Phil 2233 Notes, Fall 2005: I. KNOWLEDGE OF THE PAST

advertisement
Phil 2233 Notes, Fall 2005:
I. KNOWLEDGE OF THE PAST
1. A skeptical argument:
A. The nature of knowledge.
What we know is normally expressed in terms of declarative
sentences, often set in a “that” clause:
I know (that) Jean Chretien is Prime Minister.
I know (that) the room we’re in is TH 103.
Tony Blair knows (that) his government exaggerated the
evidence of Iraqi WMDs.
The “that” clause indicates what the subject of the sentence
knows (sometimes we say it specifies the content of what’s
known).
Knowledge must meet three conditions:
i.
Belief: To know something, you must believe it—if you
don’t believe these things, you don’t know them.
(Sometimes, when the evidence is finally in, someone will
say “I just knew it,” when in fact they weren’t at all sure in
advance. But this is not literally true—they didn’t know
(perhaps they suspected, or feared, or hoped; but none of
these states of mind constitutes knowledge).
ii.
Truth: We cannot know what isn’t true. We can believe
things that are false, of course, and be very confident
about them, too (and sometimes people express such
confidence by saying that they know), but in the end, if it
isn’t true, we don’t know it, no matter how confident we
are about it.
iii.
Justification: We cannot know something just by guessing
at the truth and getting lucky. We need to have good
(sufficient) reason for our beliefs before they can count as
knowledge.
These three conditions are required, but not sufficient for
knowledge, as a very interesting class of examples illustrates. But
our argument doesn’t need a definition of knowledge—all it
requires is some details about the justification condition.
Justification, it is often said, cannot be circular. That is, if I know
that p (for some declarative sentence p), then I need a justification
for p; last time we described this justification as some further
declarative sentences that I know and that constitute(s) a sufficient
justification for p. Call these sentences q, q',… Then my
justification for p is circular if my justification for one or more of
q, q',… depends on being justified in believing p.
The problem is obvious—if my justification of p depends on my
being justified in believing q, and my justification of q depends on
my being justified in believing q, then both justifications are in
question. I haven’t really justified either yet.
B. Evidence about the past.
There are all kinds of things we claim to know about the past; if
we’re right about these claims, there must be some way of
justifying them.
Consider a few things I know about the past:
I had a bagel for breakfast this morning.
John F. Kennedy was assassinated in 1963.
Aristotle taught Alexander the Great.
The Sumerians developed the cuneiform system of writing.
Glaciers spread across northern Europe and North America
within the last 20,000 years.
Trilobites were arthropods that thrived during the Paleozoic era.
How do I justify these claims? The answers are obvious at
first—I remember eating the bagel, and I remember JFK’s
assassination and that it took place during the year I was in
grade 2. I’ve read about Aristotle and his place in the
Macedonian court, which is well-known to ancient history types
and well documented. Similarly, I’ve read about cuneiform
writing (and its roots in simple accounting techniques). I’ve
both directly seen evidence of glaciation (striations, polished
pavements, kettle lakes, moraines, drumlins, etc.) and read a lot
about the evidence, its history, and the conclusions geologists
have drawn from it. Finally, I’ve seen fossils of trilobites and
read about them as well; their place in the geological column is
secure, and their status as arthropods is likewise firmly
established.
Is there anything in common amongst these justifications?
Here’s something—they all depend on assumptions about
processes linking the present situation to some fact(s) about the
past:
When I talk about what I remember, I assume that there is some
process connecting my present memory to the facts I am
remembering. Something happened (I ate a bagel, JFK was
assassinated) and I was or became aware of it (my senses
detected it, in the first case; I heard it over the radio in the
second), and that awareness left a persistent trace (a memory)
that remains in me today, a trace that somehow encodes
information about what happened that I am able to understand.
When I talk about documented historical events, different
processes are involved, but the same pattern occurs: a present
trace encodes information about a past event because of a
process linking that event (or certain facts about it) to the
present existence and features of the trace.
And the same goes for geological evidence and the conclusions
we draw from it.
C. The Skeptical Argument.
Now we’re ready to see the problem. Processes take place over
time—any knowledge we have about processes is knowledge about
the past. So to justify my knowledge claims about the past, I need
more than just knowledge of the present traces. I also need
knowledge of the processes that have produced them—that is, I
need knowledge of the past to justify knowledge claims about the
past. We are stuck in a circle—to justify a claim about the past, we
must already know something about the past. So if our knowledge
about the past must be founded exclusively on knowledge about
the present, we can’t have any knowledge of the past at all.
This leads directly to a kind of radical skepticism about the past:
No matter what the present facts we now observe are, we cannot
justify inferring anything about what has happened before now.
D. Responding to skeptics.
This kind of circle is a familiar phenomenon. It goes back more
than 2000 years, to Diodorus, who proposed an argument called
“the wheel”. For Diodorus, the puzzle was about knowledge of
truth. There are certain sentences that we take to be true. If
someone asks us how we can tell that they are true, we need to
provide some criterion that we use to separate truths from
falsehoods. But if they then ask how we know that this is a good
criterion, we are faced with a difficult choice:
Either there is some higher-order (meta-) criterion that this
criterion satisfies (and they will surely ask how we know that
criterion is a good one), or the criterion is a good one because it
reliably picks out sentences we know are true (and they will surely
point out that we invoked the criterion in the first place to defend
our commitment to the sentences, so surely it’s cheating if we now,
circularly, appeal to the sentences to defend the criterion).
Here is a diagnosis of the problem: If we assume that we’re
starting from zero, we’re in trouble. And so we should be. If we
don’t have any idea of what’s true, of what it takes to be true or
how to tell if a sentence is true, then we can’t “bootstrap” our way
to a defensible account. You can’t find something unless you have
some idea of how to tell when you’ve found it! But if we start
with a reasonable (if partial, flawed and fallible) account of some
things that are true and some ways of deciding what’s true and
what isn’t, then we’ve got something we can build on.
The same goes for knowledge of the past. If we start with no idea
of what processes have taken place in the past, then nothing we
observe now can justify conclusions about the past. But if we start
with some ideas about what processes have produced various
features of the world around us, and compare and test these against
the “traces” we find in the world around us, then we can evaluate
our claims about the past.
The evaluation turns on the coherence and vindication of the
accounts we give. Coherence requires a nice fit between all the
various aspects of the story we tell—what we observe about
present traces and what we say about the processes they are traces
of must fit together; the traces must be the sort of traces (and
arranged in the sorts of ways) we would expect the processes to
produce. (This is the virtue that drives theoretical work, in which
we try to invent hypotheses that will explain the evidence.)
Vindication requires a little bit more—to vindicate our views, we
must apply them successfully—for instance, by making predictions
about features of the traces we expect to observe, given our present
commitments about processes, and then observe the predicted
features. (This is the virtue that drives empirical work, in which
we test the hypotheses theorists come up with.)
Processes:
Let’s list some of the main processes that are central to geological
thinking about the earth. As we read through the history, we’ll see
many examples of models of these processes (and others, which
have fallen by the wayside in geological thinking), together with
arguments about how well or poorly these models fit the evidence
(i.e. traces) they are meant to account for:
Erosion (See especially Hutton, Lyell)
Sedimentation/Deposition (see Steno)
Formation of rocks & minerals (see Werner, von Buch, Desmarest,
Hutton)
Fossilization (see Steno, Leonardo,
Volcanic activity (see Guettard, Desmarest, …)
Orogeny (mountain building) (see de Beaumont, Agassiz, Darwin)
Formation of valleys (see Scrope, Cuvier (?), de Beaumont, …)
Glaciation (see Peraudin, Venetz, Charpentier, Agassiz, et al.)
…
Many authors have drawn lines between what are called the
historical sciences and the lab, or theoretical sciences. The
historical sciences include geology together with paleontology and
the application of evolution to account for the history of life, while
the lab sciences include physics and chemistry. Creationists have
emphasized this distinction, arguing that there is a greater degree
of uncertainty about the results of the historical sciences because
(at least in most cases) “there are no witnesses to the events they
describe,” while the theories of the lab sciences are directly tested
(even “proven”) by measurements we can witness today.
But this is a serious mistake, as a little thought will tell you. For
instance, we have a powerful theory of electromagnetic radiation,
embodied in Maxwell’s equations. This theory is a paradigmatic
example of a lab science. We have tested this theory over and over
again, and (aside from the subtleties of quantum mechanics, which
lead to important modifications for small-scale processes) it is as
well-confirmed and reliable as any theory we have. One important
result of this theory is that the speed of light turns out to be a fixed
quantity, related to the electromagnetic properties of empty space.
Similarly, the constancy of spontaneous decay rates for various
isotopes is a well-established lab result, backed up by a rich
theoretical understanding of the quantum mechanical processes
involved in such decay.
But as soon as we apply these important results from the lab
sciences to the world around us, we obtain results that many
creationists find unacceptable—and a typical response is to reject
these well-supported lab results. Thus, to explain the fact that stars
millions of light years away are visible to us, some creationists
have proposed that the speed of light was much higher in the past
than it is today (others have proposed a version of the 5-minute
hypothesis, in which the light we see from these stars was already
almost here when the universe was created about 6,000 years ago).
And to explain away the results of radioactive dating (which, I
might add, coheres beautifully with the independently developed
geological time scale), some propose that decay rates are not
constant after all, and were much higher in the past.
In fact, the purported epistemic difference between the lab and the
historical sciences is an illusion. First, with respect to testing our
ideas about the past, witnesses are nothing special. In fact, the
memories of witnesses are often less reliable (by the standards of
coherence and vindication) than other sorts of evidence—the
evidence of fingerprints and DNA tests in criminal investigations;
the evidence of archeological investigations (which sometimes
serve to correct historical reports)…
Second, the lab sciences progress by developing extremely general
theories of the basic processes that go on in the world—processes
like gravitational accelerations and mechanical motions in general,
nucleosynthesis in stars (pioneered by Fred Hoyle), chemical
changes and their relation to temperatures, pressures and other
important physical variables, and so on. Such theories are always
vulnerable, since they make so many predictions about what has
gone on, and what will go on in the future. Our conviction that
these theories are well tested and reliable reflects a reasonable
standard of proof for science—but it’s a standard that the historical
sciences of geology and biology also meet.
The standard I’m talking about here combines two features: 1.
Coherence, that is, the processes and principles we accept must fit
together coherently with the observations we make (and viceversa: observations are not open-and-shut proofs; in fact they are
often unreliable, and we spend a lot of time establishing the
reliability of various kinds of observations, for example, when we
develop new instruments for particle physics. 2. Vindication;
when we reason with our commitments and reach conclusions
about what we should expect to observe or to happen in various
new circumstances, we are vindicated when our expectations are
fulfilled (when our intervention is successful).
Download