green numbers give full details.     |    back to list of philosophers     |     unexpand these ideas

Ideas of Gilbert Harman, by Text

[American, b.1938, Professor at Princeton University.]

1970 Induction
§IV p.150 If you would deny a truth if you know the full evidence, then knowledge has social aspects
     Full Idea: If one reads of a genuine assassination, but then fails to read the reports next day which untruthfully deny the event, one probably does not know of the event. But we must conclude that knowledge has a further 'social aspect'.
     From: report of Gilbert Harman (Induction [1970], §IV) by Ernest Sosa - The Raft and the Pyramid Appx
     A reaction: I doubt if this is enough to support an externalist account of defeasibility. Wise people don't 'know' of an event after one report. For 24 hours the Royalists thought they had won Marston Moor! You know he's dead when you see the Zapruder film.
1973 Thought
p.81 In negative coherence theories, beliefs are prima facie justified, and don't need initial reasons
     Full Idea: According to Harman's negative coherence theory it is always permissible to adopt a new belief - any new belief; because beliefs are prima facie justified you do not need a reason for adopting a new belief.
     From: report of Gilbert Harman (Thought [1973]) by J Pollock / J Cruz - Contemporary theories of Knowledge (2nd) §3.4.1
     A reaction: This must be placed alongside the fact that we don't usually choose our beliefs, but simply find ourselves believing because of the causal impact of evidence. This gives an unstated rational justification for any belief - something caused it.
Pref p.-4 We see ourselves in the world as a map
     Full Idea: Our conception of ourselves in the world is more like a map than a story.
     From: Gilbert Harman (Thought [1973], Pref)
     A reaction: Dennett offer the 'story' view of the self (Ideas 7381 and 7382). How do we arbitrate this one? A story IS a sort of map. Maps can extend over time as well over space. I think the self is real, and is a location on a map, and the hero of a story.
2.2 p.28 People's reasons for belief are rarely conscious
     Full Idea: The reasons for which people believe things are rarely conscious.
     From: Gilbert Harman (Thought [1973], 2.2)
     A reaction: Probably correct. The interesting bit is when they bring the beliefs into consciousness and scrutinise them rationally. Philosophers routinely overthrow their natural beliefs in this way.
3.2 p.38 Could a cloud have a headache if its particles formed into the right pattern?
     Full Idea: If the right pattern of electrical discharges occurred in a cloud instead of in a brain, would that also be a headache?
     From: Gilbert Harman (Thought [1973], 3.2)
     A reaction: The standard objection to functionalism is to propose absurd implementations of a mind, but probably only a brain could produce the right electro-chemical combination.
3.3 p.41 Defining dispositions is circular
     Full Idea: There is no noncircular way to specify dispositions; for they are dispositions to behave given certain situations, and the situations must be include beliefs about the situation, and desires concerning it.
     From: Gilbert Harman (Thought [1973], 3.3)
     A reaction: This is nowadays accepted dogmatically as the biggest objection to behaviourism, but it could be challenged. Your analysis may begin by mentioning beliefs and desires, but if you keep going they may eventually fade out of the picture.
3.6 p.47 Reasoning might be defined in terms of its functional role, which is to produce knowledge
     Full Idea: Reasoning could be treated as a functionally defined process that is partly defined in terms of its role in giving a person knowledge.
     From: Gilbert Harman (Thought [1973], 3.6)
4.3 p.61 Speech acts, communication, representation and truth form a single theory
     Full Idea: The various theories are not in competition. The theory of truth is part of the theory of representational character, which is presupposed by the theory of communication, which in turn is contained in the more general theory of speech acts.
     From: Gilbert Harman (Thought [1973], 4.3)
     A reaction: Certainly it seems that the supposed major contenders for a theory of meaning are just as much complements as they are competitors.
5.1 p.71 Truth in a language is explained by how the structural elements of a sentence contribute to its truth conditions
     Full Idea: A theory of truth for a language shows how the truth conditions of any sentence depend on the structure of that sentence. The theory will say, for each element of structure, what its contribution is.
     From: Gilbert Harman (Thought [1973], 5.1)
     A reaction: This just seems to push the problem of truth back a stage, as you need to know where the truth is to be found in the elements from which the structure is built.
5.2 p.75 Logical form is the part of a sentence structure which involves logical elements
     Full Idea: The logical form of a sentence is that part of its structure that involves logical elements.
     From: Gilbert Harman (Thought [1973], 5.2)
5.2 p.76 A theory of truth in a language must involve a theory of logical form
     Full Idea: Some sort of theory of logical form is involved in any theory of truth for a natural language.
     From: Gilbert Harman (Thought [1973], 5.2)
5.3 p.78 Ambiguity is when different underlying truth-conditional structures have the same surface form
     Full Idea: Ambiguity results from the possibility of transforming different underlying truth-conditional structures into the same surface form.
     From: Gilbert Harman (Thought [1973], 5.3)
     A reaction: Personally I would call a 'truth-conditional structure' a 'proposition', and leave it to the philosophers to decide what a proposition is.
5.4 p.81 Many predicates totally resist translation, so a universal underlying structure to languages is unlikely
     Full Idea: There are many predicates of a given language that resist translation into another language, …so it is unlikely that there is a basic set of underlying structures common to all languages.
     From: Gilbert Harman (Thought [1973], 5.4)
     A reaction: Not convincing. 'Structures' are not the same as 'predicates'. Once a language has mapped its predicates, that blocks the intrusions of differently sliced alien predicates. No gaps.
6.3 p.92 Our underlying predicates represent words in the language, not universal concepts
     Full Idea: The underlying truth-conditional structures of thoughts are language-dependent in the sense that underlying predicates represent words in the language rather than universal concepts common to all languages.
     From: Gilbert Harman (Thought [1973], 6.3)
6.4 p.94 Sentences are different from propositions, since two sentences can express one proposition
     Full Idea: 'Bob and John play golf' and 'John and Bob play golf' are equivalent; but if they were to be derived from the same underlying structure, one or the other of Bob and John would have to come first; and either possibility is arbitrary.
     From: Gilbert Harman (Thought [1973], 6.4)
     A reaction: If I watch Bob and John play golf, neither of them 'comes first'. A proposition about them need not involve 'coming first'. Only if you insist on formulating a sentence must you decide on that.
6.5 p.97 Are there any meanings apart from in a language?
     Full Idea: The theory of language-independent meanings or semantic representations is mistaken.
     From: Gilbert Harman (Thought [1973], 6.5)
     A reaction: This would make him (in Dummett's terms) a 'philosopher of language' rather than a 'philosopher of thought'. Personally I disagree. Don't animals have 'meanings'? Can two sentences share a meaning?
6.5 p.98 The analytic/synthetic distinction is a silly division of thought into encyclopaedia and dictionary
     Full Idea: No purpose is served by thinking that certain principles available to a person are contained in his internal encyclopaedia - and therefore only synthetic - whereas other principles are part of his internal dictionary - and are therefore analytic.
     From: Gilbert Harman (Thought [1973], 6.5)
     A reaction: If it led to two different ways to acquire knowledge, then quite a lot of purpose would be served. He speaks like a pragmatist. The question is whether some statements just are true because of some feature of meaning. Why not?
6.7 p.104 Analyticity is postulated because we can't imagine some things being true, but we may just lack imagination
     Full Idea: Analyticity is postulated to explain why we cannot imagine certain things being true. A better postulate is that we are not good at imagining things.
     From: Gilbert Harman (Thought [1973], 6.7)
6.7 p.105 Only lack of imagination makes us think that 'cats are animals' is analytic
     Full Idea: That 'cats are animals' is often cited as an analytic truth. But (as Putnam points out) the inability to imagine this false is just a lack of imagination. They might turn out to be radio-controlled plastic spies from Mars.
     From: Gilbert Harman (Thought [1973], 6.7)
6.8 p.109 There is only similarity in meaning, never sameness in meaning
     Full Idea: The only sort of sameness of meaning we know is similarity in meaning, not exact sameness of meaning.
     From: Gilbert Harman (Thought [1973], 6.8)
     A reaction: The Eiffel Tower and le tour Eiffel? If you want to be difficult, you can doubt whether the word 'fast' ever has exactly the same meaning in two separate usages of the word.
7.2 p.119 If you believe that some of your beliefs are false, then at least one of your beliefs IS false
     Full Idea: If a rational man believes he has at least some other false beliefs, it follows that a rational man knows that at least one of his beliefs is false (the one believed false, or this new belief).
     From: Gilbert Harman (Thought [1973], 7.2)
8.1 p.127 Any two states are logically linked, by being entailed by their conjunction
     Full Idea: Any two states of affairs are logically connected, simply because both are entailed by their conjunction.
     From: Gilbert Harman (Thought [1973], 8.1)
10.1 p.157 You don't have to accept the conclusion of a valid argument
     Full Idea: We may say "From P and If-P-then-Q, infer Q" (modus ponens), but there is no rule of acceptance to say that we should accept Q. Maybe we should stop believing P or If-P-then-Q rather than believe Q.
     From: Gilbert Harman (Thought [1973], 10.1)
10.2 p.159 Induction is an attempt to increase the coherence of our explanations
     Full Idea: Induction is an attempt to increase the explanatory coherence of our view, making it more complete, less ad hoc, more plausible.
     From: Gilbert Harman (Thought [1973], 10.2)
10.4 p.164 Coherence avoids scepticism, because it doesn't rely on unprovable foundations
     Full Idea: Scepticism is undermined once it is seen that the relevant kind of justification is not a matter of derivation from basic principles but is rather a matter of showing that a view fits in well with other things we believe.
     From: Gilbert Harman (Thought [1973], 10.4)
     A reaction: I would (now) call myself a 'coherentist' about justification, and I agree with this. Coherent justification could not possibly deliver certainty, so it must be combined with fallibilism.
10.4 p.166 We don't distinguish between accepting, and accepting as evidence
     Full Idea: There is no distinction between what we accept as evidence and whatever else we accept.
     From: Gilbert Harman (Thought [1973], 10.4)
10.4 p.168 Deductive logic is the only logic there is
     Full Idea: Deductive logic is the only logic there is.
     From: Gilbert Harman (Thought [1973], 10.4)
11.2 p.179 Inference is never a conscious process
     Full Idea: Inference is never a conscious process.
     From: Gilbert Harman (Thought [1973], 11.2)
12.1 p.189 You have to reaffirm all your beliefs when you make a logical inference
     Full Idea: Since inference is inference to the best total account, all your prior beliefs are relevant and your conclusion is everything you believe at the end. So, you constantly reaffirm your beliefs in inference.
     From: Gilbert Harman (Thought [1973], 12.1)
12.1 p.190 Memories are not just preserved, they are constantly reinferred
     Full Idea: I favour the inferential view of memory over the preservation view. …One constantly reinfers old beliefs.
     From: Gilbert Harman (Thought [1973], 12.1)
     A reaction: This has a grain of truth, but seems a distortion. An image of the old home floats into my mind when I am thinking about something utterly unconnected. When we search memory we may be inferring and explaining, but the same applies to searching images.
1974 The Inference to the Best Explanation
p.1 Best Explanation is the core notion of epistemology
     Full Idea: Gilbert Harman introduced the term 'inference to the best explanation', and argued that it is the core notion of epistemology.
     From: report of Gilbert Harman (The Inference to the Best Explanation [1974]) by J.J.C. Smart - Explanation - Opening Address p. 01
     A reaction: Hard to assess that, but it sounds right. I'm a fan of coherence theories of justification, and also coherence theories of explanation, and there is a neat package there somewhere.
1983 Human Flourishing, Ethics and Liberty
9.2.1 p.156 What counts as 'flourishing' must be relative to various sets of values
     Full Idea: If we base our ethics on human flourishing, one implication would seem to be moral relativism, since what counts as 'flourishing' seems inevitably relative to one or other set of values.
     From: Gilbert Harman (Human Flourishing, Ethics and Liberty [1983], 9.2.1)
     A reaction: This remark seems to make the relativist assumption that all value systems are equal. For Aristotle, flourishing is no more relative than health is. No one can assert that illness has an intrinsically high value in human life.
9.2.2 p.156 Basing ethics on flourishing makes it consequentialist, as actions are judged by contributing to it
     Full Idea: Basing ethics on human flourishing tends towards utilitarianism or consequentialism; actions, character traits, laws, and so on are to be assessed with reference to their contributions to human flourishing.
     From: Gilbert Harman (Human Flourishing, Ethics and Liberty [1983], 9.2.2)
     A reaction: This raises the question of whether only virtue can contribute to flourishing, or whether a bit of vice might be helpful. This problem presumably pushed the Stoics to say that virtue itself is the good, rather than the resulting flourishing.
1986 Change in View: Principles of Reasoning
1 p.4 Implication just accumulates conclusions, but inference may also revise our views
     Full Idea: Implication is cumulative, in a way that inference may not be. In argument one accumulates conclusions; things are always added, never subtracted. Reasoned revision, however, can subtract from one's view as well as add.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 1)
     A reaction: This has caught Harman's attention, I think (?), because he is looking for non-monotonic reasoning (i.e. revisable reasoning) within a classical framework. If revision is responding to evidence, the logic can remain conventional.
1 p.5 The rules of reasoning are not the rules of logic
     Full Idea: Rules of deduction are rules of deductive argument; they are not rules of inference or reasoning.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 1)
     A reaction: And I have often noticed that good philosophing reasoners and good logicians are frequently not the same people.
1 p.8 The Gambler's Fallacy (ten blacks, so red is due) overemphasises the early part of a sequence
     Full Idea: The Gambler's Fallacy says if black has come up ten times in a row, red must be highly probable next time. It overlooks how the impact of an initial run of one color can become more and more insignificant as the sequence gets longer.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 1)
     A reaction: At what point do you decide that the roulette wheel is fixed, rather than that you have fallen for the Gambler's Fallacy? Interestingly, standard induction points to the opposite conclusion. But then you have prior knowledge of the wheel.
2 p.12 It is a principle of reasoning not to clutter your mind with trivialities
     Full Idea: I am assuming the following principle: Clutter Avoidance - in reasoning, one should not clutter one's mind with trivialities.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 2)
     A reaction: I like Harman's interest in the psychology of reasoning. In the world of Frege, it is taboo to talk about psychology.
2 p.15 If there is a great cost to avoiding inconsistency, we learn to reason our way around it
     Full Idea: We sometimes discover our views are inconsistent and do not know how to revise them in order to avoid inconsistency without great cost. The best response may be to keep the inconsistency and try to avoid inferences that exploit it.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 2)
     A reaction: Any decent philosopher should face this dilemma regularly. I assume non-philosophers don't compare the different compartments of their beliefs very much. Students of non-monotonic logics are trying to formalise such thinking.
2 p.19 We strongly desire to believe what is true, even though logic does not require it
     Full Idea: Moore's Paradox: one is strongly disposed not to believe both P and that one does not believe that P, while realising that these propositions are perfectly consistent with one another.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 2)
     A reaction: [Where in Moore?] A very nice example of a powerful principle of reasoning which can never be captured in logic.
2 p.20 Logic has little relevance to reasoning, except when logical conclusions are immediate
     Full Idea: Although logic does not seem specially relevant to reasoning, immediate implication and immediate inconsistency do seem important for reasoning.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 2)
     A reaction: Ordinary thinkers can't possibly track complex logical implications, so we have obviously developed strategies for coping. I assume formal logic is contructed from the basic ingredients of the immediate and obvious implications, such as modus ponens.
3 p.23 High probability premises need not imply high probability conclusions
     Full Idea: Propositions that are individually highly probable can have an immediate implication that is not. The fact that one can assign a high probability to P and also to 'if P then Q' is not sufficient reason to assign high probability to Q.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 3)
     A reaction: He cites Kyburg's Lottery Paradox. It is probable that there is a winning ticket, and that this ticket is not it. Thus it is NOT probable that I will win.
4 p.29 In revision of belief, we need to keep track of justifications for foundations, but not for coherence
     Full Idea: The key issue in belief revision is whether one needs to keep track of one's original justifications for beliefs. What I am calling the 'foundations' theory says yes; what I am calling the 'coherence' theory says no.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 4)
     A reaction: I favour coherence in all things epistemological, and this idea seems to match real life, where I am very confident of many beliefs of which I have forgotten the justification. Harman says coherentists need the justification only when they doubt a belief.
7 p.65 Coherence is intelligible connections, especially one element explaining another
     Full Idea: Coherence in a view consists in connections of intelligibility among the elements of the view. Among other things these included explanatory connections, which hold when part of one's view makes it intelligible why some other part should be true.
     From: Gilbert Harman (Change in View: Principles of Reasoning [1986], 7)
     A reaction: Music to my ears. I call myself an 'explanatory empiricist', and embrace a coherence theory of justification. This is the framework within which philosophy should be practised. Harman is our founder, and Paul Thagard our guru.
1987 (Nonsolipsistic) Conceptual Role Semantics
12.1 p.206 Meaning from use of thoughts, constructed from concepts, which have a role relating to reality
     Full Idea: Conceptual role semantics involves meanings of expressions determined by used contents of concepts and thoughts, contents constructed from concepts, concepts determined by functional role, which involves relations to things in the world.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1)
     A reaction: This essay is the locus classicus for conceptual-role semantics. Any attempt to say what something IS by giving an account of its function always feels wrong to me.
12.1 p.206 Some regard conceptual role semantics as an entirely internal matter
     Full Idea: I call my conceptual role semantics 'non-solipsistic' to contrast it with that of authors (Field, Fodor, Loar) who think of conceptual role solipsistically as a completely internal matter.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1)
     A reaction: Evidently Harman is influenced by Putnam's Twin Earth, and that meanings ain't in the head, so that the conceptual role has to be extended out into the world to get a good account. I prefer extending into the language community, rather into reality.
12.1.1 p.207 Take meaning to be use in calculation with concepts, rather than in communication
     Full Idea: (Nonsolipsistic) conceptual role semantics is a version of the theory that meaning is use, where the basic use is taken to be in calculation, not in communication, and where concepts are treated as symbols in a 'language of thought'.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1.1)
     A reaction: The idea seems to be to connect the highly social Wittgensteinian view of language with the reductive physicalist account of how brains generate concepts. Interesting, thought I never like meaning-as-use.
12.1.2 p.208 Concepts in thought have content, but not meaning, which requires communication
     Full Idea: Concepts and other aspects of mental representation have content but not (normally) meaning (unless they are also expressions in a language used in communication).
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1.2)
     A reaction: Given his account of meaning as involving some complex 'role', he has to say this, though it seems a dubious distinction, going against the grain of a normal request to ask what some concept 'means'. What is 'democracy'?
12.1.2 p.208 Mastery of a language requires thinking, and not just communication
     Full Idea: If one cannot think in a language, one has not yet mastered it. A symbol system used only for communication, like Morse code, is not a language.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1.2)
     A reaction: This invites the question of someone who has mastered thinking, but has no idea how to communicate. No doubt we might construct a machine with something like that ability. I think it might support Harman's claim.
12.1.3 p.3 The use theory attaches meanings to words, not to sentences
     Full Idea: A use theory of meaning has to suppose it is words and ways of putting words together that have meaning because of their uses, not sentences.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1.3)
     A reaction: He says that most sentences are unique, so cannot have a standard use. Words do a particular job over and over again. How do you distinguish the quirky use of a word from its standard use?
12.1.4 p.210 If one proposition negates the other, which is the negative one?
     Full Idea: A relation of negation might hold between two beliefs without there being anything that determines which belief is the negative one.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.1.4)
     A reaction: [He attributes this thought to Brian Loar] This seems to give us a reason why we need a semantics for a logic, and not just a structure of inferences and proofs.
12.2.2 p.212 Reasoning aims at increasing explanatory coherence
     Full Idea: In reasoning you try among other things to increase the explanatory coherence of your view.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.2.2)
     A reaction: Harman is a champion of inference to the best explanation (abduction), and I agree with him. I think this idea extends to give us a view of justification as coherence, and that extends from inner individual coherence to socially extended coherence.
12.2.2 p.212 We have a theory of logic (implication and inconsistency), but not of inference or reasoning
     Full Idea: There is as yet no substantial theory of inference or reasoning. To be sure, logic is well developed; but logic is not a theory of inference or reasoning. Logic is a theory of implication and inconsistency.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.2.2)
     A reaction: One problem is that animals can draw inferences without the use of language, and I presume we do so all the time, so it is hard to see how to formalise such an activity.
12.2.2 p.213 I might accept P and Q as likely, but reject P-and-Q as unlikely
     Full Idea: Principles of implication imply there is not a purely probabilistic rule of acceptance for belief. Otherwise one might accept P and Q, without accepting their conjunction, if the conjuncts have a high probability, but the conjunction doesn't.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.2.2)
     A reaction: [Idea from Scott Soames] I am told that my friend A has just won a very big lottery prize, and am then told that my friend B has also won a very big lottery prize. The conjunction seems less believable; I begin to suspect a conspiracy.
12.2.4 p.215 Reality is the overlap of true complete theories
     Full Idea: Reality is what is invariant among true complete theories.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.2.4)
     A reaction: The sort of slogan that gets coined in the age of Quine. The whole manner of starting from your theories and working out to what we think reality is seems to be putting the cart before the horse.
12.2.6 p.217 Reason conservatively: stick to your beliefs, and prefer reasoning that preserves most of them
     Full Idea: Conservatism is important; you should continue to believe as you do in the absence of any special reason to doubt your view, and in reasoning you should try to minimize change in your initial opinions in attaining other goals of reasoning.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.2.6)
     A reaction: One of those principles like Ockham's Razor, which feels right but hard to justify. It seems the wrong principle for someone who can reason well, but has been brainwashed into a large collection of daft beliefs. Japanese soldiers still fighting WWII.
12.3.3 p.221 The content of thought is relations, between mental states, things in the world, and contexts
     Full Idea: In (nonsolipsistic) conceptual role semantics the content of thought is not in an 'intrinsic nature', but is rather a matter of how mental states are related to each other, to things in the external world, and to things in a context understood as normal.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.3.3)
     A reaction: This is part of Harman's functional view of consciousness, which I find rather dubious. If things only have identity because of some place in a flow diagram, we must ask why that thing has that place in that diagram.
12.3.3 p.223 The way things look is a relational matter, not an intrinsic matter
     Full Idea: According to functionalism, the way things look to you is a relational characteristic of your experience, not part of its intrinsic character.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.3.3)
     A reaction: No, can't make sense of that. How would being in a relation determine what something is? Similar problems with the structuralist account of mathematics. If the whole family love some one cat or one dog, the only difference is intrinsic to the animal.
12.3.4 p.224 There is no natural border between inner and outer
     Full Idea: There is no natural border between inner and outer.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.3.4)
     A reaction: Perhaps this is the key idea for the anti-individualist view of mind. Subjectively I would have to accept this idea, but looking objectively at another person it seems self-evident nonsense.
12.3.4 p.226 We can only describe mental attitudes in relation to the external world
     Full Idea: No one has ever described a way of explaining what beliefs, desires, and other mental states are except in terms of actual or possible relations to things in the external world.
     From: Gilbert Harman ((Nonsolipsistic) Conceptual Role Semantics [1987], 12.3.4)
     A reaction: If I pursue my current favourite idea, that how we explain things is the driving force in what ontology we adopt, then this way of seeing the mind, and taking an externalist anti-individualist view of it seems quite attractive.
1990 The Intrinsic Quality of Experience
p.459 Qualities of experience are just representational aspects of experience ('Representationalism')
     Full Idea: Harman defended what came to be known as 'representationalism' - the view that qualitative aspects of experience are nothing other than representational aspects.
     From: report of Gilbert Harman (The Intrinsic Quality of Experience [1990]) by Tyler Burge - Philosophy of Mind: 1950-2000 p.459
     A reaction: Functionalists like Harman have a fairly intractable problem with the qualities of experience, and this may be clutching at straws. What does 'represent' mean? How is the representation achieved? Why that particular quale?
1995 Rationality
1.2 p.22 You can be rational with undetected or minor inconsistencies
     Full Idea: Rationality doesn't require consistency, because you can be rational despite undetected inconsistencies in beliefs, and it isn't always rational to respond to a discovery of inconsistency by dropping everything in favour of eliminating that inconsistency.
     From: Gilbert Harman (Rationality [1995], 1.2)
     A reaction: This strikes me as being correct, and is (I am beginning to realise) a vital contribution made to our understanding by pragmatism. European thinking has been too keen on logic as the model of good reasoning.
1.3 p.23 Ordinary rationality is conservative, starting from where your beliefs currently are
     Full Idea: Ordinary rationality is generally conservative, in the sense that you start from where you are, with your present beliefs and intentions.
     From: Gilbert Harman (Rationality [1995], 1.3)
     A reaction: This stands opposed to the Cartesian or philosophers' rationality, which requires that (where possible) everything be proved from scratch. Harman seems right, that the normal onus of proof is on changing beliefs, rather proving you should retain them.
1.4.5 p.31 Induction is 'defeasible', since additional information can invalidate it
     Full Idea: It is sometimes said that inductive reasoning is 'defeasible', meaning that considerations that support a given conclusion can be defeated by additional information.
     From: Gilbert Harman (Rationality [1995], 1.4.5)
     A reaction: True. The point is that being defeasible does not prevent such thinking from being rational. The rational part of it is to acknowledge that your conclusion is defeasible.
1.4.5 p.32 All reasoning is inductive, and deduction only concerns implication
     Full Idea: Deductive logic is concerned with deductive implication, not deductive reasoning; all reasoning is inductive
     From: Gilbert Harman (Rationality [1995], 1.4.5)
     A reaction: This may be an attempt to stipulate how the word 'reasoning' should be used in future. It is, though, a bold and interesting claim, given the reputation of induction (since Hume) of being a totally irrational process.
1.5.2 p.33 A coherent conceptual scheme contains best explanations of most of your beliefs
     Full Idea: A set of unrelated beliefs seems less coherent than a tightly organized conceptual scheme that contains explanatory principles that make sense of most of your beliefs; this is why inference to the best explanation is an attractive pattern of inference.
     From: Gilbert Harman (Rationality [1995], 1.5.2)
     A reaction: I find this a very appealing proposal. The central aim of rational thought seems to me to be best explanation, and I increasingly think that most of my beliefs rest on their apparent coherence, rather than their foundations.
1.5.2 p.34 Enumerative induction is inference to the best explanation
     Full Idea: We might think of enumerative induction as inference to the best explanation, taking the generalization to explain its instances.
     From: Gilbert Harman (Rationality [1995], 1.5.2)
     A reaction: This is a helpful connection. The best explanation of these swans being white is that all swans are white; it ceased to be the best explanation when black swans turned up. In the ultimate case, a law of nature is the explanation.
1999 Moral Philosophy meets social psychology
10.1 p.165 Maybe consequentialism is a critique of ordinary morality, rather than describing it
     Full Idea: Consequentialism may be put forward not as an attempt to capture intuitive folk morality but rather as a critique of ordinary tuitions.
     From: Gilbert Harman (Moral Philosophy meets social psychology [1999], 10.1)
     A reaction: It is certainly true that most people are concerned with why an action was performed, and (after initial anger) are prepared to forgive an unintended disaster. We have no moral objections to earthquakes, which have bad consequences.
10.1 p.165 Maybe there is no such thing as character, and the virtues and vices said to accompany it
     Full Idea: It may be the case that there is no such thing as character, no ordinary character traits of the sort people think there are, none of the usual moral virtues and vices.
     From: Gilbert Harman (Moral Philosophy meets social psychology [1999], 10.1)
     A reaction: This would be a devastating fact for virtue theory, if it were true. I don't believe it. He thinks patterns of behaviour result from circumstances, but we give accurate and detailed pictures of people's characters (esp. in novels).
10.2 p.167 If a person's two acts of timidity have different explanations, they are not one character trait
     Full Idea: If Herbert is disposed to not speak in history class (but not other subjects), and explanation of this is different from his avoidance of roller coaster rides, then these two dispositions are not special cases of a single character trait.
     From: Gilbert Harman (Moral Philosophy meets social psychology [1999], 10.2)
     A reaction: A basic Harman argument for denying the existence of character (and hence of virtues). I just say that character traits are more complex than his caricature of them. If I keep imagining disaster and humiliation for myself, that is a character trait. p.176 Virtue ethics might involve judgements about the virtues of actions, rather than character
     Full Idea: There are variants of virtue ethics that do not require character traits in the ordinary sense. For example, moral thinking might be explicated by appeal to judgements about whether particular actions are just or courageous or whatever.
     From: Gilbert Harman (Moral Philosophy meets social psychology [1999],
     A reaction: A very interesting proposal (from Judith Jarvis Thomson). This would flatly reject Aristotle, and one presumes that the judgement about the virtue of the action would largely be a matter of pondering cultural conventions (or, perhaps, consequences).