29 ideas
| 4036 | What matters is not how many entities we postulate, but how many kinds of entities [Armstrong, by Mellor/Oliver] |
| Full Idea: Armstrong argues that what matters is not how few entities we postulate (quantitative economy), but how few kinds of entities (qualitative economy). | |
| From: report of David M. Armstrong (Properties [1992]) by DH Mellor / A Oliver - Introduction to 'Properties' §9 | |
| A reaction: Is this what Ockham meant? Armstrong is claiming that the notion of a 'property' is needed to identify kinds. See also Idea 7038. |
| 15754 | Without properties we would be unable to express the laws of nature [Armstrong] |
| Full Idea: The ontological correlates of true law-statements must involve properties. How else can one pick our the uniformities which the law-statements entail? | |
| From: David M. Armstrong (Properties [1992], 1) | |
| A reaction: I'm unconvinced about the 'laws', but I have to admit that it is hard to know how to describe the relevant bits of nature without some family of concepts covered by the word 'property'. I'm in favour of taking some of the family into care, though. |
| 4034 | Whether we apply 'cold' or 'hot' to an object is quite separate from its change of temperature [Armstrong] |
| Full Idea: Evading properties by means of predicates is implausible when things change. If a cold thing becomes hot, first 'cold' applies, and then 'hot', but what have predicates to do with the temperature of an object? | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: A clear illustration of why properties are part of nature, not just part of language. But some applications of predicates are more arbitrary than this (ugly, cool) |
| 8535 | To the claim that every predicate has a property, start by eliminating failure of application of predicate [Armstrong] |
| Full Idea: Upholders of properties have been inclined to postulate a distinct property corresponding to each distinct predicate. We could start by eliminating all those properties where the predicate fails to apply, is not true, of anything. | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: This would leave billions of conjunctional, disjunctional and gerrymandered properties where the predicate applies very well. We are all 'on the same planet as New York'. Am I allowed to say that I 'wish' that a was F? He aims for 'sparse' properties. |
| 8537 | Tropes fall into classes, because exact similarity is symmetrical and transitive [Armstrong] |
| Full Idea: Exact similarity is a symmetrical and transitive relation. (Less than exact similarity is not transitive, even for tropes). So the relation of exact similarity is an equivalence relation, partitioning the field of tropes into equivalence classes. | |
| From: David M. Armstrong (Properties [1992], §2) | |
| A reaction: Armstrong goes on the explore the difficulties for trope theory of less than exact similarity, which is a very good line of discussion. Unfortunately it is a huge problem for everyone, apart from the austere nominalist. |
| 8538 | Trope theory needs extra commitments, to symmetry and non-transitivity, unless resemblance is exact [Armstrong] |
| Full Idea: Trope theory needs extra ontological baggage, the Axioms of Resemblance. There is a principle of symmetry, and there is the failure of transitivity - except in the special case of exact resemblance. | |
| From: David M. Armstrong (Properties [1992], §2) | |
| A reaction: [see text for fuller detail] Is it appropriate to describe such axioms as 'ontological' baggage? Interesting, though I suspect that any account of properties and predicates will have a similar baggage of commitments. |
| 8539 | Universals are required to give a satisfactory account of the laws of nature [Armstrong] |
| Full Idea: A reason why I reject trope theory is that universals are required to give a satisfactory account of the laws of nature. | |
| From: David M. Armstrong (Properties [1992], §2) | |
| A reaction: This is the key thought in Armstrong's defence of universals. Issues about universals may well be decided on such large playing fields. I think he is probably wrong, and I will gradually explain why. Watch this space as the story unfolds... |
| 8529 | Deniers of properties and relations rely on either predicates or on classes [Armstrong] |
| Full Idea: The great deniers of properties and relations are of two sorts: those who put their faith in predicates and those who appeal to sets (classes). | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: This ignores the Quine view, which is strictly for ostriches. Put like this, properties and relations seem undeniable. Predicates are too numerous (gerrymandering) or too few (colour shades). Classes can have arbitrary members. |
| 8532 | Resemblances must be in certain 'respects', and they seem awfully like properties [Armstrong] |
| Full Idea: If a resembles b, in general, they resemble in certain respects, and fail to resemble in other respects. But respects are uncomfortably close to properties, which the Resemblance theory proposes to do without. | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: This is a good objection. I think it is plausible to build a metaphysics around the idea of respects, and drop properties. Shall we just talk of 'respects' for categorising, and 'powers' for causation and explanation? Respects only exist in comparisons. |
| 8530 | Change of temperature in objects is quite independent of the predicates 'hot' and 'cold' [Armstrong] |
| Full Idea: To appreciate the implausibility of the predicate view, consider where a thing's properties change. 'Hot' becomes applicable when 'cold' ceases to, ..but the change in the object would have occurred if the predicates had never existed. | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: They keep involving secondary qualities! Armstrong is taking a strongly realist view (fine by me), but anti-realists can ignore his argument. I take predicate nominalism to be a non-starter. |
| 8536 | We want to know what constituents of objects are grounds for the application of predicates [Armstrong] |
| Full Idea: The properties that are of ontological interest are those constituents of objects, of particulars, which serve as the ground in the objects for the application of predicates. | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: Good. This is a reversal of the predicate nominalist approach, and is a much healthier attitude to the relationship between ontology and language. Value judgements will be an interesting case. Does this allow us to invent new predicates? |
| 8531 | In most sets there is no property common to all the members [Armstrong] |
| Full Idea: Most sets are uninteresting because they are utterly heterogeneous, that is, the members have nothing in common. For most sets there is no common property F, such that the set is the set of all the Fs. | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: One might link the interesting sets together by resemblance, without invoking the actual existence of an item F which all the members carry (like freemasons' briefcases). Personally I am only really interested in 'natural' sets. |
| 15753 | Essences might support Resemblance Nominalism, but they are too coarse and ill-defined [Armstrong] |
| Full Idea: A sophisticated Resemblance theory can appeal to the natures of the resembling things, from which the resemblances flow. The natures are suitably internal, but are as coarse as the things themselves (and perhaps are the things themselves). | |
| From: David M. Armstrong (Properties [1992], 1) | |
| A reaction: Note that this is essentialism as an underpinning for Resemblance Nominalism. His objection is that he just can't believe in essences, because they are too 'coarse' - which I take to mean that we cannot distinguish the boundaries of an essence. |
| 19527 | We don't acquire evidence and then derive some knowledge, because evidence IS knowledge [Williamson] |
| Full Idea: When we acquire new evidence in perception, we do not first acquire unknown evidence and then somehow base knowledge on it later. Rather, acquiring new is evidence IS acquiring new knowledge. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.4) | |
| A reaction: This makes his point much better than Idea 19526 does. |
| 19528 | Knowledge is prior to believing, just as doing is prior to trying to do [Williamson] |
| Full Idea: Knowing corresponds to doing, believing to trying. Just as trying is naturally understood in relation to doing, so believing is naturally understood in relation to knowing. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.4) | |
| A reaction: An interesting analogy. You might infer that there can be no concept of 'belief' without the concept of 'knowledge', but we could say that it is 'truth' which is indispensible, and leave out knowledge entirely. Belief is to truth as trying is to doing? |
| 19529 | Belief explains justification, and knowledge explains belief, so knowledge explains justification [Williamson] |
| Full Idea: If justification is the fundamental epistemic norm of belief, and a belief ought to constitute knowledge, then justification should be understood in terms of knowledge too. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.5) | |
| A reaction: If we are looking for the primitive norm which motivates the whole epistemic game, then I am thinking that truth might well play that role better than knowledge. TW would have to reply that it is the 'grasped truth', rather than the 'theoretical truth'. |
| 19530 | A neutral state of experience, between error and knowledge, is not basic; the successful state is basic [Williamson] |
| Full Idea: A neutral state covering both perceiving and misperceiving (or remembering and misrembering) is not somehow more basic than perceiving, for what unifies the case of each neutral state is their relation to the successful state. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.5-6) | |
| A reaction: An alternative is Disjunctivism, which denies the existence of a single neutral state, so that there is nothing to unite the two states, and they don't have a dependence relation. Why can't there be a prior family of appearances, some of them successful? |
| 19531 | Internalism about mind is an obsolete view, and knowledge-first epistemology develops externalism [Williamson] |
| Full Idea: A postulated underlying layer of narrow mental states is a myth, whose plausibility derives from a comfortingly familiar but obsolescent philosophy of mind. Knowledge-first epistemology is a further step in the development of externalism. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.6) | |
| A reaction: Williamson is a real bruiser, isn't he? I don't take internalism about mind to be obsolescent at all, but now I feel so inferior for clinging to such an 'obsolescent' belief. ...But then I cling to Aristotle, who is (no doubt) an obsolete philosopher. |
| 19536 | Knowledge-first says your total evidence IS your knowledge [Williamson] |
| Full Idea: Knowledge-first equate one's total evidence with one's total knowledge. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.8) | |
| A reaction: Couldn't lots of evidence which merely had a high probability be combined together to give a state we would call 'knowledge'? Many dubious witnesses confirm the truth, as long as they are independent, and agree. |
| 19526 | Surely I am acquainted with physical objects, not with appearances? [Williamson] |
| Full Idea: When I ask myself what I am acquainted with, the physical objects in front of me are far more natural candidates than their appearances. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.3) | |
| A reaction: Not very impressive. The word 'acquainted' means the content of the experience, not the phenomena. Do I 'experience' the objects, or the appearances? The answer there is less obvious. If you apply it to colours, it is even less obvious. |
| 22200 | If you eliminate the impossible, the truth will remain, even if it is weird [Conan Doyle] |
| Full Idea: When you have eliminated the impossible, whatever remains, however improbable, must be the truth. | |
| From: Arthur Conan Doyle (The Sign of Four [1890], Ch. 6) | |
| A reaction: A beautiful statement, by Sherlock Holmes, of Eliminative Induction. It is obviously not true, of course. Many options may still face you after you have eliminated what is actually impossible. |
| 19534 | How does inferentialism distinguish the patterns of inference that are essential to meaning? [Williamson] |
| Full Idea: Inferentialism faces the grave problem of separating patterns of inference that are to count as essential to the meaning of an expression from those that will count as accidental (a form of the analytic/synthetic distinction). | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.6) | |
| A reaction: This sounds like a rather persuasive objection to inferentialism, though I don't personally take that as a huge objection to all internalist semantics. |
| 19535 | Internalist inferentialism has trouble explaining how meaning and reference relate [Williamson] |
| Full Idea: The internalist version of inferentialist semantics has particular difficulty in establishing an adequate relation between meaning and reference. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.6) | |
| A reaction: I would have thought that this was a big problem for referentialist semantics too, though evidently Williamson doesn't think so. If he is saying that the meaning is in the external world, dream on. |
| 19533 | Inferentialist semantics relies on internal inference relations, not on external references [Williamson] |
| Full Idea: On internalist inferential (or conceptual role) semantics, the inferential relations of an expression do not depend on what, if anything, it refers to, ...rather, the meaning is something like its place in a web of inferential relations. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.6) | |
| A reaction: Williamson says the competition is between externalist truth-conditional referential semantics (which he favours), and this internalist inferential semantics. He is, like, an expert, of course, but I doubt whether that is the only internalist option. |
| 8533 | Predicates need ontological correlates to ensure that they apply [Armstrong] |
| Full Idea: Must there not be something quite specific about the thing which allows, indeed ensures, that predicates like 'underneath' and 'hot' apply? The predicates require ontological correlates. | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: An interesting proposal, that in addition to making use of predicates, we should 'ensure that they apply'. Sounds verificationist. Obvious problem cases would be speculative, controversial or metaphorical predicates. "He's beneath contempt". |
| 4035 | There must be some explanation of why certain predicates are applicable to certain objects [Armstrong] |
| Full Idea: When we have said that predicates apply to objects, we have surely not said enough. The situation cries out for an explanation. Must there not be something specific about the things which allows, indeed ensures, that these predicates apply? | |
| From: David M. Armstrong (Properties [1992], §1) | |
| A reaction: A nice challenge to any philosopher who places too much emphasis on language. A random and arbitrary (nominalist?) language simply wouldn't work. Nature has joints. |
| 19532 | Truth-conditional referential semantics is externalist, referring to worldly items [Williamson] |
| Full Idea: Truth-conditional referential semantics is an externalist programme. In a context of utterance the atomic expressions of a language refer to worldly items, from which the truth-conditions of sentences are compositionally determined. | |
| From: Timothy Williamson (Knowledge First (and reply) [2014], p.6) | |
| A reaction: I just don't see how a physical object can be part of the contents of a sentence. 'Dragons fly' is atomic, and meaningful, but its reference fails. 'The cat is asleep' is just words - it doesn't contain a live animal. |
| 8541 | Regularities theories are poor on causal connections, counterfactuals and probability [Armstrong] |
| Full Idea: Regularity theories make laws molecular, with no inner causal connections; also, only some cosmic regularities are manifestations of laws; molecular states can't sustain counterfactuals; and probabilistic laws are hard to accommodate. | |
| From: David M. Armstrong (Properties [1992], §2) | |
| A reaction: [very compressed] A helpful catalogue of difficulties. The first difficulty is the biggest one - that regularity theories have nothing to say about why there is a regularity. They offer descriptions instead of explanations. |
| 8540 | The introduction of sparse properties avoids the regularity theory's problem with 'grue' [Armstrong] |
| Full Idea: Regularity theories of laws face the grue problem. That, I think, can only be got over by introducing properties, sparse properties, into one's ontology. | |
| From: David M. Armstrong (Properties [1992], §2) | |
| A reaction: The problem is, roughly, that regularities have to be described in language, which is too arbitrary in character. Armstrong rightly tries to break the rigid link to language. See his Idea 8536, which puts reality before language. |