|
Notes on Hume
Sandra LaFave These notes are divided into two main parts. Read the first part as you read the assigned portions of the text. Read the second with the Palmer book.
I. Notes on Assigned Portions of Hume's Enquiry
Section II “Of the Origin of Ideas” Hume distinguishes impressions and ideas. Our thought is not unbounded. We can combine, transpose, augment, and diminish “the materials afforded by the senses,” but “all our ideas or more feeble perceptions are copies of our impressions or more lively ones.” (11) Two arguments show this: (1) no one can find a counterexample; (2) blind people can’t form ideas of color, deaf people of sounds, etc. Many ideas are confused; some are altogether bogus. Impressions, by contrast, are “strong and vivid” (13). Since impressions are far easier to know with certainty, Hume devises his test for the legitimacy of ideas: find the corresponding impression(s). If any idea has no corresponding impressions, it probably comes not from our experience, but from our imagination combining legitimate simpler impressions into bogus composites (e.g., the “golden mountain”). Section III “Of the Association of Ideas” We find that ideas are associated in the mind according to three principles: resemblance, contiguity, and cause and effect. These are natural, built-in principles that we can’t override. Section IV “Skeptical Doubts …” — Part I Hume’s Fork: “All the objects of human reason or enquiry may naturally be divided into two kinds, to wit, Relations of Ideas, and Matters of Fact.” (15) Relations of Ideas correspond to what Leibniz called analytic statements; Matters of Fact correspond to what Leibniz called synthetic statements. Relations of Ideas are “intuitively or demonstratively certain”, “discoverable by the mere operation of thought” (15); their denials are inconceivable. E.g., math statements and math proofs. These propositions don’t depend “on what is any where existent in the universe” (15) — i.e., they are true a priori. They give no information about the world. Matters of Fact are statements whose denials are conceivable and not self-contradictory, e.g., “The sun will rise tomorrow”. They are known by “the present testimony of our senses” and “the records of our memory” (16). We also reason about sense impressions and memories. For example, we often draw conclusions that go beyond immediate sense data and memory. For example, I believe my friend is in France because I have a letter from her with a French postmark in which she writes, “Greetings from Paris!”, etc. Now, according to Hume, “all reasonings concerning matter of fact seem to be founded on the relation of Cause and Effect. By means of that relation alone we can go beyond the evidence of our memory and senses.” (16) E.g., we think the French postmark on our friend’s letter, and her remarks, are caused by her actually having sent the letter from France. So, how do we reason about causes and effects? Only, Hume says, from experience, i.e., from what has happened in the past, e.g., the famous billiard ball example (18). We have prior sense impressions to support the constant conjunction of causes and effects, as well as the temporal priority of the cause (49-53). But we can know nothing more. We can note with certainty only past correlations of events, e.g., whenever I got letters with French postmarks in the past, they turned out to be sent from France; whenever I ate bread in the past, I was nourished. But what caused these past correlations? Were laws of nature responsible (e.g., “bread is nourishing for humans”)? We cannot know: “these ultimate springs and principles are totally shut up from human curiosity and enquiry”. (19) Conclusions about matters of fact that are beyond immediate sense and memory — especially predictions about the future, e.g., that bread will continue to nourish me — do not follow necessarily from experience, then, and are therefore not really reasonable. They depend on two additional, merely customary, notions: (1) that there is a necessary connection between causes (e.g., eating bread) and effects (being nourished), and (2) that the future will resemble the past. But these last two ideas are completely without supporting impressions. “As to past Experience, it can be allowed to give direct and certain information of those precise objects only, and that precise period of time, which fell under its cognizance: But why this experience should be extended to future times, and to other objects ... this is the main question on which I would insist.” (21) Per Hume, it’s not even likely that future will resemble past (22-23). To recap: Cause and Effect is one of the three non-overridable natural principles of association of ideas. It is the only one relevant to making predictions about future matters of fact. We know of causes and effects only from experience. Experience covers only what has happened in the past or present; it does not extend to the future. We think we can nonetheless predict what will happen in the future because we think there are unchanging laws of nature that establish necessary connections between causes and effects. We also believe the future will resemble the past, in regularly following these so-called laws. The only legitimate ideas are those supported by impressions. There are no impressions to support the idea of necessary connections between causes and effects; there are no impressions to support the idea that the future will resemble the past. Therefore, what’s happened in the past or present (our experience) is completely irrelevant to knowledge of cause-and-effect matters. We really can’t be said to know anything certain about the future. (Think about repercussions for science and ordinary life if Hume is right.) Section V “Sceptical Solution of … Doubts” “All inferences from experience ... are effects of custom, not of reasoning.” (28) “Custom, then, is the great guide of human life. It is that principle alone, which renders our experience useful to us, and makes us expect, for the future, a similar train of events with those which have appeared in the past.” (29)
II. Hume ’s Fork, the Problem of Induction, and Logical Positivism Some Basic Concepts of Logic Reasoning is the process of deriving a claim (the conclusion) from other claims (the premises) using principles of logic. Math proofs are good examples of reasoning. An argument is a piece of reasoning. An argument consists of a set of statements. The arguer claims that one of those statements (the conclusion) follows from, or is supported by, the other claim(s) (the premises). In a deductive argument, the arguer claims that the conclusion must be true if the premises are true. The argument has correct deductive logic if the arguer’s claim is correct: that is, if the conclusion really must be true if the premises are true. An argument has incorrect deductive logic if the arguer’s claim is false: that is, if the conclusion isn’t necessarily true if the premises are true. Logic is unaffected by false premises. Logic assumes the premises are true. Deductive logic asks only one question: must the conclusion be true if (assuming) the premises are true? Whether the premises are true is not a logical matter; it is a factual matter. In an inductive argument, the arguer claims that the conclusion is probable, or likely to be true, if the premises are true. The inductive argument has correct inductive logic if the arguer’s claim is correct: that is, if the conclusion really is likely if the premises are true. An argument has incorrect inductive logic if the arguer’s claim is false: that is, if the conclusion isn’t likely if the premises are true. Logic is unaffected by false premises. Logic assumes the premises are true. Inductive logic asks only: is the conclusion likely if (assuming) the premises are true? Whether the premises are true is not a logical matter; it is a factual matter. Mathematical inferences are generally deductive. Inferences about the world, especially scientific inferences about causality and the future, are generally inductive. Employing Hume’s terminology, an inductive argument goes from immediate impressions to other non-immediate ideas. A priori and a posteriori The expressions “a priori” and “a posteriori” can be used as adjectives or adverbs; they can modify a noun (such as “knowledge,” “statement,” or “claim”), a verb (such as "know") or an adjective (such as “true” or “false”). The “a” in these expressions is the Latin preposition meaning “from.” So “a priori” means “from before [observation]” and “a posteriori” means “from after [observation]”. The expressions “a priori” and “a posteriori” describe how we know the truth or falsity of a statement. A statement is true or false a priori if no observation or experiment is required to determine if it is true or false. Examples of a priori statements are mathematical assertions, statements true or false by definition, and logical truths and falsehoods. We “just know” when some claims are a priori true or false. For example, we “just know” that the same statement cannot be both true and false in the same sense at the same time (a rule of logic called the law of non-contradiction). A statement is true or false a posteriori if observation or experiment is required to determine if it is true or false; we don’t “just know” it. Examples of a posteriori statements are statements about the world, e.g., “Dogs are carnivores” or “Ottawa is the capitol of Canada.” Analytic and Synthetic Analytic statements are a special class of a priori statements. In analytic statements, the predicate concept adds nothing to the subject concept, e.g., “Bachelors are unmarried,” or “The red house is red.” Synthetic statements are a special class of a posteriori statements. In synthetic statements, the predicate concept adds something to the subject concept (the two concepts are synthesized), e.g., “The red house is owned by a dentist.” Hume’s Fork According to Hume, legitimate reasoning has just two possible kinds of subject matter: 1. Relations of Ideas (e.g., math, logic) or 2. Matters of Fact (e.g., empirical matters). Reasoning about relations of ideas is analytic and a priori. Reasoning about matters of facts is synthetic and a posteriori. For Hume, any legitimate statement is either analytic a priori or synthetic a posteriori. According to Hume, analytic a priori statements – the kind we use when we reason about relations of ideas – tell us nothing about the world; they tell us only about how we think and use language. Thus, according to Hume, the only statements than can tell us anything about the world are synthetic a posteriori. And according to Hume, if a statement is synthetic a posteriori, it must be grounded in impressions (sense data or passion). If no impressions support a synthetic statement, the statement is bogus superstition, and should be rejected. In other words, Hume’s fork has two tines. Legitimate statements
are either
Thus, statements are either analytic a priori (in
which case they tell us nothing about the world), OR they are synthetic
a posteriori (in which case they must be supported by impressions).
For Hume, there are no other legitimate possibilities. Hume’s fork means that statements about matters of fact always
require empirical support; we can never “just know” them. This is why
Hume criticizes the Ontological Argument, which attempts to prove that
the claim “God exists” is true a priori. For Hume, no claim about
existence can be a priori, since whether or not something exists
is a matter of fact, and thus must be known a posteriori. Hume’s Fork does not necessarily plunge us into skepticism
about morality, since for Hume, morality is a matter of the passions,
and passions are one of the sources of impressions. So to say “Stealing
is wrong” simply means “I feel stealing is wrong”; but what if everybody
feels the same way? Then morality is a set of objective facts about human
feeling based on common human nature. Hume’s Fork and the Problem of Induction Reasoning about relations of ideas is typically deductive,
so in mathematics, conclusions of logically correct arguments follow necessarily
from their premises. But according
to Hume, math statements don’t tell us anything about the world. Although
the conclusions of mathematical arguments are certain if the premises
are true, we never have assurance that
the premises are true (since for Hume legitimate ideas come only from
impressions and we do not have impressions of the objects of mathematics).
Reasoning about matters of fact, by contrast, is inductive.
Therefore, conclusions of logically correct arguments about matters of
fact are at best likely if the premises are true. We cannot, however,
derive any trustworthy conclusions
about the world based on inductive reasoning. This is because, according
to Hume all reasonings about matters of fact are based on the relation
of cause and effect, which in turn is based on two other presuppositions:
(1) that there is a necessary connection between causes and effects; and
(2) that the future will resemble the past. But we do not have impressions
of necessary connection between cause and effect (we only have impressions
of constant conjunction and temporal priority of the “cause”); nor do
we have impressions to support the claim that the future will resemble
the past. Therefore, all reasonings about matters of fact are in fact
unsupported by impressions and thus not legitimate – a great blow to science!
This line of argument is often called Hume’s Problem of Induction. Until
Hume, philosophers thought that conclusions of good inductive arguments
were likely if the premises were true. According to Hume, we have no guarantee
based on reason that they’re even likely. Rather, we rely on custom and
habit. Logical
Positivism Logical Positivists are 20th-century heirs of
Hume. The logical positivists propose the
Verificationist Criterion of Meaningfulness. This is a test of meaningfulness
for synthetic statements (statements about matters of fact). It is a 20th-century
reworking of Hume’s Fork, except the subject matter now is explicitly
meaningfulness versus nonsense. The criterion is as follows: a synthetic statement is meaningful
(not nonsense) if and only if you know (or you can imagine) what impressions
would verify it and what impressions would falsify it. If you don’t know the verification and falsification conditions
for synthetic statements, then you are talking nonsense. In other words, the logical positivist fork has three tines.
Statements are either There are no other possibilities, according to the
positivists. Like Hume, logical positivists proceed to apply the verificationist
criterion of meaningfulness to religious assertions. Unlike Hume, they
also apply it to ethical assertions. They hope thereby to reduce those
assertions to nonsense. For example, religious people usually believe
that the statement “God exists” expresses a fact about reality. If the
statement expresses a fact (and not just a claim about how we think or
use language), then the statement “God exists” is not analytic. So it
must be either synthetic or nonsense. If the statement “God exists” is
synthetic, then it can be verified or falsified empirically; i.e., there
are sense impressions that support or refute it.
What impressions verify or falsify it? I hope you begin to
see the problem. According to the ordinary Western account of God, no
impressions verify or falsify the claim “God exists”. If no impressions
verify or falsify the claim “God exists,” then the claim falls into the
nonsense category. This upsets some people.
Sample Essay Questions
|