Conceptual Revolutions

by Paul Thagard

In this path-breaking work, Paul Thagard draws on the history and philosophy of science, cognitive psychology, and the field of artificial intelligence to develop a theory of conceptual change capable of accounting for all major scientific revolutions. The history of science contains dramatic episodes of revolutionary change in which whole systems of concepts have been replaced by new systems. Thagard provides a new and comprehensive perspective on the transformation of scientific conceptual systems.


Thagard examines the Copernican and the Darwinian revolutions and the emergence of Newton's mechanics, Lavoisier's oxygen theory, Einstein's theory of relativity, quantum theory, and the geological theory of plate tectonics. He discusses the psychological mechanisms by which new concepts and links between them are formed, and advances a computational theory of explanatory coherence to show how new theories can be judged to be superior to previous ones.

Genres: Philosophy, Science, History

Paperback

First published February 19, 1992

Reviews

[Liquidlasagna · Rating 1 out of 5]

Honestly, i think the guy is a fraud

"advances a computational theory of explanatory coherence to show how new theories can be judged to be superior to previous ones"

basically a computer scientist who think that you can show some theories are better than others with computer modelling...

carefully lay out your propositions

but some are labelled evidence
others are labelled hyptheses
others are labelled as 'facts'

and then you turn the crank

.........

tweak your propositions just right, and you get the right printout to show something is 'superior'

he thinks this can be extended to legal reasoning, and not just empirical and analytic thing...

since you try to shape moral statements into some legalistic 'analytical form' to solve dilemmas of a legal nature

which is nothing more than programming a fancy version of the game CLUE

get 17 propositions

E5 Calderwood changed his story several times
E8 Dotson said Olgivie is a liar
E3 Calderwood said that he saw a patrol car pull over a Volkswagen like Knott's near I-15
E9 Anderson and Schwartz saw scatches on Peyer's dace the night of the killing
E14 Bloodstains found on Knott's clothes matched Peyer's blood

sighs

Hypothesis that peyer is guilty

proposition-3 fibers from Peyer's uniform were transferred to knott
proposition-1 Payer killed Knott
proposition-7 Peyer had a bloody nightstick
proposition-8 Anderson was having personal problems when first interrogated

Hypotheses that Peyer is innocent

proposition-18 Peter is a good man
proposition-16 Martin lied
proposition-17 Anderson was mistaken about the nightstick
proposition-15 Payer's scratches came from a fence
proposition-14 Ogilvie lied
proposition-14A Ogilvie is a liar
proposition-13 The six fibers flated around in the police evidence room
proposition-11 someone other than Peyer killed Knott
proposition-12 Calderwood made his story up

I think anyone can see how stilted and artificial this is
and how adding the right proposition or leaving out the right proposition
can just have something fall apart

toss in some probabilities with some chemistry and you can get some inflated pomp about hypotheses for and against Lavoisier or Coprernicus

and then add in all the promise of 'machine learning'

And then magically Colossus: The Forbin Project and chatgpt will solve everything flawlessly

toss in some hype for machine learning
and profound implications for psychology

"social psychology is a very rich domain for a theory of explanatory coherence..."

"so far, my discussion of the psychological revenace of ECHO has been merely suggestive, showing that explanatory coherence judgements may be plausibly considered to contribute to impostant kinds on inferential behavior... a defense of ECHO as a psychological model, however, will require controlled experiments that provide a mich finer-grained evaluation of the theory of explanatory coherence."

"Fortunately, there appears to be great potential for testing explnatory coherence theory and the ECHO model by comparing the performance of human subjects with ECHO-based predictions about qualitative and quantative features of the acceptance and rejection of hypotheses."

It smells like so much horse crap

A computer scientist trying to be Thomas Kuhn
and being as tricky as a Chomsky


.........

and well his own work admits that the 'software' is only as good as the 'principlesin that theory'

Much of what he's blathered about elsewhere seems to be that he's got grand plans of computational models of actual neurons in the brain and other bullshit,

and well when it's dullsville in the computer science lab, he just goes off to youtube
going for low-hanging fruit

like the worst of those skeptics

Youtube: Paul Thagard on Why Astrology is a Pseudoscience

when he's not blathering on about dangerous misunformation and conspiracy theories under his bed, and under every rock

and he's even hysterical about 'thought experiments' being dangerous

Einstein must LOVE this guy

"Paul Thagard has recently argued that thought experiments are dangerous and misleading"


The world can only stand so many bad philosophers, and so many bad computer scientists

and a special level of hell is needed for bad philosopher-computer scientists

...........

one review of this book said this, where thagard wants to kick Kuhn in the nuts

"One of Thagard's most compelling insights is that scientific paradigms are not in fact "incommensurable," as Kuhn suggested, and that conceptual shifts are therefore essentially irrational events."

"As he sees it, there is enough continuity in scientific revolutions to justify the claim that the adoption of new views is a rational process"

and Thagard will use bad analogies like saying, no it's not like a religious conversion like Kuhn would think of a scientific revolution, it's like learning a second language.

suuuuuure

.............

"According to Thagard, each of these cases demonstrates that conceptual revolutions involve radical transformations of whole systems of concepts and laws, not the mere revision of beliefs or theories."

sure sure

concepts - abstract ideas
laws - a statement that always seems to be true

not the mere revision of
beliefs - acceptance that a statement is true
theories - ideas intended to explain something

I think he's playing obscurantist word games

just trying to say, it's about scientific laws, something grander than mere scientific theories

..............

And it seems other people have issues with Thagard

quote

.......
The funny thing is that I quoted Thagard in my piece.

I dismissed him offhandedly because he was dismissive of Peterson, but upon reflection, I’ve come to realize that isn’t enough.

While there’s certainly a correlation between dismissiveness and bad arguments, that’s not why Thagard’s argument is bad.

nfortunately, I didn’t realize that until my second read-through of his piece, “Jordan Peterson’s Flimsy Philosophy of Life.”

So, if I want to show that Thagard is wrong, I have to take him seriously.

So, without further ado, Paul Thagard is wrong about Jordan Peterson because he doesn’t know how to read... in the literary theory sense of the phrase.

After reading 12 Rules for Life, I would argue that Peterson’s thesis is that “understanding the deeper meanings present in myth can improve your life and make you happy.”

It’s a philosophical conclusion, but it’s deeply grounded in literary theory.

Paul Thagard doesn’t get archetypes.

Or, at least, he doesn’t understand the fundamental difference between reading scripture in a literalistic sense and reading it in a literary sense where one is looking for underlying meaning.

“But philosophers since Plato have recognized many problems with basing ethics on religion.”

What’s particularly interesting about this is that Thagard seems to be locked into the state of mind that there are only two ways to approach religion, which happen to be the ones we discussed earlier. In his eyes, you can either apply “comfirmational-instructional” or “disconfirmational” theories.

Thagard finds the fact that Peterson isn’t applying disconfirmational theory a serious problem.

This reveals his total ignorance of literary theory.

Worse, Thargard seems to miss the breadth of stories that Peterson uses in his arguments.

Religious myths are only a subset of the stories he examines.

After seeing how many works Peterson considers, it’s hard to escape the conclusion that Thagard’s stance is reductionist.

At our least generous, we have to wonder if Thagard actually read Peterson’s book.

But, Thagard literally says “Peterson’s answer looks to religion.”

That claim is at least misleading, but it may be more accurate to say that it’s outright false.
............

Yeah, gentle readers, he doesn't pass the smell test in oh so many ways

[Craine · Rating 5 out of 5]

I am disappointed this book is not more well known as it bridges much of the gap left within a Kuhnian philosophy of science where what exactly constitutes a scientific revolution is explored as well as looking at the formation of concepts and giving a very interesting exploration of the use of computer software (ECHO) to determine scientific theory candidates. The leniency of cognitive science, philosophy and computer science in this treatment is truly magnificent and something which I hope will be more considered in the future (keeping in mind that this book was written over 20 years ago!).

(Note I don't like the star rating and as such I only rate books based upon one star or five stars corresponding to the in my opinion preferable rating of thumbs up/down. This later rating system encourages in my opinion the degree to which the reader is likely to read a review instead of merely glancing at the number of stars)

[Sanane · Rating 5 out of 5]

How can we define and quantify a scientific revolution, what does a revolution do to the current paradigms and accepted theories? He presents an objective way to quantify this question using networks and optimization. Then, he proceeds to demonstrate this in concrete examples, like Newtonian, Darwinian, geological etc.

Also he discusses the meaning of concepts like explanation, coherence, rejection, substitution etc.

Highly recommended to anyone who is involved in any kind of scientific activity.