a reference to the variety of individual concepts ‘connected

with’ (see below) Sextus and Adam. Even though our subsystem of indi-

vidual concepts was carved out without appeal to counterpart theory, we

can, nevertheless, still speak with the counterpart theorists! To do so, we

first define ‘the concept of Adam-at-w’:

c

w

a

=

df

?x(Concept(x)&?F(xF ≡|=

w

Fa))

So, if w

α

is the actual world, the concept of Adam-at-w

α

(i.e., c

w

α

a

)is

provably identical to the concept of Adam (c

a

). Moreover, we may now

say precisely when two individual concepts c and c

prime

are counterparts:

Counterparts(c,c

prime

)=

df

?u?w

1

?w

2

(c =c

w

1

u

&c

prime

=c

w

2

u

)

In other worlds, individual concepts c and c

prime

are counterparts whenever

there is an ordinary object u and there are worlds w

1

and w

2

such that

c is the concept of u-at-w

1

and c

prime

is the concept of u-at-w

2

.Thus,when

Leibniz talks of the ‘many possible Adams’, we take him to be referring

to the many di?erent concepts of Adam-at-w (for the various w). These

are all counterparts of one another. There is nothing mysterious at all

about this notion of counterpart. Of course, our notion of counterpart is

31 A (Leibnizian) Theory of Concepts

an equivalence condition, and so di?ers from that of Lewis. But, to my

knowledge, there is nothing in Leibniz’s work that would rule this out.

(We will employ our notion of counterpart further in the next section,

when we consider the question of how our Leibnizian modal metaphysics

of individual concepts accounts for contingency.)

We concludethis sectionby notingthat ourreconstructionofLeibniz’s

modal metaphysics is based on a general theory of relations and the rela-

tional properties definable in terms of such a theory. (Readers unfamiliar

with the theory of abstract objects will find a very brief discussion of its

theory of relations in Appendix I.) The Leibnizian justification for relying

on a general theory of relations comes from the July 14, 1686 letter to

Arnauld:

I say that the concept of an individual substance involves all its

changes and all its relations, even those which are commonly called

extrinsic, that is to say which pertain to it only by virtue of the

general interconnection of things, and in so far as it expresses the

whole universe in its own way...

68

Thequestionhereis, whatdidLeibnizmeanby‘involves’ifnot‘contains’?

Even if Leibniz takes relations to be abstract, ideal entities, it would seem

that they play role in the background of his metaphysics of modality.

69

In any case, the appeal to a general theory of relations facilitates our

reconstruction of a Leibnizian metaphysics.

§9: Containment and the Fundamental Theorem

In this section, we continue to use the variable u to range over ordinary

objects. In what follows, we analyze the Leibnizian notionof concept con-

68

This is the translation in Ishiguro [1972], p. 99. One may wish to check the source

G.ii 56, for it might be thought that Leibniz didn’t intend to mention ‘relations’ in

this passage, given that Morris and Parkinson translate the classical French word

‘denominations’ as the English ‘denominations’ rather than as ‘relations’. In PW 62,

they translate the passage from G.ii 56 as follows:

...I say that the individual substance includes all its events and all its

denominations, even those which are commonly called extrinsic (that is,

they belong to it only by virtue of the general interconnexion of things and

because it expresses the whole universe in its own way), ...

However, Morris and Parkinson, in note ‘g’ placed at the word ‘extrinsic’ in the above

passage, indicate that relations are indeed intended. See PW 248, note g.

69

For further confirmation, see D’Agostino [1976].

Edward N. Zalta 32

tainmentusingthe notionofcontainmentthatwedefined atthebeginning

of Section 5. That is:

The concept x contains the concept z

is analyzed as:

x followsequal z

Now Leibniz analyses:

Alexander is a king

in the following terms:

The concept of Alexander contains the concept king

On our theory of concepts, this Leibnizian analysis becomes:

c

a

followsequal c

K

So truth is a matter of the subject concept containing the predicate con-

cept.

The more modern analysis of ‘Alexander is a king’, of course, is: Ale-

xander exemplifies (the property of) being king; i.e., Ka. But consider

the following fact:

Theorem 38: Fu≡ c

u

followsequal c

F

(The proof is in Appendix II.) This establishes that our Leibnizian anal-

ysis of “The concept of Alexander contains the concept king” as c

a

followsequal c

K

is equivalent to the more modern analysis of ‘Alexander is a king’ as Ka.

These results present us with a defense of a distinction that Leibniz

draws in Article 13 of the Discourse on Metaphysics and in a letter to

Hessen-Rheinfels. Recall that Arnauld charged that Leibniz’s analyses

turn contingent truths into necessary truths. Arnauld worried that the

contingent statement that ‘Alexander is a king’ is analyzed in terms of

the necessary truth ‘the concept of Alexander contains the concept king’.

Leibniz defends himself against Arnauld’s charges by distinguishing abso-

lute necessity from ‘hypothetical necessity’.

70

Although Leibniz explains

70

In Article 13, Leibniz says:

To give a satisfactory answer to it, I assert that connexion or sequence is

of two kinds. The one is absolutely necessary, whose contrary implies a

33 A (Leibnizian) Theory of Concepts

this distinction in terms of the di?erence between what God is free to do

absolutely and what God is contrained to do given previous choices, there

is a distinctive way of explaining ‘hypothetical necessities’ in the present

framework.

The explanation begins by pointing to the fact that any proof of

c

a

followsequal c

K

would have to rest on the contingent premise that Alexander

is a king (i.e., it rests on Ka). Just consider the fact that if we are to

cite Theorem 38 to prove that c

a

followsequal c

K

, we must assume the contingent

premise Ka. Moreover, it is important to observe that (it is a theorem

that) if x followsequal y,thena50(x followsequal y), for the logic of encoding (Appendix I)

guarantees that encoded properties are necessarily encoded. So, as an

instance of this result, we know: if c

a

followsequal c

K

,thena50(c

a

followsequal c

K

). But since

a proof of the antecedent would rely on a contingent fact, it follows that

the proof of the consequence would as well. So we have a case where the

proof of a necessary truth depends on a contingent assumption. It may

be that this is the way to understand ‘hypothetical necessities’.

71

(Some reader might find it important to observe here that Theorem

38 is a logical truth that is not necessary. To see why, reconsider the

Lemma preceding Theorem 28 (i.e., c

u

F ≡ Fu) more carefully. Note that

the Lemma itself is a an example of a logical truth that is not necessary.

It is a logical truth because it is derivable from our logical axioms and

rules alone. But it is not metaphysically necessary. To see why, suppose

that Kais true (at the actual world) and consider a world, say w

1

,where

Alexander is not a king. At w

1

, Ka is false. But note that at such a

world, c

a

K is true. That’s because c

a

, the object that (at the actual

world) encodes all and only the properties that Alexander actually ex-

emplifies, encodes its properties rigidly and so encodes K even at w

1

.

72

contradiction; this kind of deduction holds in the case of eternal truths,

such as those of geometry. The other is only necessary by hypothesis (ex

hypothesi), and so to speak by accident; it is contingent in itself, since

its contrary does not imply a contradiction.

This is the translation in PW 23-24. The source is G.iv 437. In a letter to Hessen-

Rheinfels, Leibniz says:

These last words contain the proof of the consequence; but it is very clear

that they confuse necessitatem ex hypothesi with absolute necessity.

This is the translation in PW 49. The source is G.ii 18.

71

This analysis corrects an error in Zalta [1983], where it was claimed (p. 90) without

qualification that c

a

K (which was formalized in [1983] as ‘ˉaK’) is a necessary truth.

72

The description that serves to define c

a

is a rigid designator and so still denotes

at w

1

the object x that satisfies ‘?F(xF ≡ Fa)’ at the actual world.

Edward N. Zalta 34

So w

1

is a world where the left condition (c

u

F) of the Lemma is true

but the right condition (Fu) is false. Hence the Lemma is not a neces-

sary truth. Similarly, Theorem 38 is a logical truth that is not necessary.

Even though it is logically required (provable) that Ka is equivalent to

c

a

followsequal c

K

, this equivalence is not necessary in the metaphysical sense. For

a fuller discussion of such logical truths that are not necessary, see Zalta

[1988b].)

We turn next to the proof of the fundamental theorem of Leibniz’s

metaphysics of individual concepts. This theorem can be stated in one

of two equivalent ways: (1) if F is a property that ordinary object u

exemplifies but might not have exemplified, then (i) the concept of u

contains the concept F and (ii) there is a (complete individual concept

which is a) counterpart of the concept of u which doesn’t contain the

concept F and which appears at some other possible world, or (2) if F

is a property that ordinary object u doesn’t exemplify but might have

exemplified, then (a) the concept of u doesn’t contain the concept F and

(b) there is a (complete individual concept which is a) counterpart of the

concept of u which both contains the concept F and which appears at

some other possible world. In order to prove this theorem, we first draw

attention to two consequences of our work so far. The first is that it is

an immediate consequence of Theorem 38 and the Lemma to Theorem 28

that individual concept c

u

encodes a property F i? c

u

contains c

F

:

Theorem 39a: c

u

F ≡ c

u

followsequal c

F

However,asomewhatmoregeneralresultisnowprovable,namely, forany

individual concept x, x encodes a property F i? x contains the concept

c

F

:

Theorem 39b: IndividualConcept(x) →?F(xF ≡ x followsequal c

F

)

(The proof is in Appendix II.)

With these two facts in hand, we proceed to establish the fundamental

theorem of Leibniz’s metaphysics of concepts. Consider a simple claim

that expresses a contingent fact, for example, that Alexander is a king

but might not have been:

Ka&a51?Ka

Now from the first conjunct, we know (by Theorem 38) that the concept

of Alexander (c

a

) contains the concept king:

35 A (Leibnizian) Theory of Concepts

c

a

followsequal c

K

Moreover, it is a theorem of world theory (Appendix III) that the second

conjunct a51?Kais equivalent to:

?w(|=

w

?Ka)

Since there is some world where Alexander is not a king, pick one, say

w

1

. We can easily establish (from work in the previous section) that

the concept of Alexander-at-w

1

(c

w

1

a

) exists and is an individual concept.

We also know that c

w

1

a

is a counterpart of c

a

(in the sense defined in

the previous section). We furthermore know that c

w

1

a

appears at w

1

,

by the definition of appearance and the fact (provable from definitions)

that Alexander realizes c

w

1

a

at w

1

. Moreover, we know (as a provable

consequence of our world theory) that w

1

is not the actual world w

α

,

since the proposition that Alexander is a king is true at w

α

and not at

w

1

. We also know that since (by definition) c

w

1

a

encodes exactly the

properties that Alexander exemplifies at w

1

, it follows that c

w

1

a

fails to

encode the property of being a king. From this, it follows (by Theorem

39bandthefactthatc

w

1

a

isanindividualconcept)that c

w

1

a

failstocontain

the concept king (c

K

). So, assembling what we know and generalizing on

w

1

, the following is a consequence of the fact that Alexander is a king but

might not have been:

The (complete, individual) concept of Alexander contains the con-

cept king, but there is a (complete, individual concept which is a)

counterpart of the concept of Alexander which doesn’t contain the

concept king and which appears at some other possible world.

If we generalize on Alexander and the property of being a king, then we

have, in essence, just established the following fundamental theorem of

Leibniz’s metaphysics of individual concepts:

Fundamental Theorem of Leibniz’s Metaphysics of Concepts:

Theorem 40a:(Fu&a51?Fu) → [c

u

followsequal c

F

&

?x(Counterparts(x,c

u

)&x negationslashfollowsequal c

F

& ?w(wnegationslash=w

α

&Appears(x,w)))]

Theorem 40b:(?Fu&a51Fu) → [c

u

negationslashfollowsequal c

F

&

?x(Counterparts(x,c

u

)&x followsequal c

F

&?w(wnegationslash=w

α

&Appears(x,w)))]

Thus, from a simple, pre-theoretic modal statement which asserts that a

certain fact about an ordinaryobject is contingent, we canprove the facts

Edward N. Zalta 36

that aresupposed to obtain in a Leibnizian metaphysics ofindividual con-

cepts. Indeed, these are the very facts that are stipulated (in set-theoretic

terms) as truth conditions for such statements in the possible-world se-

mantics developed by other commentators on Leibniz.

73

It should now be

clear why we have called this the fundamental theorem of Leibniz’s meta-

physics of individual concepts—the particular systems of set-theoretic,

possible world semantics constructed by other commentators to model

Leibniz’s metaphysics were designed expressly to assign the very truth

conditions for statements of contingency that are derived in Theorem 40.

Takentogether,Theorem40andthe other theoremsprovedinthis section

and in the previous section establish that our metaphysics of individual

concepts has the distinguishing features of Leibniz’s metaphysics of indi-

vidual concepts. These results also show that our system of (syntactically

second-order) quantified modal logic with encoding and comprehension

over abstract objects o?ers a genuine alternative to set theory and pos-

sible world semantics, at least insofar as we are interested in formulating

an overarchingsystem that unifies Leibniz’s logic and metaphysics of con-

cepts.

Now although any discussion following the proof of the fundamental

theoremisbound tobeanti-climactic, it maystillbe ofinteresttoindicate

briefly that Leibniz may have anticipated not only Montague’s [1974]

subject-predicate analysis of basic sentences of natural language, but also

the idea of a generalized quantifier. Recall that Montague was able to

giveauniformsubject-predicateanalysisofafundamental classofEnglish

sentences by treating such noun phrases as ‘Adam’ and ‘every person’ as

sets of properties. He supposed that the proper name ‘Adam’ denotes the

set of all and only the properties F that the individual Adam exemplifies

and supposed that the noun phrase ‘every person’ denoted the set of all

and only the properties F that every person exemplifies. Then, English

sentences such as ‘Adam is happy’ and ‘Every person is happy’ could be

givena subject-predicate analysis: such sentences are true i? the property

denoted by the predicate ‘is happy’ is a member of the set of properties

denoted by the subject term.

73

See, in particular, Mondadori [1973] (p. 250) and Fitch [1979] (pp. 300-303). The-

orem 40 also seems to capture the main idea developed in Mates [1968], though for

Mates, you have to drop the clause concerning ‘counterparts’ (since he doesn’t use

them). There are other minor di?erences as well.

37 A (Leibnizian) Theory of Concepts

With this in mind, it seems worthwhile to point out that quantified

sentences such as ‘Every person is happy’ and ‘Some person is happy’ can

be given an analysis in terms of Leibniz’s containment theory of truth.

The similarity to the Montagovian analysis should then be obvious. Let

us define ‘the concept every G’(‘c

?G

’) and ‘the concept some G’(‘c

?G

’)

as follows:

c

?G

=

df

?z(Concept(z)&?F(zF ≡?x(Gx → Fx)))

c

?G

=

df

?z(Concept(z)&?F(zF ≡?x(Gx&Fx)))

Then, if we let G be thepropertyofbeingaperson(‘P’), we may analyze:

Every person is happy

Some person is happy

respectively, as:

c

?P

followsequal c

H

c

?P

followsequal c

H

Notice that if there are at least two distinct ordinary things exemplifying

G, then neither c

?G

nor c

?G

are individual concepts.

Conclusion

The results described in the previous section are somewhat di?erent from

the ones we developed in [1983]. In that earlier work, certain analyses

were somewhat simpler. The Leibnizian concept G was identified with

the property G and the Leibnizian claim that ‘the individual concept

of u contains the concept G’ was analyzed more simply as ‘c

u

G’. But

under that scheme, concept containment became a connection between

two di?erent kinds of things (namely, A-objects and properties). Clearly,

though, Leibnizian concept containment and concept inclusion are sup-

pose to connect things of a single kind, namely, concepts. That idea is

preserved in the way we have analyzed containment and inclusion in the

present work—the notions of containment and inclusion that we use both

in the logic and metaphysics of concepts are defined as conditions that

apply to concepts (i.e., A-objects). Of course there is a simple thesis that

connects the present analysis with the earlier analysis, namely, Theorem

Edward N. Zalta 38

39a. Thisthesisshowsthattheformeranalysisisequivalenttothepresent

one—our Leibnizian analysis of ‘Alexander is a king’ in [1983] (p. 90) as

c

a

K is equivalent to our present analysis of this claim as: c

a

followsequal c

K

.

The consistency of the theory of abstract objects immediately estab-

lishes the consistency of our Leibnizian logic of concepts, metaphysics

of individual concepts, and containment theory of truth. Moreover, the

truth of the theory of (Leibnizian) concepts, and its consistency with our

naturalized world view, depends on the truth of the principles of the the-

ory of abstract objects and their consistency with naturalism. This latter

theory is at least consistent with naturalism, or so we have recently ar-

gued.

74

We shall not argue here that the theory itself is true, though the

fact that it validates a Leibnizian metaphysics should constitute further

evidence in favor of the theory.

Appendix I: A Sketch of the Theory

The metaphysical theory of abstract objects is a stated in terms of a lan-

guage containing two basic forms of predication. Encoding predication

contrasts with the traditional exemplification mode of predication. That

is, in addition to the traditional form of predication ‘x exemplifies F’

(‘Fx’), we also have the new form of predication ‘x encodes F’(‘xF’).

Whereasordinaryobjectsonlyexemplifyproperties,abstractobjectsboth

exemplify and encode properties. And whereas ordinary objects are iden-

tified whenever they necessarily exemplify the same properties, abstract

objects are identified whenever they necessarily encode the same proper-

ties.

To state the theory of abstract objects more precisely, we utilize two

primitive predicates: a 1-place predicate E! (to denote the property of

being concrete), and a 2-place predicate =

E

(to denote the relation of

identity on ordinary objects). We then say that x is ordinary (‘O!x’) i?

it is possible that x is concrete. We also say that x is abstract (‘A!x’) i?

x is not the kind of object that could be concrete. In formal terms:

O!x =

df

a51E!x

A!x =

df

?a51E!x

Now we may state the six basic principles (axioms and definitions) of our

theory in terms of these notions:

74

See Linsky and Zalta [1995].

39 A (Leibnizian) Theory of Concepts

Principle 1: O!x →a50??FxF

Principle 2: x=

E

y ≡ O!x&O!y &a50?F(Fx≡ Fy)

Principle 3: ?x(A!x&?F(xF ≡ φ)), where φ has no free xs

Principle 4: x=

A

y =

df

A!x&A!y &a50?F(xF ≡ yF)

Principle 5: x=y =

df

x=

E

y ∨ x=

A

y

Principle 6: x=y ≡ (φ(x,x) ≡ φ(x,y)), where φ(x,y)istheresult

of substituting y for one or more occurrences of x in φ(x,x), pro-

vided that y is substitutable (in the usual sense) for x at each such

occurrence.

Principle 1 tells us that ordinary objects necessarily fail to encode prop-

erties, and Principle 2 axiomatizes the notion of identity that pertains

to ordinary objects: ordinary objects x and y are identical

E

i? necessar-

ily, they exemplify the same properties. Principle 3 is the comprehension

principleforabstractobjects(‘A-objects’). Foranyconditionφonproper-

ties F, this principle has an instance that asserts that there is an abstract

object that encodes just the properties satisfying φ. Principle 4 defines

a notion of identity for abstract objects: abstract objects x and y are

identical

A

i? necessarily, they encode the same properties. Principle 5

just defines a notion of identity for any objects x and y in the domain of

objects (since every object is either ordinary or abstract, this constitutes

a completely general definition of object identity). Finally, Principle 6

asserts that identical objects may be substituted for one another in any

context φ.

75

The comprehension principle asserts the existence of a wide variety of

abstract objects, some of which are complete with respect to the proper-

ties they encode, while othersareincomplete in this respect. For example,

one instance of comprehension asserts there exists an abstract object that

encodes just the properties Clinton exemplifies. This object is complete

because Clinton either exemplifies F or exemplifies the negation of F,for

every property F. Another instance of comprehension asserts that there

75

Ishouldmention that thisprinciplecan beformulatedinmoregeneral termssothat

identical properties F and G can also be substituted for one another in any context.

If we suppose that α and β are either both object variables or both relation variables,

then our system includes the following more general formulation of the substitution

principle: α=β ≡ (φ(α,α) ≡ φ(α,β)).

Edward N. Zalta 40

is an abstract object that encodes just the two properties: being blue and

being round. This object is incomplete because for every other property

F, it encodes neither F nor the negation of F. But though abstract ob-

jects may be partial with respect to their encoded properties, they are all

complete with respect to the properties they exemplify.Inotherwords,

the following principle of classical logic is preserved: for every object x

and property F, either x exemplifies F or x exemplifies the negation of

F (i.e, ?F?x(Fx ∨

ˉ

Fx)).

76

Abstract objects are simply di?erent in kind from ordinary objects:

the latter are not the kind of thing that could encode properties; the for-

mer are not the kind of thing that could be concrete. Moreover, abstract

objects necessarily fail to exemplify the properties of ordinary objects—

theynecessarilyfailtohaveashape, theynecessarilyfailtohaveatexture,

they necessarily fail to reflect light, etc. Consequently, by the classical

laws of complex properties, abstract objects necessarily exemplify the

negations of these properties. But the properties abstract objects encode

are more important than the properties they necessarily exemplify, since

the former are the ones by which we individuate them.

The six principles listed above are cast within the framework of a

classical modal (S5 with Barcan formulas) second-order predicate logic.

77

Moreover, this logic is extended by the logical axiom for encoding and

the axioms that govern the two kinds of complex terms: (rigid) definite

descriptions of the form ?xφ and λ-predicates of the form [λy

1

...y

n

φ].

The logical axiom for encoding is:

Logic of Encoding: a51xF →a50xF

Intuitively, this says that if an object encodes a property at any possible

world, it encodes that property at every world; thus facts about encoded

76

In this principle,

ˉ

F is still defined as [λy ?Fy]. This λ-expression is governed by

the usual principle (see below in the text):

[λy ?Fy]x ≡?Fx

Note that encoding satisfies classical bivalence: ?F?x(xF ∨?xF). But the incom-

pleteness of abstract objects is captured by the fact that the following is not in general

true: xF ∨ x

ˉ

F.

77

By including the Barcan formulas, this quantified modal logic is the simplest one

available—it may interpreted in such a way that the quantifiers ?x and ?F range over

fixed domains, respectively. See Linsky and Zalta [1994], in which it is shown that this

simplest quantified modal logic, with the first and second order Barcan formulas, is

consistent with actualism.

41 A (Leibnizian) Theory of Concepts

properties are not relativized to any circumstance. This axiom and the

definitionofidentityforabstractobjectsjointlyensurethatthe properties

encoded by an abstract object are in some sense intrinsic to it.

We conclude this summary of the system by describing the axioms

governing the complex terms and some simple consequences of the fore-

going. A standard free logic governs the definite descriptions, along with

a single axiom that captures the Russellian analysis: an atomic or (de-

fined) identity formula ψ containing ?xφ is true i? there is something y

such that: (a) y satisfies φ, (b) anything satisfying φ is identical to y,and

(c) y satisfies ψ. In formal terms, this becomes:

Descriptions: ψ

?xφ

y

≡?x(φ&?z(φ

z

x

→ z=x)&ψ

x

y

), for any atomic

or identity formula ψ(y)inwhichy is free.

To keep the system simple, these definite descriptions are construed as

rigid designators, and so this axiom is a classic example of a logical truth

that is contingent. Thus, the Rule of Necessitation may not be applied

to any line of a proof that depends on this axiom.

78

The final element of the logic concerns complex relations. They are

denoted in the system by terms of the form [λy

1

...y

n

φ] meeting the con-

dition that φ have no encoding subformulas. These λ-predicates behave

classically:

79

λ-Equivalence:[λy

1

...y

n

φ]x

1

...x

n

≡ φ

x

1

,...,x

n

y

1

,...,y

n

Inlessformalterms: objectsx

1

,...,x

n

exemplifytherelation[λy

1

...y

n

φ]

i? x

1

,...,x

n

satisfy φ. A comprehension schema for relations is a simple

consequence of λ-Equivalence.

Inwhatfollows,itprovesuseful toappealtosomesimpleconsequences

of our system. Let us define ‘there is a unique x such that φ’(‘?!xφ’) in

the usual way:

?!xφ =

df

?x?y(φ

y

x

≡ y=x)

78

Further examples and discussion of such logical truths that are not necessary may

be found in Zalta [1988b].

79

I should mention here that this logic of relations issupplemented by precise identity

principles that permit necessarily equivalent properties, relations, and propositions to

be distinct. These principles are expressed in terms of a basic definition of property

identity: F =G ≡ a50?x(xF ≡ xG). For a more detailed explanation of this principle

and the definitions of relation identity and proposition identity, the reader may consult

the cited works on the theory of abstract objects.

Edward N. Zalta 42

Then we may appeal to Principles 3 and 5 to prove the following more

exact comprehension principle for A-objects:

Principle 3

prime

: ?!x(A!x&?F(xF ≡ φ)), where φ has no free xs

The proof of this principle is simply this: by Principle 3, we know there

is an A-object that encodes just the properties satisfying φ; but there

couldn’t be two distinct A-objects that encode exactly the properties sat-

isfying φ, since distinct A-objects have to di?er by at least one encoded

property.

Consequently, for each formula φ thatcanbeusedtoproducean

instance of Principle 3

prime

, the following is true:

Corollary to Principle 3

prime

: ?yy=?x(A!x&?F(xF ≡ φ))

We are therefore assured that the following description is always well-

defined:

?x(A!x&?F(xF ≡ φ))

Such descriptions will be used frequently in what follows to define various

Leibnizian concepts. Indeed, they are governed by a simple theorem that

plays a role in the proof of most of the theorems which follow:

A-Descriptions: ?x(A!x&?F(xF ≡ φ))G ≡ φ

G

F

In other words, the A-object that encodes just the properties satisfying

φ encodes property G i? G satisfies φ. This theorem is easily derivable

from the Descriptions axiom described above.

Appendix II: Proofs of Selected Theorems

The proofs in what follows often appeal to the Principles described in

Appendix I and to the theorems of world theory discussed in Appendix

III.

? Proof of Theorem 4: We prove the concepts in question encode the

same properties. (←) Assume c

G

⊕c

H

P. We need to show:

?x?F(xF ≡ G?F ∨ H?F)P

So by A-Descriptions, we must show: G?P ∨ H?P. By hypothesis,

we know:

43 A (Leibnizian) Theory of Concepts

?x?F(xF ≡ c

G

F ∨ c

H

F)P

But by A-Descriptions, it follows that c

G

P ∨ c

H

P. Now, for disjunctive

syllogism, suppose c

G

P. Then by definition of of c

G

, it follows that:

?x?F(xF ≡ G?F)P

So by A-Descriptions,weknowG?P. And by similar reasoning, if c

H

P,

then H?P. So by disjunctive syllogism, it follows that G?P ∨ H?P,

which is what we had to show.

(→) Assume:

?x?F(xF ≡ G?F ∨ H?F)P

We want to show: c

G

⊕c

H

P. By the definition of real sum, we have to

show:

?x?F(xF ≡ c

G

F ∨ c

H

F)P

By A-Descriptions, we therefore have to show that c

G

P ∨ c

H

P.By

applying A-Descriptions to our hypothesis, though, we know:

G?P ∨ H?P

So, for disjunctive syllogism, suppose G?P. Then, by A-Descriptions:

?x?F(xF ≡ G?F)P

That is, by definition of the concept G,weknow:c

G

P. By similar rea-

soning, if H ?P,thenc

H

P. So by our disjunctive syllogism, it follows

that c

G

P ∨ c

H

P, which is what we had to show. trianglerighttriangleleft

? Proof of Theorem 7:(→) Assume (x⊕y)⊕zP. Then, by definition

of ⊕,wehave:

?w?F(wF ≡ x⊕yF ∨ zF)P

This, by A-Descriptions, entails:

x⊕yP ∨ zP

Expanding the left disjunct by the definition of ⊕,wehave:

?w?F(wF ≡ xF ∨ yF)P ∨ zP

And reducing the left disjunct by applying A-Descriptions,wehave:

Edward N. Zalta 44

(xP ∨ yP) ∨ zP

This, of course, is equivalent to:

xP ∨ (yP ∨ zP)

Applying A-Descriptions in the reverse direction to this line, we obtain:

xP ∨ ?w?F(wF ≡ yF ∨ zF)P

By definition of ⊕,thisisequalto:

xP ∨ y⊕zP

And by another application of A-Descriptions, this becomes:

?w?F(wF ≡ xF ∨ y⊕zF)P

At last, by definition of ⊕,wereach:

x⊕(y⊕z)P

So if (x⊕y)⊕z encodes P,sodoesx⊕(y⊕z). (←) To show that (x⊕y)⊕z

encodes P given that x⊕(y⊕z)encodesP, reverse the reasoning. trianglerighttriangleleft

? Proof of Theorem 12i: Suppose x precedesequal y and x negationslash= y. To show that

y negationslashprecedesequal x, we need to find a property F such that yF &?xF.But,ifxnegationslash=y,

either there is a property x encodes y doesn’t, or there is a property y

encodes that x doesn’t. But, since x precedesequal y,itmustbethelatter.trianglerighttriangleleft

? Proof of Theorem 15i: Assume?z(z precedesequal x ≡ z precedesequal y) toshowthat x=y,

i.e., that for an arbitrary property P,thatxP ≡ yP.(→) Assume xP,

and for reductio, assume ?yP. To reach a contradiction, simply consider

the concept c

P

. c

P

precedesequal x and ?(c

p

precedesequal y), contrary to hypothesis. (←)By

analogous reasoning. trianglerighttriangleleft

? Proof of Theorem 16i: Assume that ?z(x precedesequal z ≡ y precedesequal z) to show that

x=y, i.e., that for an arbitrary property P,thatxP ≡ yP.(→) Assume

xP, and for reductio, assume ?yP. To reach a contradiction, consider the

fact that y must satisfy our initial hypothesis: x precedesequal y ≡ y precedesequal y.Butwe

know that y precedesequal y by the reflexivity of inclusion. So x precedesequal y, contradicting

the fact that xP &?yP.(←) By analogous reasoning. trianglerighttriangleleft

? Proof of Theorem 25:(?) Assume x precedesequal y.

a) Suppose x = y. By the idempotency of ⊕, x⊕x = x,inwhichcase,

45 A (Leibnizian) Theory of Concepts

x⊕x = y. So, we automatically have ?z(x⊕z = y).

b) Suppose x negationslash= y.Thensincex precedesequal y, we know there must be some

properties encoded by y which are not encoded by x. Consider, then, the

object that encodes just such properties; i.e., consider:

?w?F(wF ≡ yF &?xF)

Call this object ‘w’ for short (we know such an object exists by the ab-

straction schema for A-objects). We need only establish that x⊕w = y,

i.e., that x⊕w encodes the same properties as y.(→) Assume x⊕wP

(to show: yP). By definition of ⊕ and A-Descriptions, it follows that

xP ∨ wP.IfxP, then by the fact that x precedesequal y, it follows that yP.On

the other hand, if wP, then by definition of w, it follows that yP &?xP.

So in either case, we have yP.(←) Assume yP (to show x⊕wP). The

alternatives are xP or ?xP.IfxP,thenxP ∨ wP,sobyA-Descriptions:

?z?F(zF ≡ xF ∨ wF)P

So, by the definition of ⊕,wehavex⊕wP. Alternatively, if ?xP,thenwe

have yP &?xP. So by definition of w, wP, and by familiar reasoning, it

follows that x⊕wP. Combining both directions of our biconditional, we

have established that x⊕wP i? yP, for an arbitrary P.Sox⊕w = y,and

we therefore have ?z(x⊕z = y).

(?) Assume ?z(x⊕z = y). Callsuch anobject ‘w’. To show x precedesequal y,

assume xP (to show yP). Then, xP ∨ wP,whichbyA-Descriptions and

the definition of ⊕, entails that x⊕wP. But by hypothesis, x⊕w = y.So

yP. trianglerighttriangleleft

? Proof of Theorem 26:(?) Assume x precedesequal y.So?F(xF → yF). To

show that x⊕y = y, we need to show that x⊕y encodes a property P i?

y does. (→) So assume x⊕yP. Then, by definition,

?z?F(zF ≡ xF ∨ yF)P

By A-Descriptions, it then follows that xP ∨ yP.ButifxP, then by the

fact that x precedesequal y, it follows that yP. So both disjuncts lead us to conclude

yP.(←) Assume yP.ThenxP ∨ yP.SobyA-Descriptions,

?z?F(zF ≡ xF ∨ yF)P

In other words, x⊕yP.

Edward N. Zalta 46

(?) Assume that x⊕y = y.

80

To show that x precedesequal y, assume, for an

arbitrary property P,thatxP (to show: yP). Then xP ∨ yP.By

A-Descriptions, it follows that:

?z?F(zF ≡ xF ∨ yF)P

By definition of ⊕, it follows that x⊕yP. But given that x⊕y = y,it

follows that yP. trianglerighttriangleleft

? ProofofTheorem7ofWorldTheory(See Appendix III): The

proof is simplified by citing the following theorem of object theory:

81

Consequence of λ-Equivalence: a50([λy p]x ≡ p)

Now to prove the theorem, consider an arbitrary proposition r,worldw,

and object a. Then, by the Consequence of λ-Equivalence, a50([λy r]a ≡

r). Now, for the left-right direction, suppose |=

w

r. Then, since a50(r →

[λy r]a), it follows by Theorem 4 of World Theory (Appendix III), that

|=

w

[λy r]a. Analogous reasoning establishes the theorem in the right-to-

left direction. trianglerighttriangleleft

? Proof of Theorem 28: Consideranarbitraryconceptofanindividual,

say, c

a

, for some ordinary individual a. To show that c

a

is an individual

concept, we show that c

a

appears at the actual world w

α

.

82

Pick an

arbitrary property, say P. By the Lemma just stated in the text, c

a

P ≡

Pa. But, it is a consequence of the definition of the actual world w

α

that

|=

w

α

Pa ≡ Pa. So by properties of the biconditional, |=

w

α

Pa≡ c

a

P.

Since P was arbitrary, it follows that RealizesAt(a,c

a

,w

α

). Generalizing

on a, it follows that c

a

appears at w

α

. trianglerighttriangleleft

? Proof of Theorem 30: Suppose x appears at w. So some ordinary

object, sayb, realizesx atw; i.e., ?F(|=

w

Fb≡ xF). We wantto show, for

an arbitrary proposition q,thatx[λy q]i?w[λy q]. (→) Assume x[λy q].

80

If we allow ourselves an appeal to Theorem 25, we are done. For it follows from

this that ?z(x⊕z = y), which by Theorem 25, yields immediately that x precedesequal y.

81

The logical axiom λ-Equivalence is:

[λy

1

...y

n

φ]x

1

...x

n

≡ φ

x

1

,...,x

n

y

1

,...,y

n

The consequence of this logical axiom just cited in the text is derived by: letting n=1,

letting φ be the propositional variable ‘p’, and applying the Rule of Necessitation to

the result. Notice that the variable ‘y’ doesn’t appear free in the simple variable ‘p’,

and so the result of substituting the variable ‘x’for‘y’inφ (i.e, φ

x

y

)isjustφ itself.

82

The definition of w

α

is: ?z?F(zF ≡?p(p & F =[λyp])). It follows that a proposi-

tion p is true in w

α

i? p is true; i.e., that |=

w

α

p ≡ p.

47 A (Leibnizian) Theory of Concepts

So, by definition of b, |=

w

[λy q]b. And by Theorem 7 of World The-

ory (Appendix III), it follows that |=

w

q, i.e., w[λy q]. (←) Reverse the

reasoning. trianglerighttriangleleft

? Proof of Theorem 31: Assume c is an individual concept. So it ap-

pears at some world, say, w

1

. For reductio, assume that c also appears at

w

2

, w

2

negationslash=w

1

. Since the worlds are distinct, there must be some proposi-

tion true at one but not the other (by the definition of A-object identity

and the fact that worlds only encode properties of the form [λy p]). So

without loss of generality, assume that |=

w

1

p and negationslash|=

w

2

p.Sinceworlds

are maximal (by Theorem 2 of World Theory) (Appendix III), it follows

that |=

w

2

?p. But, by the previoustheorem, c mirrorsw

1

, since itappears

there. So since |=

w

1

p,weknowc[λy p]. But c also mirrors w

2

, since it

appears there as well. So, from our last fact, it follows that |=

w

2

p.This

contradicts the possibility of w

2

(Theorem 3 of World Theory). trianglerighttriangleleft

? Proof of Theorem 36: Suppose c is a individual concept. Then c

appears at some world, say w

1

, and some ordinary object, say b,real-

izes c at w

1

. Consider an arbitrary property P. By a theorem of logic:

a50?y(Py ∨?Py). Given a50?y(

ˉ

Py≡?Py), we know that a50?y(Py ∨

ˉ

Py).

So by Theorem 5 of World Theory (Appendix III), ?w?y(|=

w

(Py∨

ˉ

Py)).

Instantiating to w

1

and b, we therefore know: |=

w

1

(Pb ∨

ˉ

Pb). Now if

Pb∨

ˉ

Pbis true at w

1

, it is a simple exercise to show that either |=

w

1

Pb∨

|=

w

1

ˉ

Pb. But since b realizes c at w

1

(i.e., c encodes all and only the prop-

erties that b exemplifies at w

1

), it follows that cP ∨?c

ˉ

P. And since P

was arbitrary, c is complete. trianglerighttriangleleft

? Proof of Theorem 37: Suppose c is an individual concept. To show

that c is singular, pick an arbitrary world, say w

1

, and assume that there

areordinaryobjects, saya andb, which both realizec atw

1

. For reductio,

suppose that bnegationslash=a.Thensincea and b are distinct ordinary objects, we

know that both a=

E

a and b=

E

b.

83

Now it is a fact of object theory

that:

84

u=

E

u →a50(u=

E

u)

83

It is an immediate consequence of the principle defining =

E

described at the outset

of the paper that, for any ordinary object u, u=

E

u.

84

If u =

E

v,thengiventhatO!u is defined as a51E!u, one can, in S5, derive the

necessitation of each conjunct in the definition of =

E

(Principle 2 in Appendix I). So,

since all three conjuncts are necessary, the conjunction is necessary. Thus, a50(u=

E

v).

Edward N. Zalta 48

It therefore follows that a50(a=

E

a)anda50(b=

E

b). Now the following is a

logical theorem of object theory:

85

Consequence of λ-Equivalence: a50([λy y=

E

u]u ≡ u=

E

u)

So, in particular, a50([λy y=

E

a]a ≡ a=

E

a), and a similar claim holds for

b. So by a simple principle of modal logic, we know both a50[λy y=

E

a]a

and a50[λy y =

E

b]b. So by Theorem 5 of World Theory (Appendix III),

it follows that |=

w

1

[λy y =

E

a]a and |=

w

1

[λy y =

E

b]b. But, if the

former, then since a realizes c at w

1

, we know that c[λy y =

E

a]. But

since b realizes c at w

1

, |=

w

1

[λy y=

E

a]b. Now by Theorem 5 of World

Theory and the above Consequence of λ-Equivalence, we also know that

|=

w

1

([λyy=

E

a]b ≡ b=

E

a). So since worldsare modally closed (Theorem

4 of World Theory), it follows that |=

w

1

b=

E

a.So,a51b=

E

a,andbyfacts

governing =

E

, it follows that b=

E

a. But, by the definition of identity

(Principle 5, Appendix I), if b =

E

a,thenb = a, which contradicts our

hypothesis. trianglerighttriangleleft

? Proof of Theorem 38:(→) Assume Qa,wherea is an ordinary

object and Q an arbitrary property. To show c

a

followsequal c

Q

, we need to show

?F(c

Q

F → c

a

F). So assume c

Q

P, for an arbitrary property P.Thenby

definition of c

Q

and A-Descriptions,weknowQ ? P.SofromQa and

Q ? P it follows that Pa. So, by the definition of c

a

and A-Descriptions,

it follows that c

a

P.

(←) Assume c

a

followsequal c

Q

(to show: Qa). But by A-Descriptions and the

fact that Q ? Q, it follows that

?z?F(zF ≡ Q ? F)Q

So, by definition of c

Q

, c

Q

Q. But, by hypothesis, c

a

followsequal c

Q

,andsoit

follows that c

a

Q. But by definition of c

a

and A-Descriptions,itnow

follows that Qa. trianglerighttriangleleft

? Proof of Theorem 39b: Suppose a is an individual concept. Then a

appears at some world, say w

1

. So some ordinary object, say b, realizes

a at w

1

; i.e., ?F(|=

w

1

Fb≡ aF). Now we want to show, for an arbitrary

property, say Q,thataQ ≡ a followsequal c

Q

.

(→) Assume aQ.Wewanttoshow?F(c

Q

F → aF), so where P is

arbitrary, assume c

Q

P (to show: aP). By the definition of c

Q

, it follows

85

Since =

E

is a primitive relation symbol, we may use it to construct complex

relations. So [λy y=

E

u] is an acceptable λ-expression.

49 A (Leibnizian) Theory of Concepts

that Q ? P, i.e., a50?x(Qx → Px). Now by Theorem 5 of World Theory

(Appendix III), it follows that ?w(|=

w

?x(Qx → Px)). And in particular,

it follows that |=

w

1

?x(Qx → Px). Now since a encodes precisely what b

exemplifies at w

1

, we know in particular that |=

w

1

Qb ≡ aQ. But since

we have assumed, aQ, it follows that |=

w

1

Qb.NowsinceQb is true at

w

1

, and it is also true at w

1

that ?x(Qx → Px), it follows by the modal

closure of worlds, that |=

w

1

Pb. In which case, we know that aP,which

is what we had to show.

(←) Simple exercise. (Hint: Consider the proof of Theorem 38.) trianglerighttriangleleft

Appendix III: The Basic Theorems of World Theory

In this Appendix, we review the notion of a possible world that is defin-

able in the theory of abstract objects and review the theorems governing

this notion. The notions of ‘possible world’ and ‘truth at a world’ have

been made formally precise elsewhere in work on the theory of abstract

objects.

86

So we only sketch the basic results (definitions and theorems)

that are needed to formally representLeibniz’s modal metaphysics of con-

cepts.

We begin by extending the notion of an object encoding a property to

that of an object encoding a proposition. If we treat propositions as 0-

place properties, then an object x may encode the proposition p in virtue

ofencoding the complex propositionalproperty being such that p. We will

symbolize such a propositional property as ‘[λyp]’andsotorepresentthe

fact that object x encodes proposition p,wewrite:x[λy p].

87

We then

say that a world is any A-object x which might have encoded all and only

true propositions.

88

Using the variable ‘w’ to range over A-objects that

satisfy the definition of a world, we may then say that that a proposition

p is true at world w (‘|=

w

p’) just in case w encodes p:

|=

w

p =

df

w[λy p]

The consequences of these definitions constitute ‘world theory’. The the-

orems of world theory that play a role in what follows are:

89

86

See Zalta [1993] and [1983], Chapter IV.

87

The propositional property [λy p] is logically well-behaved despite the vacuously

bound λ-variable y. It is constrained by the ordinary logic of complex predicates,

which has the following consequence: x exemplifies [λy p]i?p, i.e., [λy p]x ≡ p.

88

In formal terms: World(x)=

df

a51?p(x[λy p] ≡ p).

89

In what follows, we always give the symbol |=

w

the narrowest possible scope; for

example, ‘|=

w

p → p’istobereadas:(|=

w

p) → p.

Edward N. Zalta 50

1. There is a unique actual world.

?!w?p(|=

w

p ≡ p)

2. Every world is maximal.

?p(|=

w

p∨|=

w

?p)

3. Every world is possible.

??p,q(?a51(p&q)& |=

w

p & |=

w

q)

4. Every world is modally closed.

|=

w

p & a50(p → q) →|=

w

q

5. A proposition is necessarily true i? true at all worlds.

a50p ≡?w(|=

w

p)

6. A proposition is possible i? there is a world where p is true.

a51p ≡?w(|=

w

p).

7. For any object x: a proposition p is true at world w if and only if

at w, x exemplifies being such that p.

?x(|=

w

p ≡|=

w

[λy p]x)

TheproofofTheorems1–6maybefoundelsewhere.

90

The proof of

Theorem 7 is given in Appendix II.

The picture that emerges from this theory of worlds can be described

in Leibnizian language. Each of these possible worlds exists (i.e., our

quantifier ‘?’ rangesoverthem all). God entertainedthem all anddecided

which one was the best. Then to ‘actualize’ the best possible world, God

made it the case that p, for each proposition p encoded in that world.

Appendix IV: Original Texts of the Cited Passages

In this Appendix, we provide the original texts for those scholars who

wish to check and confirm the accuracy of the translations and citations

whichappearin ourmain text. We begin with the sourcesinthe Gerhardt

volumes and end with the sources in the Couturat volume.

90

See Zalta [1983] and [1993].

51 A (Leibnizian) Theory of Concepts

Gerhardt

G.ii 18:

Ces dernieres paroles doivent contenir proprement la preuve de la con-

sequence, mais il est tres manifeste, qu’elles confondent necessitatem ex

hyposthesi avec la necessit′e absolue.

G.ii 20:

Car par la notion individuelle d’Adam j’entends certes une parfaite

representation d’un tel Adam qui a de telles conditions individuelles et

qui est distingu′eparl`a d’une infinit′e d’autres personnes possibles fort

semblables, mais pourtant di?erentes de Luy ... Il y a un Adam possible

dont la posterit′e est telle, et une infinit′e d’autres dont elle seroit autre

n’est il pas vray que ces Adams possibles (si on les peut appeller ainsi)

sont di?erens entre eux, et que Dieu n’en a choisi qu’un, qui est justement

le nostre?

G.ii 44:

Car tous les predicats d’Adam dependent d’autres predicats du m?eme

Adam, ou n’en dependent point. Mettant donc `a part ceux qui dependent

d’autres, on n’a qu’`a prendre ensemble tous les predicats primitifs pour

former la notion complete d’Adam su?sante `a en deduire tout ce qui luy

doit jamais arriver, ...

G.ii 56:

[...et ce n’est que dans ce sens que] je dis que la notion de la sub-

stance individuelle enferme tous ses evenemens et toutes ses denomina-

tions, m?eme celles qu’on appelle vulgairement extrinseques (c’est `a dire

qui ne luy appartiennent qu’en vertu de la connexion generale des choses

et de ce qu’elle exprime tout l’univers `asamaniere),...

G.iv 432:

Il est bien vray, que lorsqueplusieurspredicatss’attribuent `aunm?eme

sujet, et que ce sujet ne s’attribue plus `a aucun autre, on l’appelle sub-

stance individuelle.

G.iv 433:

Ilfautdoncconsiderercequec’estqued’estreattribu′everitablement`a

un certain sujet. ... Ainsi il faut que le terme du sujet enferme tousjours

celuy dupredicat, en sorteque celuyquientendroitparfaitementla notion

du sujet, jugeroit aussi que le predicat luy appartient.

...

Edward N. Zalta 52

Celaestant, nouspouvonsdirequelanatured’une substanceindividu-

elle ou d’un estre complet, est d’avoir une notion si accomplie qu’elle soit

su?sante `a comprendre et `a en faire deduire tous les predicats du sujet

`a qui cette notion est attribu′ee. Au lieu, que l’accident est un estre dont

la notion n’enferme point tout ce qu’on peut attribuer au sujet `aquion

attribue cette notion. Ainsi la qualit′edeRoyqueappartient`a Alexandre

le Grand, faisant abstraction du sujet n’est pas assez determin′ee `aun

individu, et n’enferme point les autres qualit′es du m?eme sujet, ny tout ce

que la notion de ce Prince comprend, ...

G.iv 434:

De plus toute substance est comme un monde entier et comme un

miroir de Dieu ou bien de tout l’univers, qu’elle exprime chacune `asa

fa?con...

...

Ainsi l’univers est en quelque fa?con multipli′e autant de fois qu’il y a

de substances, et la gloire de Dieu est redoubl′ee de m?eme par autant de

representations toutes di?erentes de son ouvrage.

G.iv 437:

...pour y satisfaire solidement, je dis que la connexion ou consecution

est de deux sortes, l’une est absolument necessaire, dont le contraire im-

plique contradiction, et cette deduction a lieu dans les verit′es ′eternelles,

comme sont celles de Geometrie; l’autre n’est necessaire qu’ex hypothesi,

et pour ainsi dive par accident, et celle est contingente en elle m?eme, lors

que le contraire n’implique point.

G.vi 107:

... et qu’il y a une infinit′e de Mondes possibles, dont il faut que Dieu

ait choisi le meilleur, ...

G.vi 363:

Je vous en montreray, o`u se trouvera, non pas tout `afaitlem?eme

Sextus que vous av′es vu (cela ne se peut, il porte tousjours avec luy

ce qu’il sera) mais des Sextus approchans, qui auront tout ce que vous

connoiss′es d′eja du veritable Sextus, mais non pas tout ce qui est d′eja

dans luy, sans qu’on s’en appercoive, ny par consequent tout ce qui luy

arrivera encor. Vouis trouver′es dans un monde, un Sextus fort heureux

et elev′e, dans un autre un Sextus content d’un etat mediocre, ...

G.vi 615-616:

53 A (Leibnizian) Theory of Concepts

Or, comme il y a une infinit′e des Univers possibles dans les Id′ees de

Dieu et qu’il n’en peut exister qu’un seul, il faut qu’il y ait une raison

su?sante du choix de Dieu, qui le determine `a l’un plust?ot qu’a l’autre.

G.vii 21:

Car si nous l’avions telle que je la con?cois, nous pourrions raisonner

en metaphysique et en morale `a pue pres comme en Geometrie et en

Analyse, parce que les Caracteres fixeroient nos pens′ees trop vagues et

trop volatiles en ces matieres, o`u l’imagination ne nous aide point, si ce

ne seroit par le moyen de caracteres.

G.vii 200:

Quofacto, quandoorienturcontroversiae,nonmagisdisputationeopus

erit inter duos philosophos, quam inter duos Computistas. Su?ciet enim

calamos in manus sumere sedereque ad abacos, et sibi mutuo (accito si

placet amico) dicere: calculemus.

G.vii 218:

ab est a sive (omne) animal rationale est animal.

a est a sive (omne) animal est animal.

G.vii 222:

... nihil referre sive dicas ab sive dicas ba,...

...Repetitio alicujus literae in eodem termino inutilis est et su?cit

eam reineri semel, exempli causa aaseu homo homo.

G.vii 223:

d est c,ergobd est bc rursus per priora.

G.vii 230:

Axioma 1. Si idem secum ipso sumatur, nihil constituitur novum, seu

A⊕A ∞ A.

G.vii 236:

Defin. 1. Eadem seu coincidentia sunt quorum alterutrum ubilibet

potest substitui alteri salva veritate. Exempli gratia, Triangulum et Tri-

laterum,... A∞ B significat A et B esse eadem, ...

Propos. 1. Si A sit ∞ B, etiam B erit ∞ A.

Prop. 2. Si A non ∞ B, etiam erit B non ∞ A.

Prop. 3. Si A ∞ B, et B ∞ C, erit A ∞ C.

Coroll. Si A ∞ BetB∞ CetC∞ D, erit A ∞ D.

G.vii 237:

Edward N. Zalta 54

Prop. 4. Si A ∞ BetBnon∞ C, erit A non ∞ C.

Def. 3. A inesse in L seu L continere A idem est ac pro pluribus inter

quae est A simul sumtis coincidens poni L.

Axiom 1. A⊕A ∞ A.

Axiom 2. B⊕N ∞ N⊕B.

Prop.5.SiAestinB,etsitA∞ C, etiam C est in B.

Prop.6.SiCestinB,etsitA∞ B, etiam C erit in A.

G.vii 238:

Prop. 7. A est in A.

Prop. 8. A est in B, si A ∞ B.

Prop. 9. Si A ∞ B, erit A⊕C ∞ B⊕C.

Scholium: Haec propositio converti non potest, multoque minus duae

sequentes, et infra in probl. quod est prop.23docebitur modus instantiam

reperiendi.

Prop. 10. Si A ∞ LetB∞ M, erit A⊕B ∞ L⊕M.

Scholium: Haec propositio converti non potest, neque enim si sit A⊕B

∞ L⊕MetA∞ L, sequitur statim esse B ∞ M;...

G.vii 239:

Prop. 12. Si B est in L, erit A⊕BinA⊕L.

Prop. 13. Si L⊕B ∞ L, erit B in L.

Prop. 14. Si B est in L, erit L⊕B ∞ L.

G.vii 240:

Prop. 15. Si A est in B et B est in C, etiam A est in C.

Corollary to Prop. 15: Si A⊕N est in B, etiam N est in B.

Prop.17.SiAestinBetBestinA,eritA∞ B.

G.vii 241:

Prop. 18. Si A est in L et B est in L, etiam A⊕BeritinL.

Prop. 20. Si A est in M et B est in N, erit A⊕BinM⊕N.

Couturat

C 51:

Omnem propositionem veram categoricam 〈a?rmativam 〈universa-

lem〉〉, nihil aliud significare quam connexionem quandam inter Pr?dica-

tum et subjectum 〈in casu recto de quo hic semper loquar〉, ita scilicet ut

55 A (Leibnizian) Theory of Concepts

pr?dicatum dicatur inesse subjecto 〈vel in subjecto contineri, eoque vel

absolute etin se spectato, vel certe [in aliquo casu] seu in aliquoexemplo〉,

seu ut subjectum dicto modo dicatur continere pr?dicatum : hoc est ut

notio subjecti 〈vel in se, vel cum addito〉 involvat notionem pr?dicati, ...

C 235:

(6) AA ∞ A

(7) AB ∞ BA

C 260:

(14) AA idem est in hoc calculo quod A.

C 262:

(7) AA idem est quod A.

C 366:

(18) Coincidunt A et AA, ...

C 396:

189. Principia ergo h?c erunt: 〈prim`o〉 aa = a (unde patet etiam non

b =nonb, si ponamus non b = a).

C 421:

(3) A ∞ AA.

(4) AB ∞ BA seu transpositio nil nocet.

Bibliography

Bealer, G., 1982, Quality and Concept, Oxford: Clarendon

Bealer, G., 1998, ‘A Theory of Concepts and Concept Possession’, Philo-

sophical Issues 9(=Concepts, E.Villanueva(ed.), Atascadero: Ridge-

view): 261-301

Casta?neda, H., 1990. ‘Leibniz’s Complete Propositional Logic’, Topoi,

9/1: 15-28

Casta?neda, H., 1976, ‘Leibniz’s Syllogistico-Propositional Calculus’, No-

tre Dame Journal of Formal Logic, 17: 481-500

Chierchia, G. and Turner, R., 1988, ‘Semantics and Property Theory’,

Linguistics and Philosophy, 11 (August): 261-302

Cocchiarella, N., 1978, ‘On the Logic of Nominalized Predicates and its

Philosophical Interpretations’, Erkenntnis 13: 339-369

D’Agostino, F., 1976, ‘Leibniz on Compossibility and Relational Predi-

cates’, The Philosophical Quarterly, 26/103: 125-38

Edward N. Zalta 56

Fitch, G., 1979, ‘Analyticity and Necessity in Leibniz’, Journal of the

History of Philosophy, 17/1: 29-42; page reference is to the reprint in

Gottfried Wilhelm Leibniz: Critical Assessments,Volume1,R.Wool-

house (ed.), London: Routledge, 1994, pp. 290-307

Ishiguro, H., 1972, Leibniz’s Philosophy of Logic and Language, London:

Duckworth

Ishiguro, H., 1990, Leibniz’s Philosophy of Logic and Language,Cam-

bridge: Cambridge University Press, second edition

Kauppi, R., 1967, Einf¨uhrung in die Theorie der Begri?ssysteme (mono-

graph), in Acta Universitatis Tamperensis,Ser.A,Vol.15,Tampere:

Tampereen Yliopisto

Kauppi, R., 1960,

¨

Uber die Leibnizsche Logik, Helsinki: Suomalaisen Kir-

jallisuuden Kirjapaino Oy

Leibniz, Gottfried Wilhelm:

C Opuscules et fragments in′edits de Leibniz, L. Couturat (ed.),

Paris, 1903

CA Correspondence with Arnauld,translatedbyH.T.Mason,Man-

chester: Manchester University Press, 1967

G Die philosophischen Schriften von Gottfried Wilhelm Leibniz,

Volumes i – vii, C. I. Gerhardt (ed.), Berlin, 1875–90

GI Generales Inquisitiones de Analysi Notionum et Veritatum,

F. Schupp (ed.), Hamburg, 1982

LP Logical Papers, G. H. R. Parkinson (ed. and trans.), Oxford:

Clarendon, 1966

PW Philosophical Writings,G.H.R.Parkinson(ed.),M.Morrisand

G. Parkinson (trans.), London: Dent & Sons, 1973

T Theodicy, translatedbyE.Huggard,NewHaven: YaleUniversity

Press, 1952

Lenzen, W., 1990, Das System der Leibnizschen Logik, Berlin: W. de

Gruyter

Lewis, D., 1968,‘CounterpartTheoryandQuantifiedModalLogic’, Jour-

nal of Philosophy, 68: 113-126

Linsky, B., and Zalta, E., 1995, ‘Naturalized Platonism vs. Platonized

Naturalism’,The Journal of Philosophy, xcii/10(October1995): 525-

555

57 A (Leibnizian) Theory of Concepts

Linsky, B., and Zalta, E., 1994, ‘In Defense of the Simplest Quantified

Modal Logic’, Philosophical Perspectives 8: 431-458

Lloyd, G., 1978, ‘Leibniz on Possible Individuals and Possible Worlds’,

Australasian Journal of Philosophy, 56/2: 126-142; page reference

is to the reprint in Gottfried Wilhelm Leibniz: Critical Assessments,

Volume 1, R. Woolhouse (ed.), London: Routledge, 1994, pp. 366-385

Mates, B., 1986, The Philosophy of Leibniz: Metaphysics and Language,

Oxford: Oxford University Press

Mates, B., 1972, ‘Individuals and Modality in the Philosophy of Leibniz’,

Studia Leibnitiana, 4: 81-118

Mates, B., 1968, ‘Leibniz on Possible Worlds’, in Logic, Methodology,

and Philosophy of Science III, B. van Rootselaar and J. Staal (eds.),

Amsterdam: North Holland; page references are to the reprint in

Leibniz: A Collection of Critical Essays, H. Frankfurt (ed.), Notre

Dame: Notre Dame University Press, 1976 (Garden City: Anchor

Books, 1972)

Menzel, C., 1986, ‘A Complete, Type-Free “Second Order” Logic and

Its Philosophical Foundations’, Center for the Study of Language and

Information, Technical Report #csli–86–40, Stanford University

Mondadori, F., 1973, ‘Reference, Essentialism, and Modality in Leibniz’s

Metaphysics’, Studia Leibnitiana, 5/1: 74-101;pagereferenceisto the

reprint in Gottfried Wilhelm Leibniz: Critical Assessments,Volume

1, R. Woolhouse (ed.), London: Routledge, 1994, pp. 230-255

Mondadori, F., 1975, ‘Leibniz and the Doctrine of Inter-World Identity’,

Studia Leibnitiana, 7/1: 21-57; page reference is to the reprint in

Gottfried Wilhelm Leibniz: Critical Assessments,Volume1,R.Wool-

house (ed.), London: Routledge, 1994, pp. 256-289

Montague,R., 1974,Formal Philosophy: Selected Papers of Richard Mon-

tague, Richmond Thomason (ed.), New Haven: Yale University Press

Parsons, T., 1978, ‘Nuclear and Extranuclear Properties, Meinong, and

Leibniz’, Nous 12: 137-151

Parsons, T., 1980, Nonexistent Objects, New Haven: Yale University

Press

Peacocke,C.,1991,‘TheMetaphysicsofConcepts’,Mind, C/4(October):

525-546

Rescher, N., 1954, ‘Leibniz’s Interpretation of His Logical Calculi’, in

Journal of Symbolic Logic, 19/1 (March): 1-13

Russell, B., 1900, A Critical Exposition of the Philosophy of Leibniz,

Edward N. Zalta 58

Cambridge: Cambridge University Press

Swoyer, C., 1998, ‘Complex Predicates and Theories of Properties’, Jour-

nal of Philosophical Logic, 27: 295-325

Swoyer, C., 1996, ‘Theories of Properties: From Plenitude to Paucity’,

Philosophical Perspectives, 10: 243-264

Swoyer, C., 1995, ‘Leibniz on Intension and Extension’, Nous 29/1: 96-

114

Swoyer, C., 1994, ‘Leibniz’s Calculus of Real Addition’, Studia Leibni-

tiana XXVI/1: 1-30

Vailati, E., 1986, ‘Leibniz on NecessaryandContingentPredication’, Stu-

dia Leibnitiana, 18/2: 195-210

Williamson, T., 1998, ‘Bare Possibilia’, Erkenntnis 48: 257-273

Wilson, M., 1979, ‘Possible Gods’, Review of Metaphysics, 32/4: 717-733

Zalta, E., 1999, ‘Natural Numbers and Natural Cardinals as Abstract

Objects: A Partial Reconstruction of Frege’s Grundgesetze in Object

Theory’, Journal of Philosophical Logic, 28/6: 619-660

———, 1993, ‘Twenty-Five Basic Theorems in Situation and World The-

ory’, Journal of Philosophical Logic, 22: 385-428

———, 1988a, Intensional Logic and the Metaphysics of Intentionality,

Cambridge, MA: MIT/Bradford

———, 1988b, ‘Logical and Analytic Truths That Are Not Necessary’,

The Journal of Philosophy, 85/2: 57-74

———, 1983, Abstract Objects: An Introduction to Axiomatic Meta-

physics, Dordrecht: D. Reidel