I’m thinking about writing another book. I won’t disclose the title or topic yet, but there’s a bit of research for it I think can be usefully crowdsourced, and may also give a clue about the book for those of you interested.
I’ve written before about the difference between descriptive and generative theories. To recap and simplify, a descriptive theory accounts for what is; a generative theory finds causal regularities beneath a descriptive account and predicts consequences not yet observed.
Now I want to zero in on a parallel difference among entire sciences. Some scientific fields – like, say, evolutionary biology – are tremendously productive of models and insights that can be applied elsewhere. On the other hand, some other sciences – like, say, astronomy – seldom export ideas or models.
Note that while it is appropriate to think of sciences that export lots of ideas as ‘generative’, the class of sciences that don’t are not merely descriptive. Astronomy, for example, has lots of generative theory inside it; astrophysics, for example makes predictions about stellar spectra and elemental abundances. But astronomy as a whole is not generative because none of its theory really informs anything outside astronomy.
So I’m going to start with a (non-exhaustive) list of scientific fields, indicating roughly how generative I think they are and what if anything they export. I invite additions and corrections from my readers.
Evolutionary biology – extremely generative. Principal exports: evolution by selective pressure on random variation, adaptive radiation, genetic drift, mutation, and many others.
Mathematics – not an empirical science but extremely generative nevertheless. All kinds of abstract mathematical models end up suggesting applied-math models of the real world with interesting testable consequences.
Economics – highly generative. Principal exports: supply/demand equilibria, satisfaction under constraint, implicit knowledge, deadweight losses, search costs, coordination overhead, rational ignorance.
Linguistics – not very generative at all except for a small corner near psycholinguistics that exports some provocations about the relationship between thought and representation.
Physics – highly generative. Principal exports: conservation laws, principles of least action, entropy, state spaces, symmetry and symmetry-breaking, energy levels.
Astronomy – gorgeous, but almost completely non-generative.
Geology, zoology, agronomy – non-generative
I’m not sure I’d use the word “sciences” for this. I think I’d go more with “disciplines”. I don’t think Economics and Linguistics (much as practitioners of said disciplines might disagree!) are sciences. They certainly bring more rigor to the game than, say, psychology or sociology.
But … I dunno. I’d love to see rebuttals to my position. Maybe I’m completely out to lunch.
Just a nit. Feel free to ignore me.
However, I don’t think my point takes away from your larger thesis. This is going to be a fascinating conversation.
Is it really true that astronomy is non-generative?
The way we came up with the idea of dark matter and dark energy was to observe the macro-scale behaviour of the universe through telescopes, and conclude that such behaviour was not consistent with our existing theories of physics.
We are now trying to detect dark matter right here on earth.
So, astronomy led to a generative theory that has deeply affected physics.
Hmm. I’m now thinking that I was too hasty and have misunderstood what you are trying to say.
One comment about economics is that it has often been constrained by the choice of metaphor used by economists. For example, classical economists’ focus on general equilibria derived from the prevalence of metaphors taken from thermodynamics and classical kinematics. More modern economic theory often depends on biological metaphors such as contagion and activation thresholds.
For reference, see Richard Bronk, The Romantic Economist.
Hm. It seems to me there’s an aspect of all this you’re not capturing, either by design or otherwise: sciences (or disciplines, if you accept NukemHill’s argument) have value not only in that they generate predictions about the state of the universe, but also in that they test predictions made by others. Astronomy isn’t generative in itself, but it does provide a rich ground for predictions made by relativity theory, as one example that spring immediately to mind.
It seems (and perhaps this is a tautology, but I think it’s a useful clarification for me at least to make it explicit) the highly generative fields are those whose empirical results lead to mathematical models that map well onto other areas.
Eric: Astronomy is incredibly generative; you have to go back a little deeper into the fundamentals to see it.
1) Astronomy generated the theory that natural events could be predicted with precision. This is fundamental to the creation of agriculture.
2) Astronomy generated Keplerian mechanics, and from Kepler, Newtonian mechanics, because of a revolution in the models.
3) A discrepancy in astronomical observation of Mercury’s orbit was a generative cause of the theory of relativity.
Focus not on the stars, oh man of categorical mein. Astronomy is the observational science of gravity and its secondary and tertiary effects.
Eric: Where would you rate chemistry in this spectrum of fields?
How about the studies of electrical inductance on the human brain?
Though my degree is in Linguistics, I have to agree that Linguistics is not terribly generative outside of itself. OTOH this doesn’t mean that it has no value or is somehow “lesser” than other, more generative sciences. I don’t think that you meant to suggest that, but it would be easy for a casual reader to walk away with that impression.
Alas, the kind of linguistic theory that’s most likely to be exported to other sciences is the flaky Whorfian hypothesis which is rejected by most actual linguists.
Astronomy = the hyper generative field of optics and the semi-generative realm of general relativity. Also huge push in early number crunching computers, modern CCD arrays. Please do not ignore the EE crazy aspect of radio astronomy, everything from DSP to interferometry to planetary radar to exotic microwave stuff. Here’s a terrible phrase “Astronomers were THz before THz were cool”… How about adaptive optics?
You’d have a lot more fun going 2-D. There is NOTHING in astronomy that is generative until at least mid-level undergrad, and it seems everything in astronomy is generative above grad student level. On the other hand, some grade school kid can do sorta-generative work with selective breeding of plants or the famous fruit fly mutation studies.
>Mathematics – not an empirical science but extremely generative nevertheless.
That is true from a formalist view of math, but from a more applied view, I think math can be empirical. Instead of the deduction from axioms that the formalist focuses on, consider it the search for more interesting or useful axioms.
I submit that geology *is* generative. And in fact is the genesis of of other generative disciplines. It was the study of geology which broke down the Creationist timeline. It was geology which helped Darwin synthesize a framework for the timelines involved in what we call genetic mutation and evolution. Finches don’t grow different beaks in short timeframes.
And geology has provided lots of evidence and theories concerning global climate especially more recent climate events which understanding has changed perceptions about population movements and pre-history. Geological fact structures also underlie theories and models of supervolcanoes and extinction events.
For just one example, the history of man in the Americas. Where from? Asia. How? land bridge. What land bridge? Low sea levels due to glacial ice. How did man bypass the ice? Ice-free passageway to warmer lands. All of these factual items rest on geological research. And When? Pick a date, any date, there are a number of candidates awaiting further anthropological research.
Some of those sciences are actually subcategories of larger ones. An example is zoology, a subcategory of biology. You can argue that Darwin was a zoologist that (mostly) used his observations of animals to generate his theory of evolution; zoology is generative after all.
…and how about geology and Wegener’s plates?
It is inaccurate to class any science as non-generative. Some are less or more generative, but likely none is generatively sterile. A continuum exists.
Also, formal notations usually or always aid expressing and understanding models, theories, etc. A notation denoting generativeness would be useful. Here is a first attempt at a generativeness notation.
astronomy > fusion > nuclear energy > reactors, bombs
astronomy > need for accurate models > calculus > technology as we know it
agronomy > growing crops > growing capital via investments
agronomy > harvesting > selling investments
biology > evolution > as stated in initial post
biology > contagion > economic-business models
biology > biotic taxonomy > other taxonomies-ontologies
biology > autotrophs > manufacturing
biology > heterotrophs > equity traders
biology > parasites > criminals, politicians
biology > disease > viruses > cultures > computer viruses > Microsoft Windows
Hope this helps.
>It is inaccurate to class any science as non-generative. Some are less or more generative, but likely none is generatively sterile. A continuum exists.
That is true in an absolute sense, but not useful. Some sciences are not generative enough to be interesting for my purpose in the book.
Also, your notion of “generative” is much looser than I’m going after. It’s true that astronomy historically exported the need for accurate models and mathematics, but I’m not interested in historical generation. I’m interested in the extent to which understanding the theory of field A helps a person have creative insights about field B.
Interesting observation: a theory can be generative in your sense without necessarily being true. For example, as several people have already pointed out historically astronomy was extremely generative, note that at the time there was no real distinction between astronomy and astrology.
Another example: there have already been several breakthroughs in areas of mathematics based on insight from string theory, regardless of whether ST turns out to be true.
>Another example: there have already been several breakthroughs in areas of mathematics based on insight from string theory, regardless of whether ST turns out to be true.
That’s a good example of generativity in my strict sense, but none of the examples yet proposed for astronomy really are. In order for astronomy to be generative in my sense, there would have to be some X we want to study that is enough like celestial mechanics that knowing celestial mechanics helps you have insights about X.
Isn’t that exactly what Ken Burnside has demonstrated though?
Understanding astronomy leads to the realisation that nature is regular and predictable. This leads to new insights in other fields because we start to study those fields through the lens of regular law.
>Astronomy generated the theory that natural events could be predicted with precision.
I’ve already noted that I’m not interested in historical generation.
What is the cut-off point here? Are you just wanting to look at the last ten years? Last fifty?
>What is the cut-off point here? Are you just wanting to look at the last ten years? Last fifty?
The relevant question is roughly this: if you want to study scientific fields in order such that the ones that give you the most help understanding fields other than themselves come first, what’s the right order?
@NukemHill: I suspect you don’t know what Linguistics is.
More generally (and ironically), 19th-c. Linguistics exported/generated ideas which were imported into evolutionary biology (e.g. evolution, genetic relationships between species). Darwin’s models were influenced by linguistic models of relationships between languages.
>Darwin’s models were influenced by linguistic models of relationships between languages.
I find this claim both interesting and plausible. Can you point at evidence?
Zoology has resulted in the discovery and refinement of principles which have implications for medicine, materials science and manufacturing, and (most recently) robotics. Does that count?
I will add to the chorus of those who think astronomy is generative. The only way to test some of the deep, fundamental physics is through high energy accelerators or through cosmology.
The answer to your question on this subject is echoed at http://xkcd.com/435/
Well, if you had not stipulated ‘scientific’ fields then I would make the case for certain areas of philosophy: logic, philosophy of science, history of science, epistemology.
Then probably mathematics, if you count that as a science.
>Well, if you had not stipulated ‘scientific’ fields then I would make the case for certain areas of philosophy: logic, philosophy of science, history of science, epistemology.
We can include those on the list, given my purpose.
>Then probably mathematics, if you count that as a science.
Recall that I described it as “extremely generative” in the OP.
It’s starting to look like this xkcd strip may be the most relevant to ESR’s project.
I’m not so sure that linguistics is almost entirely descriptive. Some examples:
— Probabilistic Context Free Grammars for modeling protein folding.
— The Chomsky hierarchy of formal languages is important for many core parts of computer science, from programming language design to writing efficient regular expression parsers.
— Trying to model the syntax-semantics interface has led to a lot of cross-pollination in symbolic logic, particularly substructural logics. I don’t know if this is generative in the sense that you mean, but there can’t be that many fields of logic research where e.g. the order of premises matters.
— Observations from Linguistics (such as Zipfs’ law that most word types are rare, but most word tokens are attributable to a few word types) were influential in the development of information theory.
I imagine that discoveries in Astronomy in the area of accretion disks and planet formation could inform Geology.
And maybe Biology/Exobiology (habitable zone orbits, etc).
As a physics major I feel like you don’t understand astronomy. Most people assume it is the study of stars. There is actually a lot of things inside of astronomy. However it is a fairly generative field for the rest of physics as well as personnel applications.
Most fundamental proof for General Relativity comes from astronomy and several famous astronomers were instrumental in correcting its constants and calling Einstein bad names in papers when he tried to restrict the theory.
The applications from that alone include GPS satellites, Atomic clocks and most atomic theory regarding binding energy and why stuff doesn’t explode.
Honestly I’m not an astronomy major so I don’t know a whole lot but there are lots of items that are generated from astronomy. It’s not a new field however so people forget a lot of the things it produces. It is also an extremely tiny field with a small number of people inside of it.
I keep hearing people calling (Sapir-)Whorf “flaky” and similar epithets, but none of them ever explain what’s wrong with the idea that the vocabulary you have to express ideas influences your ability to form ideas.
To me, one of the biggest problems with linguistics is that it assumes that language is only for communicating ideas between people, and rejects the notion that it is involved in the formulation of those ideas within the mind of a single person. Where is the justification for this? I think it’s self-evident that we think in a language, and that a language optimized for a certain kind of thinking will thereby encourage it.
“I think it’s self-evident that we think in a language, and that a language optimized for a certain kind of thinking will thereby encourage it.”
We don’t always think in a language, at least not always in the one we speak. I’m thinking of Kekule’s dream, where the chemical structure of benzene came to him as a picture of atoms chasing each other around. There’s also intuition, where we have no idea of where the idea came from.
@meat.paste and syskill:
I’d be more inclined to go with this one:
I think that logic has got to be the number one most generative field, and the one that will best help you understand every scientific endeavour.
Without deductive logic you can’t understand how to take a hypothesis and turn it into an appropriate test. You can’t understand the scientific method. You have to be able to understand ‘if my hypothesis is true, then x must also be true, and y must not be true. Therefore if I find that x is not true or that y is true, then my hypothesis must be false.’
As part of my degree I had two years of formal logic, and even that short introduction to the subject has proved invaluable as I have branched out into other fields. Having a solid grounding in logical fundamentals really gives you a leg up when reading scientific papers and deciding whether the researchers have established rigorous protocols.
I agree that evolutionary theory is highly generative outside its own domain. This is because it really is not restricted to the biological domain; it’s a special case of dynamics. And dynamics applies to…
Well, what doesn’t it apply to?
For a fun pop-level work that touches on this, try Kevin Kelly’s “Out of Control.” He discusses computer modelling that “evolves” corporate logos, works of art, programs, etc.
A similar point could be made about the concept of equilibrium in economics, particularly the uber-concept, Nash Equilibrium and its numerous variants. The core idea there is “mutual best response.” Of course this comes up in any domain in which intelligent – or even not particularly intelligent – agents interact.
@The monster: When people say the Sapir-Whorf hypothesis is “flaky,” they usually mean it’s hard to come up with some kind of concrete, testable prediction. It might seem intuitively appealing, but our view of the relationship between conceptual structure and linguistic structure is still too hazy to make clear conclusions.
There are justifications for viewing language as about communication. Here are a couple:
— Florian Jaeger and Alice Turk, among others, have been arguing that language is structured to be efficient in the sense of the Noisy Channel Theorem, and thus is efficient for communicating.
— There are evolutionary reasons to believe that language developed as an incrementally more efficient communication system rather than a sudden, symbolic instantiation of thought. This indicates that the primary focus of language is not to enable thought, although there could of course still be a tight relationship between language and thought.
Eric: You have a bias regarding astronomy and geology.
Your categorization is less about “generative models” – which astronomy abounds with – and more about “can I run a controlled experiment on this in human-scale time frames?”
This is the underlying rubric of your “will thinking about field X give me insight into unrelated field Y?” criteria.
It’s hard to identify insights from field X when you can’t do repeatable tests. One of the reasons why you’re getting the minor amount of flack over astronomy from me and others is that your categories appear arbitrary and non-useful.
Astronomy is like LISP. It may not be the first discipline you study seriously – but it’s one that, after picking up another discipline or two, gives you the widest acceleration in capabilities.
More concrete generative theories of astronomy:
1) Possible Holographic nature of the universe.
2) The measure of the rate of expansion of the universe, and a timescale for the change of physical constants in the Standard Model
3) The second order detection of non-baryonic and non-fermionic matter through gravitic lensing.
4) Cosmic Microwave Background measurements and their implications for the structure of the universe.
5) Exoplanet detection and formation.
Astronomy isn’t the study of stars. It’s the study of gravity, down to second and third order effects.
>Astronomy isn’t the study of stars. It’s the study of gravity, down to second and third order effects.
An oversimplification, but a nicely provocative one. I love astronomy – I spent many happy hours with my telescope sketching the Moon and looking at planets and nebulae and clusters when I was a teenager, and read lots of books about it. I know about as much about the field as most non-specialists ever retain, I think. That’s exactly why I judge it exports very little.
You do a good job of listing interesting some interesting things one can learn studying astronomy. But it’s not exactly the kind of coupling I’m looking for. Once I know about evolution under selection in biology, I’m equipped to notice the same sorts of things happening in systems of memes and computer programs – evolutionary biology exports not just observations we can use to check theories in other fields, it exports patterns that are armatures for entire theories. Once I know about least-action and equipotential principles in physics, there are analogies with generative power as far away as psychology and mass social behavior. What are the patterns that astronomy exports?
>> me: Darwin’s models were influenced by linguistic models of relationships between languages.
>esr: I find this claim both interesting and plausible. Can you point at evidence?
(1) Atkinson, Q. and Gray, R.D. (2005). Curious parallels and curious connections: Phylogenetic thinking in biology and historical linguistics. Systematic Biology, 54(4), 513-526. [ http://language.psy.auckland.ac.nz/publications/index.php?pub=AtkinsonAndGray2005SystBiol ]
(2) O’Hara, Robert J. 1996. Trees of history in systematics and philology. Memorie della Società Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano, 27(1): 81–88. [ http://rjohara.net/cv/1996-milan ]
(3) Nathalie Gontier. “Depicting the Tree of Life: the Philosophical & Historical Roots of Evolutionary Tree Diagrams”. Evolution, Education & Outreach 4.1 (2011). [ http://vub.academia.edu/NathalieGontier/Papers/878107/Depicting_the_Tree_of_Life_the_Philosophical_and_Historical_Roots_of_Evolutionary_Tree_Diagrams ]
(4) Maybe also: Stephen Alter. 1998. Darwinism and the Linguistic Image: Language, Race, and Natural Theology in the Nineteenth Century. Baltimore: The Johns Hopkins University Press. [ Amazon link: http://www.amazon.com/Darwinism-Linguistic-Image-Nineteenth-Intellectual/dp/0801858828 ] — I haven’t read this one, but there’s a review here ( https://whewellsghost.wordpress.com/2010/09/17/darwin-and-the-linguistic-image/ ) which makes it sound promising….
In fact, linguistics in part inherits the notion of genetics trees from stemmatics, which was interested in reconstructing “good” versions of manuscripts.
I would also add that “Linguistics” is a pretty far-reaching term, covering everything from formal semantics—which is tied up with logic, and thus I think would count as generative—to philology/historical linguistics, mentioned above.
>In fact, linguistics in part inherits the notion of genetics trees from stemmatics, which was interested in reconstructing “good” versions of manuscripts.
Very interesting, thank you. I skimmed the O’Hara paper based on its title and am satisfied that there is some actual there there. I’ll read it in full, and the others.
It occurs to me, thinking about stemmatics, that an earlier pattern for that kind of diagramming might have been genealogical charts.
re: the xkcd comic.
There is an interesting point there in that it is often true that field B (e.g. steam engine engineering) is technically a special case of Science A (e.g. physics), but one that is both interesting and impossible to model using the standard tools of A. We invent a new set of tools, and call it B. But often a generalization of those tools (in this case, thermodynamics) is useful more widely in A and even in systems outside field A entirely.
I seem to remember reading that models of textual corruption and correction were useful in genetics. Or was it the other way around? That is, models used to study e.g., the introduction of errors into, and the elimination of errors from, successive generations of handcopied Bibles, were also useful in studying genetic mutations and the correction thereof.
Doesn’t this go against your assertion earlier that you are not interested in historical generation?
The ideas about evolutionary dynamics that originated with biology are no longer its exclusive preserve. It is not necessary to study biological evolution today in order to understand behaviour of an evolutionary computer algorithm, for example. That knowledge has been incorporated into computer science.
The fact that these ideas came from biology is merely an historical contingency. They only belong to biology in the sense that biologists happened to be the first people to discover them.
As another example, spherical trigonometry was developed, in part, by astronomers (who were also usually mathematicians). That knowledge has gone back ‘upstream’ to mathematics, so that today we don’t have to study the special case of astronomy in order to understand trigonometry. Nevertheless, certain aspects of spherical trig were ‘exported’ from astronomy.
It seems to me that your question is actually inherently an historical one. In order to understand what disciplines are the most generatively fertile we have to look at their history.
Sorry if this seems as if I am being deliberately difficult, but I am just trying to better define the question.
>Doesn’t this go against your assertion earlier that you are not interested in historical generation?
No, because I’m trying to grasp the way concepts like evolution under variation with selection radiate out from their home disciplines today, not how it happened historically. You say
>The fact that these ideas came from biology is merely an historical contingency. They only belong to biology in the sense that biologists happened to be the first people to discover them.
but I think you are incorrect. People who think about evolution in genetic programming use metaphors from biology. It doesn’t happen the other way around – you don’t see biologists adopting the jargon and tools of genetic programming to explain biology. The reason for this may be historically contingent, but the asymmetry itself is a fact about present time and what I want to focus on.
Let me draw aside the veil a bit. The question I’m actually considering is: how can an aspiring polymath self-train most efficiently? That is, which scientific disciplines are not only important in themselves but also export conceptual toolkits that will give the fastest running start in other sciences?
This is why I’ve been focusing on evolutionary biology, economics, physics, and mathematics in this discussion (and I should have mentioned computer science too). I’ve read voraciously in more other sciences than I could easily list, but my direct experience is that the ones I just listed are the ones that reliably export tools I can use in multidisciplinary problems.
I’m not suggesting that language is not used for communication, but that the focus on it solely as a means of communication, ignoring its role in high-order cognition, is silly. Yes, language evolved to communicate ideas between people, but it also made it possible for them to think about more complex ideas than they could possibly conceive of without language.
But in both cases, we have to translate those non-linguistic forms of thinking into some concrete symbolic representation in order to do high-level thinking.
In a very real sense, the kind of thinking on which science depends involves a process of a scientist “communicating with himself”: He must formally state his hypothesis in order to construct experiments that can either confirm or reject it. Until the idea is fixed symbolically in either words or some kind of specialty language such as chemical notation, there is no way the idea can really be used.
I’ve always thought that science is the lens through which we focus an image of reality in our minds….tirelessly grinding and polishing the lens to bring reality into sharper focus and to correct flaws in our perception. Mathematics is our current lingua franca toolkit for unambiguously communicating what we ‘see’ to others.
Your project sounds intriguing, ESR…
I think I would quibble about economics. Can you give an example of supply/demand equilibrium, for example, being used in another field? Satisfaction under constraints seems more like a mathematical or physical idea to me; the Hamiltonian formulation of classical mechanics, for example, predates (I think) use of such ideas in economics. Economics has very wide application to all sorts of human activities, so the things it describes are very often applicable, which can ‘fake’ generative-ness; we can all see these things in our own lives. But that’s not the same thing as making concepts useful in other sciences.
I would also point out that astrophysics often informs particle physics, as in the various theories of what dark matter consists of – astrophysical observation constrains what models are possible; conversely the models often inform what astrophysicists look for. I think I would be careful in making sweeping statements about this sort of thing from the outside; modern science is fractal and has lots of little details that only those who live it every day are aware of.
It almost seems to me that what you are looking at is rather the familiarity that a well-informed layman has with the concepts of each field, rather than their influence on other fields.
>But that’s not the same thing as making concepts useful in other sciences.
The distinction is important only if you’re excluding sociology, psychology, and anthropology as sciences – those being the places where exports from economics often have considerable explanatory power.
Hmmmm… Might I suggest Computer Science?
While, strictly speaking, this is a mathematics discipline, I’ve find it highly generative in enabling me to find and predict patterns in various systems.
Not sure this is what you’re looking for, but I think this is the “discovery” that Wolfram made and attempted to formalize.
I see now why you were opposed to astronomy. Even though it has clearly exported important theories, those theories have been entirely assimilated into other disciplines.
I think the reason you are happier to include evolutionary biology in your curriculum (if I can call it that) is that this is a recent discipline, and its insights are less-well assimilated into higher level fields. Such fields are however emerging (see the literature on complex adaptive systems) and if you were to write this book in another century or two you might have a similar attitude to evolutionary biology as you do to astronomy.
Currently there is still value in studying evolutionary biology in itself.
We have already said mathematics is important but I would stress certain subfields, including statistics, probability (including bayes), calculus, and linear algebra. Also, information theory.
> The distinction is important only if you’re excluding sociology, psychology, and anthropology as sciences – those being the places where exports from economics often have considerable explanatory power.
As a physicist, I am of course obligated by guild rules to exclude those fields as sciences. That aside, can you give a specific example of an economic concept with explanatory power in, say, anthropology? It seems to me that there is an important difference between anthropologists doing economy and anthropologists doing anthropology informed by economics, although the line is not necessarily easy to draw.
>That aside, can you give a specific example of an economic concept with explanatory power in, say, anthropology?
Read my papers and see :-). My discussion of the origin of property rights in Homesteading the Noosphere is particularly relevant. It explains, for example, why the !Kung San have ownership customs with respect to waterholes but not random patches of arid scrubland. This led into my discussion of project ownership among hackers.
I wrote “Read my papers and see [economic ideas with explanatory power in anthropology]”
After thinking about it for a while, I’ve decided that it would be misleading to leave you with the notion that I think this kind of idea transfer is common. It isn’t. Anthropologists as a group not only don’t grok economics, they’re rather hostile to grokking economics for reasons similar to those that cause their friction with the evolutionary-bio crowd.
The kind of analysis I did in HtN is generally something you see from economists trespassing on the anthropogists’ turf. David D. Friedman does it. Not many others try.
Oh, and for god’s sake put Kuhn on the reading list.
>Oh, and for god’s sake put Kuhn on the reading list.
Absolutely not. Kuhn was deeply, deeply wrong. See http://esr.ibiblio.org/?p=40 for discussion
Purely as a historical footnote, I would point out that the entire credit account of evolutionary biology transfers over to economics, because Darwin’s key theoretical concept was a generalization of Malthus.
Biblical criticism has been highly generative, as Nietzsche pointed out. All the skills that go into establishing the correct form of a text, or determining what meanings can reasonably be attributed to it—which is much of literary criticism and history—were first developed for use on the Bible. Admittedly that’s not a “science” but I think it’s clearly a Wissenschaft.
Meteorology is turning out to be generative in a surprising way: It gave us the first example of computational intractibility, which led to sensitive dependence on initial conditions as a concept, and to the realization of the limits of predictability of dynamical systems, all of which applies directly, for example, to economics.
> My discussion of the origin of property rights in Homesteading the Noosphere is particularly relevant.
I read the relevant chapter; I concede the point.
>Anthropologists as a group not only don’t grok economics, they’re rather hostile to grokking economics for reasons similar to those that cause their friction with the evolutionary-bio crowd.
A more mainstream example that somewhat supports your point is Netting’s Cultural Ecology. I read the second edition in 1994, and though the title says “ecology”, which is the in-thing with anthropologists, the content was actually mostly economics as applied to pre-industrial, and mostly pre-agricultural, groups.
Interestingly, I have browsed several “Cultural Ecology” books since, but none of them were as clear (or as small) as Netting’s, and about two-thirds of them were mostly attempts at “anthropologic justification” of environmentalism.
Can you recommend a good survey of philosophy of science? I have read in several, but every one the author had particular sciences he disparaged for non-scientific reasons. For example, Philip Kitcher wrote some readable books, but he hated sociobiology and it showed clearly in his works.
>Can you recommend a good survey of philosophy of science?
Reichenbach’s “The Rise of Scientific Philosophy” influenced me a lot in my childhood and teens; it dates from 1951. Bertrand Russell’s “History of Western Philosophy” introduced me to Peircean fallibilism. That one is available for download. These sources are old, but the unfortunate truth is that more recent work improves on them very little. In some respects the entire field is still struggling to catch up to where C.S. Peirce was in 1890.
Later I was very influenced by General Semantics, which is more or less applied philosophy of science recast as a quasi-therapeutic discipline. Any of the introductions such as Chase’s The Tyranny of Words or Johnson’s People in Quandaries cast a remarkable amount of light on philosophical problems through their advocacy of rigorous language analysis.
I wrote: “These sources are old, but the unfortunate truth is that more recent work improves on them very little.”
One of my long-term to dos is to write a book on the philosophy of science. The analytic tradition that runs from Hume and Locke through Russell, Peirce, the Vienna Circle, and their successors has never been properly married to modern neuroscience and evolutionary psychology. It is possible I might be able to do something significant in this line.
” The question I’m actually considering is: how can an aspiring polymath self-train most efficiently? That is, which scientific disciplines are not only important in themselves but also export conceptual toolkits that will give the fastest running start in other sciences?”
Firstly, it seems your definition of polymath excludes many non-science fields, but the first comment already addressed “disciplines” vs. “sciences.” (Even though I actually strongly disagree with NukemHill’s assessment of the playing field.) Does a polymath not train in historical knowledge?
Secondly, I don’t find exportation of theory inherently the best indicator of an ability to “jump start” knowledge acquisition. I would argue that, if you have the stomach for it, philosophical, linguistic, and even literary theory (those which directly address cognition and creativity) can be better “kickers” to enhanced knowledge than those that are exported across fields (and essentially remain the previously acquired knowledge with new applications).
Thirdly, while some of the theories you’ve mentioned truly move across and inform other fields, some merely seem like loose instructional analogies. For example, evolutionary biology certainly informs, maybe even subsumes, much of psychology, but the idea that physics can “teach” psychoanalytical transference, just as an example, rather than merely serving as a useful analogy seems like it could lead to a highly flawed education in psychology.
Fourth, the idea that there is a “ideal” order or preference to knowledge acquisition seems profoundly flawed, particularly if you are less focused on the “purer” “sciences.”
What about music? Does the practice of it allow, or encourage, greater functionality in different mental disciplines at a later point? (Or does this not count as a discipline for your purposes?)
@esr: I’m starting to see what you’re talking about. You’re a math major. You’re looking for new equations (that you’ve not seen before) that could be applied outside their normal fields of applicability. This is fine and dandy, but don’t classify sciences as ‘generative’ only if they provide them. You only see half of science. Theory is nothing without experiment, and vice versa. Astronomy continues to generate new theories all over the place by producing observations that can’t be explained when they are made. Michaelson and Morley reported that they couldn’t detect any change in the speed of light regardless of the motion of the Earth; that one generated a lot of things, and led to Einstein and relativity. It’s all science, and just as generative as the output of a Newton or a Darwin.
@William H. Stoddard Agreed on Malthus’s import to evolutionary theory.
Disagree that the entirety of literary criticism arises out of biblical criticism. The key developments in literary criticism occurred between the turn of the century (Propp) and the unfortunate decline of post-structuralism into cultural studies in the 70s and 80s; not only did these developments occur in areligious and atheistic settings and from like-minded thinkers, the theories were not well-served by or focused on “significant” texts. In fact, I was specifically thinking of Barthe’s birthing of post-structuralism in “S/Z” — a theory and method of analysis I would rate very highly in terms of “kick starting” knowledge acquisition, while not being highly exportable, in my comment above.
Economics is the imperial science, because it has something to say to every other science, and most things end up being a special case of one economic principle or another.
I think of descriptive theory as being primarily backward focused, i.e. offering an explanation of past events leading to present conditions. Conversely, generative theory is more future focused and assumes a fundamental causal relationship over time. Chaos Theory plays havoc with both modalities (as in, we cannot accurately predict weather patterns more than a week in advance). And yet, in fractal dynamics there is order and repetition over scale, and endless allegories in nature. I also think that organic and inorganic sciences are fundamentally different. The physical universe behaves according to immutable laws, and we humans merely seek to better understand the world as it exists. The behavior of life forms is intrinsically guided by survival and reproductive advantages, and in the case of human life, by volitional factors.
> how can an aspiring polymath self-train most efficiently?
Nanotechnology advocate K. Eric Drexler has also addressed this need, in at least two brief essays:
How to Understand Everything (and why)
How to Learn About Everything
If any of this is useful to you, Drexler will likely let you use it freely. He has done some work in open source molecular modeling software.
Hope this helps.
>If any of this is useful to you, Drexler will likely let you use it freely.
I have no doubt of that at all. Eric Drexler and I have been friends for decades :-)
The bullet points under “How To Learn Everything” look particularly useful.
“What are the patterns that astronomy exports?”
Astronomy -> Astrology (in essence, a subfield of fractal cosmology) -> Jungian Psychology -> all sorts of predictive applications.
Science hasn’t seriously tested it probably because of deep-seated institutional aversion to the very idea that it isn’t ancient superstition of some kind.
Does programming count? I remember the philosophers debate about whether existence is a real predicate, i.e. what you say about a nonexistent object (“the king of France is bald”) is false the same way as a false statement about an existing object? And I thought to myself, man, every object-oriented programmer solves it like five seconds.
Math was mentioned, but I would especially emphasize the statistics part of it esp. methodologies of sampling etc. This helped me a lot not to buy any random pop-sci psychology article of the “study suggests people find more happiness from experiences than from possessions” kind.
Economics – I would say it is a two-bladed stone. When I discovered it (it being basic incentive analysis, cost-benefit analysis and everything that comes from that, the basic stuff, not the behavioral econ etc.) felt like this is the philosopher’s stone that explains everything in human behavior and history. Now I am mostly trying to unlearn it and stop thinking like an economist, as it is greatly different from how everyday people actually think and given that the economy is based on such decisions… in the economist’s world insurance companies wouldn’t exist. Neither would lottery. And then this breaks down the theories of comparative advantage etc. etc.
I read The Tyranny of Words in the early to mid 1990s and one of Hayakawa’s books in the late 1980s, but they struck me as a combination of fairly obvious and overly ambitious. Among other problems, the whole General Semantics idea is largely an extension of the strong Sapir-Whorf mistake.
Your mention of Peirce, though, reminded me of one I started to read that looked good, but I got distracted, Susan Haack’s Evidence and Inquiry. I need to run it down and finish reading it.
>the whole General Semantics idea is largely an extension of the strong Sapir-Whorf mistake.
It can look that way at first, but actually not. The actual GS insight is that Whorfian effects become important at the point where reasoning takes on a linguistic representation, whenever that is. We know that point is reached when people communicate via language, so the intersubjective social reality mediated by language is subject to Whorfian effects (and thus benefits from the GS critique) even if private thought is not.
And I must tell you that based on my experience as a crib bilingual who has spoken four languages I don’t think strong Sapir-Whorf is a mistake at all. I have actually felt the effect on my thoughts when I code-shift between languages, and if the reports I’ve heard from other polylinguals are to be believed this is an experience we all have. It is also relevant that I can think things in mathematical notation or Python that I cannot think in any natural language, or vice-versa.
can you give a specific example of an economic concept with explanatory power in, say, anthropology?
Nash Equilibrium. When I was a grad student (in Economics) I had a talk with an Anthropology grad student about this very topic. Apparently it comes up a lot in Anth. That’s what she told me, anyway.
It would be difficult to imagine a setting with purposeful agents interacting in which NE wouldn’t be useful. After all, purposeful agents best-respond to their environment, part of which is the behavior of other agents.
In fact, there’s something of a war between mathematicians and economists over who “owns” Nash. (That’s how I feel, anyway, whenever I encounter a mathematician calling Nash a mathematician or game theory a branch of math!)
OTOH, regarding optimization theory, I think we have to concede that it’s a subdivision of mathematics, properly speaking. Economists use it, of course. But I doubt that whoever first thought of setting a derivative to zero (Newton or Leibniz or whoever) was an economist.
in the economist’s world insurance companies wouldn’t exist.
This is a strange statement. In fact, first year grad school we went over formal conditions in which a person would pay some positive amount to avoid bveing subjected to an even-money gamble. Basically, it just requires a certain kind of risk-aversion. This is all very standard, actually.
You’re closer to being on target when you mention lotteries. These things taken together – insuring to avoid risk and at the same time gambling by buying lottery tickets – are regarded as a puzzle by economists interested in such topics.
“Basically, it just requires a certain kind of risk-aversion.”
And not even irrational risk aversion, it just requires a certain shape to the utility gained from income.
Biology and Chemistry ran into each other in unexpected ways when people realized that they could synthetically produce biological-compounds.
Does that make either science a generative science? Chemistry generated a lot of theory that was applicable to biology, and later medicine (among other fields).
Tim: Whatever current literary criticism and theory are doing with texts, they still have to have a text to work on. There’s a whole set of redactive skills that go into establishing what “the text” actually is, which is applicable to more than literary issues; I’ve edited a lot of history of mathematics papers, and historians put some effort into establishing the actual wording of documents, especially those from before printing.
Though I suppose literary analysis could be done on whatever a given instance of a text actually says, including whatever mutations have occurred as it circulated. That would certainly better fit oral culture, though it might go against the grain of scribal and print cultures. But the idea of the true version of a work still seems to be with us—it’s what impels the whole idea of a “director’s cut” in film, for example.
I recall a television series that explored how no progress in human thought occurs in isolation and that serendipity is stranger than we can imagine:
I disagree about geology and think that no advance in ANY discipline occurs in isolation.
Geology > stratigraphy & continental drift > biological concepts of evolution > etc etc
A couple of things to say to this. For one, what is the justification for your deciding on using concepts from science to understand other fields? Why would this be better than just trying to understand a new topic via a metaphor of a more familiar topic? In other words, are concepts taken from a generative science to understand another just metaphors which aid understanding, or is there some deeper sense in which they apply to the second field? You wrote:
“Once I know about least-action and equipotential principles in physics, there are analogies with generative power as far away as psychology and mass social behavior.”
Right, analogies. How seriously do you wish to take these analogies, and what makes that approach better than just using some non-science analogy?
Furthermore, do you think your idea here could be used to sort fields in order of fundamentality? Might it be correct to say that biology is more fundamental than computer science because they stand in the sort of asymmetrical relationship you mentioned? Interestingly, if you buy this argument it could potentially monkey around with the order we typically imagine the sciences being in, with physics being more fundamental than chemistry, which is more fundamental than biology. The order would get mixed up if an ostensibly less fundamental science, like biology, were in fact more generative than presumably more fundamental ones, like chemistry. In addition, it could be possible for fields to reverse positions, resulting in a field like computational biology.
>Why would this be better than just trying to understand a new topic via a metaphor of a more familiar topic?
Because sometimes you get homologies that are almost like isomorphisms in mathematics – analogies which are more than just suggestive. An example from my own experience was when I realized that hacker ownership customs behave very much like common-law land tenure because they solve a similar adaptive problem.
>Furthermore, do you think your idea here could be used to sort fields in order of fundamentality? Might it be correct to say that biology is more fundamental than computer science because they stand in the sort of asymmetrical relationship you mentioned?
I think it might, yes, though I would not try to overinterpret what that ordering means. Some of it is just historical contingency.
Well, as long as you aren’t trying to write fiction…
Mathematics is so fertile it’s probably worth breaking down into pieces. Information theory is one that I’ve found surprisingly broad applications of that most people miss because they don’t know anything about it. The full richness of chaos theory is over my head, but simply knowing about strange attractors and the basic how and why of where they show up turns out to be incredibly useful across a wide variety of fields. What I literally learned in differential equations hasn’t been all that useful to me personally, but the principles of what can be done and seen with such dynamic systems has been useful. Etc.
Somewhat sadly, the “standard loadout” of mathematics taught in school is rather weak in this department. Not strictly speaking useless, but there’s nothing in the standard curriculum to correct a fundamentally static view of reality across a wide variety of disciplines, except a smidge of introduction to the simplest possible differentials.
“The relevant question is roughly this: if you want to study scientific fields in order such that the ones that give you the most help understanding fields other than themselves come first, what’s the right order?”
That’s easy: there is no order, in my opinion. You study a little of one, then a little of another, then back to the first, and so on. Sciences tend to jump around. See David Burke’s Connections. (BioBob mentions the TV series; I have the book.) I believe this to be true even if you’re trying to embark on a crash course in polymathy. You could stay pure in one field before going to the next, but there are too many nodes where theory from one discipline is groundable in application in another, permitting further progress in theory. The side trips would be too hard to ignore after a while.
I can see economics as a generative theory. Pretty basic one, too. It all starts with understanding the concept of a trade. A little later, risk.
WRT astronomy: Eric, do you define astronomy to include research on Mars? If not, where does it fit in your taxonomy? Assuming I understand what you mean by generativity, Mars research is an example of how time-dependent this is: it isn’t spawning much outside of Mars research yet, but give it a generation or two…
…not that Mars research would be of much direct help to a present-day polymath in training, granted.
esr: > It occurs to me, thinking about stemmatics, that an earlier pattern for that kind of diagramming might have been genealogical charts.
It’s possible; the question is when did genealogical charts as such arise? A lot of early genealogical information tends to be pretty linear (e.g. so-and-so, son of such-and-such, son of someone-or-other etc.).
My understanding is that there seems to be a good deal of “sharing” of paradigms/ideas/concepts between linguistics/philology-as-textual-criticism, linguistics/philology-as-science, and biology during the 19th-c.
From the 20th-c., I think the interaction between linguistics and computer science is important, with the understanding of natural languages as symbolic languages, and discussion of the nature of the computation which must underlie the psychological component of language (including the Chomsky Hierarchy of languages).
esr: > And I must tell you that based on my experience as a crib bilingual who has spoken four languages I don’t think strong Sapir-Whorf is a mistake at all.
I’m still somehow agnostic about Sapir-Whorf. The thing though is that some studies which purport to provide evidence for Sapir-Whorf turn out to conflate various factors which render their results as questionable at best.
I get what you mean on the impressionistic side, with code-switching from one language to another, but I’m not sure that evidence for a strong version of Sapir-Whorf.
>The thing though is that some studies which purport to provide evidence for Sapir-Whorf turn out to conflate various factors which render their results as questionable at best.
I understand this; I’ve read some of the critiques. It’s a very difficult effect to study. Still, having read some of the relevant literature I get a pretty strong impression that the current strong rejection of Sapir-Whorf is driven at least as much by the internal politics of the linguistics field as by anything the evidence says.
I think one of the reasons academics reject Sapir-Whorf is because of the obvious logical implication: If language affects thought, being able to think in some languages can be an advantage to thinking in certain ways. And that violates the fundamental principle of Multiculturalism: All cultures are equal,(except for the White Heterosexual Anglophone Male-dominated Western Civilization, which is of course the root of all evil in the world) and anything that says otherwise must be false.
Meanwhile, they are forever telling us that certain words are DoublePlusUnGoodThink, and ban them, as if they know damned well that language influences thought. When they start acting like they don’t believe that, I’ll believe they don’t believe it.
>> how can an aspiring polymath self-train most efficiently?
It seems to me that the important thing is generative ideas, rather than fields of study based on generative ideas. Evolution is at the basis of biology, and it’s worth doing some study to see how evolution plays out in different areas, but has biology got other generative ideas?
Maybe the next question is “what do you want to be a polymath for?”. Because the world is fascinating? Because you want to come up with the next generative idea?
If it’s the latter, you may need to spend a lot of time getting familiar with some region of knowledge which doesn’t have a generative idea. I also suggest that cultivating the ability to know when you’re surprised (something specific doesn’t fit your theories) or confused (something is so far outside your theories that they need major revision)..
Possibly of interest: Root-Bernstein’s work (Sparks of Genius, Discovering) about high level creativity. One of his points is that important new work can come out of trying to solve a practical problem in a field which isn’t one’s original specialty.
TomC: These things taken together – insuring to avoid risk and at the same time gambling by buying lottery tickets – are regarded as a puzzle by economists interested in such topics.
Why shouldn’t people insure against risk to avoid the possible consequences of bad luck and buy lottery tickets to incur those of good luck? Is there a problem with modeling it?
How many dinosaur species do we really know about?, or some modest generativenss by way of common sense.
@The Monster: I’m pretty sure that injunctions against slurs (racial and otherwise) has:
1) nothing to do with a fear that language influences thought.
2) everything to do with letting people from historically oppressed regions of society work and play without having images of violent oppression suddenly brought up.
Anyway, the Sapir-Whorf hypothesis is about structural aspects of language (such as the presence or absence of noun classes, or compulsory marking of tense or aspect), not individual lexical items.
>but has biology got other generative ideas?
Actually, it is a rich source of ideas on systems in general. Not just evolution, but also development, ecology, and even the physiology of organ systems and endocrinology within one body.
So are these “images” that are “suddenly brought up” by the mere utterance of a banned word or phrase… THOUGHTS?
What “images of violent oppression” are “suddenly brought up” by the University of North Dakota’s intercollegiate competition teams using the name “Fighting Sioux”? (Are similar images brought up by “UND Fighting ___” when the ND is Notre Dame, and the blank is filled in with “Irish” instead of Sioux?)
You could make a case for history as being one of those academic disciplines* that is generative. Econometrics is one whole discipline that had at least as much input from economic historians as from economists. Archaeology is another that used to be part of history. Textual criticism, another discipline that started with historians trying to understand their sources. The whole business of source analysis and identifying and overcoming bias.
As a historian (of hacker culture) yourself, I’m surprised that you haven’t considered it.
The problem, of course, with history is that it’s huge. The amount of available knowledge is probably comparable to the whole of science. And science as a whole is far more generative than history as a whole.
*Mostly used to avoid the argument about whether history is a science or not.
> ….I get a pretty strong impression that the current strong rejection of Sapir-Whorf is driven at
> least as much by the internal politics of the linguistics field as by anything the evidence says.
Perhaps, there are certainly internal politics to the field. But my impression has generally been that the divide is in part “cultural”, in that people on the more humanities-oriented side of linguistics have been interested in finding evidence for Sapir-Whorf, while those on the more cognitive-science-side of linguistics are more doubtful that the “surfacy” differences between languages would have any deep effect on other cognitive processes. (Though I’m sure there are exceptions to this cultural generalisation.)
>I think one of the reasons academics reject Sapir-Whorf is because of the obvious logical >implication: …. violates the fundamental principle of Multiculturalism
Linguists are generally more interested in the idea of languages are being equal/equivalent. And this equivalence is about structural issues; obviously there are social and political advantages to speaking English (Spanish, Mandarin, Hindi etc.) over Navajo (Basque, Tlingit etc.), but that’s part of the socio-side, and nothing to do with structural properties of the languages.
>Meanwhile, they are forever telling us that certain words are DoublePlusUnGoodThink, and
>ban them, as if they know damned well that language influences thought. When they start
>acting like they don’t believe that, I’ll believe they don’t believe it.
That’s an entirely different topic. The Sapir-Whorf hypothesis has to do with the cognitive effects on person X caused (or not) by person X speaking Language Y. In this second sense that you bring up, obviously language (as speech) can influence thought. If it didn’t communication wouldn’t work. If I say or write “ice cream”, that will stimulate you to think of ice cream. But that’s not what we’re talking about.
“Academics” are hardly a group that behave a unified fashion. But, in any case, I don’t who it is that you thinking of, in terms of academics, who plans to ban words.
I will be very interested to read what you have to say about economics self-education. I often find myself wanting to have a deeper and more evidence-based understanding of the subject, but I just don’t have any idea where to begin. Whenever I look into the subject I immediately find myself lost in a sea of ideologies and mutually-contradictory theories.
I would like to know to what extent economics can be considered a science, and which parts of it are the best substantiated.
>I often find myself wanting to have a deeper and more evidence-based understanding of the subject, but I just don’t have any idea where to begin.
David D. Friedman’s Hidden Order: The Economics of Everyday Life. Best introduction ever.
>I would like to know to what extent economics can be considered a science, and which parts of it are the best substantiated.
Well…that’s complicated, but here’s a good first approximation. Microeconomics is a science that is well substantiated, macroeconomics is bad voodoo. Friedman’s book teaches microeconomics. Work labeled “neoclassical”, “Chicago school” or “Austrian” is likely to be sound, with one significant limitation; rationality assumptions sometimes get them in trouble. The most interesting work going on now integrates those schools with results from experimental behavioral economics.
Work labeled “Keynesian” is probably bunk. Work labeled “Marxist” is certainly bunk. Many schools of economics. and almost the entirety of macroeconomics, are basically nothing but GIGO designed to rationalize political ratfucking. Ignore those, and ignore the fairly pointless theological disputes between the Chicago school and the Austrians.
Sure, but if your notion of the Sapir-Whorf hypothesis is that “listeners access the meanings of words that they attend to,” then (your notion) of the Sapir-Whorf hypothesis is trivially true.
Nobody is trying to tell you that you aren’t allowed to say “fighting Sioux.” Using an ethnic caricature as an athletic team mascot does quite a bit more than lead to people saying “fighting Sioux.” When people argue against these mascots, it is not the bigram “Fighting Sioux” itself that they take issue with.
Do you really need me to give some examples of violent images evoked by slurs? Racial slurs do evoke images of lynchings, burning crosses, and police brutality tageting racial minorities.
Again, it is not the bigram “Fighting ___” that people take issue with. If the Notre Dame had been founded by a government or other body engaged in a campaign towards the near-genocide of the Irish, and if racial stereotypes of the Irish still had the emotional punch today that they did in the 19th century, then I’m sure the mascot would be controversial.
I’m interested, and rather surprised, to hear that the hypothesis is not fashionable within linguistics. When I was an undergrad philosophy student this idea was almost an accepted dogma within the department (and, it seemed to me, within philosophical academia generally). Obviously there is some level of brain activity that does not require language. But this pre-linguistic aspect of ‘thought’ is not (to my understanding) the subject of Sapir-Whorf. The hypothesis refers to higher-level symbolic reasoning.
This is one reason why it is so hard to explain advanced concepts in subjects like quantum theory to non-numerical people. The ideas just cannot be expressed in English because its vocabulary and grammar are inadequate. You can make analogies and simplifications, but these are always only approximations to the truth, which can only be accurately and precisely expressed in the language of mathematics.
Also, this is why the idea of ‘newspeak’ in Orwell has such power. The idea there is that by removing words from the language, you actually remove the ability to think about the concepts those words represent.
>>Maybe the next question is “what do you want to be a polymath for?”. Because the world is fascinating? Because you want to come up with the next generative idea?
I’m probably as much of a polymath as you’ll find. I’d qualify as an expert in two or three areas, but I can talk to people in many disiplines — understand them and co-ordinate with them. Help interpete between different groups. Do impedance matching between different systems.
A pretty good shot at being a joat, although a few I can claim mastery.
Sapir-Whorf isn’t that controversial, it’s just the subject of disagreement. If you think that Noam Chomsky rules liberal academia, you might believe that academia hates Sapir-Whorf. If you pay attention, less so. Consider George Lakoff, for example.
>This is one reason why it is so hard to explain advanced concepts in subjects like quantum
>theory to non-numerical people. The ideas just cannot be expressed in English because its
>vocabulary and grammar are inadequate. You can make analogies and simplifications, but
>these are always only approximations to the truth, which can only be accurately and precisely
>expressed in the language of mathematics.
But that’s not an example comparing natural languages to natural languages. I agree that what you say is true, but I’m not convinced it’s relevant to evaluating Sapir-Whorf. Likewise, one can say that there are concepts which are easily to communicate visually than verbally etc.
>Also, this is why the idea of ‘newspeak’ in Orwell has such power. The idea there is that by
>removing words from the language, you actually remove the ability to think about the concepts
>those words represent.
As much as I like Orwell, I’m not really convinced by this argument.
(1) There’s not a Great Dictionary in the Sky that one can excise words from. If the Academie Francaise declares that “le hotdog” is not a word, and bans its inclusion in printed dictionaries, this doesn’t prevent the word from existing in the mental lexicons of French speakers.
(2) I can think about things I don’t have words for. And (see (1)), where convenient, I can always add a new word if I need a “handle” for easily accessing some concept.
I’m not really sure why this is a relevant distinction.
When I said that the newspeak idea had ‘power’ I meant literary or dramatic power. It has that power because I think it appeals to something we all feel instinctively to be true.
However, although you are right that you cannot literally expunge words from a language at will, you can (if you have a lot of influence over the culture, using tools like mass-media) exert a great deal of control over the way people use language in real life. It’s not as clear cut as Orwell makes it (he is going to a dramatic extreme for literary purposes) but you can bring about a situation where certain words are not used.
Adrian Smith: Why shouldn’t people insure against risk to avoid the possible consequences of bad luck and buy lottery tickets to incur those of good luck? Is there a problem with modeling it?
Alas, I don’t really know the details because I’m not an economist interested in such things. Here’s a brief discussion of the topic, with some links to the seminal paper and other stuff, at Marginal Revolution:
“It’s not as clear cut as Orwell makes it (he is going to a dramatic extreme for literary purposes) but you can bring about a situation where certain words are not used.”
You seem to be confusing a fiction with a belief. The fiction of 1984, if read properly, tells us that Orwell believes exactly the opposite. He doesn’t believe a gov’t can attempt to control thought by controlling language and be successfyl doing so. He may believe that a theoretical gov’t may try to do so, but he certainly doesn’t believe it can achieve the desired affect. In Orwell’s worldview, an old paperweight can reignite the suppressed knowledge of the individual.
If you want to talk generative theories I’d say the king of them all is statistics. Every thinking person should get a good handle on statistics and statistical reasoning. Hell, statistics is practically a must if you care about your own enlightened freedom: how much of our culture is based on anticipation of extremely unlikely events whether that anticipation be hope (winning the lottery) or fear (terrorist attack)?
Excellent, thanks. I’ll check that out.
Many schools of economics. and almost the entirety of macroeconomics, are basically nothing but GIGO
Speaking as an economist, and one who would probably be classified as “macro,” I’m going to have to take issue with this. Macro is much more empirical than in the bad old days. Any particular model might be nonsense – and “All models are false,” etc. – but there is a large set of empirical results now. More importantly, a set of methods for obtaining results.
As to the word “Keynesian,” this word is so smeared out in popular discourse that it’s almost meaningless. As politicians use the word it essentially means “government intervention to stabilize the macroeconomy.” That’s not what it means to macroeconomists, though. At least, that’s not my sense of how my colleagues use the word. I would say the core connotation of “Keynesian” now is probably that prices are sticky, especially the price of labor. This has a certain implication, to wit that markets (especially the lbr market) don’t continuously clear. Note, though, that the existence of a problem does not imply that there is a solution, or in particular that the gov’t has the solution. That would be arguing
(1) You’re sick.
(2) Therefore I can cure you.
That wouldn’t necessarily follow, of course.
I find it fascinating what a lag there is between what’s going on in the field and the layperson’s perception of what’s going on. The average layperson seems (assuming they have any awareness at all) to think that Keynesian/Monetarist debates are still a big deal, or something. Hasn’t been relevant in at least thirty years, kids. No one cares about that shit anymore, because everyone agrees on the method of stochastic dynamic general equilibrium modeling.
>Macro is much more empirical than in the bad old days.
You say this, and I’ll even agree it’s true. Yet, what’s the standard joke? Ah yes. “Macroeconomists predicted nine of the last five recessions.”
About the most positive thing I can say about the predictive power of macroeconomics is that it’s slightly less bogus than climate modelling. And if you think that’s damning with faint praise, you’re right.
My point in that last sentence being this: When two people disagree about something, if they agree on the method for resolving the disagreement, then the disagreement won’t persist very long or be much of a big deal.
Megan McArdle put it even better than that. “Paul Krugman, who predicted eight of the last none recessions…”
That said, that “standard joke” is itself still (I claim) a lagging indicator of macro’s bad old days. That said, I also agree that it’s only slightly less bogus than climate modeling.
Meanwhile, sticky prices are, to be more specific, a key theme of New Keynesian economics, in case anyone wants to look into it further. (I’ve lacked time, or else I’d read up on the Austrian school response to this.)
…nine of the last five…
Funny, but you can’t cite a joke as evidence for your position. Of course economists have predicted 9 of the last 5 recessions. That’s because there’s always some maverick economist predicting a recession. By that standard, “economists” have actually predicted 17,382 of the last 2 recessions.
Look, the full quote that I was responding to was this: “Many schools of economics. and almost the entirety of macroeconomics, are basically nothing but GIGO designed to rationalize political ratfucking.”
(Pained look.) No; the vast majority of it is what non-economists would probably regard as agonizingly uninteresting stuff about whether interest rates and NNP are co-integrated of order 2 or whatever. Non-economists tend to focus on the political stuff (e.g. Krugman) because that’s what’s shoved in their faces, and I suppose because for most of them, the political aspects are the only ones they find interesting.
Can someone enhance their creativity through a structured course of study? And in particular, by studying an optimum sequence of scientific disciplines.
Think of athletics as an analogy. Perhaps you are born with genetics that are favorable to superior athletic ability. An unstructured life experience may result in a pretty good outcome (such as becoming a star athlete in high school), but is that good enough. Structured workouts, enhanced diet, knowledgeable coaching, and environmental factors that stimulate competitiveness may be employed to achieve a much higher level of athletic success.
Are you planning a “how to” book for mental athletes?
>Are you planning a “how to” book for mental athletes?
Something close to that, yes.
I’m surprised about your placement of physics. And mathematics
If I understand your criteria of evaluation correctly (that a “more generative” field is one that produces more models that are deeper than a set of equations of state, and eventually lead back to the equations of state being observed), then phyiscs has been *very* generative. In our thermodynamics class, the professor told us: “First we will be learning thermodynamics as it has been discovered – via empiricism. Then we will derive all the same things starting from nothing but a few assumptions about atoms, and logically necessary information theory. It is possible for an alien from another universe that has never heard about atoms or thermal processes to derive our ideal gas law.”
In fact, if the hype that physicists put out about their discipline is anything to go by, (sometimes I feel uncomfortably far from the empirical world), you can use the basic laws of physics to eventually, with enough olympic application of math and computing, generate any other aspect of our classical world.
“It is possible for an alien from another universe that has never heard about atoms or thermal processes to derive our ideal gas law.”
Maybe possible…but not bloody likely. Those sorts of derivations are only possible if you know where you’re going from the start. Theoreticians need experiments and other theories to guide them. Newton could never have produced his mechanics without Galileo’s experiments and Kepler’s laws. Einstein would not have produced relativity without Maxwell’s equations.
The empirical approach lies at the heart and soul of science.
My general attitude is that you should begin by learning the field you are going to work in. Cladistics is going to revolutionize (and hopefully revitalize) historical linguistics; it’s already started to do so. Unfortunately, the first stage of this process was a host of non-linguists, including the Atkinson and Gray mentioned above (and before them Cavalli-Sforza) doing a vast amount of plausible but wrong work because of their ignorance of fundamental historical linguistics. In cladistic terms, they were trying to do phenotypic analysis without knowing what counts as a distinct state of a character in linguistics. For example, Armenian erku doesn’t look much like English two, which is much more like Spanish duo/dua. But all of them are derived regularly from a common ancestor, which means they are useless for classifying Indo-European languages. Malay dua, on the other hand, is sheer coincidence, a factor that non-linguists don’t make enough allowance for. (There are good coincidences like this even within Indo-European, like English much and Spanish mucho (regularly from Latin multum), or German haben ‘have’ and Latin habere, where the true equivalent of haben is capere, or Persian bad and English bad.
John Cowan, one canonical example of that sort of thing is Mbabaram, an extinct and almost totally unique Australian Aborigine language (i.e., no ties to other Australian languages are known with certitude to exist) in which the word for “dog” is “dog”. Pronounced exactly the same. Linguistic experts have concluded that it’s Just One of Those Things. This example is often cited as a warning against assuming a genetic relationship between words in different languages just because they sound alike.
Sorry, I mixed up Forster and Toth with Atkinson and Gray. My remarks were meant to apply to Forster and Toth. I think Atkinson and Gray are wrong too, but I can’t justly accuse them of profound ignorance of linguistics.
Jeff, the reason I didn’t use that example is that I can never remember how to spell “Mbabaram”, and usually come out with Mbarabam or the like.
John, if it helps any I think of a scene from the movie Surf Ninjas, in which the main character pays tribute to a visiting monk named “Baba Ram”, to the tune of “Barbara Ann” by the Beach Boys.
I’ve always thought that Heinleins troika, Math, Language, History was good advice.
I wouldn’t just include “certain areas of philosophy”, I would include philosophy as a general enterprise. Although practically useless in and of itself, the mere existence of philosophy has generated tons of spinoffs. Every science we know today has spun off from philosophy, often in unpredictable ways. The first economist for example, Adam Smith, was a professor of moral philosophy—not of epistemology as an 18th-century Hayek might have been, and certainly not of public finance or administrative law. Philosophy frequently does the job where more plausible disciplines do not. And you never know which specific corner of philosophy a new field will grow out from.
What does philosophy export? I think the most important candidates are mental habits such as:
searching for the truth rather than presuming we already possess it and only have to teach it. (That’s what theology and accounting do.)
asking truly open-ended questions;
giving every candidate answer a chance, ignoring any social taboos around it;
discarding every candidate answer, no matter how socially-approved-of, if it’s inconsistent internally or with observation.
This list is likely to be incomplete. And if its bullet points seem obvious, that’s the point, kind of. They are obvious to us netizens because we, as a culture, have already been infected by philosophy’s generative memes. Other cultures haven’t been infected in that way—ask any Arabian peasant—and so they are finding it enormously difficult to catch up in the sciences. Any sciences. It’s almost like a controlled experiment.
PS: I apologize for assuming that list tags were supported in comments. My bad.
I disagree with macroeconomics being “bad Voodoo” and with Keynesian economics being “probably bogus”. Nevertheless, I will resist the temptation to argue this point. In my experience, American conservatives and libertarians tend to have utterly false preconceptions of what Keynesian economists are actually saying. I’m afraid that identifying and rebutting these preconceptions would take the thread too far off-topic for too long. Hence, I’ll just say “I disagree”, and leave it at that.
Back on topic, though, I agree that microeconomics is much more generative than macroeconomics. Apply microeconomics to utiles (Bentham) or wealth (Posner), and you get a positive account of legal norms. Apply microeconomics to genes (Dawkins), and you get the selfish-gene model of biological evolution. In turn, apply this microeconomical selfish-gene model to memes (Dawkins, Dennett, Blackmore), and you get a comprehensive account of human consciousness and culture. (Its full implications haven’t been fleshed out yet, but I can’t wait to see them when somebody does.) Macroeconomics is different. Bogus or not, it’s always limited to the economy of actual countries in which actual people sell goods, services, and investments to each other. That alone justifies Eric’s focus on microeconomics for his project.
Geology is generally pretty “sterile”, in that most of its generativity is in the past (though it took a *long* time for Continental Drift to be accepted – books in my childhood hadn’t fully accepted it, and I’m under 50). There is one concept which has found wider application outside geology, though I’m not sure it actually started there: limit equilibrium. Systems tend to remain stable in low-entropy situations until suddenly, they’re not. In geology, the classic example is slope stability – a stable slope can have a *lot* of potential energy, but won’t fail until something pushes it over the edge. It’s kind of like chemical activation energy, only for more complicated systems.
I wonder whether the generative or descriptive nature of any branch of knowledge mostly depends on the attitude of its key players and on circumstantial events. Also, whether it defines the degree of “creativity” of such a branch of knowledge. If anything, a branch of knowledge classifying as descriptive would be a very strong indicator of crisis (=separation, decision, judgement) before a new generative stage.
Fab letter to Dodd by the way.
I realize I’m late to the game, and might be laughed out of town for this. Since you didn’t dismiss psychology as a science, I thought I’d point this out.
Psychology is so borad that it almost can’t be generitive, because understanding how people think is useful anywhere people are thinking. The theories of psychology that work, such as those of Pavlov, B.F. Skinner, or Piaget, can be applied in literally any field related to people. Yet all of these uses are still considered psychology.
For example, understanding Skinner’s reinforcement models would be extremely useful in designing a good video game, but the use would still be considered psychological- the models are being used directly to influence human responses. But it’s useful for a game designer to understand.
It’s also the answer to what one of the other users said (I can’t find the post to quote it) about economic models not being applicable to the actual behavior in the marketplace. . . hence why advertising relies on psychology so heavily, and corporations rely so heavily on advertising. Insurance and lotteries may not make sense with economic models, but they certainly make sense with psychological ones.
(Sorry to double post, but an excellent example occurred to me. Understanding cognitive bias is necessary to analyzing the validity of experiments in almost any science that isn’t math based.)
“Insurance […] may not make sense with economic models, but they certainly make sense with psychological ones.”
I’d ask you to explain, but I think I already get it. The single most ridiculous idea I have seen from otherwise intelligent people on the internet is the claim that risk-aversion (which is what drives people to buy insurance) is irrational.
One of the most valuable things I got out of the engineering education was an appreciation for how each model of how the world works is only an approximation of the truth, founded on assumptions, and of limited usefulness.
I’m not sure how you would teach this to yourself. My personality and habits of thought were already like that to begin with, but the experiences of going through an engineering program changed my mindset significantly.
I think it was being asked over and over again to state the assumptions I was making and to be explicitly aware of the conditions in which it would not work, even when just doing calculations on paper. Then, doing projects where even though I had carefully applied what we had been taught so far, the project failed because of factors we didn’t consider (in safe classroom settings, where this kind of failure is part of the point of doing the project.)
Programmers think about edge conditions (cases when things won’t work) and they have to deal with other people’s design decisions. So there is some overlap with what programmers learn. But “This API won’t let me do X” is not the same thing as “Physics won’t let me do X”. “If I make a mistake, someone will file a bug report, and I can fix it in the next update” is not the same as “If I make a mistake hundreds of people could die.”
I learned a lot from studying engineering, and I’m oversimplifying, but four types of experience come to mind as something that might be valuable to someone wanting to learn things in other fields:
1) Asking these two questions over and over in a variety of contexts: What are the assumptions? Where will this model work reasonably well and where will it fail?
2) Doing projects where you come up against the limits of what is possible in nature, and where the consequences matter.
3) Analyzing engineering and design failures.
4) Learning when you can trust your life to something that is only an approximation of the truth. (And being aware that *all* models used by engineers are only approximations.)
Some types of programming projects might work for the project part, but not just any project would do. Software for a medical device would work, or other software that has life-or-death consequences.
The reason I think it is generative is that when I encounter other areas of study, I find myself applying this mindset. I feel like studying engineering has given me a practical grounding in philosophy, or at least better bullshit detection on philosophy, and I often find myself wishing that people in the humanities would bother to ask themselves what their assumptions are and in what cases their theory would not work. Too often, their approach seems sloppy to me. In general, people don’t bother to think through what the assumptions and limits of their ideas are, and they default to thinking their theory is universally applicable and founded on unshakeable truths. Usually, they are wrong.