"What is Science?" FAQ DRAFT III for your review
Marty Fouts
A couple of years ago I posted a couple of drafts of a "What is
Science?" FAQ and they were discussed here. Time has flown, and my
past has come back to haunt me: Adam Marczyk remembered the FAQ and
suggested that I finish it for the talkorigins web site
I've made a few small changes, and it still needs more on the
relationship between science and religion, but, such as it is, here it
is. Please comment here. I'll incorporate changes, and when the
discussion seems to have settled, submit it to the web site as a FAQ.
You can find the current draft at http://www.fogey.com/whatisscience.html
until it is on the web site.
Thanks for your help,
Marty
----- START: What is Science FAQ Draft III -----
Welcome to talk.origins!
Copyright © 1999-2002 by Marty Fouts
[Please note that Marty Fouts is the author of this specific file and not the maintainer of this web site. If you have questions or comments about things you read on this page, feel free to contact Marty. If you have comments about anything else on this site, either send them to the author of the page in question, or respond through the feedback page.]
Q: What is science?
A: "Science" is a term that describes three things:
o The history of the human curiosity about how the universe operates.
o A wide range of methods used to observe, describe, predict, and
explain the measurable interactions of the observable universe.
o The body of knowledge gained by using those methods.
It is the method used that makes "scientific knowledge" scientific,
not the subject being studied. Any subject that a scientific method
can be applied too can yield scientific knowledge.
Unfortunately, this also means that any subject, even one we
normally think of as a scientific subject, can yield "unscientific
knowledge".
Q: What is the "scientific method"?
A: Although much has been made of the scientific method, there is not
a single method. Scientific methods overlap those of other human
endeavors, and few, if any, scientists are concerned with all areas
of science.
Scientists at various times observe, describe, predict, and
explain. A scientific method is a method of performing those four
tasks in such a way as to be objective and accurate. Different
subject matter requires somewhat different methods. The four
activities of science are:
o Observation: Scientists observe the universe to see how it
operates. The raw data of science is accumulated through
observation. The methods of observation range from going into the
field and recording what is there through controlled laboratory
experiments.
o Description: Scientists describe what they see in various ways,
especially through articles in scientific journals. Scientific
description attempts to be precise, accurate, and
objective. Experiments are described in enough detail that
another scientist could perform them and obtain the same
results. Scientists share their raw data with other scientists by
describing it, often in mathematical form.
o Prediction: Scientists test their understanding of how the
universe works by predicting how it will behave.
o Explanation: Scientists are not satisfied with knowing how the
universe works, they also want to know why it works the way it
does. They test their understanding of why the universe works by
proposing explanations.
It is not unusual to lump together the methods of observation and
description under the heading of experiment and the methods of
prediction and explanation under the heading of theory, although
this division is problematic.
Q: What makes a method scientific?
A: The principle contribution of science to the human quest for
knowledge is in objectifying the four categories of method.
An objective approach to knowledge is one in which two observers
using the same methods will arrive at the same knowledge, since the
knowledge depends only on the observation and not on the
observer. Clearly there are no purely objective methods, although
scientific methods strive to be as objective as humanly possible.
The idea of independently repeating an experiment to show that the
observation is objective is useful, especially in laboratory
sciences, but taken alone, it only guarantees that two scientists
arrived at the same result. It does not guarantee that the method
itself is valid.
To reduced the risk of performing bad experiments, scientists
describe their methods to other scientists, who look for errors in
the approach. The formal method of doing this is by publishing
articles in peer-reviewed journals.
Q: How is science made objective?
A: The way in which science has objectified the quest for knowledge
is, in principle, through the application of quantification. To the
extent that it can be measured it can be made objective.
Quantification requires two things:
o Standards are established so that measurements of similar
quantities can be accurately compared. If a scientist anywhere in
the world measures the length of something and reports it,
scientists everywhere know what that length is because there is a
standard for measuring lengths.
o Tools are used to improve the range, repeatability and accuracy
of measurements. Rather than estimate the length, the scientist
will measure it with a ruler of some sort. Whenever possible,
scientists develop tools to make their measurements.
Quantification is not enough to guarentee objectivity, since two
scientists will differ in the way they use the measurment tools,
and so some bias may creep into an observation. Scientists guard
against bias by verifying each other's work, through peer-review of
scientific literature, and by repeating experiments.
Q: Can you sum up 'scientific method', again, please?
A: The literature on scientific method, then, can be seen as a
discussion of the ways in which scientists attempt to objectify the
four activities of science. For instance:
o Observation: through the use of precise measurements, careful
experimental design and control, and repeatability. Observation
is the basis of science.
o Description: through the use of comparison to objective
standards, most notably, again, measurements. Descriptions are
built upon observations.
o Prediction: through the use of mathematical description of
interactions coupled with the design of experiments intended to
test predictions. Predictions are attempts to generalize
descriptions so that they may be extrapolated to new situations.
o Explanation: through the use of peer-review and the demand that
explanations account for all known observations and make testable
predictions that allow them to be differentiated from other
explanations.
Q: How do various scientific methods differ?
A: Some branches of science can rely on laboratory experiments more
easily than others, or have more control over the experiments they
perform. Physics, for example, can more often do its experiments in
a controlled lab, than can anthropology. Some branches can rely
more on quantitative descriptions while others rely more on
qualitative descriptions. The same is true for predictions.
Q: How does science add to our knowledge?
A: The branch of philosophy that concerns itself with how we obtain
knowledge is called epistemology. One of the concerns of
epistemology is determining whether a particular method of
obtaining knowledge can obtain certainty by giving us absolute
assurance of the truth of the knowledge.
Science does not provide such certainty because it would be be
necessary to have observed everything over all time to be able to
completely describe all knowledge. To the extent that scientific
knowledge is incomplete it must remain uncertain. This is true of
any program that wants to be scientific.
The resulting epistemology of western science is often called
empirical pragmatism. This philosophy is grounded in the
understanding that all scientific knowledge is provisional
knowledge and that any scientific knowledge may be rendered
obsolete by a future observation. It recognizes that, in order to
make any progress at all, science must take as given certain
assumptions that can not be validated, but that have been very
reliable for a long time, and so, will be used until they are
invalidated.
Some key aspects of this epistemology are
o objectivity: the requirement that the observations of science be
objective.
o reductionism: the belief that problems can sometimes be
subdivided into smaller problems and the larger problem's
solution can be discovered by solving the smaller problems in
turn,
o scientific induction: the belief that a sufficient number of
observations of similarity can be used to generalize,
o extrapolation: the belief that certain observations that have
been true in the past will remain true in the future.
o consistency: the requirement that any new scientific knowledge
must be consistent with what is already known.
Each of these beliefs has served science well, and each is
constantly tested against the known observations.
Q: What is the role of consistency in science?
A: Whenever scientists make a new observation or propose a new
explanation, they require that it be consistent with existing
observations and explanations. Knowledge is consistent so long as
reasoning about it doesn't lead to contradictory conclusions.
When an observation is made that seem inconsistent with what is
already known, the knowledge is carefully checked. In some cases,
an error is found in the observation and it is discarded. In
others, the observation leads to new understanding and the existing
knowlege is updated to reflect the understanding. For example, the
behavior of some physics experiments was not consistent with
classical physics, leading Einstein to propose that light traveled
in descrete bundles, resulting in quantum mechanics being
discovered.
Q: What is the difference between a theory and a fact?
A: In science, a fact is an observation or series of observations, or
a description of a series of observations. Facts are answers to
what is kinds of questions, and are the part of science that can be
the most objective. A theory is an attempt to explain known facts
and usually involves predictions about future
observations. Theories are answers to how does kinds of questions.
Q: What is the role of falsifiability?
A: Consider a theory that makes a prediction. By testing such a
prediction, we can determine if the theory is accurate or not. A
theory that makes testable predictions is said to be falsifiable. It
is common among scientists to accept as scientific only those theories
that are falsifiable in this sense.
A common misconception is the belief that a theory can stand or
fall based on a single falsifying experiment. In practice, the
result of performing a falsifying experiment can have one of
several consequences:
o The experiment confirms the prediction: Scientists tend to become
more confident in theories as such experiments are performed.
o The experiment does not confirm the prediction: Scientists tend
to modify their theories to accomodate the new information. It is
extremely rare for a theory to be completely abandoned as a
result of such an experiment.
o The experiment is inconclusive: Life goes on.
Thus, while a theory can be described as falsifiable, it is
probably more accurate to think of it as testable and to recognize
that the result of a failed test is, more often than not, a
modified theory.
Q: What is the difference between a theory and a law?
A: A law in science is an empirical relationship between measurable
quantities. Newton's "law" of gravity is such an formulation. Laws
often hold only in special cases. Laws are often descriptions, and
frequently stated as mathematical relationships. A theory is an
attempt to explain and predict. Theories often incorporate laws.
Q: But what about the scientists?
A: A significant source of confusion in discussions of the philosophy
and history of science is to intermix the sociology of scientists
with the epistemological method of science. Scientists are humans
and suffer from all of the strengths and weaknesses of humans, and
so the quest for scientific knowledge has a very human history. But
scientific epistemology, which is the theory that results from the
practice of science, is an abstraction of human though rather than
a sequence of human actions.
Q: Is science a religion?
A: No. However, many people have made a religion out of a certain set
of believes in what science can and can not do. To those people,
the use of science has taken a religious role.
Q: Is science the only epistemology necessary?
A: No. Science tells us what we can do with the universe, it can not
tell us what we should do.
Acknowledgements:
Comments have been provided by [if you have commented and I have
neglected you, please send me email and I will rectify my error]
fo...@aol.comeatspam (Foggg)
agi...@linguist.umass.edu (Andre G Isaak)
wil...@wehi.edu.au (John Wilkins)
c...@tiac.net (Richard Harter)
----- END: What is Science FAQ Draft III -----
Mike the Vike
"Marty Fouts" <usenet_p...@yahoo.com> wrote in message
news:uvg7on...@yahoo.com...
>
[snip]
> Q: What is the difference between a theory and a fact?
>
> A: In science, a fact is an observation or series of observations, or
> a description of a series of observations. Facts are answers to
I suggest adding a quote here:
> "what is" kinds of questions, and are the part of science that can be
> the most objective. A theory is an attempt to explain known facts
> and usually involves predictions about future
and here:
> observations. Theories are answers to "how does" kinds of questions.
>
[snip]
Thanks for the excellent discussion of epistemology.
Richard Morey
>
Good FAQ. Some suggestions:
"too" should be "to" in the sentence "Any subject that a scientific method
can be applied too can yield scientific knowledge."
"guarentee" should be "guarantee" and "measurment" should be "measurement"
in the sentence "Quantification is not enough to guarentee objectivity..."
The sentence "the belief that a sufficient number of observations of
similarity can be used to generalize..." should be reworded to "the belief
that a sufficient number of similar observations can be used to
generalize...."
The sentence "the belief that certain observations that have been true in
the past will remain true in the future" does not, I think, convey what you
mean. Do you mean that "the belief that given the same (or nearly the same)
circumstances, the same (or nearly the same) observations could be made in
the future."? An observation is not true or untrue, and if accurate, cannot
be "unaccurate" in the future.
"though" should be (I think) "thought" in the sentence "...is an abstraction
of human though rather than a sequence of human actions."
"believes" should be "beliefs" in the sentence "However, many people have
made a religion out of a certain set of believes in what science can and can
not do."
In the "Is science the only epistemology necessary?" section, you say "No."
and then you say that it does not tell us what we should do. This, however,
is not an epistemelogical question; it is a question for ethics. They are
different branches of philosophy. If you still mean "No," you'd need to
change that.
Once more thing: you mention the idea of controling experiments, but don't
describe the idea at all. I think a section addressing experimental design
would be a positive thing.
Good FAQ.
Richard Morey
-----------------------
Graduate Research Assistant
Cognition and Neuroscience,
University of Missouri-Columbia
Matt Silberstein
<usenet_p...@yahoo.com>:
>
>A couple of years ago I posted a couple of drafts of a "What is
>Science?" FAQ and they were discussed here. Time has flown, and my
>past has come back to haunt me: Adam Marczyk remembered the FAQ and
>suggested that I finish it for the talkorigins web site
>
>I've made a few small changes, and it still needs more on the
>relationship between science and religion, but, such as it is, here it
>is. Please comment here. I'll incorporate changes, and when the
>discussion seems to have settled, submit it to the web site as a FAQ.
>
>You can find the current draft at http://www.fogey.com/whatisscience.html
>
>until it is on the web site.
>
>Thanks for your help,
>
>Marty
>
>----- START: What is Science FAQ Draft III -----
>
>Welcome to talk.origins!
>Copyright Š 1999-2002 by Marty Fouts
>
>[Please note that Marty Fouts is the author of this specific file and not the maintainer of this web site. If you have questions or comments about things you read on this page, feel free to contact Marty. If you have comments about anything else on this site, either send them to the author of the page in question, or respond through the feedback page.]
>
>
>
>Q: What is science?
>
>A: "Science" is a term that describes three things:
>
> o The history of the human curiosity about how the universe operates.
>
> o A wide range of methods used to observe, describe, predict, and
> explain the measurable interactions of the observable universe.
>
> o The body of knowledge gained by using those methods.
>
> It is the method used that makes "scientific knowledge" scientific,
> not the subject being studied. Any subject that a scientific method
> can be applied too can yield scientific knowledge.
>
> Unfortunately, this also means that any subject, even one we
> normally think of as a scientific subject, can yield "unscientific
> knowledge".
I am not happy with this phrasing. First, it is not clear what
"unscientific" means here. Do you mean something like "life is
good" or like "cold weather leads to disease". Second, do
subjects yield knowledge or does the study of the subject yield
the knowledge? I suspect you mean the latter and mean that one
can study something in an unscientific manner.
[snip]
>Q: Is science a religion?
>
>A: No. However, many people have made a religion out of a certain set
> of believes in what science can and can not do. To those people,
> the use of science has taken a religious role.
And the companion question: does science deny the existence of
God.
>Q: Is science the only epistemology necessary?
>
>A: No. Science tells us what we can do with the universe, it can not
> tell us what we should do.
And the companion question: doesn't science say that nature is
all there is?
[snip]
AR
>
> A: The literature on scientific method, then, can be seen as a
> discussion of the ways in which scientists attempt to objectify the
> four activities of science. For instance:
>
> o Observation: through the use of precise measurements, careful
> experimental design and control, and repeatability. Observation
> is the basis of science.
>
> o Description: through the use of comparison to objective
> standards, most notably, again, measurements. Descriptions are
> built upon observations.
>
> o Prediction: through the use of mathematical description of
> interactions coupled with the design of experiments intended to
> test predictions. Predictions are attempts to generalize
> descriptions so that they may be extrapolated to new situations.
>
> o Explanation: through the use of peer-review and the demand that
> explanations account for all known observations and make testable
> predictions that allow them to be differentiated from other
> explanations.
This seems a bit strict. Even Newton's theories of motion were known at the time
to not account for all known celestial observations (though the reason was not known). If I'm
not mistaken, and I could easily be, the discovery of Pluto was made during
a search for a "planet" that would account for "unexplained" perturbations in the
orbit of other planets. However, Pluto was not large enough nor did it have the
correct orbit to explain these purterbations.
>
>
> Q: How do various scientific methods differ?
>
> A: Some branches of science can rely on laboratory experiments more
> easily than others, or have more control over the experiments they
> perform. Physics, for example, can more often do its experiments in
> a controlled lab, than can anthropology. Some branches can rely
> more on quantitative descriptions while others rely more on
> qualitative descriptions. The same is true for predictions.
>
> Q: How does science add to our knowledge?
via theories. These are the "knowledge" or explanations of science
>
>
> A: The branch of philosophy that concerns itself with how we obtain
> knowledge is called epistemology. One of the concerns of
> epistemology is determining whether a particular method of
> obtaining knowledge can obtain certainty by giving us absolute
> assurance of the truth of the knowledge.
>
> Science does not provide such certainty because it would be be
> necessary to have observed everything over all time to be able to
> completely describe all knowledge. To the extent that scientific
> knowledge is incomplete it must remain uncertain. This is true of
> any program that wants to be scientific.
I don't understand how completeness has much to do with it. Let's pretend that
Newton could have observed every object in the universe from the beginning
of time. So what? He simply would have had a few more unexplainable observations
for his theory. That wouldn't necessarily lead to a better theory, or to the theory
of relativity, or whatever may come next, or to the "truth".
>
>
> The resulting epistemology of western science is often called
> empirical pragmatism. This philosophy is grounded in the
> understanding that all scientific knowledge is provisional
> knowledge and that any scientific knowledge may be rendered
> obsolete by a future observation. It recognizes that, in order to
> make any progress at all, science must take as given certain
> assumptions that can not be validated, but that have been very
> reliable for a long time, and so, will be used until they are
> invalidated.
>
> Some key aspects of this epistemology are
>
> o objectivity: the requirement that the observations of science be
> objective.
> o reductionism: the belief that problems can sometimes be
> subdivided into smaller problems and the larger problem's
> solution can be discovered by solving the smaller problems in
> turn,
Isn't this problem specific? Why does science, generally, require this?
>
> o scientific induction: the belief that a sufficient number of
> observations of similarity can be used to generalize,
This seems to be an issue of theory not observation. That is, we
believe we can generalize from a to b because we believe the
explanation of what is occuring in a applies to the conditions of b.
Saying that a fact is an observation, seems to minimize the word fact,
and elevate the meaning of an observation. It doesn't seem to me that
observations can be divorced from theory, at least not as cleanly as this
may imply. I don't see how an observation can be dissociated from theory.
Even our senses are not objective. My visual cortex is in the dark, yet it
"sees" light. But it doesn't really see anything. It just interprets electical
and chemical signals it receives and I interpret these as colors, shades, etc.
So what one takes as the "observation"of sight, is to some extent dependent
on one's "theory" of how the mind works, and "sees". Normally we don't
think about this, because most of our "mental models" of teh senses are that
they are direct, i.e., as if I actually do "see". Similarly, many scientific
conclusions are base on comparsons of means (e.g., the mean height of
american adult males). Yet the value which is the mean may not actually exist
in teh population that is being studied. Thus it could never be observed.
So, is the mean an observation even though one could not observe it?!
It's a theoretical construct that we call data. My point is simply that the distinction
between observation or fact? can probably not be so cleanly made. They are
very intertwined adn not clearly distinct entities.
>
>
> Q: What is the role of falsifiability?
>
> A: Consider a theory that makes a prediction. By testing such a
> prediction, we can determine if the theory is accurate or not. A
> theory that makes testable predictions is said to be falsifiable. It
> is common among scientists to accept as scientific only those theories
> that are falsifiable in this sense.
Again, this distinction maynot be as clear as it sounds here. Maybe if
you could present an example of each type, it would help the reader.
Adam Marczyk
news:uvg7on...@yahoo.com...
I'd like to make a few comments:
> ----- START: What is Science FAQ Draft III -----
>
> Welcome to talk.origins!
> Copyright Š 1999-2002 by Marty Fouts
>
> [Please note that Marty Fouts is the author of this specific file and not the
maintainer of this web site. If you have questions or comments about things you
read on this page, feel free to contact Marty. If you have comments about
anything else on this site, either send them to the author of the page in
question, or respond through the feedback page.]
>
>
>
> Q: What is science?
>
> A: "Science" is a term that describes three things:
>
> o The history of the human curiosity about how the universe operates.
>
> o A wide range of methods used to observe, describe, predict, and
> explain the measurable interactions of the observable universe.
>
> o The body of knowledge gained by using those methods.
>
> It is the method used that makes "scientific knowledge" scientific,
> not the subject being studied. Any subject that a scientific method
> can be applied too can yield scientific knowledge.
>
> Unfortunately, this also means that any subject, even one we
> normally think of as a scientific subject, can yield "unscientific
> knowledge".
I might add "if studied in an unscientific way" just to make the point
completely clear.
> Q: What is the "scientific method"?
>
> A: Although much has been made of the scientific method, there is not
> a single method. Scientific methods overlap those of other human
> endeavors, and few, if any, scientists are concerned with all areas
> of science.
>
> Scientists at various times observe, describe, predict, and
> explain. A scientific method is a method of performing those four
> tasks in such a way as to be objective and accurate. Different
> subject matter requires somewhat different methods. The four
> activities of science are:
>
> o Observation: Scientists observe the universe to see how it
> operates. The raw data of science is accumulated through
> observation. The methods of observation range from going into the
> field and recording what is there through controlled laboratory
> experiments.
Maybe "to" instead of "through" there, to go with the "from"?
> o Description: Scientists describe what they see in various ways,
> especially through articles in scientific journals. Scientific
> description attempts to be precise, accurate, and
> objective. Experiments are described in enough detail that
> another scientist could perform them and obtain the same
> results. Scientists share their raw data with other scientists by
> describing it, often in mathematical form.
>
> o Prediction: Scientists test their understanding of how the
> universe works by predicting how it will behave.
>
> o Explanation: Scientists are not satisfied with knowing how the
> universe works, they also want to know why it works the way it
> does. They test their understanding of why the universe works by
> proposing explanations.
I think that last sentence is a bit unclear. How about: "They test their
understanding of why the universe works as it does by proposing explanations
and deriving predictions from them"?
> It is not unusual to lump together the methods of observation and
> description under the heading of experiment and the methods of
> prediction and explanation under the heading of theory, although
> this division is problematic.
Why is it problematic? Might want to expand on that a bit.
> Q: What makes a method scientific?
>
> A: The principle contribution of science to the human quest for
> knowledge is in objectifying the four categories of method.
>
> An objective approach to knowledge is one in which two observers
> using the same methods will arrive at the same knowledge, since the
> knowledge depends only on the observation and not on the
> observer. Clearly there are no purely objective methods, although
> scientific methods strive to be as objective as humanly possible.
I don't know about "clearly" - I think it could at least in principle be argued
that there are purely objective methods. How about: "Scientific methods strive
to be as objective as humanly possible, although the ideal of perfect
objectivity is rarely attained in practice"?
> The idea of independently repeating an experiment to show that the
> observation is objective is useful, especially in laboratory
> sciences, but taken alone, it only guarantees that two scientists
> arrived at the same result. It does not guarantee that the method
> itself is valid.
>
> To reduced the risk of performing bad experiments, scientists
> describe their methods to other scientists, who look for errors in
> the approach. The formal method of doing this is by publishing
> articles in peer-reviewed journals.
>
> Q: How is science made objective?
>
> A: The way in which science has objectified the quest for knowledge
> is, in principle, through the application of quantification. To the
> extent that it can be measured it can be made objective.
I'm a little wary of this part, because it seems to be implying that
non-numerical sciences (such as most of geology and biology) are not objective,
and I don't think that's the case. I think this might do with some revision.
For *all* known observations? I think the Kuhnians will choke on that. ;) Maybe
if you drop the "all"? - that's just my opinion, though.
"Beliefs"? Doesn't sound right. Maybe "principles"?
Rare, perhaps, but it certainly does happen. I wouldn't make the wording so
strong. I would say that there are three outcomes here:
-The experiment's result seriously contradicts a crucial tenet of the theory
and forces the theory to be discarded entirely. This usually doesn't happen with
one experiment, but may happen if evidence begins to pile up against the theory.
-The experiment's result contradicts a non-crucial part of the theory, and
so the theory is revised or modified to accommodate it.
-The experiment's result is shelved as an anomaly until more data or more
advanced understanding is available.
> o The experiment is inconclusive: Life goes on.
>
> Thus, while a theory can be described as falsifiable, it is
> probably more accurate to think of it as testable and to recognize
> that the result of a failed test is, more often than not, a
> modified theory.
Again, I'd make the wording not quite so strong here - at least temper it with
the recognition that theories can be and often are outright falsified by failed
tests.
> Q: What is the difference between a theory and a law?
>
> A: A law in science is an empirical relationship between measurable
> quantities. Newton's "law" of gravity is such an formulation. Laws
> often hold only in special cases. Laws are often descriptions, and
> frequently stated as mathematical relationships. A theory is an
> attempt to explain and predict. Theories often incorporate laws.
>
> Q: But what about the scientists?
>
> A: A significant source of confusion in discussions of the philosophy
> and history of science is to intermix the sociology of scientists
> with the epistemological method of science. Scientists are humans
> and suffer from all of the strengths and weaknesses of humans, and
> so the quest for scientific knowledge has a very human history. But
> scientific epistemology, which is the theory that results from the
> practice of science, is an abstraction of human though rather than
> a sequence of human actions.
>
> Q: Is science a religion?
>
> A: No. However, many people have made a religion out of a certain set
> of believes in what science can and can not do. To those people,
> the use of science has taken a religious role.
I think this part could use expansion, as I commented in my e-mail. Why isn't
science a religion - what qualities differentiate the two?
> Q: Is science the only epistemology necessary?
>
> A: No. Science tells us what we can do with the universe, it can not
> tell us what we should do.
Room for more expansion here as well, I think.
--
And I want to conquer the world,
give all the idiots a brand new religion,
put an end to poverty, uncleanliness and toil,
promote equality in all of my decisions...
--Bad Religion, "I Want to Conquer the World"
http://www.ebonmusings.org ICQ: 8777843 PGP Key ID: 0x5C66F737
R. Baldwin
> The resulting epistemology of western science is often called
> empirical pragmatism. This philosophy is grounded in the
> understanding that all scientific knowledge is provisional
> knowledge and that any scientific knowledge may be rendered
> obsolete by a future observation. It recognizes that, in order to
> make any progress at all, science must take as given certain
> assumptions that can not be validated, but that have been very
> reliable for a long time, and so, will be used until they are
> invalidated.
Nice.
Suggest changing "any scientific knowledge may be rendered obsolete by
a future observation" to "any scientific knowledge may be rendered
obsolete by future observations."
John Wilkins
> A couple of years ago I posted a couple of drafts of a "What is
> Science?" FAQ and they were discussed here. Time has flown, and my
> past has come back to haunt me: Adam Marczyk remembered the FAQ and
> suggested that I finish it for the talkorigins web site
>
> I've made a few small changes, and it still needs more on the
> relationship between science and religion, but, such as it is, here it
> is. Please comment here. I'll incorporate changes, and when the
> discussion seems to have settled, submit it to the web site as a FAQ.
>
> You can find the current draft at http://www.fogey.com/whatisscience.html
>
> until it is on the web site.
>
> Thanks for your help,
>
> Marty
>
> ----- START: What is Science FAQ Draft III -----
>
> Welcome to talk.origins!
> Copyright © 1999-2002 by Marty Fouts
>
> [Please note that Marty Fouts is the author of this specific file and not
> the maintainer of this web site. If you have questions or comments about
> things you read on this page, feel free to contact Marty. If you have
> comments about anything else on this site, either send them to the author
> of the page in question, or respond through the feedback page.]
>
>
>
> Q: What is science?
>
> A: "Science" is a term that describes three things:
/ at least four
>
> o The history of the human curiosity about how the universe operates.
>
> o A wide range of methods used to observe, describe, predict, and
> explain the measurable interactions of the observable universe.
>
> o The body of knowledge gained by using those methods.
o A social institution or set of institutions in which
professionals make a career.
[JW: It is important to recognise that science is a culture and society.
It is a *human* activity after all. Science is not an abstraction devoid
of politics, like, say, revelation is supposed to be.]
>
> It is the method used that makes "scientific knowledge" scientific,
> not the subject being studied. Any subject that a scientific method
> can be applied too can yield scientific knowledge.
>
> Unfortunately, this also means that any subject, even one we
> normally think of as a scientific subject, can yield "unscientific
> knowledge".
>
> Q: What is the "scientific method"?
>
> A: Although much has been made of the scientific method, there is not
> a single method. Scientific methods overlap those of other human
> endeavors, and few, if any, scientists are concerned with all areas
> of science.
[JW: Isn't this a contradiction to the claim above that method makes
science scientific?]
[JW: as are all such taxonomies - science is what it is; we divide it up
to teach, explain or otherwise do what suits us for our own purposes. We
should be careful to indicate that - there's nothing holy about how we
describe science]
>
> Q: What makes a method scientific?
>
> A: The principle contribution of science to the human quest for
> knowledge is in objectifying the four categories of method.
>
> An objective approach to knowledge is one in which two observers
> using the same methods will arrive at the same knowledge, since the
> knowledge depends only on the observation and not on the
> observer. Clearly there are no purely objective methods, although
> scientific methods strive to be as objective as humanly possible.
Another name for this is "intersubjectivity", or "crossbearings".
>
> The idea of independently repeating an experiment to show that the
> observation is objective is useful, especially in laboratory
> sciences, but taken alone, it only guarantees that two scientists
> arrived at the same result. It does not guarantee that the method
> itself is valid.
>
> To reduced the risk of performing bad experiments, scientists
> describe their methods to other scientists, who look for errors in
> the approach. The formal method of doing this is by publishing
> articles in peer-reviewed journals.
>
> Q: How is science made objective?
>
> A: The way in which science has objectified the quest for knowledge
[JW: I objewct to words like "objectified". If we mean, "the way science
gets it right" or "ensures it is realistic" or whatever, then say so.
"Objectified" is a word used by people who also use words like
"structuralism" "modernism" and "tropes" and deny the existence of
authors...]
[JW: Strictly speaking, pragmatism is the view that truth is, for all
intents and purposes, the same as what works out in practice (hence the
name).]
>
> Some key aspects of this epistemology are
>
> o objectivity: the requirement that the observations of science be
> objective.
> o reductionism: the belief that problems can sometimes be
> subdivided into smaller problems and the larger problem's
> solution can be discovered by solving the smaller problems in
> turn,
[JW: No. Absolutely not. This is decomposition of problems into smaller
subsets, in the hope that by solving simpler problems we can solve the
whole. Reductionism is the view that theories at one level are to be
fully accounted for by lower level theories (i.e., biology in terms of
chemistry, and chemistry in terms of physics). A theory is "reduced" to
another iff all the elements of the reduced theory are comprised of the
elements of the reducing theory (hence also, Newton reduces to Einstein
and QM).]
[JW: The term "law" was introduced into science in the 17thC, when it
was thought that God's laws provided the basis for the regularity of
nature. However, nothing is a law that is *not* a mathematical
relationship - the technical term for which is "model". Theories are
typically collations of models of this kind, but some theories do not
have mathematical models at first (Darwin's theory of natural selection
being a case in point). I think it worth saying that "law" is just a way
of saying "model", and that nothing much rests on the word.]
>
> Q: But what about the scientists?
>
> A: A significant source of confusion in discussions of the philosophy
> and history of science is to intermix the sociology of scientists
> with the epistemological method of science. Scientists are humans
> and suffer from all of the strengths and weaknesses of humans, and
> so the quest for scientific knowledge has a very human history. But
> scientific epistemology, which is the theory that results from the
> practice of science, is an abstraction of human though rather than
> a sequence of human actions.
>
> Q: Is science a religion?
>
> A: No. However, many people have made a religion out of a certain set
> of believes in what science can and can not do. To those people,
beliefs
> the use of science has taken a religious role.
This is sometimes
known as "scientism".
>
> Q: Is science the only epistemology necessary?
>
> A: No. Science tells us what we can do with the universe, it can not
> tell us what we should do.
[JW: an epistemology is a philosophy of how knowledge is gathered. It is
not a moral notion. I think what you are getting at is this:
Q: Is science the only way of gathering knowledge (the only possible
epistemology)?
A: No. It cannot give moral knowledge, knowledge of things that might
exist beyond the empirical world, or knowledge of values.
Note that I avoided saying "natural world" where I instead said
"empirical world"]
>
> Acknowledgements:
>
> Comments have been provided by [if you have commented and I have
> neglected you, please send me email and I will rectify my error]
>
> fo...@aol.comeatspam (Foggg)
> agi...@linguist.umass.edu (Andre G Isaak)
> wil...@wehi.edu.au (John Wilkins)
> c...@tiac.net (Richard Harter)
>
> ----- END: What is Science FAQ Draft III -----
Despite my quibbles here, this is a very creditable effort, Marty. I am
impressed.
--
John Wilkins
Sweet Analytics, 'tis thou hast ravished me [Marlowe's Faust]
Sven Silow
> o Observation: Scientists observe the universe to see how it
> operates.
Even if I sometimes express myself the same way (or use 'reality'
instead of 'universe') I think it might give an impression that all
scientists are astronomers. The sentence needs an addition of
'details', 'parts' or such.
Sven
Richard Morey
> "Objectified" is a word used by people who also use words like
> "structuralism" "modernism" and "tropes" and deny the existence of
> authors...]
>
BobAlmighty
Looks pretty good; I haven't read through it all, but this stood out to me:
> Q: Is science a religion?
>
> A: No. However, many people have made a religion out of a certain set
> of believes in what science can and can not do. To those people,
> the use of science has taken a religious role.
I think "believes" should be "beliefs".
Anyways, good job!
- Bob
Marty Fouts
AR <roto...@pitt.edu> writes:
[snip]
I'm incorporating those changes that I understand and agree
with. Comments here are to try to find common ground.
> > Q: How does science add to our knowledge?
>
> via theories. These are the "knowledge" or explanations of science
It seems to me that theories are only part of the knowledge of
science. Much of the knowledge is found in the observations as well,
for example in fields like geology, where much of the literature is
descriptive in nature.
[snip]
> > A: The branch of philosophy that concerns itself with how we obtain
> > knowledge is called epistemology. One of the concerns of
> > epistemology is determining whether a particular method of
> > obtaining knowledge can obtain certainty by giving us absolute
> > assurance of the truth of the knowledge.
> >
> > Science does not provide such certainty because it would be be
> > necessary to have observed everything over all time to be able to
> > completely describe all knowledge. To the extent that scientific
> > knowledge is incomplete it must remain uncertain. This is true of
> > any program that wants to be scientific.
>
> I don't understand how completeness has much to do with it. Let's
> pretend that Newton could have observed every object in the universe
> from the beginning of time. So what? He simply would have had a few
> more unexplainable observations for his theory. That wouldn't
> necessarily lead to a better theory, or to the theory of relativity,
> or whatever may come next, or to the "truth".
We're looking at different parts of the problem. You're saying,
correctly, that one might not have a theory that explains all that one
has observed. What I'm trying to get at is that you can never claim to
have a *complete* theory because you will never have a *complete* set
of observations and you'll never have that.
[snip]
> > o reductionism: the belief that problems can sometimes be
> > subdivided into smaller problems and the larger problem's
> > solution can be discovered by solving the smaller problems in
> > turn,
>
> Isn't this problem specific? Why does science, generally, require
> this?
Hopefully, when I address other concerns about this paragraph, the
problem here will go away.
>
> >
> > o scientific induction: the belief that a sufficient number of
> > observations of similarity can be used to generalize,
>
> This seems to be an issue of theory not observation. That is, we
> believe we can generalize from a to b because we believe the
> explanation of what is occuring in a applies to the conditions of b.
if every dog I see has 4 legs, then I generalize to the assumption
that all dogs have 4 legs. It is from the observations that the
generalization comes. (Of course, as soon as I see a 3 legged dog I
have to modify the generalization...)
[snip]
> > A: In science, a fact is an observation or series of observations, or
> > a description of a series of observations. Facts are answers to
> > what is kinds of questions, and are the part of science that can be
> > the most objective. A theory is an attempt to explain known facts
> > and usually involves predictions about future
> > observations. Theories are answers to how does kinds of questions.
>
> Saying that a fact is an observation, seems to minimize the word
> fact, and elevate the meaning of an observation. It doesn't seem to
> me that observations can be divorced from theory, at least not as
> cleanly as this may imply. I don't see how an observation can be
> dissociated from theory. [snip] My point is simply that the
> distinction between observation or fact? can probably not be so
> cleanly made. They are very intertwined adn not clearly distinct
> entities.
I'm not sure what you want me to say differently.
[snip]
Marty Fouts
"Adam Marczyk" <ebon...@hotmailNOTexcite.com> writes:
I've clipped out most of the material (having added the changes you
suggested) and only left the areas where more discussion is needed.
> Marty Fouts <usenet_p...@yahoo.com> wrote in message
> news:uvg7on...@yahoo.com...
[snip]
> > It is not unusual to lump together the methods of observation
> > and description under the heading of experiment and the methods
> > of prediction and explanation under the heading of theory,
> > although this division is problematic.
>
> Why is it problematic? Might want to expand on that a bit.
Probably, but I'm going to take the easy way out and drop the last
clause ;0
[snip]
> > Q: How is science made objective?
> >
> > A: The way in which science has objectified the quest for knowledge
> > is, in principle, through the application of quantification. To the
> > extent that it can be measured it can be made objective.
>
> I'm a little wary of this part, because it seems to be implying that
> non-numerical sciences (such as most of geology and biology) are not
> objective, and I don't think that's the case. I think this might do
> with some revision.
In my opinion, to the extent that such sciences aren't quantified,
they aren't objective. But I'm open to persuasion. How else is
objectivity approached?
[snip]
> I wouldn't make the wording so strong. I would say that there are
> three outcomes here:
> -The experiment's result seriously contradicts a crucial tenet
> of the theory and forces the theory to be discarded
> entirely. This usually doesn't happen with one experiment, but
> may happen if evidence begins to pile up against the theory.
> -The experiment's result contradicts a non-crucial part of the
> theory, and so the theory is revised or modified to accommodate
> it.
> -The experiment's result is shelved as an anomaly until more
> data or moreadvanced understanding is available.
>
I agree with you, but I'm not sure we want to get so far into a
discussion of the structure of theories. Do you think it's worth
adding a description of theories and their components?
Anyway, here's my current suggested alternative:
--- new wording
A common misconception is the belief that a theory can stand or fall
based on a single falsifying experiment. In practice, the result of
performing a falsifying experiment can have one of several
consequences:
o The experiment confirms the prediction: Scientists tend to become
more confident in theories as such experiments are performed.
o The experiment does not confirm the prediction: Scientists tend to
modify their theories to accomodate the new information. Sometimes
it is not possible to make such modification and the theories are
abandoned.
o The experiment is inconclusive: Life goes on.
Thus, while a theory can be described as falsifiable, it is probably
more accurate to think of it as testable and to recognize that the
result of a failed test is, more often than not, a modified theory.
--- new wording
[snip]
> > Q: Is science a religion?
> >
> > A:n No. However, many people have made a religion out of a certain set
> > of believes in what science can and can not do. To those people,
> > the use of science has taken a religious role.
>
> I think this part could use expansion, as I commented in my
> e-mail. Why isn't science a religion - what qualities differentiate
> the two?
I agree with you. I'm just not sure what to put in yet. I figured we
could get agreement on the rest of it while I worked out this section...
>
> > Q: Is science the only epistemology necessary?
> >
> > A: No. Science tells us what we can do with the universe, it can not
> > tell us what we should do.
>
> Room for more expansion here as well, I think.
OK.
Marty Fouts
> Marty Fouts <usenet_p...@yahoo.com> wrote:
[snip]
> > Q: What is the "scientific method"?
> >
> > A: Although much has been made of the scientific method, there is not
> > a single method. Scientific methods overlap those of other human
> > endeavors, and few, if any, scientists are concerned with all areas
> > of science.
>
> [JW: Isn't this a contradiction to the claim above that method makes
> science scientific?]
Nah. The claim above was that methodS make science scientific. the
above is only trying to say that there's more than one of them...
there was a number agreement problem, but i'll fix that in the next
draft.
> > A: The way in which science has objectified the quest for knowledge
>
> [JW: I objewct to words like "objectified". If we mean, "the way science
> gets it right" or "ensures it is realistic" or whatever, then say so.
> "Objectified" is a word used by people who also use words like
> "structuralism" "modernism" and "tropes" and deny the existence of
> authors...]
Huh. I thought 'objectified' was used by feminists. but I get the
point and have fixed the text.
[snip]
> > The resulting epistemology of western science is often called
> > empirical pragmatism. This philosophy is grounded in the
> > understanding that all scientific knowledge is provisional
> > knowledge and that any scientific knowledge may be rendered
> > obsolete by a future observation. It recognizes that, in order
> > to make any progress at all, science must take as given certain
> > assumptions that can not be validated, but that have been very
> > reliable for a long time, and so, will be used until they are
> > invalidated.
>
> [JW: Strictly speaking, pragmatism is the view that truth is, for all
> intents and purposes, the same as what works out in practice (hence the
> name).]
I thought pragmatism was only caring about what works out in practice
and forgetting about 'truth'. Is there a better wording for the above
paragraph?
[snip]
> > Q: What is the difference between a theory and a law?
> >
> > A: A law in science is an empirical relationship between measurable
> > quantities. Newton's "law" of gravity is such an formulation. Laws
> > often hold only in special cases. Laws are often descriptions, and
> > frequently stated as mathematical relationships. A theory is an
> > attempt to explain and predict. Theories often incorporate laws.
>
> [JW: The term "law" was introduced into science in the 17thC, when it
> was thought that God's laws provided the basis for the regularity of
> nature. However, nothing is a law that is *not* a mathematical
> relationship - the technical term for which is "model". Theories are
> typically collations of models of this kind, but some theories do not
> have mathematical models at first (Darwin's theory of natural selection
> being a case in point). I think it worth saying that "law" is just a way
> of saying "model", and that nothing much rests on the word.]
Hmm... I'm not sure it's this simple. Certainly, to a mathematician,
'mathematical relationship' is *not* the equivalent of 'model',
although models contain collections of relationships.
I found a definition of 'scientific law' at
http://www.sasked.gov.sk.ca/docs/chemistry/mission2mars/contents/glossary/s.htm
that seems to be what I'm trying to get at:
Scientific law: a statement describing an observed
regularity. Laws describe rather than explain. A law may be
either quantitative (including measurement) or qualitative
(general characteristics). It must describe evidence that has
been gathered using acceptable scientific standards of
reproducibility.
What do you think?
> > ----- END: What is Science FAQ Draft III -----
>
> Despite my quibbles here, this is a very creditable effort, Marty. I am
> impressed.
Thank you John. That is high praise indeed, and if it is deserved, it
is because I have, in the past few years, learned a great deal from
our exchanges.
Marty
Brian M. Scott
<usenet_p...@yahoo.com> wrote:
>john.w...@bigpond.com (John Wilkins) writes:
>> Marty Fouts <usenet_p...@yahoo.com> wrote:
[...]
This has the virtue of covering the usage in historical linguistics,
which I consider a science, albeit an historical rather than an
experimental one: the normal relationship between PIE and PGmc
consonants is described by Grimm's Law as modified by Verner's Law.
These are purely descriptive.
[...]
Brian
John Wilkins
But one could, in principle, give a mathematical description of Grimm's
Law or any other regularity, or it is not a regularity at all.
As to a model being a mathematical relationship or not - there are, of
course, other criteria (being interpretable, for a start). But that
discussion has no place here.
Welumen
"Marty Fouts" <usenet_p...@yahoo.com> wrote in message
news:uvg7on...@yahoo.com...
>
> o Observation: Scientists observe the universe to see how it
> operates. The raw data of science is accumulated through
> observation. The methods of observation range from going into the
> field and recording what is there through controlled laboratory
> experiments.
The last bit here could be rephrased slightly to remove a bit of ambiguity.
"The methods of observation range from going into the field and recording
what is there, to controlled laboratory experiments."
> Q: What is the role of consistency in science?
>
> A: Whenever scientists make a new observation or propose a new
> explanation, they require that it be consistent with existing
> observations and explanations. Knowledge is consistent so long as
> reasoning about it doesn't lead to contradictory conclusions.
>
> When an observation is made that seem inconsistent with what is
> already known, the knowledge is carefully
"When an observation is made that seem[s] inconsistent..."
>
> Q: What is the difference between a theory and a fact?
>
> A: In science, a fact is an observation or series of observations, or
> a description of a series of observations. Facts are answers to
> what is kinds of questions, and are the part of science that can be
> the most objective. A theory is an attempt to explain known facts
> and usually involves predictions about future
> observations. Theories are answers to how does kinds of questions.
Could be modified to reduce ambiguity. "Facts are answers to "what is" kinds
of questions..." "Theories are answers to "how does" kinds of questions."
Welumen
"Marty Fouts" <usenet_p...@yahoo.com> wrote in message
news:uvg7on...@yahoo.com...
>
> o Observation: Scientists observe the universe to see how it
> operates. The raw data of science is accumulated through
> observation. The methods of observation range from going into the
> field and recording what is there through controlled laboratory
> experiments.
The last bit here could be rephrased slightly to remove a bit of ambiguity.
"The methods of observation range from going into the field and recording
what is there, to controlled laboratory experiments."
> Q: What is the role of consistency in science?
>
> A: Whenever scientists make a new observation or propose a new
> explanation, they require that it be consistent with existing
> observations and explanations. Knowledge is consistent so long as
> reasoning about it doesn't lead to contradictory conclusions.
>
> When an observation is made that seem inconsistent with what is
> already known, the knowledge is carefully
"When an observation is made that seem[s] inconsistent..."
>
> Q: What is the difference between a theory and a fact?
>
> A: In science, a fact is an observation or series of observations, or
> a description of a series of observations. Facts are answers to
> what is kinds of questions, and are the part of science that can be
> the most objective. A theory is an attempt to explain known facts
> and usually involves predictions about future
> observations. Theories are answers to how does kinds of questions.
Could be modified to reduce ambiguity. "Facts are answers to "what is" kinds
of questions..." "Theories are answers to "how does" kinds of questions."
Welumen
"Marty Fouts" <usenet_p...@yahoo.com> wrote in message
news:uvg7on...@yahoo.com...
>
> o Observation: Scientists observe the universe to see how it
> operates. The raw data of science is accumulated through
> observation. The methods of observation range from going into the
> field and recording what is there through controlled laboratory
> experiments.
The last bit here could be rephrased slightly to remove a bit of ambiguity.
"The methods of observation range from going into the field and recording
what is there, to controlled laboratory experiments."
> Q: What is the role of consistency in science?
>
> A: Whenever scientists make a new observation or propose a new
> explanation, they require that it be consistent with existing
> observations and explanations. Knowledge is consistent so long as
> reasoning about it doesn't lead to contradictory conclusions.
>
> When an observation is made that seem inconsistent with what is
> already known, the knowledge is carefully
"When an observation is made that seem[s] inconsistent..."
>
> Q: What is the difference between a theory and a fact?
>
> A: In science, a fact is an observation or series of observations, or
> a description of a series of observations. Facts are answers to
> what is kinds of questions, and are the part of science that can be
> the most objective. A theory is an attempt to explain known facts
> and usually involves predictions about future
> observations. Theories are answers to how does kinds of questions.
Could be modified to reduce ambiguity. "Facts are answers to "what is" kinds
of questions..." "Theories are answers to "how does" kinds of questions."
John Wilkins
> john.w...@bigpond.com (John Wilkins) writes:
>
> > Marty Fouts <usenet_p...@yahoo.com> wrote:
>
....
> > > The resulting epistemology of western science is often called
> > > empirical pragmatism. This philosophy is grounded in the
> > > understanding that all scientific knowledge is provisional
> > > knowledge and that any scientific knowledge may be rendered
> > > obsolete by a future observation. It recognizes that, in order
> > > to make any progress at all, science must take as given certain
> > > assumptions that can not be validated, but that have been very
> > > reliable for a long time, and so, will be used until they are
> > > invalidated.
> >
> > [JW: Strictly speaking, pragmatism is the view that truth is, for all
> > intents and purposes, the same as what works out in practice (hence the
> > name).]
>
> I thought pragmatism was only caring about what works out in practice
> and forgetting about 'truth'. Is there a better wording for the above
> paragraph?
Pragmatism was originally a view that truth is what works, and that
remains the philosophical interpretation*. In common use, it means just
"what works". However, if you are going to characterise the epistemology
of science, you had best say it is empirical in nature. Scientists do
not have a Union Epistemology. Some are rationalists, some are
inductivists, some are Popperians, and so help me, some are still
logical positivists. But the one common thread is that science is
empiricallly driven, as the rest of the paragraph says.
May I suggest this? -
"As a result, science is founded upon an empirical, fallibilistic
epistemology. It is driven by the data and the understanding that all
scientific knowledge is to some degree provisional, and may be rendered
incomplete, or obsolete, by later empirical observation. [The last
sentence is fine.]
....
* That is to say, William James' reinterpretation of Peirce's semiotic
account was. In any case, the view now known as pragmatism is very much
of this view - cf. Rescher.
AR
Marty Fouts wrote:
> AR <roto...@pitt.edu> writes:
>
> [snip]
>
> I'm incorporating those changes that I understand and agree
> with. Comments here are to try to find common ground.
>
> > > Q: How does science add to our knowledge?
> >
> > via theories. These are the "knowledge" or explanations of science
>
> It seems to me that theories are only part of the knowledge of
> science. Much of the knowledge is found in the observations as well,
> for example in fields like geology, where much of the literature is
> descriptive in nature.
I'm not very familiar with the science of geology. But I would assume
that the observations are theory laden, and it is only with the theory
that the observations have ANY! meaning (see my example below).
In and of themselves, "observations" aren't science. Just knowing
what to observe, requires theory of some sort.
> [snip]
>
>
> > >
> > > o scientific induction: the belief that a sufficient number of
> > > observations of similarity can be used to generalize,
> >
> > This seems to be an issue of theory not observation. That is, we
> > believe we can generalize from a to b because we believe the
> > explanation of what is occuring in a applies to the conditions of b.
>
> if every dog I see has 4 legs, then I generalize to the assumption
> that all dogs have 4 legs. It is from the observations that the
> generalization comes. (Of course, as soon as I see a 3 legged dog I
> have to modify the generalization...)
It seems to me that the observation that you have made is a "live" "creature"
with 4 "legs".
All of the words that I've put in quotes seem to me to be theory (more
or less, some may be models depending on how you want to slice it). There are
many living creatures, some have no legs. You believe that the next "dog"
you see will have 4 legs, because of your theory/model of dog--what
you believe to be a dog. This of course is developed from observations, the
theory and observations are not completely separate entities.
It's not because of the observations that you think the next dog will have 4
legs, it's because of
your model of dog. You've defined dog, this leads you to believe it has 4 legs
and so will the next. If the next doesn't, as in your example, and you had no
theory of dog, then it
wouldn't be a dog. It's you model that tells you it's a dog, not your
observation
of the number of legs it has, or whether it has a tail, 2 ears . .. Thus,
contrary to what you've
said, you wouldn't have to modify your generalization, it's a dog due to your
conceptualization of dog, not your observaton per say.
(As an aside, based on your theories/models of the live creatures you know,
you could develop generalizations about other live creatures that you see,
even after very few observations
of these new creatures. The reason is because you think that certain
conceptual
elements/models of the live creatures that you know well, are relevant to
creatures
you've seen rarely. The generalization is based on concepts/theory, otherwise
there would
really be no reason to generalize.)
Some dogs are male some female. If your generalization were only based on
observations, you might very well say all dogs are male, all dogggs are
female.
And you'd be correct. It is because of your theory of dog that you don't say
this.
Observation doesn't solve this issue, how you conceptualize the observations
determines generalization, and that's theory.
>
>
> [snip]
>
> > > A: In science, a fact is an observation or series of observations, or
> > > a description of a series of observations. Facts are answers to
> > > what is kinds of questions, and are the part of science that can be
> > > the most objective. A theory is an attempt to explain known facts
> > > and usually involves predictions about future
> > > observations. Theories are answers to how does kinds of questions.
> >
> > Saying that a fact is an observation, seems to minimize the word
> > fact, and elevate the meaning of an observation. It doesn't seem to
> > me that observations can be divorced from theory, at least not as
> > cleanly as this may imply. I don't see how an observation can be
> > dissociated from theory. [snip] My point is simply that the
> > distinction between observation or fact? can probably not be so
> > cleanly made. They are very intertwined adn not clearly distinct
> > entities.
>
> I'm not sure what you want me to say differently.
I appologize, what I wrote was not very clear at all.
My impression is that you want to make a clean distinction between observation
and theory. That is, a valid and reliable observation is a fact. Where as a
theory is a concept.
I don't think that such a distinction is realistic. I gave an example of the
mean/average
which was an observation, but could not be observed, because it didn't
actually exist
in the population under study. This piece of data, the average, is a
conceptual entity, which
to me makes it theory, or at least "theoretical", so to speak.
Hopefully by continuing I won't cloud what I'm trying to say. A fact is
subjective, in the
sense that it is what "we" accept as truth, because the preponderance of
theory and
observation is overwhelming. And in many cases it may in fact be the truth.
But there
certainly are no criteria (scientific or otherwise) for determining
objectively what is fact
or when an observation becomes a fact.
Regards
Brian M. Scott
>> >john.w...@bigpond.com (John Wilkins) writes:
>> >> Marty Fouts <usenet_p...@yahoo.com> wrote:
>> [...]
>But one could, in principle, give a mathematical description of Grimm's
>Law or any other regularity, or it is not a regularity at all.
I'm not sure that we have the same understanding of 'mathematical'.
My notion of mathematics is pretty broad, but I don't see any useful
way to describe Grimm's Law mathematically. Does calling a consistent
relationship between an 'input' and an 'output' a function really
constitute in any useful sense a mathematical description? Even if
you think that it does, you might want to consider that most readers
of the FAQ will probably understand 'mathematical' as 'quantitative',
which I think is definitely too limiting.
[...]
Brian
Stephen Poley
<usenet_p...@yahoo.com> wrote:
>Please comment here. I'll incorporate changes, and when the
>discussion seems to have settled, submit it to the web site as a FAQ.
Looks good. A few suggestions here that haven't (as far as I can see)
already been raised.
>----- START: What is Science FAQ Draft III -----
>
>Welcome to talk.origins!
>Copyright © 1999-2002 by Marty Fouts
>
>[Please note that Marty Fouts is the author of this specific file and not the maintainer of this web site. If you have questions or comments about things you read on this page, feel free to contact Marty. If you have comments about anything else on this site, either send them to the author of the page in question, or respond through the feedback page.]
Not sure if this is necessary, but doubtless it will be rendered
consistent with the rest of the site in due course.
<snip>
>Q: What makes a method scientific?
>
>A: The principle contribution of science to the human quest for
> knowledge is in objectifying the four categories of method.
>
> An objective approach to knowledge is one in which two observers
> using the same methods will arrive at the same knowledge, since the
> knowledge depends only on the observation and not on the
> observer. Clearly there are no purely objective methods, although
> scientific methods strive to be as objective as humanly possible.
>
> The idea of independently repeating an experiment to show that the
> observation is objective is useful, especially in laboratory
> sciences, but taken alone, it only guarantees that two scientists
> arrived at the same result. It does not guarantee that the method
> itself is valid.
>
> To reduced the risk of performing bad experiments, scientists
> describe their methods to other scientists, who look for errors in
> the approach. The formal method of doing this is by publishing
> articles in peer-reviewed journals.
Actually the journals don't have to be peer-reviewed, although in recent
decades almost all journals have been. I suggest rewording the last
sentence as something like: "The formal method of doing this is by
publishing articles in journals. Obvious errors and unimportant articles
are normally weeded out by a process of peer review before publication
(and this is nowadays necessary to keep the number of articles
manageable)".
<snip>
>Q: Can you sum up 'scientific method', again, please?
>
>A: The literature on scientific method, then, can be seen as a
> discussion of the ways in which scientists attempt to objectify the
> four activities of science. For instance:
>
> o Observation: through the use of precise measurements, careful
> experimental design and control, and repeatability. Observation
> is the basis of science.
>
> o Description: through the use of comparison to objective
> standards, most notably, again, measurements. Descriptions are
> built upon observations.
Worth mentioning that descriptions wherever possible make use of
generally-accepted terminology; where new terms are needed they are
carefully defined.
<snip>
>Q: What is the difference between a theory and a fact?
>
>A: In science, a fact is an observation or series of observations, or
> a description of a series of observations. Facts are answers to
> what is kinds of questions, and are the part of science that can be
> the most objective. A theory is an attempt to explain known facts
> and usually involves predictions about future
> observations. Theories are answers to how does kinds of questions.
As this is one of the most widely misunderstood points it's perhaps
worth a bit more explanation. I suggest adding:
' Note that scientific theories are firmly based on facts. This differs
from the everyday use of the word "theory", which has a meaning closer
to the scientific term "hypothesis" or even "speculation". Statements
such as "evolution is just a theory" rest on a misunderstanding of what
a theory is.
Note that the progressive nature of science means that theories can
become the facts on which new theories are based. For example, the
discovery of a number of anomalous geological formations in Europe led
to the theory of ice ages to explain them. The occurrence of ice ages is
now so solidly supported by observation that it is regarded as fact, and
various astronomical hypotheses have been proposed to explain the fact.
Ice ages are thus both theory and fact. '
<snip>
>Q: But what about the scientists?
>
>A: A significant source of confusion in discussions of the philosophy
> and history of science is to intermix the sociology of scientists
> with the epistemological method of science. Scientists are humans
> and suffer from all of the strengths and weaknesses of humans, and
> so the quest for scientific knowledge has a very human history. But
> scientific epistemology, which is the theory that results from the
> practice of science, is an abstraction of human though rather than
> a sequence of human actions.
Maybe worth adding something along the following lines:
' Claims are sometimes made about scientific results being influenced by
social attitudes - and thus by implication being unreliable. While
social attitudes and interests may well influence which areas of science
receive most attention, they do not influence the results (except
perhaps briefly, if they allow an occasional dubious paper to make it
into print, and with the obvious exception of sociology itself). '
>Q: Is science a religion?
>
>A: No. However, many people have made a religion out of a certain set
> of believes in what science can and can not do. To those people,
> the use of science has taken a religious role.
"many people"? Not sure exactly who you are thinking of here -
scientologists and such-like?. I think this answer is a little too open
to misunderstanding. My suggestion:
"No. There are varying views as to what constitutes a religion, but
under no reasonable definition of religion is science a religion, or
even similar to one. (However some people have made a religion out of a
set of ill-founded beliefs about what science is and does. To those
people, the use of science has taken a religious role.)"
<snip>
>
>----- END: What is Science FAQ Draft III -----
>
Last point: it is maybe worth mentioning somewhere, as it's often
misunderstood, that while mathematics is very important for science, it
is not itself a branch of science.
Stephen Poley
Barendrecht, Holland
John Wilkins
I agree with your last comment, but IMM a model is of the form
a = f(x)
where the function describes the relation. If you can say that phoneme
replacement occurs at some recognisable frequency, then you have
sketched a function, even if it is not quantitative. Let's not forget
that natural selection was nonquantitative for the first 70 years of its
existence as well... But it was in-principle quantifiable. It just took
a genius like Fisher and the particulate model of heredity to make it
so... Is it my fault that those poor linguists haven't been able to
specify their entities well enough to measure and model Grimm's Law
properly?
>
> [...]
>
> Brian
Marty Fouts
sbp...@xs4all.nl (Stephen Poley) writes:
[as with all responses, I've snipped the parts that I've used to
update the FAQ and left in only the bits I'd like to discuss]
[snip]
> >Q: But what about the scientists?
> >
> >A: A significant source of confusion in discussions of the philosophy
> > and history of science is to intermix the sociology of scientists
> > with the epistemological method of science. Scientists are humans
> > and suffer from all of the strengths and weaknesses of humans, and
> > so the quest for scientific knowledge has a very human history. But
> > scientific epistemology, which is the theory that results from the
> > practice of science, is an abstraction of human though rather than
> > a sequence of human actions.
>
> Maybe worth adding something along the following lines:
>
> ' Claims are sometimes made about scientific results being
> influenced by social attitudes - and thus by implication being
> unreliable. While social attitudes and interests may well influence
> which areas of science receive most attention, they do not influence
> the results (except perhaps briefly, if they allow an occasional
> dubious paper to make it into print, and with the obvious exception
> of sociology itself). '
I'm not sure how controversial what I'm about to say is, but I don't
agree. I frankly think that social attitudes have a major influence on
scientists and have a huge influence on the results.
While the competition among scientists tends to eventually correct the
biases, the biases seem to be there and seem to be an ongoing part of
the contribution of scientists to civilization.
Look, for example, at how readily scientists have supressed negative
information in drug studies, and the resulting consequence. While it
would be tempting to say that such scientists are rare and such
consequences unusual, I don't think that the history of science in the
service of commerce would bear out that claim.
[snip]
Marty Fouts
AR <roto...@pitt.edu> writes:
[snip]
> My impression is that you want to make a clean distinction between
> observation and theory. That is, a valid and reliable observation is
> a fact. Where as a theory is a concept. I don't think that such a
> distinction is realistic. I gave an example of the mean/average
> which was an observation, but could not be observed, because it
> didn't actually exist in the population under study. This piece of
> data, the average, is a conceptual entity, which to me makes it
> theory, or at least "theoretical", so to speak.
I want a range from 'pure' observation at one extreme to 'pure' theory
at the other but I don't want to burden a simple FAQ with explaining
that gamut.
I accept that beyond a certain simple level observations are
theory-based, I do not believe that all observation arises from
theory.
What theory lies behind observational astronomy, as demonstrated by
calendars?
> Hopefully by continuing I won't cloud what I'm trying to say. A fact
> is subjective, in the sense that it is what "we" accept as truth,
> because the preponderance of theory and observation is
> overwhelming. And in many cases it may in fact be the truth. But
> there certainly are no criteria (scientific or otherwise) for
> determining objectively what is fact or when an observation becomes
> a fact.
Here we disagree completely, and I'm not sure how to resolve the
disagreement in the context of the FAQ. For the moment, I'm going to
take the view that this distinction is too subtle for a 'what is
science' FAQ and that if I'm wrong what I've written is a 'suitable
lie for beginners' and so punt on trying to accomodate your view in
the FAQ.
I accept that the distinction between 'observation' and 'fact' is not
clear and may not be worth making but I disagree with the assertion
that all observations are theory based.
Since we can't hope to resolve this in a simple FAQ, I'd like to just
leave the issue lie dormant in the FAQ, but I'd be happy to continue
discussing the topic in a broader context.
my own belief is that observations without theory can come first. One
notices that the nile floods periodicially and that the the flooding
has some consistency with the configuration of stars in the sky before
one conceives of the idea of 'year', i think.
marty
Stephen Poley
<usenet_p...@yahoo.com> wrote:
>sbp...@xs4all.nl (Stephen Poley) writes:
>
>[as with all responses, I've snipped the parts that I've used to
>update the FAQ and left in only the bits I'd like to discuss]
Glad some of them were of use.
I think it's fair to say that such cases are (unfortunately) not so very
unusual. My contention would be that having an ineffective drug on the
market for a few years counts as a minor and brief hiccup in the context
of the whole science of mammalian physiology. (On the other hand I dare
say it doesn't feel very minor if you're one of the poor sods who gets
lumbered with the drug and suffers the side-effects without any
benefits.)
*However*, having thought about it a bit further, I agree that this
suggestion of mine goes down a path that isn't really suitable for this
FAQ, and I'm entirely happy for you to drop it.
Stephen Poley
Barendrecht, Holland
AR
Marty Fouts wrote:
> AR <roto...@pitt.edu> writes:
>
> [snip]
>
> > My impression is that you want to make a clean distinction between
> > observation and theory. That is, a valid and reliable observation is
> > a fact. Where as a theory is a concept. I don't think that such a
> > distinction is realistic. I gave an example of the mean/average
> > which was an observation, but could not be observed, because it
> > didn't actually exist in the population under study. This piece of
> > data, the average, is a conceptual entity, which to me makes it
> > theory, or at least "theoretical", so to speak.
>
> I want a range from 'pure' observation at one extreme to 'pure' theory
> at the other but I don't want to burden a simple FAQ with explaining
> that gamut.
>
> I accept that beyond a certain simple level observations are
> theory-based, I do not believe that all observation arises from
> theory.
>
> What theory lies behind observational astronomy, as demonstrated by
> calendars?
This is a very good example. As is the one you provide below. It really
seems to get at the heart of the issue. My bias is that first one would
have
to notice recurrences or at least conceptualize the idea and that would
drive further observation. Especially for the development of a calendar.
However, there is no question, especially with your flooding example
below, that one might be able to notice the association betwen
flooding and the stars before one develops a speculation about what
may be going on, or what to observe. I would think that first the observer
would have to have a "notion" of the nature of the star patters: they vary
through the night (east to west), and vary over time (height in the sky)
before one could notice a connection between the flooding and the
patterns (given the inherent "movement" of the stars through the night
and the seasons, they are always in different places if you haven't
realized
these basic "principles" of their movement). But maybe not. Either way,
I agree the issues are not appropriate for a FAQ, as you say.
>
>
> > Hopefully by continuing I won't cloud what I'm trying to say. A fact
> > is subjective, in the sense that it is what "we" accept as truth,
> > because the preponderance of theory and observation is
> > overwhelming. And in many cases it may in fact be the truth. But
> > there certainly are no criteria (scientific or otherwise) for
> > determining objectively what is fact or when an observation becomes
> > a fact.
>
> Here we disagree completely, and I'm not sure how to resolve the
> disagreement in the context of the FAQ. For the moment, I'm going to
> take the view that this distinction is too subtle for a 'what is
> science' FAQ and that if I'm wrong what I've written is a 'suitable
> lie for beginners' and so punt on trying to accomodate your view in
> the FAQ.
>
> I accept that the distinction between 'observation' and 'fact' is not
> clear and may not be worth making but I disagree with the assertion
> that all observations are theory based.
>
> Since we can't hope to resolve this in a simple FAQ, I'd like to just
> leave the issue lie dormant in the FAQ, but I'd be happy to continue
> discussing the topic in a broader context.
I think you are right that it maybe more than should be in a FAQ.
Brian M. Scott
Wilkins) wrote:
>> >> >john.w...@bigpond.com (John Wilkins) writes:
>> >> >> Marty Fouts <usenet_p...@yahoo.com> wrote:
> a = f(x)
I can't agree. On the one hand, this is too narrow: a mathematical
model may have to specify relationships amongst many variables. And
on the other hand it is arguably too broad for just the reason that I
gave before: yes, you can do it with Grimm's Law, say, but it's not
clear that you've gained anything by using the word 'function' to
describe the relationship.
>If you can say that phoneme
>replacement occurs at some recognisable frequency, then you have
>sketched a function, even if it is not quantitative.
I don't quite understand what you mean here. Oversimplifying a bit,
what you actually have is a very consistent relationship (when you
throw in Verner's Law) with a relatively small number of exceptions
that are generally to be explained individually. The replacement
frequency can be taken to be just about 100%, though later changes may
obscure this. There simply isn't a quantitative element here: you
simply have PIE */g-/ > Gmc */k-/, for instance.
>Let's not forget
>that natural selection was nonquantitative for the first 70 years of its
>existence as well... But it was in-principle quantifiable. It just took
>a genius like Fisher and the particulate model of heredity to make it
>so... Is it my fault that those poor linguists haven't been able to
>specify their entities well enough to measure and model Grimm's Law
>properly?
Again, this doesn't quite seem to make sense. What do you understand
Grimm's Law to say?
Brian
Richard Harter
As a small comment here, I suggest that there are subtly different
senses of the word 'theory' at play here. There is a sense in which
all facts are theory-laden, i.e., one needs models of what counts and
does not count in observations, classfication schemes, et cetera.
However I submit that this usage of 'theory' is different from the
usage in 'scientific theory'. Crudely, one sense of the word is
concerned with what counts as facts, the other with what explains
facts.
More subtly, the issue is the difference between problematic and
unproblematic data. The investigations of science do not depend upon
having absolute truth everywhere. Instead they proceed by a process
of incremental refinement. In the particular arena of interest all is
open to question, all needs must be explained. Such arenas of
inquiry, however, are supported by "unproblematic background data".
That is, in any particular investigation we assume that the knowledge
schemas of the rest of the science enterprise can be relied upon
without calling them into question. Sometimes, it is true,
"background data" is called into question during an investigation, but
only if it is directly relevant.
Richard Harter, c...@tiac.net,
http://home.tiac.net/~cri, http://www.varinoma.com
Small children should be raised in the country, far from the centers of
civilization. More importantly, they should be kept there until grown.
Richard Harter
(John Wilkins) wrote:
>[JW: I objewct to words like "objectified". If we mean, "the way science
>gets it right" or "ensures it is realistic" or whatever, then say so.
>"Objectified" is a word used by people who also use words like
>"structuralism" "modernism" and "tropes" and deny the existence of
>authors...]
I use words such as "structuralism" "modernism" and "tropes"; however
I do not deny the existence of authors except, perhaps, that of
Wilkins.
John Wilkins
> On Thu, 11 Jul 2002 11:34:24 +0000 (UTC), john.w...@bigpond.com
> (John Wilkins) wrote:
>
>
>
> >[JW: I objewct to words like "objectified". If we mean, "the way science
> >gets it right" or "ensures it is realistic" or whatever, then say so.
> >"Objectified" is a word used by people who also use words like
> >"structuralism" "modernism" and "tropes" and deny the existence of
> >authors...]
>
> I use words such as "structuralism" "modernism" and "tropes"; however
> I do not deny the existence of authors except, perhaps, that of
> Wilkins.
>
It's true. You have uncovered my nasty little secret - I lack existence.
I tried to be real, you know, but when I realised that the Ontological
Proof failed to work for God, I also realised that it failed to work for
me as well, and hence the Cogito also failed, since Ens is a
precondition for that inference. I am saddened, but then again,
strangely not.
Marty Fouts
wil...@wehi.edu.au (John Wilkins) writes:
That's too bad, because we would really have missed you had you been
here and gone
John Wilkins
> >> >> >ossary/s.htm that seems to be what I'm trying to get at:
f(x) is a generic term. It does not mean that there cannot be a number
of further functions decomposable in that model. It just means that
there is a determinate relation, more or less, between the input and the
output.
>
> >If you can say that phoneme
> >replacement occurs at some recognisable frequency, then you have
> >sketched a function, even if it is not quantitative.
>
> I don't quite understand what you mean here. Oversimplifying a bit,
> what you actually have is a very consistent relationship (when you
> throw in Verner's Law) with a relatively small number of exceptions
> that are generally to be explained individually. The replacement
> frequency can be taken to be just about 100%, though later changes may
> obscure this. There simply isn't a quantitative element here: you
> simply have PIE */g-/ > Gmc */k-/, for instance.
A binary relation is still a relation, but I get what you mean here -
the sort of mathematical relation is pretty sparse in its detail. But
then I would say that Grimm's Law is hardly a law in the scientific
sense - at best it is a simple summary of observation. There is no
theoretical constraint upon further phoneme changes.
>
> >Let's not forget
> >that natural selection was nonquantitative for the first 70 years of its
> >existence as well... But it was in-principle quantifiable. It just took
> >a genius like Fisher and the particulate model of heredity to make it
> >so... Is it my fault that those poor linguists haven't been able to
> >specify their entities well enough to measure and model Grimm's Law
> >properly?
>
> Again, this doesn't quite seem to make sense. What do you understand
> Grimm's Law to say?
Not much. I never did linguistics. Just that there are some general
transformations of phonemes over time.
Brian M. Scott
Wilkins) wrote:
>> >> >> >john.w...@bigpond.com (John Wilkins) writes:
>> >> >> >> Marty Fouts <usenet_p...@yahoo.com> wrote:
>> > a = f(x)
My objection is to the notion that it necessarily makes sense to talk
about input and output, and now that I've thought about it a bit more,
I realize that my previous objection didn't go far enough. I consider
it entirely acceptable to say that the Euclidean geometry of R^3 is a
decent mathematical model of physical space on an everyday scale, for
instance; where is the functional relationship there? Similarly, the
use of (combinatorial) graphs to model all sorts of things (the London
Underground, acquaintance networks, communication networks, scheduling
dependencies for final exams, etc.) doesn't seem to fit your
definition.
>> >If you can say that phoneme
>> >replacement occurs at some recognisable frequency, then you have
>> >sketched a function, even if it is not quantitative.
>> I don't quite understand what you mean here. Oversimplifying a bit,
>> what you actually have is a very consistent relationship (when you
>> throw in Verner's Law) with a relatively small number of exceptions
>> that are generally to be explained individually. The replacement
>> frequency can be taken to be just about 100%, though later changes may
>> obscure this. There simply isn't a quantitative element here: you
>> simply have PIE */g-/ > Gmc */k-/, for instance.
>A binary relation is still a relation, but I get what you mean here -
>the sort of mathematical relation is pretty sparse in its detail. But
>then I would say that Grimm's Law is hardly a law in the scientific
>sense - at best it is a simple summary of observation. There is no
>theoretical constraint upon further phoneme changes.
It can't be a simple summary of observation, since none of the
entities involved is in fact observable: they're all theoretical
constructs. It's also more than just a list of inferred changes,
since the changes themselves are structured: ignoring refinements and
boundary conditions, all aspirated, voiced stops lose the aspiration;
all unaspirated, voiced stops lose the voicing; and all unvoiced stops
lose the 'stop', i.e., become the corresponding unvoiced fricative.
[...]
Brian
John Wilkins
> On Sun, 21 Jul 2002 22:57:14 +0000 (UTC), wil...@wehi.edu.au (John
> Wilkins) wrote:
>
<snipping - the sequence should be clear between Brian and I>
> >> >I agree with your last comment, but IMM a model is of the form
>
> >> > a = f(x)
>
> >> >where the function describes the relation.
>
> >> I can't agree. On the one hand, this is too narrow: a mathematical
> >> model may have to specify relationships amongst many variables. And
> >> on the other hand it is arguably too broad for just the reason that I
> >> gave before: yes, you can do it with Grimm's Law, say, but it's not
> >> clear that you've gained anything by using the word 'function' to
> >> describe the relationship.
>
> >f(x) is a generic term. It does not mean that there cannot be a number
> >of further functions decomposable in that model. It just means that
> >there is a determinate relation, more or less, between the input and the
> >output.
>
> My objection is to the notion that it necessarily makes sense to talk
> about input and output, and now that I've thought about it a bit more,
> I realize that my previous objection didn't go far enough. I consider
> it entirely acceptable to say that the Euclidean geometry of R^3 is a
> decent mathematical model of physical space on an everyday scale, for
> instance; where is the functional relationship there? Similarly, the
> use of (combinatorial) graphs to model all sorts of things (the London
> Underground, acquaintance networks, communication networks, scheduling
> dependencies for final exams, etc.) doesn't seem to fit your
> definition.
I think, for example, that in Newton's theory, a space of *six*
dimensions is required (lessee, IIRC it's x, y, z, direction, momentum
and acceleration), but that merely establishes the variables to be used
in the actual theory. The theoretical space is not the theory, as such.
The theory is the functional relationships that obtain between any
entities in that space. I am not saying that a sheet of graph paper is a
theory. I am saying that a theory is a prediction that entities in that
space will closely follow the surfaces that the functions describe in it
(pathways on the graph paper).
For example, take the parabolic pathway of projectiles in a simple
Newtonian system. There is a relation between the mass, the angular
velocity, the direction and the acceleration field (gravitational value)
of the projectile. This is a "surface" (in 2-d) on which any projectile
will travel in the absence of perturbations. It sets up, as it were, the
default and expected range of behaviours of all projectiles, and allows
us to predict things we would not be able to do in the absence of such a
theory (that in a point centered graviational field, orbits are
possible, for example).
Take something with three variables (e.g., the Ideal Gas Laws) and a 3-d
surface arises. Take more variables and more complex dynamics are
described by the functional relations.
The London Underground may be a theorem in some multidimensional space,
I don't know. But queue theory suggests that only a circumscribed range
of dynamics are possible in such a connected graph. Trains do not pass
through each other, for example. There can be a limiting set of
functions even on something as irrational as a metropolitan train
system...
>
> >> >If you can say that phoneme
> >> >replacement occurs at some recognisable frequency, then you have
> >> >sketched a function, even if it is not quantitative.
>
> >> I don't quite understand what you mean here. Oversimplifying a bit,
> >> what you actually have is a very consistent relationship (when you
> >> throw in Verner's Law) with a relatively small number of exceptions
> >> that are generally to be explained individually. The replacement
> >> frequency can be taken to be just about 100%, though later changes may
> >> obscure this. There simply isn't a quantitative element here: you
> >> simply have PIE */g-/ > Gmc */k-/, for instance.
>
> >A binary relation is still a relation, but I get what you mean here -
> >the sort of mathematical relation is pretty sparse in its detail. But
> >then I would say that Grimm's Law is hardly a law in the scientific
> >sense - at best it is a simple summary of observation. There is no
> >theoretical constraint upon further phoneme changes.
>
> It can't be a simple summary of observation, since none of the
> entities involved is in fact observable: they're all theoretical
> constructs. It's also more than just a list of inferred changes,
> since the changes themselves are structured: ignoring refinements and
> boundary conditions, all aspirated, voiced stops lose the aspiration;
> all unaspirated, voiced stops lose the voicing; and all unvoiced stops
> lose the 'stop', i.e., become the corresponding unvoiced fricative.
Sounds are not theoretical constructs, only the ways of describing them.
However, I know zip about this field, so I'll shut up about now.
>
> [...]
>
> Brian
Brian M. Scott
>> >> > a = f(x)
But John, this is completely beside the point. The question is what
is meant by 'mathematical model'. R^3 is a mathematical model of the
physical space in which we live. And since Euclidean geometry works
very well on a normal human scale, it's a fairly good model. At no
point have I been talking about what constitutes a theory. My
objection from the beginning has been to your use of the term
'(mathematical) model'.
[...]
>> > But
>> >then I would say that Grimm's Law is hardly a law in the scientific
>> >sense - at best it is a simple summary of observation. There is no
>> >theoretical constraint upon further phoneme changes.
>> It can't be a simple summary of observation, since none of the
>> entities involved is in fact observable: they're all theoretical
>> constructs. It's also more than just a list of inferred changes,
>> since the changes themselves are structured: ignoring refinements and
>> boundary conditions, all aspirated, voiced stops lose the aspiration;
>> all unaspirated, voiced stops lose the voicing; and all unvoiced stops
>> lose the 'stop', i.e., become the corresponding unvoiced fricative.
>Sounds are not theoretical constructs, only the ways of describing them.
>However, I know zip about this field, so I'll shut up about now.
Okay, but I'll try to give a brief explanation anyway.
First, phonemes aren't sounds. In a traditional view they are classes
of sounds whose number and make-up are theory-based and depend on the
specific language (or dialect). Old English, for instance, had both
the [T] sound of <th> in <thin> and the [D] sound of <th> in <this>,
but they were allophones of a single phoneme /T/. That is, the choice
of unvoiced [T] or voiced [D] was completely determined by the
phonetic context. For instance, initially /T/ was always [T], but
intervocalically it was always [D]. The sounds [T] and [D] were both
'part of' the Old English phoneme /T/. (And even this is a bit
misleading, since the phonetic symbols [T] and [D] themselves cover a
range of pronunciations not normally distinguished except in very
narrow transcriptions. Some speakers of English make them
interdentally, with the tongue actually protruding between the upper
and lower teeth; others make them with a more retracted tongue whose
tip lies just behind the upper teeth.)
Secondly, since neither Proto-Indo-European nor Common Germanic is
attested even in writing, let alone in aural recordings, the existence
and phonetic nature of PIE */g/ are theory-driven conclusions. (And
in fact there are two competing theories about its phonetic nature.)
Brian
John Wilkins
> On Sun, 28 Jul 2002 10:05:17 +0000 (UTC), john.w...@bigpond.com
> (John Wilkins) wrote:
>
> >Brian M. Scott <b.s...@csuohio.edu> wrote:
>
....
> >I think, for example, that in Newton's theory, a space of *six*
> >dimensions is required (lessee, IIRC it's x, y, z, direction, momentum
> >and acceleration), but that merely establishes the variables to be used
> >in the actual theory. The theoretical space is not the theory, as such.
> >The theory is the functional relationships that obtain between any
> >entities in that space. I am not saying that a sheet of graph paper is a
> >theory. I am saying that a theory is a prediction that entities in that
> >space will closely follow the surfaces that the functions describe in it
> >(pathways on the graph paper).
>
> But John, this is completely beside the point. The question is what
> is meant by 'mathematical model'. R^3 is a mathematical model of the
> physical space in which we live. And since Euclidean geometry works
> very well on a normal human scale, it's a fairly good model. At no
> point have I been talking about what constitutes a theory. My
> objection from the beginning has been to your use of the term
> '(mathematical) model'.
You are going to have to elaborate for me - I'm (it transpires) dense,
not deep. Start by what you mean with R^3. So far as I understand
Euclidean geometry, it merely sets out the "sensorium" in which
Newtonian models occur. It is not a model itself, but a set of axioms.
Then tell me what you mean by "model" here. For me, a model is some
interpretable mathematical relation (interpretable meaning that you can
map the function to some data set). This is perhaps a philosophical
idiom and causing some confusion.
Huh. Interesting. I must read up on this when I get a life (starting
next year, I think)
Sven Silow
>Brian M. Scott <b.s...@csuohio.edu> wrote:
>
>> On Sun, 28 Jul 2002 10:05:17 +0000 (UTC), john.w...@bigpond.com
>> (John Wilkins) wrote:
>>
>> >Brian M. Scott <b.s...@csuohio.edu> wrote:
>>
>....
>> >I think, for example, that in Newton's theory, a space of *six*
>> >dimensions is required (lessee, IIRC it's x, y, z, direction, momentum
>> >and acceleration), but that merely establishes the variables to be used
What do you mean by this? Direction, momentum and acceleration in a 3D
space are all three dimensional vectors, not scalars, and they don't
affect the dimensionality of space. A three dimensional space has
three dimensions, not more. A thingy in space might have many degrees
of freedom - how many depends on what you want to describe. If you for
instance mean to describe an orbit of a planet/asteroid around our sun
you need seven variables (assuming m(planet)<<m(sun); G*m(sun) should
also be known, but it is a constant in our solar system) at least one
of which is a time reference. The other six can for instance be
location (x,y,z) and velocity (vx, vy, vz), or they can be mean
anomaly, length of semi-major axis, eccentricity, argument of
perihelion, longitude of ascending node and inclination towards the
ecliptic. [To be even more accurate you should include lots of
corrections due to pertutrbations from other planets - but that's not
covered in the basic version - instead you issue new orbital
elements.]
>> >in the actual theory. The theoretical space is not the theory, as such.
>> >The theory is the functional relationships that obtain between any
>> >entities in that space. I am not saying that a sheet of graph paper is a
>> >theory. I am saying that a theory is a prediction that entities in that
>> >space will closely follow the surfaces that the functions describe in it
>> >(pathways on the graph paper).
>>
>> But John, this is completely beside the point. The question is what
>> is meant by 'mathematical model'. R^3 is a mathematical model of the
>> physical space in which we live. And since Euclidean geometry works
>> very well on a normal human scale, it's a fairly good model. At no
>> point have I been talking about what constitutes a theory. My
>> objection from the beginning has been to your use of the term
>> '(mathematical) model'.
>
>You are going to have to elaborate for me - I'm (it transpires) dense,
>not deep. Start by what you mean with R^3. So far as I understand
>Euclidean geometry, it merely sets out the "sensorium" in which
>Newtonian models occur. It is not a model itself, but a set of axioms.
R^3 is a space in which every point and every vector is described by
(x,y,z) where x, y, z are real numbers (i.e. belongs to R).
Euclidean space is a space where the parallel postulate is valid
(through a point there is one and only one line parallel to any other
line that doesn't go through the point). A lot of other "everyday"
things work in Euclidean space: like the sum of the angles in a
triangle being 180 degrees, and the distance between two points being
sqrt((x1-x2)^2 + (y1-y2)^2 + (z1-z2)^2) if (x,y,z) is given for a
normated orthognal base [excuse possibly incorrect terminolgy from my
word by word translation from Swedish] etc.
>Then tell me what you mean by "model" here. For me, a model is some
>interpretable mathematical relation (interpretable meaning that you can
>map the function to some data set). This is perhaps a philosophical
>idiom and causing some confusion.
Euclidean space is a *model* of our "everyday" small scale space. The
relations between points in that space being "governed" by the laws of
Euclidean geometry - like distances, angles etc as of above. In
everyday life most of us are happy with this simple model and don't
care too much about relativity and non-Euclidean space.
Sven
Marty Fouts
wil...@wehi.edu.au (John Wilkins) writes:
[snip]
> Then tell me what you mean by "model" here. For me, a model is some
> interpretable mathematical relation (interpretable meaning that you can
> map the function to some data set). This is perhaps a philosophical
> idiom and causing some confusion.
[snip]
From a mathematician's point of view, "mathematical model" as it
applies to physics and John seems to be using the term, is best
described in a method similar to that found in
Fundamentals of Mathematics
Volume I Foundations of Mathematics: The Real Number System
and Algebra
edited by H. Behnke, F. Bachmann, K. Fladt, and W. Suss (et al)
translated by S. H. Gould
MIT Press first paperback edition 1983
of special interest to the non-mathematician is section 1, especially
1.4 "The Meaning of Mathematical Propositions" and section 3 "The
Concept of a Consequence"
to a mathematician, Chapter 10, "Some Basic Concepts of a Theory of
Structure" would be of interest.
a brief exceprt from page 6 may be fruitful for this discussion
In general, mathematicians are convinced that their propositions
are meaningful, the extreme position in this respect being that
of the so-called _formalists_, who consider mathematics to be a
mere game with symbols, the rules of which, in the last analysis,
are chosen arbitrarily (_conceptionalism_). Formalism was
introduced by Hilbert as a methodological principle whereby the
concept of a proof of consistency could be clearly stated. The
formalistic point of view can also be applied to physics if with
H. Hertz(1) we define the task of theoretical physics as follows:
"Within our own minds we create images or symbols of the external
objects, and we construct them in such a way that the logically
necessary consequences of the images are again the images of the
physically necessary consequences of the objects." In other
words, we construct a process parallel to the process of nature.
(1) Die Prinzipien der Mechanik, Ambrosius Barth, Leipzig (1894),
Introduction
or, in other words, to a mathematician, a model consists of *two*
critical things:
1) A set of mappings between the physical objects (and concepts) being
represented and the elements of an axiomatic mathematical system
2) The formal propositions of that axiomatic mathematical system.
A model is "true" if and only if every valid proposition in the
axiomatic system "has a parallel" by substition in the physical system
being modeled.
Marty Fouts
pois...@dart.se (Sven Silow) writes:
John seems to be getting at a more interesting issue, and you have to
be careful to disambiguate the use of the word 'space'. an object in
euclidian space actually possesses properties independent of its
position within that euclidian space, thus producing a state-space of
more than three variables. Such state spaces are often of use in
physics, especially in the analysis of nonlinear systems, as you are
probably aware.
Sven Silow
>John seems to be getting at a more interesting issue, and you have to
>be careful to disambiguate the use of the word 'space'. an object in
>euclidian space actually possesses properties independent of its
>position within that euclidian space, thus producing a state-space of
>more than three variables. Such state spaces are often of use in
>physics, especially in the analysis of nonlinear systems, as you are
>probably aware.
Of course, but the dimensionality of the "spatial space" is not
affected by however many state variables the points in that space
have. To *describe* the physical system we might need a *lot* of
variables (which together can be viewed as a multidimensional
"space"). But only three of those are required to describe the spatial
*position* of the point that has all the other properties.
(And I've done multidimensional analysis of plant communities - which
contained no "space" variables at all :-) .)
Sven
John Wilkins
Thank you, Marty. I was starting to think I was way off-beam.
In the 1980s, there was an attempt by Stegmüller and Sneed to formalise
the Kuhnian view of science, and I learned what little I do know of this
topic there. However, more recently and earlier (his work being spread
over the same period), Patrick Suppe (a physicist turned philosopher)
developed effectively the view of models that I gave earlier in the
thread, and I, naively, thought this was broadly accepted.
I am clearly a Hilbertian formalist (which also explains why I think of
things in terms of manifolds, which I developed independently as a
teenager before I ever heard of Hilbert, but without the slighest
rigor).
John Wilkins
Except that each variable is in effect an axis in a Hilbert space.
Brian M. Scott
Wilkins) wrote:
>Brian M. Scott <b.s...@csuohio.edu> wrote:
>> On Sun, 28 Jul 2002 10:05:17 +0000 (UTC), john.w...@bigpond.com
>> (John Wilkins) wrote:
[...]
>> >I think, for example, that in Newton's theory, a space of *six*
>> >dimensions is required (lessee, IIRC it's x, y, z, direction, momentum
>> >and acceleration), but that merely establishes the variables to be used
>> >in the actual theory. The theoretical space is not the theory, as such.
>> >The theory is the functional relationships that obtain between any
>> >entities in that space. I am not saying that a sheet of graph paper is a
>> >theory. I am saying that a theory is a prediction that entities in that
>> >space will closely follow the surfaces that the functions describe in it
>> >(pathways on the graph paper).
>> But John, this is completely beside the point. The question is what
>> is meant by 'mathematical model'. R^3 is a mathematical model of the
>> physical space in which we live. And since Euclidean geometry works
>> very well on a normal human scale, it's a fairly good model. At no
>> point have I been talking about what constitutes a theory. My
>> objection from the beginning has been to your use of the term
>> '(mathematical) model'.
>You are going to have to elaborate for me - I'm (it transpires) dense,
>not deep. Start by what you mean with R^3. So far as I understand
>Euclidean geometry, it merely sets out the "sensorium" in which
>Newtonian models occur. It is not a model itself, but a set of axioms.
A set of axioms and all of their logical consequences. That, together
with a mapping between the elements of the model and the elements of
the system being modelled is pretty much what I mean by a mathematical
model. For instance, the diagonal of a rectangular rug measuring 3 m
x 4 m *will* measure 5 m (within reasonable tolerances), just as the
model -- i.e., Euclidean geometry, in this case in R^2 -- predicts.
The ubiquitous schematics of the London Underground -- basically not
much more than labelled graphs -- model the adjacency relationships of
the stops; it's a simple model that omits all sorts of information
that might be of interest in some circumstances, but it's accurate as
far as it's intended to go.
>Then tell me what you mean by "model" here. For me, a model is some
>interpretable mathematical relation (interpretable meaning that you can
>map the function to some data set). This is perhaps a philosophical
>idiom and causing some confusion.
If I understand you correctly here, your notion of model is a proper
subset of mine.
[...]
Brian
John Wilkins
Thanks. That clarifies it.
>
> >Then tell me what you mean by "model" here. For me, a model is some
> >interpretable mathematical relation (interpretable meaning that you can
> >map the function to some data set). This is perhaps a philosophical
> >idiom and causing some confusion.
>
> If I understand you correctly here, your notion of model is a proper
> subset of mine.
Yes, I think that is correct. So let me try to explain what *I* mean by
it, at any rate, even if it is not a universal sense.
In the spaces R^2 and R^3, the set of logical consequences of the axioms
(such as the postulate of parallel lines not operatant in Riemannian and
other non-Euclidean topologies) is unrestricted - ie, it covers all
possible coordinates in the space. So it is not, sensu Wilkins, a model
of, say, the solar system or a Newtonian system in general, for these
systems operate to constrain the states a system can be in and the
dynamics between those states.
Now returning to non-literal spaces, the "state space" notion of a
theory that I adhere to - similarly a state space is the set of axioms
(in practice, the ways in which variables can vary) and all logical
consquences, but they are not models of the phenomena as such.
For example, suppose I model population genetics. I will have a set of
variables, including such things as alleles, fitness value, effective
population, and so forth. These will form a space within which
populations can vary (the fitness landscape being a subset of that
space) and time will be one of the axes there. However, a model includes
the constraints on the phenomena, and so we will not call that space a
model of population genetics until we can specify the relations between
entities in that space - and this will typically be such things as
markov walks, selection relations between two or more alleles, clonal
and sexual parentage lineages, and so on. So as far as I am concerned,
it is not a model of population genetics until we specify how things
will behave under differing conditions in that space.
We were talking past each other, I think. You undoubtedly have some
other term for what I call a model here.
>
> [...]
>
> Brian
Brian M. Scott
Wilkins) wrote:
>> [...]
>Thanks. That clarifies it.
Indeed; like Newtonian mechanics (or, more simply, Hooke's law) it is
a useful model only within limits.
>Now returning to non-literal spaces, the "state space" notion of a
>theory that I adhere to - similarly a state space is the set of axioms
>(in practice, the ways in which variables can vary) and all logical
>consquences, but they are not models of the phenomena as such.
>For example, suppose I model population genetics. I will have a set of
>variables, including such things as alleles, fitness value, effective
>population, and so forth. These will form a space within which
>populations can vary (the fitness landscape being a subset of that
>space) and time will be one of the axes there. However, a model includes
>the constraints on the phenomena, and so we will not call that space a
>model of population genetics until we can specify the relations between
>entities in that space - and this will typically be such things as
>markov walks, selection relations between two or more alleles, clonal
>and sexual parentage lineages, and so on. So as far as I am concerned,
>it is not a model of population genetics until we specify how things
>will behave under differing conditions in that space.
Well, no. But you don't have to model population genetics all at
once. Kauffman's NK model, for instance, is a mathematical model (in
my sense, at least) of a part of population genetics.
>We were talking past each other, I think. You undoubtedly have some
>other term for what I call a model here.
Maybe I'm just too tired to think straight, but at the moment I can't
come up with one. I *think* that the difference may actually just lie
in what you're willing to think of as having a mathematical model in
the first place. Putting it really crudely and informally, you want
to have 'moving parts' before you're willing to talk about a
mathematical model.
Brian
John Wilkins
> On Mon, 29 Jul 2002 05:22:50 +0000 (UTC), wil...@wehi.edu.au (John
> Wilkins) wrote:
>
> >Brian M. Scott <b.s...@csuohio.edu> wrote:
>
...
> >> If I understand you correctly here, your notion of model is a proper
> >> subset of mine.
>
> >Yes, I think that is correct. So let me try to explain what *I* mean by
> >it, at any rate, even if it is not a universal sense.
>
> >In the spaces R^2 and R^3, the set of logical consequences of the axioms
> >(such as the postulate of parallel lines not operatant in Riemannian and
> >other non-Euclidean topologies) is unrestricted - ie, it covers all
> >possible coordinates in the space. So it is not, sensu Wilkins, a model
> >of, say, the solar system or a Newtonian system in general, for these
> >systems operate to constrain the states a system can be in and the
> >dynamics between those states.
>
> Indeed; like Newtonian mechanics (or, more simply, Hooke's law) it is
> a useful model only within limits.
In fact this is true of any theoretical model no matter what the limits
are - even the best of modern physical theories (excluding mbrane
theory) have limits back to the big bang or a time shortly thereafter. I
also wonder if the laws of physics apply in black holes...
>
> >Now returning to non-literal spaces, the "state space" notion of a
> >theory that I adhere to - similarly a state space is the set of axioms
> >(in practice, the ways in which variables can vary) and all logical
> >consquences, but they are not models of the phenomena as such.
>
> >For example, suppose I model population genetics. I will have a set of
> >variables, including such things as alleles, fitness value, effective
> >population, and so forth. These will form a space within which
> >populations can vary (the fitness landscape being a subset of that
> >space) and time will be one of the axes there. However, a model includes
> >the constraints on the phenomena, and so we will not call that space a
> >model of population genetics until we can specify the relations between
> >entities in that space - and this will typically be such things as
> >markov walks, selection relations between two or more alleles, clonal
> >and sexual parentage lineages, and so on. So as far as I am concerned,
> >it is not a model of population genetics until we specify how things
> >will behave under differing conditions in that space.
>
> Well, no. But you don't have to model population genetics all at
> once. Kauffman's NK model, for instance, is a mathematical model (in
> my sense, at least) of a part of population genetics.
Any model is restricted. The issue is whether it applies when it is
supposed to. The NK model is a mathematical result that may, or may not,
apply in any given case of the genetics of a population. Similarly, the
NFL theorem <is it a theorem really?> does not apply to any given case
of genetic evolution because genetic evolution occurs over a very
restricted subdomain of the set of all possible search algorithms.
A model is particularly useful because it shows up the *anomalies* - in
other words a model sets the expectations of what should happen so that
when it doesn't, we are alerted. Another sense of surprise is realising
the actual predictions of a model - as when it is assumed that one
outcome (say, anagenetic speciation) will occur but investigation of the
model shows different (cladogenetic and abrupt speciation followed by
stasis). But primarily, once a model is bedded down, it is the anomalies
that are most surprising and worth investigating. In this way, models
determine the course of worthwhile investigation by specifying problems.
>
> >We were talking past each other, I think. You undoubtedly have some
> >other term for what I call a model here.
>
> Maybe I'm just too tired to think straight, but at the moment I can't
> come up with one. I *think* that the difference may actually just lie
> in what you're willing to think of as having a mathematical model in
> the first place. Putting it really crudely and informally, you want
> to have 'moving parts' before you're willing to talk about a
> mathematical model.
>
I thought it might be a difference of usage. I think I see what the
problem is here. Context is everything - I am interested not in
mathematical models for their own sake, but mathematical models as they
are instantiated in theories. All theories are or can be formulated as
mathematical relations; the explanatory power of a theory lies in the
deductive relations mapping to the phenomena. But what holds true in
maths does not necessarily hold true in physical science. For a start,
denumerable series tend to halt in real world cases - there is no
Entscheidungsproblem in biology, for example. We sometimes forget this
fact...