From wiki@horace.org
Jump to: navigation, search


  • Epistemology - What is knowledge
  • How human gain knowledge
  1. Perception
  2. Memory
  3. Reason (e.g. 2+2=4, no female bachelors in Surrey)
  4. Testimony
  5. Justification, irrational
  • Converse dutch book
    • dutch book - a bet guarantee to win
  • Standard model of knowledge
    • Justified, True, Belief (Plato)
  • Topics in epistemology
    • Externalism
    • Internalism
    • Skepticism
    • Foundationalism
      • Pragmatism
      • Contextualism
    • Coherentism
  • A priori knowledge
    • no perception required
  • Analytic / Synthetic distinction
  • Evidentialism
  • Limit of epistemology
    • can epistemology produce knowlegde?

Logic of Arguments

  • Validity / Soundness / Cogency
  • Compelling
    • Reasonable and possible to believe
    • evidence to believe by people not already agree with the conclusion.
  • Deduction / Induction / Abduction
  • Intuitions and Thought Experiments
  • Analytic / Synethic

Rusell: Appearance and Reality

  • In daily life, we assume as certain so many things which, on closer scrutiny, are found to be so full of apparent contradictions that only a great amount of thoughts enables us to know what it is that we may really believes
  • Russell vs. Bradley: our ordinary beliefs about the world, absolute truth, and absolute idealism.

  • Sense Data
    • One thing that Russell did share with Bradley is the view that we don't ahve immediate access to tables and chairs.
    • The color varies depending on one's perspective. Thus no two people will see exactly the same color at the same time, because no two people occupy exactly the same perspective.
    • We don't seem to have access to "the color" of "the table". the same goes for other properties of "the table". Russel runs a similar argument here for shape.
    • We only have access to how things appear to us form some or another perspective.
    • Rusell gives the name "sense data" to these appearances; sense data are the things that are immediately known in sensation.
    • And "sensation" refers to the experience of being immediately aware of these things
    • More generally, Russel distinguished among mental act (sensation), subjective content (sense-datum) and object (external, physical object).

    • Objection: Many philosophers have objected that contemporary empirical psychology has shown that this soft of "sense data" view to be seen empirically disconfirmed.
    • Is this a compelling objection?
    • Not necessarily!
    • Psychological vs. epistemological account

    • Epistemology concerns with logical dependencies.

  • The argument
    • P1. The thing we are directly aware changes, depending on our point of view.
    • P2. The real, external object (if such a thing exits) does not itself change, in relation to our point of view.
    • C. Therefore, the thing we are directly aware of is not the real, external object.


Russell ch.1 -> skeptical argument

  • What we have direct access to -> inside out argument

Descarte -> different kind of skeptical argument

  • also inside out argument
  • not base on sense data
  • ask which belief we could doubt
    • leads to doubt all our beliefs if possible
    • leads to indistinguishable idea


  • Born in 1596
  • 1607-1615 Jesuit school - Aristotle view
  • 1619 - maths inventions
  • 1629 - "The world", grand theory of everything
  • 1637 - discourse on method
  • 1641 - meditations

Descartes' problem: how do I convince an audience that the new science is not unreasonable, when the audience is unwilling to question the old scientific paradigm?

Two obstacles:

  1. religion: need to convince religion authority that he is not a heretic
  2. philosophical: the Aristotelian background philosophical view that all knowledge comes through the senses

His solution was ingenious: Mediations

  • Obstacles 1: apprise the religious ahthorities
  • Obstacles 2: "Meditations" as meditations.

On this reading, the goal is:

  1. To bring his readers to an awareness of cognitive reasons content in their own minds, through which they could then use to see the first principles of metaphysics for themselves.
  2. Thereby open them up for the new science.

2. Closer look at the text meditation 1. Here is how Descartes begins "It is now some years ... establish any firm and permanent structure in the science" -> mediator recipe

So then who does the "I" in the meditation refer to?

The stage of doubts

  1. Sensory fallibility
  2. Dream argument - Spatial, temporal properties, geometry/math.
  3. Deceiving God - misfire cognitive
  4. Evil demon - misleading you to believe the wrong is right

3. Argument for philosophical skepticism

  1. Most of my beliefs about the world depends on my senses
  2. If skepticism is true, then my sense are deceptive
  3. I can't rule out skepticism
  4. I know that p only if I can rule out ~p. (that is I can know something only if I can rule out grounds for believing otherwise.)

C. I don't know very much.

Responses to Skepticism

1. Cartesian Foundationalism

What exactly is the "I" that thinks?

  1. only a momentary thought-experience
  2. a stream of thoughts
  3. an enduring, self-subsisting object (i.e. substance) which thinks
So then how does Descartes establish his intended conclusion?
Perhaps we can reconstruct it as an argument?
  • 1. I think
  • C. I exist as a thinking substance

Missing premises?

  • 1. I think
  • 2. Whatever thinks exists
  • 3. No act or accident can exist without belong to a substance.
  • C. I exists as a thinking substance.


  1. Can anyone think of an exception to the second premises? Frictional characters?
  2. The third premise is far from obvious too
  3. Couldn't the evil demon trick out reasoning?

Solution: It's discovered by the mediator in a single cognitive act, and only subsequently analyzed in terms of some argumentative structure.

What's Descartes' general epistemological view here?

Foundationalism: The view that all knowledge and justified belief ultimately on a foundation of basic beliefs, i.e. beliefs whose justification does not depend on any other beliefs.

More precisely, it's committed to two claims:

  1. There are basic beliefs
  2. All justified beliefs are justified in relation to the those basic beliefs

Basic motivation for foundationalism: response to the regress problem.

  • P1. Either there are justified basic beliefs or each justified belief has an evidential claim that:
  1. terminates in an unjustified belief
  2. is an infinite regress
  3. is circular
  • P2. But beliefs based on unjustified beliefs are not themselves justified, ruling out (1)
  • P3. No person can have an infinite chain of beliefs, ruling out (2).
  • P4. We don't think beliefs can be justified via circular reasoning, ruling out (3)
  • C. There are at least some basic beliefs.

Classical Foundationalism: the view that all knowledge and justified belief ultimately rest on a foundation of basic beliefs, which are indubitable (infallible, known with certainity), or inferred (deduced) from basic beliefs.

"Classical Foundationalism" has often taken to be synonymous with "Cartesian Foundationalsim"

Is this right?

Descartes says: "...when I direct my attention to things which I believe myself to perceive very clearly, I am so persuaded of their truth that I let myself break out into words such as these: Let who will deceive me, He can never cause me to be nothing while I think that I am, or some day cause it to be true to say that I have never been, it being true now to say that I am, or that two and three make more or less than five, or any such thing in which I see a manifest contradiction. And, certainly, since I have no reason to believe that there is a God who is a deceiver, and as I have not yet satisfied myself that there is a God at all, the reason for doubt which depends on this opinion along is very slight, and so to speak metaphysical. But in order to able altogether to remove it, I must inquire whether there is a God as soon as the occasion present itself; and if I find that there is a God, I must also inquire whether He may be a deceiver; for without a knowledge of thse two truths I do not see that I can ever be certain of anything."

Upshot: Descarte' criterion for knowledge doesn't require absolute certainty; it only requires an irresistible cognitive assent to the proposition while directly attending to it.

So, either Classical Foundationalism <> Cartesian Foundationalism, or Descartes wasn't a Cartesian Foundationalist.

Still Descartes was a Foundationalist, since some belief require irresistible assent.

2. Russell's Foundationalism

Rusell's response to skepticis

First Attempt: The simplest way to account for the corresponding 'table appearances' that different people have, is to posit the existence of a table that's the cause of these corresponding 'table appearances.'

(And the more people observed to have corresponding 'table appearances', the greater the evidence in favor of our belief in the table.)

Is this a compelling argument?

Second Attempt: The common sense hypothesis (that there are external object) is the simplest explanation of my sense data.

  • P1. My sense data of apparent objects (e.g. cat and people) follow law-like patterns.
  • P2. The simplest (and hence best) explanation of the fact that my sense data follow these law-like patterns is that our sensations are caused by external object (i.e. objects which don't change, in relation different point of view).
  • P3. IBE: If a hypothesis H is the best explanation for some phenomena E, then H is likely to be true.
  • C. The hypothesis that there are external objects is likely true.

Problem: Can Russell help himself to the fact that sense data follow law-like regularities? Is that really the sort of thing that can established without relying on testimony (and thereby begging the question at issue)?

What is Russell's general view here?

Is he a Foundationalist?

  • Yes. Our belief in external objects originally derives from instinctive beliefs:
  • "All knowledge, we find, must be built up upon instinctive beliefs, and if these are rejected, nothing is left"
  • But not all instinctive beliefs are on a par. There's a "hierarchy" of instinctive beliefs, some of which are "much stronger" than others.

So then what kind of Foundationist is Russell?

  • Is he a Classical Foundationalist?
    • No!
  • our instinctive beliefs can be over-turned by other beliefs, i.e. if the instinctive belief clashes with other beliefs that we have reason to accept.

This sort of view is often classified as Weak (Modest, Non-classical) Foundationalsim, which is the view that:

  1. There are some basic belief which are minimally justified without reference to any other beliefs.
  2. To attain knowledge, coherence among one's beliefs is also required.

Question: How does this view answer the regress problem?

  • Or put another way: can we still make sense of a reconstruction of knowledge from the foundations, on this view?

Moore's proof of the existence of an external world

Moore's proof:

  1. Here are two hands
  2. If hands exists, then there's an external world.
  • C. So there's an external world

Moore's standards of a rigorous proof

  • An argument constitutes a proof of its conclusion iff:
  1. Its conclusion is different from each of its premises;
  2. Its conclusion follows from its premises; and
  3. Each of its premises is known.

Where might Descartes or Russell take issues with Moore's reasoning?

  • How Russell would likely interpret this argument:
  1. Here are sense-data of two hand-appearances.
  2. IF sense-data exist, then there's an external world.

C. So, there's an external world.

But now the skeptic would clearly reject 2!

So then the question is this: what does Moore do in the papper to try to cut that sort of response off?

In Kant

  • Nuumena (things in themselves) vs Phenomena
    • Nuumena (things in themselves)
transcendental idealism
no access to this
-> off limits
    • Phenomena (appearance)
Empirical realist
All objects in space & time
-> all objects of science and knowledge.
-> empirical questions
  • Truth
    • empirical level -> external object


  • Analysis
    • can be debake at a transcendental

Moore's discussion of what counts as "external".

  1. Kant on "the existence of things outside of us" and "the existence of things to be met with in space."
  2. More on questions of truth (which are settled at the empirical level) and question of analysis (which can be debated at the transcendental level).

2. Russell on the nature of matter

His next question is this: What's the nature of matter? Does the table have all the same properties that my sense data of the table has?

Color and sound properties

  • e.g. Does the table itself have color? That is, is color an objective property of the table?

Russell's answer: No. According to our best science, what we call "color" is merely a sensation produced by light ray bouncing off the table, hitting our retina, and the being processed by our visual system.

In other words, "color is not supposed by science to from any part of the world that is independent of us and our senses.'

Similar argument for sound properties.

Spatial properties

What about space? Does the table have spatial properties?

Russell's answer: yes, but not the exactly the same spatial properties exhibited in sense data

Apparent/private space vs real/public space

What can we know about his physical space?

" We can only know what is required to secure the correspondence?" What does Russell mean here?

Physical space is constructed in order to account for what different people, form different points of view, can all agree upon certain things.

Russell then goes on to say: "We can know nothing of what it is like in itself, but we can know of the sort of arrangements of physical objects which results from their spatial relations."

Similar argument for the temporal properties of objects.

More generally: "Thus we find that, although the relations of physical objects have all sorts of knowable properties, derived from their correspondence with the relations of sense-data, the physical objects themselves remains unkown in their intrinsic nature..."

This is a kind of structuralist view concerning what we can know about the real properties of thins:

Standard scientific realism: the view that the nature of the objects causing our appearances is correctly described by our best theories.

Anti-realism: denies that our vest theories give us insight into the nature of objects and the world, independently of how things appear to us.

Structural realism: our best scientific theories commit us to the reality of certain structural properties, or relations, that hold between objects.

Russell on Idealism

Idealism: the view that whatever exists is, at bottom, in some sense mental.

Russell notes that this view was very widely held at the time (it's now longer such a popular view!)

Russell's characterization of Berkeley's two main contentions:

  1. Our sense data do not have any existence independent of the mind, but must be at least partly "in" the mind.
  2. Sense data are the only things that we can establish to exist via perception, so that all objects about which we can have any knowledge are, at bottom, ideas 'in" some another mind.

Russell's characterization of Berkeley's argument for 1)

  1. The only things we have direct access to are ideas
  2. All ideas are "in" the mind, and so don't have any mind-independent existence.

C. So all knowledge concerns things which are purely mind-dependent

Russell's analysis of this argument:

  • The notion of being "in" the mind is ambiguous between:
  1. a thought (e.g. a thought of a tree)
  2. the referent of the thought (e.g. an actual tree)

Similarly, the notion of an "idea" is ambiguous between:

  1. the thing about which we're aware (e.g. a tree)
  2. the mental act of apprehending the thing (e.g. the act of apprehending a tree)

The later, 2) is undoubtly mental.

The real question is whether the former, 1) is also mental. If not, then Berkeley's argumenet depends on an equivocation between 1) and 2)

Two senses of the word "know"

  • Russell: "it is by no means a truism, and is in fact false, that we cannot know that anything exists that we do not know."
  1. knowledge that something is (or is not) the case, i.e. knowledge of truths
  2. knowledge of something by being directly acquainted with it, i.e. knowledge of things

Russell: "there is no reason why I should not know of the existence of something with which nobody is acquainted."

  • e.g. knowledge of the Big Bang

2. Knowledge by Acquaintance and knowledge by Description

Knowledge by acquaintance: knowledge of things, involve a direct relation of the mind to the object known, without any intermediary process of inference or knowledge of truths.

  • e.g. we are acquainted with sense data in this way.

Knowledge by description: knowledge of truths, involving an indirect relation of the mind to the thing which is known, via description.

  • (a.k.a. propositional knowledge")
  • e.g. knowledge that there's a physical table, which causes such-and-such sense data.

What's known via acquaintance?

  1. Our sense data
  2. Our memory
  3. Our awareness of being aware of things (via introspection)
  4. Ourselves
  5. Universals

Two kinds of descriptions:

  1. "a so-and-so" (an ambiguous description, e.g. "a man")
  2. "the so-and-so" (a definite description, e.g. "the man with the iron mask")

Rusell's theory of definite descriptions: "the so-and-so" just means "there's an object which is so-and-so, and there's only one object which is a so-and-so".

Two things that this theory of definite descriptions allows us to do:

  1. To speak of non-existent objects, i.e. to meaningfully use names without being committed to there being anything to which those names refer.
e.g. we can make statements about the fountain of youth, without being committed to there actually being a fountain of youth.
  1. It also allows to have knowledge of things that go beyond what we're directly acquainted with.
e.g. Bismarck had knowledge by acquaintance of himself, but other can still have indirect knowledge about him:
    1. Those who met him have direct acquaintance with sense data of Bismarck, and can connect that sense data to the actual person Bismarck, via description (the person to whom such-and-such sense data correspond)
    2. Those who've never met him, but can still have knowledge of him via a definite description "the first Chancellor the German Empire."

But for such descriptions to successfully refer to Bismarck, the description must ultimately to be composed of constituents with which we're acquainted.

e.g. in the case of b), we would need to be acquainted with the universals Chancellor, German, Empire, etc.

"The fundamental principle in the analysis of propositions containing descriptions is this: Every proposition which we can understand must be composed wholly of constituents with which we are acquainted."

What kind of general epistemological pictures does this suggest?

A version of foundationalism, where the foundations are constituted by beliefs concerning things we're acquainted with.

If this is right, can this view be reconciled with what Russell said earlier, at the end of chapter two, about the foundational role of instinctive beliefs?


1. THe way Russell raises the problem

"It must be known to us that the existence of some soft of thing, A, is the sign of the existence of some other thing, B, either at the same time as A or at some earlier time, as for example, thunder is a sign of the earlier existence of lightening."

To do this, we seem to rely on an inference of the following sort:

"when a thing of a certain sort has been found to associated with a thing of a certain sort B, and has never been found dissociated from a thing of sort B, the greater the number of cases in which A and B have been associated, the great is probability that they will be associated in a fresh case in which one of them is known to present."

Moreover, given a sufficient number of cases of association, the probability approaches certainty without limit.

(He also goes on to distinguish a second version of this principle: under the same circumstances, it becomes increasingly probable that A is always associated with B.)

So then the question is: what's the justification for this principle/

"The man who has fed the chicken every day throughout his life at last wrings its neck instead, showing that more refined views as to the uniformity of nature would have been useful to chicken."


Hume identifies two basic types of knowledge.

  • Relations of ideas
    • A proposition expresses a relation of ideas if and only if its denial is strictly impossible, inconceivable, or self-contradictory.

(Hume seems to regard these notions as equivalent.)

  • Matter of fact
    • A matter of fact proposition has the following feature: both it and its denial are fully conceivable, possible, and non-self-


  • E.g. consider the proposition that my apartment is blue.

So then what’s the problem?

  • The question Hume raises: How do human beings arrive at their opinions concerning unobserved matters of fact?
  • E.g. consider that proposition that all snow – and, in particular, the next piece of snow you observe – will be cold.
  • Presumably we all believe this. But what’s our justification for believing this?

3. Proposed answer #1: It somehow reduces down to a relation of ideas.

  • But this seems unlikely.
  • Human thought experiment
  • On this basis Hume asserts a general proposition: Opinions about unobserved matters of fact are somehow derived from


4. Proposed answer #2

  • It’s justified on the basis of experience. But how exactly does this go?
  • What’s needed is an argument that that justified this expectation, on the basis of past experience.
  • What could try to argue as follows:
  • P1. In our experience thus far, snow has always cold
  • C./.. Snow is always cold – or at least, the next piece of snow that we examine will be cold.
  • Is this a valid argument?
  • And notice the general form of the argument:
  • P1. In our experience thus far, all Fs are Gs.
  • C./.. All Fs are Gs – or at least, the next F we examine will be G.

5. Proposed answer #3

  • We can turn the above argument into a valid argument by adding another premise.
  • (UN) For the part, if a regularity R (e.g. All Fs are Gs) holds in our experience, then it holds in nature generally, or at least in
  • the next instance.
  • “UN” stands of the “Uniformity of Nature”.

  • The resulting argument is certainly valid. And moreover, it’s sound. But it is a compelling argument - that is, would it give

skeptic a reason to accept the conclusion?

  • The skeptical problem: The skeptic has no good reason to accept UN.

  • We might try to argue for UN as follows:
  • P1. In the past, regularities that have held in my experience have been found to hold future generally, or at least in the next


  • UN./.. in general, if a regularity holds in my experience, it holds in nature generally, or at least in the next instance.

  • But this argument is, once again, invalid! Moreover, what sort of fallacy have we committed? Begging the question issue.
  • Is there a good (i.e. non-circular) justification for UN?

Here’s an argument that there’s not:

  • P1. UN itself expresses a claim about unobserved matter of fact. So UN is not a priori. If it is known at all, it is known on the

basis of experience.

  • P2. But all “knowledge” of unobserved matters of fact is known (if it is know at all) on the basis of any inductive inference.
  • P3. But UN figures as a premise in any inductive argument.
  • C. /.. there can be no non-circular argument of UN.

6. Questions about UN: The Cautious Martian and the Counterinductivist

  • (DATA) In our extensive experience thus far, snow has always been cold.
    • We take this to support the general proposition that
  • (THEORY) Snow is always cold, or at least it will be the next time I come across it.
    • Here’s what the Cautious Martian says:

  • “I Agree with you about the DATA, but the THEORY seems like a wild, unjustified leap.”
  • What could we say in response to the Cautious Martian?
  • The Counterinductivist issimilar to the Cautious Martin, except that he holds (somewhat perversely, we think) that his

experience is always misleading. For he accepts:

    • (NN) If a regularity R holds in my experience so far, it does not hold in nature generally, not even in the next instance.
    • (NN) - If snow has always been observed to be cold in the past, then it's likely that the generalization "all snow is cold" will have exceptions to it at some point - in particular, it is likely that the next piece of snow won't be cold.

2. What is his skeptical solution?

What's the principle, or mechanism, underlying our inductive inferences about matter of fact?

  • Custom or habit.

When two ideas have been constantly conjoined in a particular order, i.e. one idea is always found to be followed by another idea, THEN custom/habit produces the 'feeling' in the mind that there's a necessary connection between the two ideas.

But is this really a solution to the problem Hume raised?

3. Three interpretations of Hume's skeptical solution:

  1. Interpretation #1: Hume is claiming that are beliefs about matters of fact are not justified
  2. Interpretation #2: Hume is just raising the problem and is agnostic about what the solution is
  3. Interpretation #3: Hume is claiming that our beliefs about matters of fact are justified, but only in a very minimal sense.

Three features of Hume's solution on this interpretation:

  1. Our habits of inductive (instinct) inference and our acceptance of UN are non-rational.
  2. But our acceptance of UN is not irrational.
  3. Our acceptance of UN is not optional.

4. Russell's solution

The principle of induction is "incapable of being proved by an appeal to experience."

Hence, we must "accept the inductive principle on the ground of its intrinsic evidence"

Russell's more general view: Induction is one of a number of "self-evident" general principles.

Others include:

  1. Laws of deductive logic: e.g. law of identity, law of non-contradiction, law of excluded middle.
  2. Besides laws governing pure mathematics and geometry\
  3. Judgements about "the intrinsic desirability of things"

such principles a prior in the sense that their justification does not depend on experience.

  • Experience may be essential to direct our attention to the truth of the principle, but experience cannot serve as evidence or justification for the principle.
  • e.g. Consider Modus Ponens
  • e.g. 2+2 = 4

Rusell: "... a certain number of instances are needed to make us think of two abstractly, rather than of two coins or two books or two people, or two of any other specified kind. But as soon as we are able to divest ourselves of irrelevant particularity, we become able to see the general principle that two and two are four; any one instance is seen to be typical, and the exmination of other instances becomes unnecessary."

However, Rusell is not committed to the view that all a priori truths are basic (i.e. self-evident) principles. e.g. some truths of arithmetic, while a prior, may be deduced from the basic laws governing arithmetic.

Basic principles all have the following property: any attempt to prove or justify the principle assumes the truth of that very principle.

Moreover, some basic principles have a higher degree of "certainty" or "self-evidence" in favor of them than other do.


1. Hume vs Kant on the nature of a prior knowledge

  • For Hume, all a priori knowledge is analytic, i.e. a mere relation of ideas.
  • Kant denies this thesis.

First, Kant gives a more specific definition of analytical: a judgment is analytic just in case that predicate term is contained the subject term.

(By contrasts, with synthetic judgments, what's contained in the predicate goes beyond what's contained in the subject.)

Second, he argues that a close analysis of mathematical judgments reveals that they are not all analytic in this sense. e.g. 5+7=12.

Kant's more general view; all objective judgments we make about the world - judgments that are capable of being true or false - presuppose that we represent things within space and time.

In addition to space and time, Kant also be believes that there are certain basic "categories", by which the human understanding frames all objective judgments about the world.

e.g. the objective judgments we make about the world also presuppose that the world has a causal structure to it.

Kant's argument here is what's known as a transcendental argument

  • Transcendental argument begins with an accepted fact concerning something we takes ourselves to know (in this case, that mathematics is known a priori), and then argues tow hat must be the case for such knowledge to be possible (for beings like us)

This argumentative strategy leads to Kant's so-called "Copernican Revolution";

Rusell vs Kant on a priori knowledge

1. The way Russell frames the problem:

a. Knwoledge mathematics and geometry is not analytic in roughly Kant's sense.

A proposition is analytic iff it opposition is self-contradictory. e.g. "A is A" counts as analytic.

Such propositions are regarded as trivial, in the sense that they don't tell us anything interesting about the world.

By contrast, propositions of mathematics and geometry do extend our knowledge and make non-trivial, substantive claims about the world. Hence, they're synthetic.

b. Knowlege in mathematics and geometry is a priori.

Same thing goes for the principle of induction and (most) proposition of logic

So, the puzzle is how we can know substantive truths about the world, in advance of our experience of the world.

e.g. We know that the "7+5=12" applies on Jupiter, even though no one has traveled there to verity it does indeed apply there.

2. Russell's assessment of Kant's solution

a. To say that logic and math are contributed by us does not seem to account for our certainty that the facts must always conform to the laws of logic and mathematics.

Btus: "our nature is just as much a fact of the existing world as anything, and there can be no certianty that it will remain constraint"

Russell is here accusing Kant of psychologism, the vie that the laws of logic or mathematics are grounded in psychological laws concerning our mental progress.

Is this a compelling refutation of Kant? How might Kant respond?

Kant did not take himself to be stating a fact, at least not in any ordinary sense of the word "fact".

Moreover, the structure that Kant identifies is necessary structures.

The assumption here is that, in order for something to be a priori, it would be necessary and fixed for all time.

b. Rusell's positive view: the laws of logic are not laws of thought; they express facts about the world.

Russell: "logic is concerned with the real world just as truly as zoology" (introduction to Mathematical Philosophy, p. 169)

The difference is just that logic is more general than other sciences.

Russell: "When we judge that two and two are four, we are not making a judgement about our thoughts, but about all actual or possible couples... Thus our a prior knowledge is applicable to whether the world may contain, both what is mental and what is non-mental"

Russell concludes that "all our a priori knowledge is concerned with entities which do not, properly speaking exist, either in the mental or in the physical world. These entities are such as can be named by parts of speech which are not substantives"

e.g. for "I am in this room" to be meaningful, it must be wholly composed of constituents of which I'm acquainted. "I" and "this room" does not pose any problem, but what about "in"?

Russell's theory of definite descriptions makes an essential, interminable use of predicates and relations. There is no way of analyzing these things away. So they have to designate something that we're truly acquainted with.

Russell accounts for these entities by appealing to an updated version of Plato's theory of forms.

The World of Universals

Predicate and relation expression designate universals (abstract objects).

e.g. "white" designates the predicate whiteness, "to the north of" designates a two-place relation which holds of pairs of objects.

Nominalism is the view that denies the existence of abstract object; there are only concrete particulars.

Russell thinks that previous attempts to maintain nominalism are guilty of ignoring relations (their importance came to the fore with advances in logic).

e.g. Russell: "If we wish to avoid the universals whiteness and triangularity ... to avoid the admission of such universals as whiteness and triangularity."

Russell thinks we're therefore committed to the reality of abstract objects. Such objects do not exist, however. They are nowhere to be found in space and time. They are also neither mental nor physical.

They merely subsist

4. Russell's solution to the problem of a priori knowledge

Russell takes universals to solve the problem of how a priori knowledge is possible.

"+" is a universal that applies not only to all observed pairs of objects, but to all existing and non-existing-but-possible pairs of objects.

Thus, "all a prirori knowledge deals exclusively with the relations of universals."

C.I. Lewis and pragmatic a priori

1. Account's we've looked at so far

2. Clarence Irving Lewis: general background

  • Lewis's dates: 1883-1964
  • Lewis like Russell, did serous work in the formal logic. Once of Lewis's contributions was to give the first systematic treament of some alternatives to classical logic. I'll have more to say about this shortly.
  • He also studied the philosophy of Kant, and was deeply influenced by the Josiah Royce.
  • For Royce, all aspect of reality, are ultimately to accounted for in terms of a single all-encompasing consciousness.
  • Lewis ultimately rejected this feature of Royce's view.
  • Lewis maintained that there's a fundamental distinction between 1) the given in experience, and 2) the concepts, by which we organize or interpret the given.
  • However, unlike Kant, Lewis did not believe that the way we categorize the data of sense experience is necessary and fixed for all time.
  • Actually, he does believe that what's a priori is necessary, but that necessary doesn't imply fixed for all time, or that there ain't any alternatives.

  • Lewis is led to this view, given the development of alternative geometries and alternative logics.
  1. Alternative logics:
  • strict implication, the diamond (possible) symbol
  1. Alternative geometries:
  • Now, history had already revealed that Kant's categories and forms of intuition were not necessary,

Part of the problem is that Kant's forms of intuition guarantee that space conform to Euclidean geometry.

However, Einstein's theory allows for space to be fundamentally "curved", via non-Euclidean geometries.

Einstein's theory also showed us an alternative to an absolute definition of "simultaneity"

So this left philosophers with three possible responses to this siutation:

  1. This shows that Kant got it all wrong
  2. Kant's basic insight was right (i.e. that the mind imposes a structure that's ncessary and fixed), but he misidentified what that is
  3. Kant is right that the mind imposes structure, but that what the structure is, is not fixe and can change over time.

Early on, Russell flirted with option #2. But, in PP, he defends option#1.

Lewis choose option #3.

3. Lewis's account of a priori knowledge

Something is a priori "not because it presecribes a from which that data of sense must fit, or anticipate some pre-established harmony of experience with the mind, but precisely because it prescribes nothing to experience."

  • Lewis is believes that all a priori knowledge is purely formal and empty of empirical content. It is true "come what may."
  • How is this different from Russell's conception of logic?
  • Logic, for Lewis, is purely formal and analytic.
  • Lewis's view here is influenced by Wittgenstein's analysis of propositional logic in his Tractatus (1921)

Lewis's proposition of ogic do not state "facts" about the world, but they are not therefore meaningless.

  • Thy're to be thought of as definitions, which we impose, or stipulate, onto the world.

Lewis: "what's a priori represents an attitude in some freely taken, a stipulation of the mind itself, and a stipulation which might be made some other way if it suited our bent of need."

The emphasis on our "our bent or need" dictating our choice is what makes it a pragmatic choice.

  • e.g. #1: the truth of the law of non-contradiction is secured by our bent for simplicity.
  • Lewis's newspaper paradox
  • We over-look such paradox, and happily maintain our belief in classical logic, for the purposes of "intellectual convenience."
  • Mill's proposed demon counter-example.
  • e.g. #2: the definition of simultaneity in physic's

The "constancy of the speed of light" principle "is in reality either a supposition nor a hypothesis about the physical nature of time, but a stipulation that I make of my own free will in order to arrive at a definition of simultaneity."

  • We cannot even "interrogate experience" without the mind posing some such definition, some classification, onto the world.
  • In fact, without this activity of the mind, "no growth in science, nor any science at all, would be conceivable."
  • e.g. #3 the definition of what counts as 'real'

why do classify some things as hallucinations" or "illusions" and then we could be led to revice our assumption that spirits can't be into each group. Those phenomena which violate laws of nature explains not only science and measurement, but also the very notation of fact.

But we could just as easily revise our law of nature, what we currently call an "illusion" is not longer classified as such.

e.g. if spiritual phenomena frequently showed up on photographs then we could be led to revises our assumption that spirits can't be photographs

In this way, the distinction between the mind and the world explains not only science and measurement, but also the very notion of a 'fact."

4. Objections to Lewis' view

  • i) Are the laws of logic really just a matter of stipulation?

e.g. the law of non-contradiction.

How plausible is to say that this is merely true because we define it to be so?

Thought-experiment. Imagine we are told by an alien race that they didn't believe in that law. It wasn't true for them. What would your response be?

  • ii) what does "true by stipulation" actually mean?

It taken literally, it wont' work. There are infinite number of logical truths. But there can only be finitely many stipulations.

We might try stipulate a finite number of logical axioms, from which all other truths of logic follow. But that won't work. Why?

If taken metaphorically, then what exactly is it a metaphor for?

  • iii) Is the view consistent?

On the one hand, there is a fundamental distinction between the mind and the world, in virtue of which knowledge, scientific progress, and facts are possible.

On other hand, we find Lewis saying things like "our categories and definitions are peculiarly social products", and that we all arrive at roughly the same decisions because of our common social history.

Kirpke: A Priori Knowledge, Necessity and Contingency

1. A priori vs Necessity

a. What does "a priori" mean?

  • That which can be known independently of experience?
  • "Can" means it is possible to know things in this way, not that they are known this way. But possible for whom? Humans? Martians? God?

Can't possibly be known empirically?

  • But this seems false. e.g. using a computing machine to determine whether such-and-such number is prime.

b. What does "necessary" mean?

  • True in all possible worlds.
  • Important point: "a priori" is an epistemological notion, whereas "necessity is a metaphysical notion."

c. The two do not mean the same thing.

  • e.g. Goldbach's Conjecture (that every even number greater than 2 is the sum of two primes): is neither necessarily true or necessarily false.

But, as of it right now, it's not known a priori. Moreover, for all we know, there might be no proof either way.

2. Analyticity and Certainty

  • a. analytic = true in virtue of meaning.
    • If follows that what's analytic is both a priori and necessary.
  • b. Something can be known a priori without being certain.
    • You've read a proof in a math book, but it's possible you've made a mistake.

  • 3. Rigid Designator
    • A designator is rigid if it designates the same thing in every possible world. E.g. Barack Obama.
    • A designator is non-rigid, or accidental, if it does not. E.g. the current president of the U.S.
    • A rigid designator of a necessary existent is called strongly rigid.
    • All proper names are rigid designators.
  • 4. Contingent a priori statements
    • Suppose we use the length of some particular stick to define what counts as one meter. Call it stick S.
    • Then, by stipulation, we know that S is one meter long. So it’s known a priori that stick S is one meter long.
    • But it’s not a necessary truth that stick S is one meter long, so long as “one meter” is regarded as rigid designator.
    • So, Kripke concludes, there can be contingent a priori truths.
  • 5. Necessary a posteriori truths
    • E.g. ‘The morning star = the evening star’ and ‘superman = Clark Kent’ are necessary truths.
    • Yet, these are both truths that are known a posteriori.
  • 6. Objections
    • Kripke makes essential uses of the notion of transworld identity to define what’s necessary.
    • But can we really know how to make sense of transworld identity? What laws are we holding fixed.


1. Quine, "Two Dogmas of Empiricism"

  • a) analytic/synthetic distinction
  • b) radical reductionism: the belief that each meaningful statement is equivalent to some logical construct upon terms which refer to immediate experience.

2. Analytic/synthetic distinction

  • What's Quine's argument? The notion of analyticity is unclear, and thus not a useful concept, from a rigious point of view.
  • Strategy: go through the most plausible attempts to explain analyticity, and undermine each

a) subject-predicate containment (or true by-dint-of-logic alone):

  1. "all unmarried males are unmarried males."
  2. "all bachelors are males"
  • need to know the meaning of terms in 2, but not 1.

b) True in virtue of meaning.

  • Relying on a notion of meaning, we can turn (2) into (1).
  • but is the notion of meaning any clearer than the notion of analyticity?
  • After all, when we make the subsitution of "unmarried males" in for "bachelor", we're setting aside all sorts of copmilcation:

e.g. the pope, my 6 year-old nephew, etc.

  • In other words, in order to gain mastery of a term's meaning, we need to have a command of contexts in which sentences using that term are uttered.
  • Thus, we're led to the following problem: the notion of meaning is no more clearer than the notion of analyticity.
  • So no useful analytic/synthetic distinction can be erected on this basis.

c) True by defintion: perhaps we can just go to dictionaries to determine the correct meanings and usages of terms in different contexts.

  • Problem 1: it's no clear that dictionaries capture all of the subtleties of native language use.
  • Problem 2: dictionaries preserve pre-existing synonymies, they don't create those synonymies.

e.g. suppose you found a dictionary which defined "human" as an "an eight-legged insect" What would your response be?

  • note: this strikes against the view that all so-called analytic truths can somehow be made true by an act of stipulation
  • note: sometimes we introduce notational abbreviations by stipulation. e.g. Let "p = John is a bachelor"
  • But that in that case, we're not appealing to pre-existing synonymies.

d) Interchangeability: two expressions are synonymous just in case they can be interchanged in all contexts without change of truth value.

Problem: this doesn't work in all cases

  • Consider:
  • (1) "Joe believes that all bachelors are bachelors"
  • (2) "Job believes that all bachelors are unmarried males"

also consider: (3) "Bachelor" has less than ten letters"

This is not synonymous with "Unmarried male has less than ten letters"!

e) Semantic rules: Perhaps the problem here is just with the vagueness of ordinary languages.

Consider a sentence like "all green things are extended". Is this analytic?

Ordinary languages doesn't seem to give a clear verdict. To overcome this ambiguity, one might introduce 'semantic rules' that give clear verdicts in all such cases.

  • e.g. we might add the rule that color predicates only be applied to things that are extended in space.

But what counts as "semantical rule"?

  • e.g. can we introduce "Kent's apartment is red" as a semantical rule? If so, then "Kent's apartment is red" would count as analytic.

But can we do this for anything.

So, without some antecendent sense of what those constraints are, we're no further ahead.

Thus, we're lead to the follow objection: The procedure of introducing semantical rules only helps if we already understand what analyticity means."

3. The other dogma: reductionism

The verification theory of meaning: "the meaning of a statement is the method of empirically confirming or disconfirming it."

"An analytic statement is the limiting case which is confirmed no matter what."

If the verification theory of meaning were an adequate account of statement synonymy, then the notion of analyticity would be saved.

Radical reductionism: every meaningful statement is translatable into a (true or false) statement about sense data.

Weaker version of reductionism: each meaningful statement, though not reducible to a statement about sense data, has a unique set of observational consequences.

Quine's view: "our statement about the external world face the tribunal of experience not individually but only as a corporate body."

The dogmas are connected. How?

If reductionism (even the weak version) is true, then we have a way of sharply distinguishing analytic statements from synthetic ones.

Empiricism without the dogmas.

What's Quine's positive conception? What imagery does he use?

  • Web of belief: draw diagram

Near the center of the web: "2+3=5" and "everything is self-identical"

Closer to the periphery: "there are brick house on Halsted Street"

Is this correct? Can you think of a belief that is absolutely unrevisable?

What's the difference between the two?

According to Quine, it's just a matter of degree, we're less likely to give up those beliefs closer to the center of our belief system because they're more entrenched; but no belief is immune from revision.

A recalcitrant experience impinges on the entire web (that is, on the entire body of science)

Two consequences of this view:

  1. No sharp analytic/synthetic distinction (and no corresponding distinction between a priori and a posteriori truths)
  2. Metaphysical question can no longer be sharply distinguished from scientific ones (note the gods of Homer remark).

How is our science, our talk of physical objects, like Homeric myth? How different? What is the point?


Carnap's response to Quine

1. Agree with Quine that

  • a) ordinary language is often vague
  • b) it doesn't determine whether or not "all green things" is analytic
  • c) the difficulty arises from the uncertainty of the word green (as used in ordinary language) not from the notion of analyticity.

2. The notion of analyticity has an exact definition one within the context of an exact language (supplemented with semantical rules), not within the context of ordinary language.

  • Exact, artificial language vs natural, ordinary language.
    • e.g. "a large omelet or 2 pancakes and hash browns."
    • e.g. "bachelor"
    • e.g. "green"

  • Explicandum: the thing to be explained
  • Elucidation explicandum: an initial demarcation of what it is that needs to be explained
  • Explicatum: that which does the explaining

Carnap accuses Quine of being unclear as to whether he's talking about the elucidation explicandum or explicatum.

  • If the former, then the demand for clear verdicts be given in all cases is unreasonable, since ordinary language is not precise enough to do that.
    • Ordinary language yields 'an understanding of analyticity, in practice clear enough for application in many cases, but not exact enough for other case for theoretical purposes."
  • If the later, then a clear verdict can be given, in terms of an exact language.

3. Quinean objection

Using an artificial language to explain analyticity only make sense if we can give a definition of analyticity that applies to all all languages whatsoever.

  • Carnap's response: This demand is "manifestly unreasonable."

What is Carnap's view here?

  • That the desired definition of analyticity will inevitably be language-relative.

4. Quinean objection

But the choice of semantical rules then seems completely arbitrary then. What makes one definition better than any other?

  • Carnap's response: The choice of rules is not arbitrary, because "we advance the claim that the defined concept embraces what philosophers have meant, intuitively but not exactly, when they speak of "analytic sentences" or, more specifically, of "sentences whose truth depends on their meanings alone and is thus independent of the contingency of facts.'

What's Carnap's view here.

  • He's rejecting the idea that there is any language-transcendent fact about whether "all green things are extended" is analytic or not.

5. Quinean objection

No statement or rule in the systems is immune from revision, both (synthetic and analytic sentences) can be revised in the light of experience.

  • Carnap's response: But there's a fundamental distinction between rules that define a language, on the one hand, and sentences which can be empirically confirmed (or discomfirmed) relative to that choice of language, on the other.

Both can be revised, but only the latter can be said to be empirically confirmed (or disconfirmed)


Coherence Theory of Empirical Justification (CTEK)

1. Bonjour's theory is advanced as an overlooked response to the regress problem:

  • P1. Either there are justified basic beliefs or each justified belief has an evidential chain that:
    • (a) terminates in an unjustified belief,
    • (b) is an infinite regress,
    • (c) is circular
  • P2. But beliefs based on unjustified beliefs are not themselves justified, ruling out (a)
  • P3. No person can have an infinite chain of beliefs, ruling out (b)
  • P4. We don't think beliefs can be justified via circular reasoning, ruling out (c)
  • C. There are at least some basic beliefs.

Bonjour is essentially responding to this argument by defending option (c) as a plausible alternative to foundationalism

2. CTEK: three main theses:

  • 1. All justification for individual empirical beliefs is inferential, i.e. there justification depends on other beliefs.
  • 2. The regress of justification does not go on forever, but, rather, circles back upon itself.
  • 3. The "primary unit" of justification is not individual beliefs, but a system of beliefs, which is justified in terms of its "internal coherence."

3. The main motivation for this CTEK

The conviction that all foundationalist views are "untenable".

  • Bongour is focusing here on foundationalist views which assert that there's a "given" in experience, e.g. that with which we're immediately acquainted.

Key claim: All beliefs about the world justified via experience could be mistaken.

At most, our experience provides a reason to believe that something is the case.

4. Responding to the Regress Problem

The foundationalist may object that CTEK leads to a vicious circularity: the justification of each beliefs will end up depending on its own logically prior justification.

But this objection rests on the assumption that inferential justification must be linear, i.e. "a linear sequence of beliefs along which warrant is transferred from the earlier beliefs in the sequence to later beliefs via the connections of inference."

CTEK distinguishes two levels at which justification takes place:

  • 1. local justification: "the justification of a particular belief, or a small set of beliefs, in the context of a cognitive system whose overall justification is taken for granted."
  • 2. global justification: "the justification of the cognitive system itself."

  • Justification can look linear if we're focusing on local justification.
  • Justification is not linear, if we're focusing on global justification.

More precisely, there are 4 steps to justification:

  • 1. local justification of a particular belief
  • 2. the coherence of the overall system of beliefs
  • 3. the justification of the overall system of beliefs
  • 4. the justification of a particular belief, by virtue of its membership in the system

The linear conception of justification arises from a neglect of steps 2 and 3.

5 Further Clarification and Responses

Coherence vs Consistency: coherence requires consistency, but not vice versa; coherence has to do with 'systemic connections between the components of a system, not just with their failure to conflict."

Some standard objections to coherentist theories:

  1. Coherenece will never suffice to pick out one system of beliefs
  2. CTEK does not allow for any input from the world, and so can't be the basis for empirical knowledge. (e.g. the man who mistook his wife for a hat)
  3. CTEK is committed to a coherence theory of truth, which is absurd.

Objection 2 confuses two different ways in which a belief may be said to be inferential:

  • a) regarding the belief's origin (causal claim)
  • b) regarding how the belief is justified or warranted (epistemic claim)

CTEK allows that beliefs may be non-inferential in sense a), which is all that's needed to respond to objection 2.

To say that a belief is non-inferential in sense a) is to say that it's cognitively spontaneous, i.e. "it is not arrived at via any sort of conscious [rational] process, but simply occurs to me, strikes me, in a coercive manner over which I have no control."

Moreover, saying that a belief is non-inferential in sense a) is consistent with saying that it's inferential sense b).

  • e.g. consider a visual belief. It's non-inferential in sense a).

But its justification can nonetheless make reference to other beliefs (i.e. beliefs about the lighting, viewing conditions, the functioning of your eyes, etc.)

The three elements of observational knowledge:

  • a) a process of some sort which produces cognitively spontaneous beliefs.
  • b) the beliefs thus produced must be "reliable with respect to the subject matter" in two respects:
    • i) it's very likely that such beliefs, when produced, and true (given that certain conditions are met)
    • ii) it's very likely that the beliefs in question would be produced, if it's true, will be produced (given the certain conditions are met)
  • c) the person must know all of these things, if only in a "rough and ready" way.


Pragmatism, Relativism, and Irrationality

1. Background

2. Basic view

  • He's using the "American Pragmatists" to motivate a critique of the essence of knowledge.
  • The problem with talk of "essences", for Rorty, is that it seems to presuppose a perspective that we don't have.
  • For Rorty, all inquiry takes places with an historically situated, context, and we don't have access to how these really are.
  • We only have access to how things seem from some point of view or perspective.
  • "There is no wholesale, epistemological, way to direct, or criticize, or underwrite, the course of inquiry."

"Let me sum up by offering a third and final characterization of pragmatism: it is the doctrine that there are no constraints on inquiry save conversational ones - no wholesale constraints derived from the nature of the objects, or of the mind, or of language, but only those retail constraints provided by the remarks of our fellow-inquirers."

  • What does he mean here?
  • We have no access to how things really are independent of our historically-situated perspective.
  • We'd like to be able to say that our theories should conform to reality, that they should be responsive to the nature of objects, of the mind, or of language.
  • But Rorty wants to dismiss talk of the nature of objects, mind and language as metaphors that attempts to refer to something that we can never have any access to or knowledge of.
  • More generally, Rorty is rejecting the idea that there are constraints that will tell us when we've reached the "truth" (in the philosopher's sense of the term).
  • "...we can make no sense of the notion that the view which can survive all objections might be false. But objections - conversational constraints - cannot be anticipated. There is no method for knowing when one has reached the truth, or when one is closer to it than before."

3. Relativism and Irrationalism

  • He's not endorsing the view that all viewpoints are equally correct.
  • Rather, the view is that all constraints on inquiry are historically and culturally situated, and any critique of those constraints can only happen within the historically-situated inquiry that we find ourselves engaged in.
  • There's just no transcendent perspective, from which we can compare our inquiry with how things really are, as a check to see if we're getting things right, or if we're on our way to formulating a true theory of the world.
  • In effect, Rorty is asking to give up on traditional epistemology.
  • He singles out foundationalism, the view that there are foundations to knowledge. But his criticism applies just as much to a coherentist view.
  • So the upshot of Rorty's view is this: the attempt to articulate a global account of knowledge, of what knowledge is, or of what constitues knowledge, is completely misguided.
  • But that doesn't mean that we have to give up on all talk of "knowledge", "truth", "rationality", etc
  • It's just that those terms, if they're not going to be mere metaphors, have to be cashed out in terms of concrete details in which those terms were first defined and developed.
  • But we run into trouble when we try to formulate a wholesale, global theory of these notions.
  • Rorty is also attacking what he calls "irrationalism", the view that there are no rational standards for assessing claims to knowledge (say, because all claims to knowledge are just culturally relative utterances).
  • Roty is objecting to any global characterization of the "essence" of knowledge.

4. Objections

  • 1. Where's the argument for this view? Indeed, Rorty himself denies that he's provided an argument for the view he's espousing. But then why should we accept his view?
  • 2. In particular, why think a global account of justification or knowledge commits us to transcendent perspective (of the bad sort)?
  • Consider the following argument:
    • P1. Everything I do is motivated by beliefs and desires that I have.
    • C. Everything I do is merely self-interested.
  • This is an invalid argument; it confuses the origin of an action with the purpose of an action.

  • Similarily:
    • P1. All standards of inquiry arise from a given culture.
    • C. All standards of inquiry are culturally relative.
  • Why does the origins of our standards of inquiry prevent us from formulating standards that could plausibly taken to apply to any rational being?

  • 3. Who's Rorty attacking? His interpretation of Kant and others in the philosophical tradition are highly questionable.

David Annis

Contextualist Theory of Epistemic Justification"

1. The basic view

  • Justification = a person being able to meet certain objections
  • Not all possible objections, but just some set of reasonable objections.
  • Which objection are reasonable? It depends on the person's goals.
  • Board: the goals are epistemic, e.g. having true beliefs and avoiding false beliefs.
  • Other possible epistemic goals: simplicity, conservation of existing beliefs, maximization of explanatory power, etc.
  • Other possible non-epistemic goals; making lots of money, being well-liked, etc.

Two classes of objections to someone (S) knowing some statement (h) is true:

  • A) that S is not in a position to know that h.
  • B) that h is false.

We don't always require reasons to support our beliefs.

  • e.g. if S claim to see a brown book across the room.

It's not reasonable to object, in this context, that we can think of a possible scenario where S's perception is misleading.\

  • All that's required is that S be able to answer current objections that arise out of his or her present situation.
  • That's consistent with the possibility of new evidence coming in, which gives us reason to doubt some assumed feature of the present situation.
    • e.g. we learn that S was under the influence of a psychedelic drug all the time.
  • In that case, we're in a new epistemic position, from which a different set of objections now count as reasonble.
  • "Thus for S to be held accountable for answering an objection, it must be manifestation of a real doubt where the doubt is occasioned by a real life situation... S is not required to respond to an objection if in general it would be assigned a low probability by the people questioning it."

  • Global objections: call into question our entire system of beliefs or a large portion of them.
  • Local objections: only call into question some specific belief.

Annis is not ruling out global objections.

  • e.g. if w all started to have an increasing number of false visual beliefs, as a result of nuclear radiation from a third world war, then we might have reason to raise a global objection to a vast tract of perceptual beliefs.

2. The social nature of justification

  • Reasonable objections to S's justification in believing h are determined in relation to an issue-context.
    • e.g. 1. If we asking a non-medically trained person whether polio is caused by a virus
      • Now suppose we ask the same question in the context of an examination of a M.D. degree.
    • e.g. 2. Suppose the issue is whether there's a link between some disease y and the presence of a certain chemical x.
      • So, to test that hypothesis, we administer a drug (which is known to supprse the presence of chemical x) on animals which have disease y.
      • Now, suppose we want to know whether this drug will cure the disease in humans without harmful effects.
      • The evidence required in this case would be more demanding.

Thus, the issue-context determines the goal of inquiry, as well as the level of understanding and knowledge required to justify a belief, relative to that goal.

  • Social information - the beliefs and information of others - also plays an important role here.
  • In particular, "actual social practices and norms of justification of a culture or a community of people" play an important role.
  • These practices and norms of determine the level of understanding and knowledge required in a given context.
  • It also determines which objections are reasonable in a given context
  • Thus, for S to be justified in believing h, he or she only needs to meet those objections which satisfy the practices and norms of the relevant community of inquirers.
  • Annis refers to this relevant community as the "objector-group", which will be a subset of the social community to which S belongs.
  • This is also a naturalized theory of justification.

3 The regress argument

  • Annis: the contextualist view offers a response to the regress problem, a response which arguably avoids collapsing into either foundationalism or coherentism.
  • e.g. suppose the Joeses are looking for a red chair to replace a broken one in their house
    • To justify his belief that a given chair is red, with the given issue-context, it's enough that they can point to a red chair a few feet away and say "that's a red chair.'
  • Thus, in the above issue-context, perceptual beliefs like "that's a red chair" are contextually basic, i.e. the appropriate objector-group does not require that S have reasons for the belief in order to be in a position to have knowledge.

Applied to the regress problem:

  • 1. The contextualist with the coherentist: there is an end to the regress of justification, so justification isn't circular.
  • 2. The contextualist also disagrees with foundationalist: these basic beliefs are only contextually basic, and so not basic in the foundationalist's sense.

4. Objection to contextualist theory

Is this really an alternative to foundationalism and coherentism?

  • 1. Is this really different than foundationalist?
  • 2. Is this really different from coherentism?
  • 3. Foundationalism and Coherentism seem to come apart in relation to the way that each theory deal with the issue of global justification.

Is this contextualist attempting straddle the fence by not taking a stand on this issue?


1. Intuition

  • "I am frightened, because he looks so threatening"
  • What's going on here?
  • Russell: We can justify some claims by an appeal to "intuition"
  • Remember that, for Russel, some beliefs are "basic" because of their "intuitive" evidence.
  • What point is Wittgenstein making here?
  • "Calling something 'the cause' is like pointing and saying: 'He's to blame!'"
  • We don't posit a faculty of "intuition" here. It's just how instinctive react, in certain circumstances.
  • It's part of what Wittgenstein calls a "language game."

Thus, the exclamation "he's to blame!" is part of praise-blame language game, and "I am frightened, because he looks so threatening" is part of a cause-effect language game.

  • W is opposing here is the idea that we have a special faculty of "intuition" that justifies our ordinary beliefs about the world.
  • The justification of a claim, on W's view, is determined in relation to the rules of a language game.

2. Doubt

  • "That's a chair"
    • Could you be mistaken about that? Yes
  • But we justify doubt in relation to other beliefs that we're not actively doubting.
  • That is, doubt seems to be defined in terms of the "rules" of a language game which aren't in doubt.
  • Can there be a language game in which we begin by doubting all the rules of that language game?
    • What would that even look like?

  • W's point seems to be: doubt is a local phenomena that is defined in terms of a background of beliefs that we take for granted.
  • From this point of view, the global skepticism that Descartes and Russell took seriously is not really a form of doubt at all; it's just an under-described language game.
  • However, W is not agreeing with Moore here that we therefore have a simple "proof" of external world.
  • Such "knowledge" is not something that one is ever taught, or finds out, or proves. It is simply part of the background against which we come to know other things.
  • Of course, we can imagine a set of circumstances in which we would doubt the existence of things like hands.
  • But W's point is that, if we spell out those circumstances in detail, we find that it's not a form of global skepticism; rather, it's just a different language game, in which different background of beliefs are taken for granted.

3. Coherentism or Foundationalism?

  • What's W's view? Coherentism? Foundationalism?
  • We seems to be rejecting the global account of knowledge that both of these views are attempting to give.
  • All justification is local, relative to a background of beliefs that are taken for granted.
  • There seems to be no way to even formulate questions of global justification, or the regress problem, on W's view.

Plato and Gettier

1. Plato's account of knowledge

- Meno dialog.

  • What is Plato's point here?
    • That true opinion is on par with knowledge as a guide to action.
    • However, Plato nonetheless sees an importance between two:

- Meno dialog. (cont.)

  • What's Plato's point here?
    • That there's an important difference between knowledge and true opinion; the former is "tied down."
    • There's some debate about what Plato means by being "tied down" or "tethered" by what he called aitias logismos (the calculation of reason, or reasoned explanation.)

So then knowledge = true opinion + a reasoned explanation to why the belief's true.

  • This is often put as follows: knowledge = justified true belief

2. Gettier: "Is Jusitified True Belief Knowledge?"

As the title suggests, the paper challenges the view that knowledge = justified true belief

Case 1. Suppose Smith has strong evidence for the following conjunction:

  • (d) Jones is the man who will get the job, and Jones has 10 coins in his pocket
    • So, Smith is justified in believing (d).
  • But (d) entails (e): the man who will get the job has 10 coins in his pocket.
    • So if Smith is justified in believing (d), he's also justified in believing (e).
  • But imagine that, unknown to Smith, it's he, and no Jones who will get the job. Moreover, unknown to Smith, he also has 10 coins in his pocket.
    • Proposition (e) then remains true. And it was already stipulated that Smith had good reason to believe it's true.
  • So, if knowledge = justified true belief, then Smith knows (e).

But it's clear that he doesn't know (e), since (e) is true in virtue of the number of coins in Smith's pocket, and Smith has no idea how many coins are in his pocket.

Case 2. Smith has strong evidence for the following proposition:

  • (f) Jones owns a Ford
    • So then he has also strong evidence to be the following disjunction:
  • (h) Either Jones own a Ford, or Brown is in Barcelona.
    • (even though he has no idea where Brown is; Smith picked the city of Barcelona at random)
  • But now imagine that Jones doesn't actually own a Ford, but is at present driving a Ford rental car.
    • Moreover, by the sheerest coincidence, and entirely unknown to Smith, Brown is in Barcelona.
  • Then(h) is true, and Smith is justified in believing it, but he doesn't know it.
  • Thus, justified true belief is not sufficient for knowledge.

Goldman and Olsson, "Reliabilism and the Value of Knowledge"

1. Some possible responses to Gettier-style counter examples

  • J*TB
  • JTB + X
    • Internalism
    • Externalism - Reliabilism.
  • TB

2. Two sense of "knowledge"

  • Knowledge in the weak sense = not being ignorant of the truth
  • Knowledge in this sens is not any more valuable than mere true belief.
  • Knowledge in the strong sense = true belief + something else that affords it more value than mere true belief
  • Specifying what this "something else" is, and how it adds value, is called the (EVOC) problem.

3. Goldman defens a view called process reliabilism:

  • S knows that p iff:
  • (1) p is true,
  • (2) S believes p to be true,
  • (3) S's belief that p was produced through a reliable process, and
  • (4) a suitable anti-Gettier clause is satisfied

Note that this view is incompatible with Plato's conception of knowledge, according to which S knows that p only if S can present the reasons why p is true.

  • e.g. #1: a child who knows that the table in front of him is red, or that his parents love him.
  • e.g. #2: a dog that knows that there's food in its bowl.
  • Do the child and the dog need to be able to "present reasons" in order to know these things?
  • According to the classic reliabilist view, these beliefs may be count as knowledge in virtue of being caused by reliable processes.
  • A process is reliable just in case it generally produces true beliefs
    • e.g. the child's belief about the color of the table is produced by a visual process that, in fact, tends to produce true beliefs.

4. Internalism vs Externalism

More generally, reliablilism is a form of externalism (which is opposed to internalism) about knowledge:

  • Internalism: knowledge requires justification and the nature of this jsutification is completely determined by a subject's internal states of or reasons.
  • Externalism: either knowledge does not required justification or the nature of justification is not completely determined by internal factors alone.

On virtue of externalism is that seems to get the child and dog cases correct.

Thus, a move that Goldman makes early in on is to distinguish two different forms of justification: weak and strong.

  • Strong justification corresponds roughly to what the internalist wants.
  • Weak justification is satisfied when the beliefs is produced by a reliable process.
  • Other forms of externalism: proper function theory (e.g. Plantinga)

e.g. S knows that p if that belief is produced by cognitive faculties that are 'functioning properly.'

6. The Swamping Objection to Reliabilism: the value of reliably-produced, true beliefs is no more than just true beliefs.

  • (S1) Knowledge equals reliably produced true belief (simple reliabilism)
  • (S2) If a given belief is true, its value will not be raised by the fact that it was reliably produced.
  • (S3) Hence: knowledge is no more valuable than unreliably produced true belief.
    • e.g. if a work of art is beautiful, its aesthetic value is not enhanced by adding that "it also has the properly of being likely to be beautiful."

Goldman rejects both (S1) and (S2)

  • He rejects (S1) by adding in an anti-Gettier clause.
  • He rejects (S2) by exploring a few different answers that the reliabilist can give to the EVOC problem.

7. The conditional probability solution to the EVOC problem.

  • Reliabilism affords extra-value, by way of the "property of making it likely that one's future beliefs of a similar kind will also be true."
  • e.g. #1: having a reliable expresso-making machine is more valuable than a machine that just happened to make a good expresso in one case, because it's more likely to produced good expressos on other occasions.
  • e.g. #2: having a reliable navigation system is more valuable than a system that just happened to pointed you in the right direction to Larissa, because it's more likely to point you in the right direction on future occasison.
  • But this solution depends on some empirical regularities being in place: non-uniqueness, cross-temporal access, learning , and generality, respectively.

In other words, "the conditional probability solution explains why reliabilist knwoledge is normally but not always more valuable than mere true beliefs."

8. Value Optimization:

Type/token distinction: e.g. distinguish the type of belief-forming process from a given instance (token) of that process.

  • It's certainly true that a token reliable process doesn't add any value to a true belief, since its value can be traced to the token true belief that it causes.
  • So then the question is: does having a true belief caused by a reliable process type add value over and above just having the true belief?

Intrinsic/Instrumental distinction

  • Some types of things have just instrumental value (e.g. aspirin)
  • Value optimization thesis: in some cases, a type of state (or process) may initially only (type-) instrumental value, but eventually acquire independent, or autonomous, value.

e.g. having good motives.

In particular, learning that a true belief was caused by a reliable process type brings some value to the table, over and above the value of having a true belief. More precisely: reliable belief forming process may initially only have instrumental value (i.e. their value is traceable to the true beliefs they produce), but they can come to be viewed as intrinsically valuable over time.

Pollock, "The Gettier Problem"

1. Initial response to Gettier's counter-examples: Smith's justification is inferred from a false belief.

  • Reply: Barn Facade example

2. Pollock's proposal: there is no true proposition Q such that if Q were added to S's beliefs than he would no longer be justified in believing that P.

  • In other words, there is not true proposition that, if known, would overturn, or defeat, his justification for believing that P.
  • But this won't do, since there could be propositions Q which satisfy Pollock's proposal but which don't defeat S's claim to knowledge.

Grabit example

3. Two sense of "should believe"

  • a) subjective sense: concerns what we should believe given what we actually do believe.
  • b) objective sense: concerns what we should believe given what is in fact true

4. Pollock's account of knowledge

  • Objective epistemic justification = (roughly) getting everything right while doing everything right. i.e.
    • a) S is subjectively justified in believing that P.
    • b) there is a set X of truth, such that from that vantage point, S's justification would not be overturned; and moreover, there is no more inclusive set Y of truths such that, from that vantage point, S's justification would be overturned.

Objective epistemic justification is a necessary condition for having knowledge.

Internalism and Evidentialism

1. Evidentialism

"the epistemic justification of a belief is determined by the quality of evidence for that belief"

  • Notice that this view is incompatible with externalism

2. Clifford

  • Clifford: 'it is always wrong, everywhere, and for anyone, to believe anything upon insufficient evidence"
    • Why might we think this view is true?
  • Give Clifford's Ship Example:

3. Pascal's Wager

4. Hospital Wager

5. William James

1. Terminology

  • Hypothesis: something that may be believed
  • Option: a decision between two hypotheses.
  • Living option: a decision between two live hypothesis
  • Live hypothesis: something that is a real candidate for belief. A hypothesis is live, we might say, for a person just in case that person lacks compelling evidence disconfirming that hypothesis, and the hypothesis has an intuitive appeal for that person.
  • Momentous option: the option may never again present itself, or the decision cannot be easily reversed, or something of importance hangs on the choice. It is not a trivial matter
  • Forced option: the decision cannot be avoided - the consequences of refusing to decide are the same as actually deciding for one of the alternative hypothese
  • Genuine option: one that's living, momentous, and forced.
  • Intellectually open: neither evidence nor argument decide the issue.

2. The main argument might be sketched as follows:

  • P1. Two alternative intellectual strategies are available:
    • Strategy A: Risk losing out on a possible (but unproven) truth/good in order to avoid error at all costs (Clifford)
    • Strategy B: Risk error for a chance at a possible truth/good (ames)
  • P2. The Agnostic (using Clifford's Rule) embodies Strategy A, and the Theist embodies Strategy B. But,
  • P3. Whenever we're dealing with a genuine option that's intellectually open (or a "GO" for short), Strategy B is at least as justifiable as Strategy A.

    • What are some contexts in which James thinks that strategy A is preferable?
      • Scientific and legal.
    • What are some contexts in which James thinks that strategy B is preferable?
      • Whether to believe in God, whether to act ethically, whether someone will reciprocate your live.
      • What is the religious option, according to James? Whether to believe (1) that the best things are eternal and (2) that we're better off now if we believe this to be true.
      • So then let's define God as a being who guarantees the truth of (1) and (2)
  • P4. The decision of whether to believe or not in God is a GO.
  • C. The Theist is at least as justified as the Agnostic.

  • Is the argument valid?
  • Is it compelling?

  • Objections:
    • 1. to (P1): There's a third option: accept A for pragmatic purpose (i.e. act as if it's true), but with hold full belief (i.e. remain agnostic about whether it really is true). Is James confused action with belief?
    • 2. to (P1) again: it assumes voluntarism, i.e. that we can choose our beliefs. But can we really choose our belief? e.g can I choose to believe that the board behind me is purple rather than white.
    • 3. to (P3): Strategy B could lead to carelessness or harm to others. Has James adequately dealt with this sort of worry?

Stephen Stitch

  • Reflective Equilibrium, Analytic Epistemology and the Problem of Cognitive Diversity

1. Outline

2. Cognition and cognitive diversity

  • As far as we know, belief acquisition may be like language acquisition in two respects:
    • 1. Their acquisition may be deeply dependent on environmental factors
    • 2. They may differ in radical ways from individual to individual, and from culture to culture.
  • If 1 and 2 are true, then we get the problem of cognitive diversity: if different cultures reason in different ways, which of these way should we prefer?

3. Possible criterion: Reflective Equilibrium

  • Reflective equilibrium: the process of bringing judgments about particular inferences and about general principles of inference into harmony with one another.
    • e.g. what justifies our judgment that killing innocent people is wrong?

Similarly: What justifies our belief in the principle of induction?

Two possible interpretations of reflective equilibrium:

  • 1. the reflective equilibrium test is constitutive of justification
  • 2. passive the reflective equilibrium tests provides evidence in favor of the judgments or principles in question being justified.

If 1, then what kind of truth is it?

  • a) a conceptual truth, i.e. it follows from the meaning of "justification"?
  • b) a non-conceptual necessary truth that's discovered a posteriori? (like water =H2O)
  • c) it's a stipulative proposal, i.e. it's not trying to analyze a pre-existing concept of justification, but proposing a new notion?

4. Does reflective equilibrium capture our notion of justification?

  • There are patently unacceptable rules of inference that pass the reflective equilibrium test
    • e.g. the gambler's fallacy

But few of us would be willing to say that if the gambler's fallacy is in reflective equilibrium for a person, then her inferences that accord with that principle are justified

  • Possible responses in defense of reflective equilibrium
    • 1. Insist that the gambler's fallacy is justified for that person
    • 2. The notion of reflective equilibrium needs to be refined.
      • a) Broaden the scope of judgments that are taken into account (via "wide reflective equilibrium")
      • b) Restrict the class of people whose equilibrium is to count in assessing the justification of an inferential principle (via "expert reflective equilibrium")

5. Neo-Goodmanian Project

  • So Goodman's reflective equilibrium test doesn't quite work.
  • but that just means we have to work harder.
  • To be justified just is to pass the tests that we invoke in assessing an inferential practice.
  • So all we need to do is to figure out what those tests are, and then we'll have figured out justification amounts to.

6. Some questionable presuppositions of the Neo-Goodmannian Project

  • 1. Justification make turn to look nothing like reflective equilibrium.
  • 2. There's only one notion of justification
  • 3. This is a coherent notion for which a set of necessary and sufficient conditions can be given.
  • 4. The test or procedure we use for assessing the justification of cognitive processes exhausted our concept of justification (i.e. our concept might not match any test/procedure)

7. Against Analytic Epistemology

  • Different theories on the table: evidentialism, reliablilism, neo-Goodmannian theory, coherentism, foundationalism, etc
  • Each give a difference account of the J-rules, i.e. the set of necessary and sufficient conditions that define "justification"

e.g. coherence theories judge the rightness of a set of J-rules by whether conformity with those rules leads to a coherent system of systems.

  • All such attempts fall within a general project of "conceptual analysis" or "conceptual explication"

Key contention: "Yet surely the evaluative practices embedded in everyday thought and language are every bit as likely as the cognitive processes they evaluate to culturally acquired and to vary from culture to culture."

  • Positive proposal: step back to acknowledge the possibility in the diversity of different notation of justification.
  • Still evaluate against background non-epistemology values.

Russell: The Limits of Philosophical Knowledge

1. Thesis:

  • "knowledge concerning the universe as a whole cannot be obtained by metaphysics."

2. Hegel

  • Russell uses Hegel (1770-1831) - or rather, a particular interpretation of Hegel - to articulate the view that philosophy can provide deep insights into the nature of reality.
  • We start with how things appear to us, which is only a partial view on reality as a whole.
  • Philosophy then gets to examine these appearances and point to various contradictions or paradoxes ("thesis" and "antithesis"), which results from the "incompleteness" of the point of view we've taken on reality.
  • To escape such contradiction, we have to find new, less incomplete ideas (issuing from a less "partial" or "incomplete" perspective), which allow us to resolve the paradox (i.e. 'synthesize" the original idea and its antithesis)
  • On this picture, the goal of philosophy is to reach a complete, paradox-free theory of reality (i.e. "Absolute Reality")

3. Russell's analysis of this view:

  • It rests on the assumption that whatever has relations to things outside of itself (i.e. has a nature that is defined in reference to other things) is self-contradictory.
    • e.g. a man's nature is constituted by his memories, his emotions, and everything else that is true about him.
  • Thus, without these other things, he could not exists, and so he's not self-subsistent (and, in that sense, not "real")
  • Thus, to regard him as "real" world be self-contradictory.

4. Russell's criticism of this view:

  • Not a truths about a thing are part of the "nature" of that thing.
    • e.g. that truth that Kent wrote black socks today is not part of Kent's "nature".
  • Moreover, we can have knowledge of a thing without knowing all the truths about that thing.
  • This is because:
    • 1. acquaintance with a thing does not involve knowledge of its relations
    • 2. knowledge of some of its relations does not entail knowledge of all of its relations or all of the truths about that thing.

5. Russell's positive view:

  • 1. Logic define what's contradictory
  • 2. Bug logic can't tell us what exists, it only tells us what's possible
  • 3. All knowledge as to what exists is based on experience, either through acquaintance or by descriptions.
  • 4. All knowledge as to what exists is, therefore, essentially fragmentary and incomplete
  • 5. Philosophy does not have any special perspective, or vantage point, from which to obtain knowledge as to what exists
  • 6. The difference between philosophy and science is simply a matter of degree: philosophy is more focused on critically examining some principles that are assumed in daily life and the sciences.

Quine: "Epistemology Naturalized"

1. Foundations of Mathematics (the epistemology of so-called analytic truths)

  • conceptual side: every concept of mathematics is reducible to a concept of logic
  • doctrinal side: every truth of mathematics is reducible to a truth of logic

2. Empirical knowledge (the epistemology of so-called synthetic truths)

  • conceptual side: every meaningful empirical concept is reducible to sense data
  • doctrine side: every empirical truth is reducible to truths about sense data
  • contextual definition vs eliminative/reductive definitions.
    • e.g. "bachelor" = "unmarried male"
    • A contextual definition: every meaningful statement containing an empirical concept is reducible to a statement that's only about some or other sense data.
  • a. On the doctrinal side: reduction of math to logic, plus set theory (at best). And the set theory is *not* logic.
  • b. On the conceptual side: concerned with definition, conceptual clarity, etc

But the reduction is only to logic, plus set theory (at best), and set theory is no more clearer than the concepts we're trying to explain.

  • e.g. axiom of choice
  • further issue: contextual definition

So then what's the point of "rational reconstructions" that make use of such contextual definitions, if they can only ever yield a partial analysis of the meaning of many empirical concepts?

To hold onto the idea that there must be *some* link between all meaningful, empirical concepts and experience.

4. Dual Containment Thesis

Psychology is contained in epistemology, and vice versa.

  • What does this mean exactly?
  • Quine's more general view: philosophy then becomes a proper part - albeit a peculiarly abstract and general part - of the natural, empirical sciences.

5. Objections to Quine's View

  • a) Where should we do this "naturalizing"?
    • Why psychology, why not sociology?
  • b) What happens to normative concepts like justified or rationally acceptable?


  • SFU PHIL201 - Spring 2011
  • Human Knowledge Classical and Contemporary Approaches 3rd Edition, Paul K. Moser and Arnold Vander Nat.
  • The Problems of Philosophy, Bertrand Russell