|
|||
 |
|||
|
|
|
|
|
|||
 |
|||
|
|
|
|
Rand and Hume: An Analysis Ayn Rand had nothing but contempt for the work of David Hume. Since she treats his mature work as nonsensical, and since she also has an earned authority based upon her positive work, anyone who is convinced that Objectivism contains fundamental truth is obliged to pick through Hume carefully, to see if his system refutes itself. An analysis of his epistemological system shows that it is self-contradictory in part, and confused in one crucial plank. A confused philosophy, though, is not invalidated by that confusion: it is amenable to reconstructive repair. An application of the same reasoning used by Leonard Peikoff on Immanuel Kant’s Analytic-Synthetic Dichotomy (Rand 1990, 88-121) has to be performed. What if this analysis is performed and Hume is found only partially wanting? And: what if elimination of a crucial confusion also eliminates most of the sophistries that are commonly associated with Hume’s epistemology? And thirdly: what if cleaned-up Humeanism reveals something oddly familiar? Not Quite A False Dichotomy, But Close The first item you see when reading the central text of Hume’s An Enquiry Concerning Human Understanding (Hume 1949, xl) is the “Author’s Advertisement,” in which Hume explicitly states that his earlier works have been superseded by this one, and that it is polemical, not analytical, to confute the arguments in Human Understanding with his earlier arguments collected in A Treatise of Human Nature. So, for the purposes of this paper, I will be taking Hume at his word and discussing the arguments in Human Understanding only. Central to Human Understanding is the division of knowledge into two spheres: “Relations of Ideas” and “Matters of Fact.” (Hume 1949, 24.) Hume defines the first concept in such a way: Relations of Ideas are “intuitively or demonstratively certain.” (Hume 1949, 24) without reference to facts of reality. He uses, as two examples, “the square of the hypotenuse is equal to the squares of the two sides” and “three times five is equal to the half of thirty.” (Hume 1949, 24.) These kinds of principles are arrived at through operations of logic, and never need a reference to sense data to prove their truth. “Matters of Fact,” on the other hand, need sense data as part of the proof. In order to show the truth of the proposition “the sun will rise to-morrow,” the demonstrator has to add the propositional equivalent of an ostensive definition, because it is possible to imagine the contrary. (Hume 1949, 25.) This kind of dichotomy, like all dichotomies, should be tested to see if it is a false one, by using the same test that Leonard Peikoff used in his refutation of the Analytic-Synthetic Dichotomy (Rand 1990, 88-121): demonstrating that any knowledge can be put on both sides of the dividing line. The test will begin with the noting that Hume has slipped in a hidden assumption: certain forms of knowledge are matched to a particular sense. A phrase such as “All triangles are round” is nonsensical on the face of it if you rely upon your visual imagination to interpret it, but if you rely upon your auditory imagination, it is imaginable. Let the words in that proposition ring in your mind without referring to a corresponding picture and you will realize it. This sense experiment would seem to be a cheat, though, because if the senses are reliable, then the use of one sense modality to imagine the truth of a proposition which cannot be imagined through another sense modality means that the senses are being used to defeat the senses through a switch of modalities. This suggests a reformulation of Hume’s definition of Relation of Ideas: “Any proposition whose falsity implies a contradiction between the imaginational product of one sense modality and that of another.” This is an expansion of “intuitively...certain.” (Hume 1949, 24.) In otherwords, the existence of more than one sense is the base of the category Relation of Ideas. Let us examine this by using Hume’s own criterion. Imagine a race of near-humans with sight only: they have an entire conceptual structure based upon the sense data of sight and sight alone. One of these near-humans spells out the proposition “Triangles are not round” to another. The second near-human looks at the argument - and then returns the sign-language equivalent of “A triangle is round” with an accompanying sight-communication which translates into: “You see what I communicated? If so, you saw it.” Note that if you collapse two sense modalities into one, a proposition that us humans consider to be nonsensical on the face of it suddenly becomes conceivable - and hence, a Matter of Fact. You don’t have to even imagine a race of near-humans with one sense and one sense alone to show this: positing a group of near-humans with four senses (touch, taste, smell and sight) enables you to turn quite a few propositions rooted in geometry, a field which Hume holds to be composed of inarguable Relations of Ideas, into arguable Matters of Fact. Just collapse sight and sound into one and you’ll “see” it. In order to refute the above idea, you have to come up with a proposition that is un-imaginable using one sense modality alone - including the use of that sense for communicative purposes. As far as collapsing Matters of Fact into Relations of Ideas is concerned, just imagine another group of near-humans - only ones with extra senses that we don’t have. Example: “radiation-itch,” which allows a near-human to detect what us humans find undetectable through our sense - radiation levels. This would make a proposition which is imaginable to us, such as “friendly plutonium,” nonsensical to them. Or: a detector of electromagnetic radiation that is used as a homing pigeon’s detector of the magnetic field of the earth. Such a sense would render the proposition “the farther away you are from the radio station, the worse your life is” a fundament of a Relation of Ideas, rather than an arguable Matter of Fact, to a near-human with that kind of sense. But note that, in order to classify all Matters of Fact into Relations of Ideas, we would have to posit a near-human with an infinite number of senses. The quantity of infinity is not imaginable. So there will always be a residuum of possible Matters of Fact. The application of the other part of the test, though, shows that there will be no Relations of Ideas if man’s senses are shrunk to one. What this shows is that Hume is an empiric. Using his scheme to evaluate itself, we wind up with: it is possible to eliminate Relations of Ideas, but not Matters of Fact. Enter the Skeptic Hume’s categorization implicitly relies upon the existence of multiple senses, all of which, he posits, are reliable. What if they’re not? An example of this would be the “unreliability of taste.” I have a Vitamin D tablet in front of me as I write this - a substance which I need to keep my body healthy. I put it in my mouth and eat it. It tastes like chalk, a substance which is not beneficial for me to eat. My sense of taste has informed me that I am eating a substance that is equivalent to chalk. And yet, my memories, paler impressions derived from my visual and auditory senses, have informed me that what I have heard and seen about vitamins is plausible enough to be communicated to me in an authoritative mode by sources I have judged to be credible. I can assure myself of their credibility by remembering everything else that they have communicated to me, and by checking what impressions I have received from them against other impressions from other sources - including direct evidence from my own senses. I don’t need to take society’s word for it. Is it thinkable for the sense of taste to be unreliable? Yes; I have concluded that a substance which tastes bad is good for me. It is also possible to show the unreliability of the sense of taste - just eat a Vitamin D tablet yourself after going without any Vitamin D long enough to develop a case of osteomalacia. De gustibus non disputatum. How did we wind up in the skeptic’s trap, and how does Hume escape from it? The Nature of Man Note that I have shown that it is possible, using Hume’s own criterion of “thinkability” (imaginability), to turn all Relations of Ideas into Matters of Fact. Hume’s own demarcation criterion defeats part of Hume’s own schema. Why does the division point between Matter of Fact and Relations of Ideas seem so intuitively certain, then? Because of Hume’s own implicit bedrock postulate: “Man is a creature with five senses, all of which are reliable, though not perfect.” It is the falsifiability of this postulate which leaves an opening for the Kantian as well as for the Pyrrhonist skeptic. Observe Hume’s argument which rules out those who deny the validity of the senses: “We need only ask such a skeptic, What his meaning is? And what he proposes by all these curious researches?...[I]f he will acknowledge anything, [he will acknowledge] that all human life must perish, were his principles universally and steadily to prevail. All discourse, all action would immediately cease; and men remain in a total lethargy, till the necessities of nature, unsatisfied, put an end to their miserable existence.” (Hume 1949, 178-9.) In otherwords: deductions from the bedrock postulate that the senses are unreliable lead to a philosophy which is anti-human - which is inconsistent with the nature of man. I should add that this argument does not invalidate the skeptic, because Hume has used an argument that is proper to discussing Matters of Fact. Then why does it ring so true? Because Hume implicitly relied upon the Law of Identity as applied to man. If this implicit use of the identity axiom is dropped, then Hume’s system collapses into: “All propositions - all of them - are merely Matters of Fact. We just assume that some of them are necessary tautologies, called Relations of Ideas, because we are creatures which rely upon all of our five senses, some of which can deliver immediate data which contradict immediate or remembered data from others. The necessary tautologies are the opposite of those ideas which are derived from contradictory sense data. This will not change as long as humans are humans.” In otherwords: the only way to come up with a proposition that is certain is to assert one whose denial confuses the senses by making the evidence of one contradict the evidence of another. An example of this kind of intuitively certain proposition would be “A triangle has three sides.” So Hume’s classificatory system is inadequate - but it does not collapse completely. The limits of his system pin him down as a “humanist empiric.” How Did Hume Arrive At This Division? Concept-Dropping Hume’s justification of the above demarcation begins with him postulating that all ideas find their origin in sense impressions. Here is his identification: “By the term impression, then, I mean all our more lively perceptions, when we hear, or see, or love, or hate, or desire, or will.” (Hume 1949, 16.) Note that this combines sense data with emotions. Hume actually assumes that human beings in fact have more than five senses; the extras are used for introspective purposes. We could ourselves assume that Hume has posited a sixth sense which is used to detect all emotions - call it “emote” - but this posited sixth sense could be broken up into several. Let’s hold it at one, though, because Hume never specifies whether one or many introspective senses exist, and the use of Occam’s Razor as a judgement tool clearly implies adding only one. So, man has six senses: five for extrospection and one for introspection. It is possible that the sixth sense I posited could be an amalgam of man’s five, which have been co-opted for introspective use. By this standard, love or hate or desire or will are percepts derived from the sense of emote. What about all “sense impressions,” though? Are they just percepts in disguise? By Rand’s definition - “A percept is a group of sensations automatically retained and integrated by the brain of a living organism” (Rand 1990, 5), they are. Hume’s description of an “Impression” is clearly consistent with Rand’s definition of “percept.” But is Rand’s definition more correct than Hume’s identification? Observe that all of Hume’s examples of sense impressions are subsumed by Rand’s definition. Also, note that any pure sensation has to be inferred: “The knowledge of sensations as components of percepts is not direct, it is acquired by man much later: it is a scientific, conceptual discovery.” (Rand 1990, 5.) Or, to put it another way, sensations have to be abstracted from percepts. Hume’s use of immediacy as a descriptor clearly implies that a sense impression is a percept, and that the equating of a sense impression with a sensation is an equivocal use of the word “sense.” Note, though, that Hume implicitly adds concepts into the category of “sense impressions” by equating a concept with the perceptual label used to identify it - an approach which Wallace Matson declares explicitly to be a better formulation of Rand’s epistemology. (Den Uyl and Rausmann 1984, 34-35.) This makes Professor Matson a Humean, and his criticism of Rand’s epistemology a Humean evaluation of her work. Unsurprisingly, the first objection that Professor Matson makes to Rand’s epistemology is that her defense of the existence axiom, the Law of Identity and “consciousness [as] the faculty of perceiving that which exists” (Rand 1990, p. 4) as axiomatic simply dismisses her critics out of hand. He continued by noting that her use of the fallacy of the stolen concept to justify the validity of the senses, which he calls a transcendental argument (Den Uyl and Rausmann 1984, 27), depends upon her proving a negative, which he calls a “task of formidable difficulty.” (Den Uyl and Rausmann 1984, 27). ) He continues with this observation: “What then can be done to overcome the initial rhetorical advantage of the inside-out philosophy [subjectivism, which starts by denying objective reality]? If there is some luminous general principle to which we can appeal, some self-evident truth that is incompatible with the Cartesian approach, I do not know what it is, and apparently no one else does either, inasmuch as three and a half centuries have failed to turn it up.” (Den Uyl and Rausmann 1984, 28). This treats axioms as if they were percepts - as if they were simply the product of sense impressions. What Professor Matson has said is that the contrary is arguable because no such sense impression exists. In his second objection, Professor Matson says that Rand’s theory of concepts contains a residuum of subjectivism, based upon this argument: “How do I know whether you know that p? By knowing whether you are in the appropriate frame of mind. But how can I find that out? I cannot, if your state of mind is something knowable only by introspection... How indeed can I know whether I know that p? Granting that I can know my own state of mind, how do I know that it is the right one for knowing that p? [Note the implicit assumption that introspective data is derived from percepts.] Because it corresponds to the fact that p? How am I going to find out whether it does? How am I going to compare a state of mind with something that is not a state of mind? [In otherwords: how do I link introspective percepts to percepts derived from extrospection?] What, indeed does it mean to talk of doing so?” This line of reasoning is later identified by Professor Matson as leading to solipsism (Den Uyl and Rausmann 1984, 29) - which is precisely what Hume observed, though Hume himself used “Pyrrhonism.” (Hume 1949, 177.) No wonder, in the third objection, to Rand’s use of the term “concept,” Professor Matson argues for the use of a collection of percepts - words - as a more sensible alternative to Rand’s concepts. (Den Uyl and Rausmann 1984, 30.) This is precisely the nub of the conflict between Hume’s system and Rand’s: A Randian would insist that Hume’s epistemology is anti-conceptual, while a Humean would respond that a “concept” is simply a ghost in the machine, and that the term “concept” refers to a special kind of percept, one distinguished by use. In otherwords, Rand’s term “concept” is simply a generalization uniting percepts of method. (Professor Matson’s recommendation that words make the best percepts of method is implicitly an argument from convenience, which itself depends upon using the nature of man as a standard of validation for epistemological principles.) Re-Casting Hume‘s Association of Ideas: “Maybe” Isn’t Good Enough After basing all knowledge in percepts, Hume discusses how those special types of percepts called “ideas” are linked together, after noting that the percepts one has sensed and/or memorized clearly show that these “ideas” are related in at least one form. He lists three: resemblance, contiguity and causation (Hume 1949, 23), and adds in a footnote that contrast, or contrariety, can be derived from these three, speculating that the proper derivation of contrasting is from both resemblance and causality. (Hume 1949, 23, n. 4.) Hume later states that the method of use of causation as an idea - inference - is rooted in Matters of Fact: “When it is asked, What is the nature of all our reasonings concerning matter of fact? the proper answer seems to be, that they are founded on the relation of cause and effect [causation]. When again it is asked, What is the foundation of all our reasonings and conclusions concerning that relation? It may be replied in one word, Experience. But if we still carry on our sifting humour, and ask, What is the foundation of all conclusions from experience? This implies a new question, which may be of more difficult solution and explication.... I say [to this question]... that, even after we have experience of the operations of cause and effect, our conclusions from that experience are not founded on reasoning, or any process of the understanding.” (Hume 1949, 32-3.) The justification of this answer takes the form of an argumento a contrario, which is contained in its entirety in Section IV, Part II. (Hume 1949, 32-41.) In essence, Hume says that cause-and-effect sequences cannot be proven because perceptual information about the future is unavailable to man. Before this argument is evaluated, it should be noted that the little sophists who give Hume a bad name base their sophistries upon Hume’s opinion that contrariety - contrasting - is derived from causality, and thus all posited dissimilarities are Matters of Fact. Hume justifies this derivation only with a “maybe.” (Hume 1949, 23, n. 4.) So, the sophist types who are commonly associated with Hume could be declared to be “bad Humeans” if contrasts deserve to be called a fourth “[principle] of connexion between ideas.” (Hume 1949, 23.), with the justification that ‘maybe’ isn’t good enough an argument to collapse the four into three in the way Hume recommended. This would give us these four relations between the percepts of method which are called “ideas”: 1. Similarity, or Resemblance: one percept is similar to the other. 2. Contiguity: one percept is associated with each other. 3. Cause and effect: one percept follows another in the course of time - that they are contiguous through time. 4. Contrariety: one percept is dissimilar to another. It is clear that Similarity is a paraphrase of “A is A” and that Contrariety is a corollary of the Law of Identity. Before looking at Association and Causation, a closer look at Hume’s bedrock postulate - the validity of sense data - should be taken. Consider this line of reasoning: 1. The senses are reliable, and any presumed unreliability of the senses is impossible to conceive of unless "unreliability” is a derivation from the norm. The ultimate justification of this is that the principle of the reliability of the senses is what every living human being, while living, acts in conformance to while they live. (Hume 1949, 178-9.) In otherwords: the postulate of the reliability of the senses is a logical derivation from the observance of the actions of living beings. This is axiomatic because a consistent denial of the senses leads to self-extinguishment of the senser. 2. Each sense impression, or percept, is unique. 3. Two uniques are comparable by the perceiver in one and only one way: equating them. In otherwords, a resort to Similarity. Association is not a comparison; it implicitly assumes that the two percepts are not similar. Otherwise, “they are together” would logically collapse into “they’re the same thing.” The postulate which makes this deduction invalid is the principle of Contrariety: “These two percepts are together, but they are not equatable, so they are Associated, not Similar. If they were equatable, they would be Similar.” 4. The Principle of Similarity can be used to equate all possible sense impressions - “everything is similar, in some way, to everything else.” 5. Declaring all sense impressions, all percepts, to be similar implies that a single percept of method can be used to denote every sense impression - such as the word “existence.” 6. Therefore, it is consistent with the doctrine of the reliability of the senses, provided that the word “existence” is used to denote all possible sense impressions, to conclude: “existence exists, and as long as humans still continue to receive sense impressions -as long as they are alive - existence will continue to exist.” Rand the Aristotlean - Meet Hume the Aristotlean Now let’s take a closer look at Hume’s fundamental postulate - the reliability of sense impressions. Since the sensory apparatus is an attribute of consciousness, Hume’s starting postulate is a derivation of “Consciousness is conscious.” Since denial of the validity of the senses, when followed consistently, means the extinguishment of consciousness through death, the denial of the validity of the senses means the denial of consciousness itself. This follows from Hume’s justification of the reliability of sense impressions. What we have in Hume is a philosopher who starts with the primacy of an attribute of consciousness - sense impressions, or percepts - and then uses the Law of Identity in disguised form as two of his four principles of association of ideas. As far as the other two - Association and Causation - it was shown above that Association depends upon Contrariety. Two percepts which are not amenable to the Principle of Similarity, but are nevertheless found together, are Associated. “This B seems to be an A, but it is not - therefore, A is not B, but A is associated with B because they are found together.” Or, to put it another way: Two existents that our senses associate with each other are not A’s, they are an A and a B. The Principle of Association is used to refine our knowledge of existence in conjunction with the Law of Identity in both forms. As far as Causation is concerned, it is, as noted above, simply the application of the Principle of Association to conjunctions in time. The above implies that all of Hume’s principles of association of ideas are derivations from Aristotle’s laws of Existence and Identity - the same laws that Rand used as the axiomatic base of her philosophy. Both Rand and Hume are students and successors of Aristotle. How Rand and Hume Differ As noted above, the fundamental difference between Rand and Hume has as its base their differing identification of the source of man’s ideas. Rand identifies them as rooted in a faculty distinctive to man called the “conceptual faculty,” which lets human being think conceptually through regarding entities as units (Rand 1990, 6), whereas Hume declares concepts to be only a special form of percept, to be distinguished from other types of percepts by the way they are used. So the difference between Hume and Rand devolves into a debate over man’s nature. Is Hume A Confused Pre-Objectivist... This speculation is based upon another, more interesting, similarity between Rand and Hume: Hume’s characterization of Cause and Effect is uncannily similar to Rand’s theory of concepts. Existents which are associated in time with other existents are subject to the idea of cause and effect. But, as Hume notes, it is impossible to prove that a cause can lead to one and only one effect, because doing so would involve proving that every possible effect bar one could not have been associated with the cause - a task which would require omniscience. He casts a causal argument as a one-many relationship - as open-ended in time. Rand’s theory of concepts imply that they are epistemologically open-ended (Rand 1990, 17-18.) She adds, though, that it is not necessary to prove once and for all whether or not every knowable existent is subsumed by a particular concept: all one has to do is to make sure that the match is consistent with the sum of one’s knowledge - and to not evade any later evidence which shows that the previous identification was wrong in retrospect, but right at the time the original connection was made. Contextual certainty, in otherwords. (Peikoff 1991, 174.) Dr. Peikoff himself notes that contextual certainty can be achieved by subsuming doubtful cases into the concept being used and excluding existents that are clearly not subsumed by that concept - in otherwords, by giving a doubtful existent the benefit of the doubt. This way, knowledge is refined, not invalidated, by future knowledge. (Peikoff 1991, 173.) Applying this criterion to Hume’s characterization of causal sequences after translating it into Humean terms gives you: “Positing a cause-and-effect relation with contextual certainty can be achieved by making sure that such a relation is not contradicted by any sense impressions which show that the cause under consideration has not led to that specific effect which you are relating to the cause. Once you have received such a sense impression, and have satisfied yourself that your senses were not temporarily fooled, you must reformulate that causal relation to be consistent with that later sense impression.” Rand’s theory of concepts and Hume’s analysis of the idea of causality are sufficiently compatible to suggest that Hume’s epistemology is proto-Objectivist. It should be of little surprise that Hume’s assumption that logic can be used non-problematically for Relations of Ideas is compatible with both his reliance upon the Law of Identity and Rand’s definition of logic as “the art of non-contradictory identification.” (Rand 1990, 36.) ...Or Is Rand A Variety Of Neo-Humean? The above compatibility has another interpretation, though: Objectivism is just a kind of neo-Humeanism. This follows from the clean-up of Hume above, and by postulating that Rand takes as axiomatic laws which can be derived from the doctrine of the reliability of the senses. To put this more elaborately: Objectivism is a variant of Humeanism which: This identification of Rand as a kind of neo-Humean is, of course, refutable.
Discuss this Article (14 messages) |
 |
 |
|
 |
|
