In this discussion we’ll look again at metaphysics, covering (in more depth) some of the same ground as the previous instalment but also considering some new aspects. In particular, we’ll study metaphysics insofar as it is the attempt to investigate Being – especially those categories into which philosophers have suggested everything that exists must fall. What are these categories, though? How do we distinguish between them? How should we characterise them? These are the kinds of questions we’ll examine, alongside classical and contemporary metaphysical problems.
Since the Ancient Greeks, Being has not been considered the same as existence. The former was understood to include not just those things that exist but also the various categories that such a thing could have: being tired or being scared, for example. Perhaps the most famous treatment of Being was in the dispute between Parmenides and Heraclitus in Plato’s dialogues. According to the former (in Kenny’s translation):
What you can call and think must Being be
For Being can, and nothing cannot, be.
From here a simple argument with devastating consequences could be developed. Being had to be
, by definition. Likewise, non-Being could not
be. These agreed, what of change
? If Being were to change, it would have to become non-Being – which cannot be. The conclusion had to be that there is no change. (A similar but later version of this argument would be used in theology: if there were a perfect being, such as God or the Absolute, how could He/it change? After all, a change from perfection would have to be to non-perfection.) We even find this problem in everyday life when asking someone “what are you thinking about?” and receiving the answer “oh, nothing”. It was relied upon by philosophers then and now asking about creation: how can nothing be
? If we agree that it cannot, how can something come from nothing? Some, like Aristotle, relied on a version of this thinking to argue that the universe had to be eternal and uncreated.
These were – and remain – difficult arguments to counter. The basis of it all, Being as fixed, was opposed by Parmenides contemporary Heraclitus, who insisted to the contrary that the fundamental nature of the universe was change
: everything is forever in flux. This was famously stated as “you cannot step into the same river twice”. These two positions formed a metaphysical battleground for subsequent philosophers. Plato tried to reconcile them, firstly by separating the universe into a realm of ideas
which was timeless (or Parmenidian) and a realm of the senses
which was in flux (or Heraclitan). This was the Theory of Ideas
, subject to critique in the middle dialogue called Parmenides
and separating the ideal
(“Good”, for example) from its approximations in the intelligible world (such as conduct we call “good”). It was eventually superseded by the theory of forms
, found in the Sophist
, which added to Being four additional forms: same, difference, motion and rest. The second of these allowed the possibility of avoiding Parmenides metaphysical straightjacket: when we talk of that which is not, we do not speak of non-Being but instead something that differs from what is. The collection of all “non-x”s, such as non-righteous, non-circular, and so on, give us non-Being, which is (by construction) just as real as the set making up Being. This clever solution nevertheless provided ample scope for continued study of Being, not the least of which was the status of universals
The Problem of Universals
Suppose we take a proposition like the following:
Suppose further that (1) is true
. Recalling our previous discussion
, the subject
of (1) is “Hugo” and the predicate
is “wise”. Moreover, we say that “Hugo” refers
to something: the subject, Hugo. What about “wise”? Does it refer to anything and, if so, what?
Although there are many species, the basic contention of metaphysical realism
is that the predicate “wise” also
refers. In other words, the truth of (1) results from a match between a linguistic
(the proposition) and non-linguistic
(the way the universe is) arrangement, the proposition in (1) picking out the circumstance that Hugo really is
wise. Following on from this, we can straightforwardly say that there must be something called “Hugo” in the world for (1) to be true. Likewise, suggests the metaphysical realist, there must be something
corresponding to what we describe by “wise”.
This is not quite the full story, however. Suppose we take another proposition:
Since the structures of (1) and (1*) are the same, the metaphysical realist concludes that both are pointing at the same thing: the quality of “being wise” that is present in both Hugo and Paul. Indeed, it is because of the existence of this quality that (1) and (1*) are true. However, what is it that the predicate in these propositions is referring to? The word “wise” does not name
a referent (the thing it is pointing to) because what is actually at issue is a more general concept: wisdom
. In that case, (1) should be read as
- Hugo typifies wisdom. (1a)
The way to understand propositions like (1), then, is to adjust them slightly in this fashion and read them as stating a match (or approximation, perhaps) between the subject and predicate. Since we want to be able to say and make sense of propositions like (1), we are committed to the metaphysical machinery that allows us to – and that, says the metaphysical realist, involves accepting that “wisdom” and other concepts actually exist.
There are many qualities that might take the place of “wisdom” in (1a) – such as folly
, to give some more realistic examples – and these are what we term universals
. A universal can be a property
(as “wisdom” functions in (1a)), but there are other possibilities. Consider:
- Hugo is a male. (2)
- Hugo is the father of Trystyn. (3)
(2) matches the subject (“Hugo”) with a kind
(“male”, “human” or “rugby player”, for instance) while (3) gives a relation
(“teacher”, “son” or “team mate”, say). For (3) to be true, are we committed to the existence of “fatherhood”? This is the question asked by the problem of universals and which metaphysical realists answer in the affirmative.
A major criticism of this account, however, is that it leads to an infinite regress. For (1) to be true, we agreed that (1a) had to also be true; that is, that Hugo typify wisdom (in the literature this is sometimes called exemplifying
). This suggests that for (1a) to be true we further require another universal – typification
– as a relation, leading to another proposition:
- The typification in (1a) is a relation. (1b)
(1b) checks whether (1a) enters into the correct kind of typification, but to be sure that (1b) is true we would need a (1c), and so on: an infinite regress. This seems to commit us to an ontology of universals piled upon universals, which is unsatisfactory even to many realists.
There are several ways around this objection. One is to bite the bullet and accept that it must be so, rather than lose the ability to make sense of propositions. Another is to deny that there is any infinite regress by saying that the initial analysis is all we need to understand what propositions mean. The subsequent levels, then, need not commit us to the existence of anything else because the process of typification does not require additional levels. This is tantamount to saying that the metaphysical realist’s account does not apply everywhere, so it has a restricted domain of validity. A third option is to suggest that (1a) is just another way of saying (1), so that the difference is only grammatical and does not require a separate round of analysis.
An altogether different complaint against the metaphysical realism we have so far considered is to ask about predicates like “married” and “unmarried”. Suppose we take the following propositions:
- Hugo is unmarried. (4)
- Hugo is married. (5)
If (4) is true then (5) is false, and vice versa (unless we adopt dialetheism
). Why, then, do we need both “married” and “unmarried” to be universals to make sense of a proposition like (4)? If we agree that some universals are superfluous, however, how do we decide which ones are necessary and which are not? Some metaphysical realists (who are usually also scientific realists
, which we will come to later) claim that the predicates we require are those needed for a final physical theory, but the objection made by Hempel’s dilemma (considered in our fourteenth piece
) makes this problematic.
Because of its association with several of Plato’s dialogues, metaphysical realism is also often called Platonism
. An area of disagreement among realists is whether all universals necessarily
exist; that is, are there universals that might have existed but do not, or do all universals exist regardless of whether we come across them in our universe? For example, take a proposition like:
- Hugo is a married bachelor. (6)
Although we might say that (6) is false (or indeed meaningless – see our seventeenth discussion
) because there can be no such thing as a “married bachelor”, our metaphysical realist reading of (6) seems to imply that “Hugo” as subject and “married bachelor” as predicate must exist and not match. We arrive at what Aristotle called a two worlds ontology
wherein some universals are typified by particular instances and some are not, and we can ask how we can ever know anything about the latter or, more importantly, how there can be any connection between the two. Thoroughgoing Platonists suggest in response that we can learn about those universals that are not typified from our experience with those that are.
The problem of universals has a long and distinguished pedigree, having been studied by philosophers through the ages. Those who have rejected universals have traditionally been called nominalists
, meaning “name”), the most famous being William of Ockham. This historical link provides us with the main objection raised against universals by nominalists (in addition to some of those already considered): the principle now known as Ockham’s Razor
According to the nominalist, we can understand propositions like those we have considered above solely by reference to particulars. Although the realist account employing universals may seem convincing, too, it requires additional
entities: universals. The role played by these in explaining propositions like (1) may be interesting but is inessential. Since universals represent a metaphysical theory and Ockham’s Razor enjoins us to accept the most parsimonious theory, universals are eliminated by nominalists.
Nominalism, then, involves a claim that a metaphysical theory is possible which only involves particulars; and this is a claim that needs to be justified. At this point the nominalists part company in offering differing accounts. An austere
form suggests that the realist’s story does not achieve anything and that propositions like (1) are irreducible. “Hugo is wise” is true because Hugo is
wise. To claim that (1) holds because it can be understood as (1a) does not achieve anything, since “Hugo typifies wisdom” means only that Hugo is wise. This apparently trivial reading is all that is required, says the austere nominalist, and the appeal to a universal is no less so.
Problems with this approach arise as soon as we consider a proposition containing abstract concepts, such as:
- Honour is praiseworthy. (7)
Here the nominalist may wish to translate
(7) to make sense of it other than by appealing to universals. Suppose we do so and take the new proposition to be irreducible:
- Honourable people are praiseworthy. (7a)
The nominalist may presume that (7a) requires no further analysis, so that it is true because honourable people are
praiseworthy. However, we could imagine a person who is honourable but also a murderer, say – a quality we would likely agree is not at all praiseworthy. Thus it is possible for (7) to be true while (7a) is not; so (7a) cannot be an accurate translation of what we mean by (7).
Another possibility is the following:
- Other things being equal, honourable people are praiseworthy. (7b)
This is easy to understand but much harder to clarify: these “other things” are precisely those the nominalist is proposing to eliminate as superfluous. How can this austere approach be simpler, then?
Another approach takes the discussion so far to be mistaken in supposing that propositions are talking about non-linguistic entities; instead, they are just linguistic expressions that we employ to talk about sentences having similar forms. This, it turns out, is much the same as the view held by Roscelin, Ockham and Abelard in the twelfth century, according to which it is only names
that can be universal – not predicates (whence nominalism
). A recent form of this meta-linguistic
nominalism was detailed by Wilfred Sellars, according to which the discussion of a universal is really only talk about linguistic expressions. For example, the use of “wise” in (1) should properly be understood as saying that all instances of this predicate are adjectives, describing a particular characteristic of Hugo. The correct way to analyse them is by their use
in language, not by reference to universals. However, critics have noted that the function of “wise” in English is the same as (translated) terms in other languages and therefore have suggested that Sellars’ account would commit him to the existence of “linguistic roles” as universals.
Much of modern nominalism derives from the insights of Wittgenstein and the suggestion that words gain their meaning from their use, as hinted at above. The realist insists that a proposition like (1) requires the existence of “wisdom” in order to make sense of it, but the nominalist can disagree and say that we know what (1) means because we learn
to understand terms like wise. To declare that Hugo is wise, then, is just to remark that his behaviour resembles that which we have come to call wise, and nothing more. The debate continues, which is why the problem of universals has held a fascination for thousands of years.
The Problem of Realism
Since the development of science, the problem of universals has taken on a new aspect. Faced with an array of scientific theories that apparently work extremely well (insofar as they make correct predictions or allow us to control phenomena), philosophers of science have asked why
this is so. Explaining this success is a genuine challenge, one possible response to which is to say that it is due to our theories accurately getting at reality.
In basic terms, there is a division between those who believe that this reality exists independently of us and those who are not so sure. According to the realist
, an explanation of planetary orbits invoking gravity works because there really are planets and a force we call gravity; and, moreover, that this is so whether we are here to notice and remark on it or not. They account for this conception by a theory of meaning much the same as that we have already covered, whereby true statements about the universe work because they get at real things – like quarks, aardvarks and philosophers.
It is important to realise that opposition to scientific realism does not consist in the denial that reality exists. This suggestion is a straw man of a complex set of arguments and it is hard to see why it should be worth considering. Instead, anti-realists maintain that what we refer to as reality
is made up at least in part by our perceptual apparatus or the way in which we experience it. They do this in a variety of ways, from the idea that esse est percipi
(“to be is to be perceived”) from Bishop Berkeley’s idealist account to the more recent semantic anti-realism of Hilary Putnam and Michael Dummett. The instrumentalist
form relates more specifically to the philosophy of science and will thus be covered in our next essay, particularly its connection with nominalism.
The realist account relies on the principle of bivalence
, according to which the reality described by a statement either obtains or it does not. This is so regardless of our epistemological capabilities: if we say that bodies are attracted according to a law of gravitation described by a certain equation, then this is either true or false in the final analysis, whether or not we can ever know it to be so. The combination of this principle and the metaphysical apparatus discussed above in the section on universals is what the realist uses to ascertain the meaning of a statement.
By contrast, the anti-realist employs a theory of meaning according to which we know the meaning of a statement insofar as we have a warrant for it; that is, we know what it would take for the statement to be considered true or false. The obvious corollary, however, is that a statement which is impossible to justify in principle
would thus violate the principle of bivalence. The rejection of this principle thus characterises the anti-realist – at least according to Dummett.
Although this excursion into the philosophy of language may seem like hair-splitting, it is easy to find propositions to illustrate the difficulty. For example, consider the following:
- Plato enjoyed whistling. (8)
For the realist, either (8) obtains or it does not. If true, its meaning is plain: Plato did enjoy whistling. This type of statement, however, is just the kind of (apparently) undecidable one that might be expected to cause us trouble. Its meaning is seemingly straightforward, but how can we say that Plato did (or did not) enjoy whistling when this is unjustifiable? We know its meaning implicitly
, says the anti-realist, but we cannot do so explicitly unless we assume its truth beforehand (thereby begging the question) or becoming trapped in an infinite regress (as before with universals).
Another version of anti-realism relies on the inscrutability of reference
. Suppose we take a word from a new language which we are trying to translate into our own. If we point to the thing we believe it to denote, say, the speaker may nod enthusiastically but the precise meaning of the word is under-determined. This is because the native speaker may understand us to mean the object as a whole, the collection of its parts, the general concept it embodies, and so on; just as if someone indicated a tiger and said “cat?”
The point of this for the anti-realist is that it suggests that a direct translation between language and reality is impossible, and that some kind of mediation
is required. If this is so, we would have to give up the idea of a mind-independent reality. Realists respond by saying that if there is an inscrutability in talking of reference in realist terms, the same must apply to anti-realist conceptual schemes. The realist can also remark that the under-determinancy of reference might apply to some
terms but not necessarily all. A distinction like this is made in the philosophy of science, which we will return to later in this series. Being the modern counterpart of the problem of universals, the problem of realism continues to be debated.
Whatever the status of universals, another issue for metaphysics is the make-up of the particulars relied upon by both realist and nominalist alike.
Bundle and Substratum Theories
Particulars are, as we have said, “things” (people, objects, critters and the like), but what can we say about their structure? One ontological theory holds that a particular is constituted by the many properties we associate with it and an underlying substratum
, existing independently of the properties overlaying it. Bundle
theorists, however, who have tended to be empiricists, disagree that any substratum exists and suggest that particulars are no more than “bundles” of their properties, arguing that substrata have no empirical content (being beyond the reach of any experience in principle) or that there is no need to posit a substratum to explain particulars.
There have been interesting objections made to bundle theories. Firstly, suppose that a particular changes
. If we believe that the particular was but a bundle of its attributes then the changed particular would no longer be identical with itself. That is, the Hugo of tomorrow is not the Hugo of today. We will return to this difficulty shortly in considering time. Secondly, however, consider a list of propositions describing Hugo:
- Hugo is cowardly. (9)
- Hugo is slow. (10)
- Hugo is boring. (11)
We have discussed above the question of what exactly each of the predicates here are picking out (“cowardice”, “slowness”, and so on), but what about Hugo, the subject
? If Hugo is no more than a collection of his attributes then we can recast each of the propositions as tautologies – “A cowardly, slow, boring … thing is cowardly” in (9), for instance; and are thus not really saying anything about Hugo. What we require, according to the substratum theorist, is something underlying all these propositions about Hugo in order to make sense of them at all.
The bundle theorist can respond that this is as much a difficulty for an account relying on substratum. Since nothing can be said about this fundamental character of a particular, beyond its attributes, we are no closer to understanding (9). Moreover, why should we presuppose that we need to know everything about a particular in order to describe it via propositions like (9) – (11)?
Another criticism of bundle theories relies on the identity of indiscernibles
. If two particulars share all their attributes then this principle states that they must be identical. In that case, if Hugo and Paul alike satisfied the propositions above we would be forced to accept that they were not distinct individuals unless we allow that there is something additional about them – their substrata. This is a much more difficult objection, one which has led to much recent work in metaphysics. Nevertheless, the bundle theorist can ask what the substratum beyond attributes can be, since it has no attributes. How can we describe it, then? We appear stuck between an inability to discern individual particulars sharing the same attributes or the impossibility of characterising the supposed substratum that distinguishes them.
To avoid this dilemma, Aristotelians have attempted to demarcate kinds
. The former are what particulars belong to
(so Hugo is a human, or a man) while the latter are what they have
(Hugo being cowardly, and so on). We could then have several instances of the same kind
(Hugo and Paul), sharing the same properties but nevertheless being distinct. Although it is difficult, perhaps, to see why this should solve the problem, the claim is that membership of a kind
is what individuates particulars. While humans share properties, it is membership of a kind that marks them out as distinct. The elaboration of the full Aristotelian account, however, is beyond the scope of this introduction.
The subject of time
throws up a host of metaphysical questions, even before we get to the physics associated with its study. Does the future already exist, along with the past? Do we “live for the moment”, as many a romantic has suggested while crooning below a balcony? If the universals we considered above really do exist, do they do so forever? What about particulars, if we happen to be nominalists: when do they come into existence and subsequently pass away? Are we the same person as we were yesterday? If not, what happened to that person?
Presentism and Eternalism
There are typically taken to be two theories of time, upon which differing ontologies are based. Presentism
is the view that only the present exists: the past has gone and the future is still to come. This kind of thinking is implicit in figures of speech, like saying “tomorrow has yet to pass”: it passes the now
and thereby becomes the past, the point at which it does so being the present. It asks how “the past” and “the future” can meaningfully be said to exist
as we do now.
The alternative view is called eternalism
and denies that there is anything (ontologically) special about “now”. When we talk of the present, we merely provide a reference point to help us say that one event happened before another; and so Galileo exists just as surely as we do, only within a different context.
Endurantism and Perdurantism
Typically associated with these theories of time are two theories of the persistence
of particulars. Endurantists
hold that the Hugo of yesterday is identical with the Hugo of today and hence the two Hugos are the same, persisting (or enduring) over time. Conversely, the perdurantist
believes the Hugos to be different, often talking of stages or of yesterday’s Hugo as a part of his temporal development. The consequences of these pairings of theories is significant: for the endurantist only the now
truly exists and hence talk of possible
worlds or states of affairs is just that. The perdurantist, on the other hand, grants no metaphysical privilege to any specific time and so possible worlds (of the past, the future or our imagination) are equally as real as the one we find ourselves in now. If there are any number of possible worlds, however, each slightly (or considerably) different to this one, what does that mean about the Hugos in them? Does it imply that each of them is real or perhaps that they are all aspects of a universal Hugo?
Since perdurantism is contrary to our commonsense view of the world, arguments against endurantism are required. They take the form of either an appeal to a four-dimensional view of existence drawn from physics, according to which existence is across time
as well as space; or the suggestion that endurantists cannot account for change
, especially when it involves the loss of a part of a whole. Perdurantists ask whether a person who has lost a leg in an accident, say, is the same person they were beforehand. If so, the suggestion is that this commits the endurantist to explaining why having legs mattered in the first place; and so on until little is left.
Whatever ontology we find tempting, the nature of time is so problematic that perhaps the search for its nature is mistaken to begin with? Even so, it is easy to see the relationship between these investigations of time, the metaphysical problems introduced above and the old Parmenides/Heraclitus dispute in Ancient Greece, which is why many philosophical questions are considered timeless
(to employ an awful pun).
Dialogue the Fifteenth
The scene: Anna and Trystyn are sat in the park, talking quietly.
It’s amazing how much hurt we do unintentionally.
Sometimes the truth takes us places we don’t want to go, I guess.
(Shaking her head…
) That’s not what I mean. We talk to one another but the translation is never quite right. We misunderstand, it gets amplified, and people get hurt.
Can it be otherwise?
Well, I wonder if there really is anything at base, grounding these things we struggle with.
I said you were wrong before – dishonest with Steven. Now I wonder what I was getting at.
I thought it was for the best at the time.
Maybe you didn’t think at all? (She sighs.
) Anyway, what is dishonest? I invoked a match between your conduct and something I called “dishonesty”, but where is it? In my head or in the world?
I don’t follow.
Suppose that it really was wrong – what you did – and not just my opinion. Where are these things? “Right conduct”, I mean – which is what some call it, I think. When I call you dishonest, is it the same dishonesty as when I charge it of someone else?
You mean a match between someone – me – and what you say about them?
I suppose so.
There’s a correspondence, they reckon, between the world and what we can say about it. Some insist that when you say “Trystyn was dishonest” you are just making a specific remark; but others that you hit on something universal – “Dishonesty” with a capital “d”, perhaps. What a person does when they’re dishonest is but a particular manifestation of a fundamental characteristic of the universe.
But where are these universals?
You want to test for them?
(There is a silence.
This is partly why they say there is no science without philosophy – or without metaphysics, really. We can say that something exists or doesn’t exist on the basis of experiment but why does experiment decide such things and what does existence mean in the first place?
Does dishonesty exist? That’s what I’m asking.
Sure, but the other question is prior. If dishonesty exists, where is it? If it is just an attribute we give to certain conduct, what makes it up? What attributes does it
have? If existence is a collection of these properties, what’s left when we take them away one by one? Maybe this universal dishonesty you’re thinking of is the sum total of behaviour we would describe by the terms that make it up? And so it goes.
I don’t know the answer to these questions. Stop trying to tie me in knots.
) You asked.
So say you don’t know.
It’s not me. I got it wrong, but you don’t need an ultimate justification for saying so. If you want to hang me from that
tree then you’ll have to be prepared to see the ground fall away beneath you.
(Another long silence.
So what do you suggest?
Hold on to something.
By Paul Newall (2005)