"It was Frege and Russell, not Wittgenstein or Quine, who began what may be called the ontology of the analytic tradition."
Jan Dejnožka, The Ontology of the Analytic Tradition and Its Origins, (p. 149)
"Russell, in Principles of Mathematics, may seem to be a radical relativist. I quote this text again:
Numbers cannot be asserted of objects, because the same set of objects may have different numbers assigned to them...; for example, one army
is so many regiments and such another number of soldiers. This view seems to me to involve too physical a view of objects: I do not consider the army to be the
same object as the regiments. (Principles of Mathematics [POM] 519).
Alternatively, in that every application of a concept 'presupposes numerical diversity', in that every entity has its own immediate identity,
Principles suggests a radical realism. The one thing Principles seems to not to be is modified realism, since Russell expressly denies the distinction between
real distinction and conceptual distinction (POM 466). But this denial seems quite disingenuous in light of his own distinction between empirical (or
actual) existence and mathematical (or logical) existence.
Surely the truth is that Principles indulges in a rich and complex modified realism. Spatial, temporal, and material points are kinds of
terms which differ only immediately. (...)
Consider also Russell's distinction in Principles between actual existence, existence, and mere nonexistent being, in order of progressively
muted substance substitutes. Empirical existents are much like Frege's concrete objects, existents are much like Frege's abstract objects. Logical existents
seem real than empirical existents, but more real than nonexistent beings. Spatiotemporal reals (points and instants) seem to be in between empirical existents
and logical existents, since empirical evidence determines the geometry actual world. Being is the general status of which the foregoing are kinds. Objects
roughly include both terms and classes as many (POM 55n). Terms are simply beings. Classes as many have mathematical existence, or better, logical
existence. Properties and relations are probably hybrid classifications, since some are empirically given and others are logico-mathematical. Possibly there
are similar gradations of ontological status among nonexistent chairs, nonexistent material points, nonexistent colors, and so on, though possibly they are all
just nonexistent entities. Russell does not address that question.
Russell rejects substances and essences in the traditional sense. But he admits six sorts of beings or substances, or substance substitutes:
(1) All entities, including both being and existence, have timeless being in 1903. (2) Universals' have being in 1912. (3) Being is general timelessness in
1914. (4) Being is logical atoms in 1918. (5) Being is object words in 1940. (6) Being is qualities (particulars, not universals) in 1940-59. 1 described these
six sorts of being in my Erkenntnis paper (Dejnožka 1990). In addition, Russell admits two substitutes for material substances: (7) Ordinary physical
things are causal lines in 1927-59 (The analysis of matter [AMA] 285; Human Knowledge: its scopes and limits [HK] 453-60, 489-90; My Philosophical development
[MPD] 146-47). 'Thus the persistence of substance is replaced by the persistence of causal laws" (AMA 285). (8) Space-time structures are what are probably
real in 1927-59 (AMA 249-57; HK 250-66, 460-75, 491-92; MPD 147-48). Russell speaks of 'substantial structures' which replace 'pieces of matter' and also of
structures of events (HK 461). Of course, (7) and (8) overlap; a causal line is an instantiated structure.
In his 1914-18 philosophy of logical fictions, in which particulars (sense-data) or perhaps simples are alone real, Russell may seem a
radical realist. Bodies, numbers, and minds (except one's own mind) are logical fictions with fictitious identities. And 'there is no such thing as a fiction'
(The Philosophy of Logical Atomism [PLA] 189). In 1919 this virtually becomes Hume's neutral monist distinction between impressions and fictions. Like Hume's
impressions, Russell's particulars are real beings. Each can logically happen to be the whole universe. But instead of admitting distinctions of reason within
lone sense-data, as Hume does within impressions, Russell admits "parts" which, if you attend to them, become 'new' data (new real beings) in their own right
(PLA 203; see On the relation of universals and particulars [RUP] 114 and An inquiry into meaning and truth [IMT] 334). Much as with Frege, this is a shifting
of phenomenological real identities over time sans any shifting of concepts. Russell assigns particulars the 'logical position' of substances (PLA 204).
Particulars are mind-independent (1a), essentially complete (1b), ultimate logical subjects of predication (1c), logically independent (1d), given in
acquaintance (l1), the unchanging building blocks in the logical construction of changes (1f), and have phenomenologically real identities as opposed to the
conceptual identities of logical fictions (1g). Criterion (1) seems fulfilled-but for radical realism, since logical fictions are said to exist only in a
purely nominal sense.
Nonetheless, I classify the 1914-18 Russell as a modified realist. For there is that exception to logical fictions, one's own mind, which
ought to be in some sense more substantival than sense-data, despite everything Russell says about sense-data as being as real as anything can be. Only the
1921 Russell's neutral monism, in which even one's own mind is a construction, seems a truly radical realism. It is also worth noting that as series of classes
of sensibilia, two constructed bodies are really distinct in sense (2) just in case they have no sensibilium in common.
The 1914-21 Russell's constructionism (this includes neutral monism), in sting unsensed sensibilia to account for perception and physical
lawfulness, a scientific explanatory realism. It is also a phenomenological realism in that se-data are physically real events. And third, it is a
methodological realism. Analyses end with sensed entities, if not with entities known to be simple.
Russell's 1927-59 representational realism meets criterion (3) of explanatory modified realism. It is a kind of scientific realism. In The
Analysis of Matter, Russell defends realism against radical reductionism. He says, "There are many possible ways of turning some things hitherto regarded as
'real' into mere laws concerning the other things. Obviously there must be a limit to this process, or else all the things in the world will merely be each
other's washing" ([AMA] 325). Russell says, "We must find some reality for the electron, or else the physical world will run through our fingers like a
jelly-fish" (AMA 319). Thus physical structures such as electrons are not mere logical fictions. Indeed, two electrons are really distinct in sense (2) if they
have no constituent event in common (AMA 288). Yet Russell reserves metaphysical status for the events which compose electrons, and ultimately for whatever
entities may comprise the final interpretation of physics (AMA 2, 9). This suggests a modified realism in which instantiated physical structures are real
facts, but are less real than any ultimate, i.e. simple, constituents they may have." (pp. 244-247)
From: Jan Dejnožka, The Ontology of the Analytic Tradition and Its Origins. Realism and Identity in Frege, Russell, Wittgenstein, and
Quine, Lanham: Littlefield Adams Books 1996. (Paperback edition reprinted with corrections, 2002; reprinted with further corrections, 2003).