Science and Natural Language

by Mark English

I once watched a nature documentary about a group of lions in the wild. At one stage a lioness and her cubs found themselves separated from the group. After a period of time during which she almost starved and her cubs were killed, she finally managed to rejoin the group. The other lions certainly recognized her and seemed to welcome her back. They sniffed her body. Who knows what they were thinking – if anything. But the point that struck me was that there was a story here that we as language users could conceptualize and communicate and which these creatures were utterly cut off from. They were tied to the present, and so couldn’t (presumably) even conceptualize time past, much less communicate facts about it.

Narrative is at the heart of language and at the heart of human psychology and culture. We define our individual selves in terms of a personal narrative or set of narratives; and our social and political views could be seen to be a function of how we see ourselves in relation to the broader community, of how our personal narratives intersect with our evolving culture and the cultures with which it interacts.

Political rhetoric is all about appealing to the basic values and hopes and fears of an audience and telling a simple story. In fact, if the current US presidential election contest is any guide, the stories are getting simpler and simpler. (About how voting for us is going to make everything right again or at least keep us safe and voting for them is going to lead to Armageddon.)

Of course, there’s something very suspect about these sorts of stories: their function is to not to describe things and events accurately but rather (in the case of personal narratives) to justify, or (in the case of political rhetoric) to change behavior. And all these narratives are language-dependent in one way or another: language and narrative have a very strong affinity.

Narrative is just one of the forms that linguistic communication can take, and there are many different kinds of narrative, each serving a different kind of (usually pragmatic or practical) function.

Science – like art and literature but in a different way – goes beyond the pragmatic and practical. It is concerned with a particular kind of understanding: understanding for its own sake. (Sure, scientific understanding is a means to many ends but it can also be conceived of as an end in itself.) And scientific activity, though it inevitably occurs within a social and so within a linguistic context and employs natural language directly (especially in its referential mode), typically goes well beyond the scope of natural language, drawing on mathematics and other invented codes and formalisms.

This obvious fact says something about how the world beyond the human, social world is, and also something about the limitations of natural language. However the world is, it is not such that it can be adequately described and explained in natural language.

This is a sobering thought but, looked at in evolutionary terms, it is hardly surprising. Language evolved over many tens of thousands of years to deal with immediate social and pragmatic problems, not scientific ones. In fact, science as we know it only really got started less than 3000 years ago and only matured into its current form in the last few hundred years.

Of course, science depends in many respects on language and on the kinds of referring, abstracting and symbol-manipulation (or morphosyntactic patterning) activities which natural language involves. The capacity for complex language changes the nature of thinking in radical ways, and there is clearly a link between, for example, natural language and artificial languages (like mathematics and logic). For example, the variables of formal logic look a lot like pronouns and its predicates look a lot like verbs or adjectives, and there are many other parallels.

Yes, certain non-linguistic animals can count and maybe subtract in an intuitive kind of way. They also enact certain processes which could be represented in terms of formal logic. But building and using explicit abstract systems can only be done by humans, and I think these formal systems are modelled to a large extent on natural language.

Even ordinary natural language involves a high degree of abstraction. You only need to look at common nouns like ‘tree’ or ‘dog’ to see that. So-called abstract nouns just take the process further (and often further from reality, as it happens). It has even been suggested (by Noam Chomsky) that language evolved primarily as a means of thinking rather than as a means of communication.

Be that as it may, it’s clear enough that without language there would have been no science and no mathematics. This is not to say that linguistically-based cultures necessarily give rise to scientific activity. In fact, throughout history the vast majority of human cultures and civilizations have risen and thrived and disintegrated or merged with others without ever producing anything remotely resembling modern science.

The British biologist Lewis Wolpert wrote a beautiful book called The Unnatural Nature of Science. I won’t try to deal with his specific claims here, but arguably science is unnatural in the sense of running against the grain of most of our innate patterns of thinking. Curiosity – which in a basic form is present in some other animals (after all, it killed the cat) – is one of the few natural drives against which science is not set.

This is just how things are. And if you want (not everyone does, of course) to have an understanding of the wider world in which we find ourselves, the world which both underlies and transcends the world of our social experience, you will have, not to leave natural language and natural language-based thinking behind – you can’t do that – but rather to recognize its inherent limitations and so the necessity of adopting other codes and modes of thinking. Unfortunately such codes and modes just don’t come naturally to most of us and have to be painstakingly learned, usually in a formal educational context and over a quite long period of time.

It is probably best just to accept the phenomenon of natural language as we find it, in all its variety and in its various aspects, and not make judgments about which aspects are more or less important. Certainly, it is a multi-faceted phenomenon which can be fruitfully approached and studied in a range of very different ways. The question as to whether some linguistic functions might be somehow more central (or primary) than others is problematic. Language is as language does, and there seems little point in trying to privilege one function over another.

What about forms? Here I have some strong – but not really worked out or developed – intuitions about the centrality of narrative.

I think it is very suggestive that when you are learning a language narrative texts are the most helpful. It is surprisingly easy to read a story in a foreign language, probably because our brains are primed for this sort of thing and project us forward into the narrative, setting up expectations which in simple stories are usually confirmed. Static description, by contrast, is harder to read. Balzac wrote novels, but they were full of description. Unlike Stendhal, say, Balzac described everything, and there are no prior expectations about, for example, how exactly a room might be furnished. You also need a wider vocabulary to deal with this kind of approach. My suggestion here is that perhaps the narrative mode is indeed primary in some sense.

Also worth noting in this connection is that narratives are often fictions – including the self-narratives we weave about our own lives. Sometimes these are quite out of kilter with reality, and could be reasonably characterized as ‘false’; but, strangely, even if they are plausible, they are never ‘true’ exactly. Unless one believes in an all-knowing God, there simply is no one true or truest narrative or interpretation of a person’s life. Similar things could be said about other kinds of narrative which are not overtly fictional, such as general history. An historical account can be false (i.e. can get things wrong) but can’t be ‘true’ in any clear and objective sense.

This is at once an obvious and a radical idea, and even a shocking one in relation to our personal narratives. It shocked me when I first cottoned on to it, as previously I had always implicitly assumed that there was a true story of an individual life more or less in the way that there is a true account of how stars are formed (or any scientifically-describable phenomenon). This contrast brings out, I think, a fundamental difference between scientific and other kinds of understanding.

Science has always had an uneasy relationship with natural language. Typically, in scientific circles, natural language is not trusted. Scientific understanding is rarely based on natural language alone, and scientists gravitate to technical language and (where appropriate) mathematics. They are, like anyone engaged in complex joint activities, dependent on natural language, but they remain wary of it. Many scientifically-oriented thinkers could be characterized as revisionists with respect to natural language which they typically see as frustratingly vague or ill-behaved. (As indeed it is for most scientific purposes.)

Communicating science to non-scientists is notoriously difficult. Popular science books are said to lose thousands of potential readers per equation, so they are often exclusively language-based. But in areas like physics, for example, such books are of little value in conveying the actual content of current theories. Stories and metaphors can take you only so far.

This is a somewhat depressing note on which to end, so let me throw in (for what it’s worth) a final thought. Imagine there are superior beings who can observe our universe, whether from within it or from outside it (whatever that might mean) doesn’t matter. My question is, what would interest them most: the physics (i.e. the underlying mechanisms); or, to use Balzac’s phrase, the human comedy (bearing in mind that ‘comedy’ had a broader meaning back then)?

For us physics is difficult, even mysterious. But for them it would be as easy and boring as, say, basic arithmetic is to us. Who is to say that these superior intelligences might not be more interested in the crazy and totally unpredictable contingencies of cultures such as ours – in our human stories – than in what the scientifically-minded amongst us are inclined see as the deep and serious stuff?

Categories: Essay

Tagged as: , ,

49 Comments »

  1. While I don’t doubt the importance of language, I have a rather different view of it than that expressed by Mark. So I guess I’ll be disagreeing a little.

    I see language as a social adaptation, that evolved to facilitate cooperatition within the social group.

    And scientific activity, though it inevitably occurs within a social and so within a linguistic context and employs natural language directly (especially in its referential mode), typically goes well beyond the scope of natural language, drawing on mathematics and other invented codes and formalisms.

    Looked at that way, the use of language for ordinary descriptions also goes beyond the scope of natural language. The explorers who landed in Australia had no words to adequately describe the unique plants and animals that they found there. So they had to invent new words. But surely it is part of the scope of natural language, that it is extensible, that we can increase our vocabulary as needed. And once we recognize that language is extensible in that way, then scientific use of language does not go beyond its scope.

    I think it goes too far to credit language with making science possible. Again, I’m not doubting the importance of language, but I think the emphasis on language misses other issues. If you could give language to the troop of lions that you mention, they would not start doing science. Indeed, most human cultures did very little science, in spite of having a natural language available.

    As a solo individual, there is no reason at all for me to engage in science. My time should be spent seeking food, water, shelter, etc, to satisfy my biological needs.

    I see the culture at large as the real source of science. Our culture has managed to change its reward system, so that I can (as part of a group) be rewarded for doing science. And yes, we might need language to make those changes in the reward system. But it is the social structure at large, including those changes to the reward system, that we should recognize as making science possible.

    Like

  2. A quaint view of science and its relationship to ordinary language and common understanding, in many ways reminiscent of logical positivism — those damned inaccurate natural languages! We’d better fix ’em up, and don’t even try using them for science until we do! — and completely unsustainable given what we now know, in the same way that logical positivism is unsustainable.

    Natural language no less “accurately” describes the world than science does. But the two describe the world in fundamentally different ways or better, in fundamentally different aspects. Natural language describes the world in intensional terms, while science does so purely extensionally. Not only is it a mistake to say that the former is “less accurate” than the latter, one could easily make the opposite case. Hence the reason for not thinking of the two in terms of relative degrees of “accuracy.”

    And of course, the clincher is that all technical dialects are ultimately held up to considerations of adequacy that are expressed in ordinary language and under common understandings. Were they not, there would be no way to distinguish legitimate from illegitimate technical expressions and performances.

    So, I’m afraid that I think you’ve pretty much got this wrong in every dimension. But it was a fine read nonetheless and interesting too.

    Liked by 1 person

  3. If a tree falls in the forest, does it make a sound?

    What is the nature of “sound”? Is it simply a vibration in the air, or the perception of it by the hearer?

    Science explores the limits of the possible and the boundaries of the probable, but it does not assign any meaning to the eventualities (it is “knowledge for its own sake”). “Meaning” is incredibly personal and contextual, confounded by factors not germane to the outcome, and so constantly subject to arbitration. This is the great goal of politics – the arbitration of meaning – and I believe that while meaning may be informed by the axioms revealed by scientific understanding, meaning will never be subjugated to science.

    Liked by 3 people

  4. If I am not mistaken, in Asimov’s “I, Robot” there is a story about a robot that could understand things that humans could not and wasn’t interested with that because that was boring, just like your last paragraph.

    (I am not really sure where I found this story, if it was in “I Robot” or elsewhere)

    Like

  5. Dan-K said,
    Natural language no less “accurately” describes the world than science does

    I agree completely. We have a wonderful world of prose and a powerful, evocative, beautiful world of poetry that thrill, inspire and awe us.

    Our literature, song and poetry say something very deep and fundamental that simply cannot be said in any other way. Science is desperately unable to describe, understand or record the ineffable beauty we find in literary prose, song and poetry.

    Science is the extension of natural language to give it more precision in certain subject areas. While it has great precision in those areas it is blind to other areas.

    Narrative is at the heart of language and at the heart of human psychology and culture.

    I agree and think this is a very important statement. We are the sum of our memories. Our memories only exist because we can string them together in a narrative.

    Like

  6. Brian,
    Science explores the limits of the possible and the boundaries of the probable

    That is one way of putting it. I think a better view is that science is the pursuit of meaning through the understanding of causation. Our lives are dominated by our perceptions of causation. Every event has a cause and every action is a cause and our lives are embedded in causal chains. For a long time we had only folk theories of causation that had little explanatory power. To exert more control over our environment we needed better theories of causation and science was the outcome. Science then supplied meaning in that it better explained and predicted causation. This is only a partial, limited form of meaning and the great danger we have today is that we mistake it for the entirety of meaning, or what we call scientism.

    Like

  7. This is a charming conceit: that higher intelligence visiting us would find nothing of interest in our best science, which is never more than the logical result of assiduous exploration of material, where they would be way ahead of us. And if they choose not to annihilate us, it may be because we can offer them Shakespeare or Mozart.

    Like

  8. Neil

    “But surely it is part of the scope of natural language, that it is extensible, that we can increase our vocabulary as needed. And once we recognize that language is extensible in that way, then scientific use of language does not go beyond its scope.”

    Adding to the lexicon doesn’t take us beyond the basic (syntactic and other) structures that define natural language. But mathematics and formal languages go beyond natural language, and you need these new structures to describe/explain many aspects of the world.

    “I think it goes too far to credit language with making science possible. Again, I’m not doubting the importance of language, but I think the emphasis on language misses other issues. If you could give language to the troop of lions that you mention, they would not start doing science. Indeed, most human cultures did very little science, in spite of having a natural language available.”

    I made this point myself in the essay (paragraph 12). I am claiming that language is a necessary but not a sufficient condition for the development of science.

    “I see the culture at large as the real source of science. Our culture has managed to change its reward system, so that I can (as part of a group) be rewarded for doing science. And yes, we might need language to make those changes in the reward system. But it is the social structure at large, including those changes to the reward system, that we should recognize as making science possible.”

    I don’t deny that the broader culture is the source of science, but I was thinking more in terms of the way having language allows us to think in certain ways (involving abstraction, symbolic representation and manipulation, etc.). It may be that language could be seen a kind of cultural byproduct of the evolution of the brain, but I think a better way to look at it is in terms of co-evolution: culture and brains influencing one another and evolving together.

    Like

  9. Dan

    … those damned inaccurate natural languages! We’d better fix ’em up, and don’t even try using them for science until we do! — and completely unsustainable given what we now know, in the same way that logical positivism is unsustainable.

    This actually misrepresents what I am saying. Some logical positivists had a naïve view of natural language and wanted to ‘fix it up’ (Korzybski comes to mind, but even he was often misrepresented). I talk about some scientifically-minded people as being linguistic revisionists, but the crucial thing is that they are – quite rightly – wary of language and the common assumptions and ways of thinking which it embodies *in the context of scientific activity*. Nobody is suggesting that ordinary language is not perfectly well adapted for dealing with daily life.

    Natural language no less “accurately” describes the world than science does. But the two describe the world in fundamentally different ways or better, in fundamentally different aspects. Natural language describes the world in intensional terms, while science does so purely extensionally.

    Science has its limits but I’m not sure I’d draw the boundary exactly where you would. (Intensional *definitions* are an intrinsic part of science, mathematics and logic.)

    Not only is it a mistake to say that [natural language] is “less accurate” than [scientific descriptions], one could easily make the opposite case. Hence the reason for not thinking of the two in terms of relative degrees of “accuracy.”

    Exactly. It’s a question of using the *appropriate* language or framework. Verbal accounts of physics are inadequate (worse than inaccurate). Likewise attempts to replace ordinary language in areas relating to making judgments about art or manners or morals.

    And of course, the clincher is that all technical dialects are ultimately held up to considerations of adequacy that are expressed in ordinary language and under common understandings. Were they not, there would be no way to distinguish legitimate from illegitimate technical expressions and performances.

    I explicitly made the point that scientific activity requires a linguistic and cultural context (or matrix).

    So, I’m afraid that I think you’ve pretty much got this wrong in every dimension.

    Which is quite an achievement when you think about it. 🙂

    Liked by 1 person

  10. Brian

    “Science explores the limits of the possible and the boundaries of the probable, but it does not assign any meaning to the eventualities (it is “knowledge for its own sake”). “Meaning” is incredibly personal and contextual, confounded by factors not germane to the outcome, and so constantly subject to arbitration. This is the great goal of politics – the arbitration of meaning – and I believe that while meaning may be informed by the axioms revealed by scientific understanding, meaning will never be subjugated to science.”

    Nobody’s talking about subjugation, are they?

    On ordinary life and politics, I agree with you: they are just not amenable to rigorous scientific treatment.

    Like

  11. Labnut

    Thanks for the comments. I like what you said to Brian about science, though I think perhaps you are putting more weight on the ‘causation’ concept than it can bear.

    “Science is desperately unable to describe, understand or record the ineffable beauty we find in literary prose, song and poetry.”

    “Desperately unable”? You seem to be personifying science in a rather curious way.

    I agree with you about literature actually – though I would probably put it a bit differently. 🙂

    Also, I think the beauty and power in literature is not in the literary work per se but rather in the vision of the author which bears within it an entire culture (as does the perspective of any one of us). Its power comes more from the embedded culture – and from human universals* – than from the individual, if that makes sense.

    * Wilfred Owen said “the poetry is in the pity”.

    Like

  12. Mark,
    narratives are often fictions – including the self-narratives we weave about our own lives. Sometimes these are quite out of kilter with reality, and could be reasonably characterized as ‘false’; but, strangely, even if they are plausible, they are never ‘true’ exactly

    there is a true account of how stars are formed (or any scientifically-describable phenomenon). This contrast brings out, I think, a fundamental difference between scientific and other kinds of understanding.

    I think that these two sentences are central to your essay. It is a remarkable fact that that we are animals that innately seek the true, the good and the beautiful and have been equipped with the means to do it. In this essay we are concerned with our need to pursue the truth. But truth is an elusive thing. We each believe that we hold the truth but that belief is continually challenged by conflicting truths in the world around us. Other people’s accounts of the truth contradict our understanding of the truth. The disorder of the natural world defies a truthful understanding.

    Thus we believed that truth existed and found that truth in ourselves, in our own internal narrative. But we did not believe that an absolute, unchanging truth existed in the world external to us. Lacking that belief, science was not possible.

    This changed in the Middle Ages under the influence of Aristotelian/Thomist philosophy with its insistence on a God that was absolute, simple, unchanging, rigorously logical and completely dependable. As this philosophical belief permeated society and took hold in our imagination it became possible to imagine a world governed by an absolute, predictable order where truth existed. Thus we were enabled to believe ‘there is a true account of how stars are formed (or any scientifically-describable phenomenon)‘. Once we believed that science was born as we started seeking the truth that we believed must exist in the external world.

    Like

  13. Mark,
    “Desperately unable”? You seem to be personifying science in a rather curious way.

    I am being purposely polemical.

    I agree with you about literature actually – though I would probably put it a bit differently

    I know you do. But I would be interested to see how you put it differently.

    I think perhaps you are putting more weight on the ‘causation’ concept than it can bear.

    Please explain.

    Like

  14. There is a way that language itself is a proto-science. We understand when a pole becomes a lever and a rock becomes a fulcrum. We have achieved a generalisation and an operational understanding. We begin to know that there is a relationship between length and the effort expended. We may go from there to see the wheelbarrow as a type of lever. What if we extend the distance from where the scissors cross to the hand and so on? There was geometry before Pythagoras and Euclid but the Greeks added the sacred significance of number. Number rules everything. Numbers are the grammar of reality. If all the numbers you have are 1,2, 3 and plenty you won’t have science yet your operational understanding may be excellent.

    Like

  15. I tend to think of “natural language” as a technical term which would describe the vast bulk of scientific language and terminology that has “evolved naturally in humans through use and repetition”. If a pathologist sees a particular appearance of dying cells and coins the term “apoptosis” (using a formalised derivation method not unique to scientists alone), this is accepted or rejected in the same way as any other neologism – do our percepts staring down a microscope agree with his descriptions and pictures, and does the word trip off the tongue (the second “p” was originally silent). Since the founding of psychology as a science, there has been plenty of scientific human-centred talk, which has filtered down into everyday speech: eg “unconscious sabotage”, “introversion”, “cognitive dissonance”, or changed the meaning of older terms eg “confabulation”.

    As to formal languages like those of mathematics, I have seen the distinction made between abstract geometry (our artificial language) and the concrete geometry that actually is out there in the world. A correspondence theory of truth might just have it that mathematical languages are the most natural languages of all…

    Like

  16. Responding to Mark:

    Adding to the lexicon doesn’t take us beyond the basic (syntactic and other) structures that define natural language. But mathematics and formal languages go beyond natural language, and you need these new structures to describe/explain many aspects of the world.

    I see that we have very different views here.

    Perhaps I should add that I am a mathematician, and that I often taught the formal language class in our computer science department (before I retired).

    To me, saying that “formal languages go beyond natural language” is like saying that ice cream goes beyond cabbage. It is comparing two things that have very little in common.

    I have always disagreed with Chomsky’s emphasis on syntactic structures. From my perspective, a natural language is primarly a semantic system, and the syntax is mainly there for disambiguation. A formal language is primarily syntactic, and doesn’t really have semantics other than the formal semantics that emerge from the syntax. Mathematics is mainly communicated in natural language, because that provides the needed semantics. We resort to formal language when we need to be very precise in expressing detail.

    I don’t deny that the broader culture is the source of science, but I was thinking more in terms of the way having language allows us to think in certain ways (involving abstraction, symbolic representation and manipulation, etc.).

    Fair enough. But I’d have to say that most of my own thinking about mathematics is non-linguistic. That includes thinking about how to prove a theorem. In the final steps, where I have to give precise details, I begin to think more linguistically. But most of the thinking is non-linguistic. And it is the enriched semantics that come from being part of a community of mathematicians, that is particularly important.

    I’m sure that how we think varies from person to person. I do know that some mathematicians seem to think more formalistically. But many don’t.

    Liked by 1 person

  17. mark: Glad to see that I somewhat misunderstood you and that we’re not particularly far apart.

    Certainly Quine thought that if a language was to be suitable for science it would have to be extensional, and I am inclined to agree with him on that.

    Like

  18. Not much time to comment tonight.

    Interesting essay. I think the interchanges between Mark and Dan, and between Mark and Neil, have helped clarify some of the important distinctions here.

    Hopefully more to say tomorrow.

    Like

  19. labnut

    I certainly agree that modern science owes much of its character, including its empiricism and engagement with the world (technological focus etc.), to medieval thought which merged classical thinking with other traditions (including the Hebraic).

    [Quoting me] “I agree with you about literature actually – though I would probably put it a bit differently.”

    I know you do. But I would be interested to see how you put it differently.

    I made a couple of points about this in that previous response. I suspect you would have a more Romantic/Platonistic view (like Iris Murdoch?) than I do.

    [Quoting me] “I think perhaps you are putting more weight on the ‘causation’ concept than it can bear.”

    Please explain.

    Well, even within science the concept of causation is sometimes questioned or contested. I see science in slightly broader terms, I think; as motivated primarily by a desire to understand, to make sense of the world using our natural faculties and involving a wide range of approaches and methods (observation, reason, experiment, modelling, etc.). Certainly looking for underlying causes is a driving force and characteristic of many areas of science.

    Liked by 1 person

  20. ombhurbhuva

    Language as (in a sense) proto-science? Yes, to the extent that it is involved in the kinds of thoughts and activities you describe.

    Such a view is quite compatible with what I am saying in the essay, I think (science building on capacities associated with natural language).

    Like

  21. davidlduffy

    Yes, scientific neologisms can be seen as a part of natural language and, as you say, they sometimes filter down into everyday speech.

    “As to formal languages like those of mathematics, I have seen the distinction made between abstract geometry (our artificial language) and the concrete geometry that actually is out there in the world. A correspondence theory of truth might just have it that mathematical languages are the most natural languages of all…”

    Nice point.

    Like

  22. Neil

    Your second comment touches on some really fundamental questions.

    … saying that “formal languages go beyond natural language” is like saying that ice cream goes beyond cabbage. It is comparing two things that have very little in common.

    Yes, I could have put it better but I think it makes more sense if you read it in context. I wrote:

    “Adding to the lexicon doesn’t take us beyond the basic (syntactic and other) structures that define natural language. But mathematics and formal languages go beyond natural language, and you need these new structures to describe/explain many aspects of the world.”

    I certainly didn’t mean to imply that a formal system is an enhanced version of natural language.

    I have always disagreed with Chomsky’s emphasis on syntactic structures.

    Notice that I said “syntactic and other” structures. These include phonological, lexical and semantic structures.

    I agree with you, however, that semantics is basic to natural language – that semantics is built in from the start. (Whereas formal systems are defined essentially in syntactic terms.)

    From my perspective, a natural language is primarily a semantic system, and the syntax is mainly there for disambiguation.

    I agree with the first part of the sentence but I think you are underplaying the creative and dynamic nature of natural language syntax here.

    [Quoting me] “I don’t deny that the broader culture is the source of science, but I was thinking more in terms of the way having language allows us to think in certain ways (involving abstraction, symbolic representation and manipulation, etc.).” Fair enough. But I’d have to say that most of my own thinking about mathematics is non-linguistic. That includes thinking about how to prove a theorem. In the final steps, where I have to give precise details, I begin to think more linguistically.

    I think there might be some confusion here. I fully accept that much of our thinking (including mathematical thinking) is not linguistic. (By linguistic I generally mean “relating to natural language”.) There is also an important distinction between intuitive thinking (where the ‘reasoning’ is unconscious) and conscious, explicit reasoning (which may or may not be natural language-based).

    Like

  23. Hi Mark, plenty to think about here.

    I agree that language is necessary (though not sufficient) for science. I’m not sure if I agree that it goes beyond natural language… though I found the arguments you and Dan have had about their relationship compelling.

    Does limiting description/concern about the world to the “extensional” mean one has utterly left natural language behind? Or does that just mean one is dealing with a limited portion of natural language (which taken as a whole can handle more)?

    Perhaps paralleling what another commenter said, for me science is an activity that forces us (or at least scientists) to build their natural language vocabulary. For those looking at a certain phenomena it becomes natural to adopt such terms and concepts.

    As an example, a car mechanic will, given their familiarity with tools and engines, have sets of terms and concepts that would be considered “natural language”, though many uninterested in car repair will never know. I don’t see how that differs in any important way from a scientist that becomes familiar with another set of tools and say biological “engines”.

    Also, differing from you in the essay, I’m not sure if I agree that math and logic are outside natural language, but rather have developed formalized symbols that represent a healthy chunk of natural language. It’s funny but I really improved in math after I realized that all the symbols were shorthand for statements about an action, relationship, or object. You can therefore translate equations back into (what I consider) “natural language.” And if you couldn’t I’m not sure what meaning they would have. This really helped me when I was dealing with quantum mechanics in Physical Chemistry. I think I would have been lost otherwise.

    As far as what science (and math and logic) tackles, I think it is about adding greater precision and finer detail to our accounts/narratives. I agree with your criticism of “causation” since it is not the real or inherent focus of science. Feynman famously skewered the idea that physicists were any better than Mayan (or Aztec?) priests in understanding causation, as opposed to developing useful models to deliver predictions we can work with. In addition, there are a number of fields that are interested in description, detection, or taxonomy and have little concern for mechanisms at all.

    Like

  24. Natural language includes both the intensional and the extensional, but when limited to the extensional, its expressive and descriptive power is tremendously diminished — can’t distinguish between “creature with a heart” and “creature with a kidney” or between “superman” and “Clark Kent.”

    Like

  25. Hi Mark (and others), your final thought experiment about what superior beings that encountered humans would be interested in is of course correct. They may have little interest in our understanding of the universe (or anyway our models of it), versus what we managed to create for ourselves with the universe.

    This experiment reminded me of two short movies (or tv episodes) I watched.

    In one, aliens traveled light years to make contact with humans for one purpose: to reassemble the original cast and crew of a show that got canceled (but they loved) so that they could finally see how it was supposed to end.

    In another piece, some super intelligence, maybe aliens, were threatening to blow up the world/annihilate humans/whatever unless humans could show them something that made them important. They had geniuses and such coming in to show all the knowledge we have and I think even some great artists too. All reduced to tears. Then some magician comes in and shows them a trick, his “magic string” which was basically some cheap novelty magic trick where there was no actual string. That did it.

    These point up one problem with this kind of thought experiment. People seem to take “superior beings” as being impressed by stuff we find important or impressive. It could be they’d be more impressed with an un-ironic Russ Meyer retrospective or Match Box cars set, than the collected works of Shakespeare or a Michelangelo exhibit.

    There is no accounting for taste.

    Liked by 1 person

  26. Hi Dan, that’s right. But then does that mean science, which in most fields is limited to extensional concerns, is not engaged in (or separated from) natural language? Or is it merely engaged in a severely limited form of natural language based on its very limited scope of activity/concern?

    This is why I brought up the comparison between a car mechanic and a scientist.

    Like

  27. [Drafted this before I saw dbholmes’s latest comment: we seem to be making a similar point.]

    Dan wrote:

    Natural language describes the world in intensional terms, while science does so purely extensionally.

    As I suggested in an earlier response, to Dan’s comment, I don’t see it like this. The key point is that people involved in scientific activity use natural language not just to coordinate their activities but as an actual part of their scientific work. In many areas of science, natural language is not enough and it needs to be supplemented by other codes, etc..

    Science is – as I see it – about finding out about (more specifically gaining objective knowledge about, that is knowledge that can be independently checked) how the world is and how it works. You use linguistic descriptions and ordinary reasoning where appropriate.

    In a later comment (responding to dbholmes) Dan writes:

    Natural language includes both the intensional and the extensional, but when limited to the extensional, its expressive and descriptive power is tremendously diminished — can’t distinguish between “creature with a heart” and “creature with a kidney” or between “superman” and “Clark Kent.”

    Again, the implication here (more explicit in the previous quote) is that the scientific use of language is limited to the extensional.

    I am not saying that science has no limits: of course it does. But science is about (certain kinds of) human understanding, and it uses all the resources of natural language (amongst other codes) to achieve its ends.

    Like

  28. Mark,
    Imagine there are superior beings who can observe our universe, whether from within it or from outside it (whatever that might mean) doesn’t matter. My question is, what would interest them most:

    The simple way to answer that question is to ask what we would do if we were that superior species. And the answer is our diversity means that we have a wide range of scientific interests. We would therefore deploy scientific teams from the full range of our scientific endeavour to study the matter. Historians and philosophers of science would study our science. Their psychologists, anthropologists and sociologists would have a field day 🙂

    Of course we are the superior beings in respect to many creatures and already do some of this as appropriate.

    Like

  29. dbholmes

    … [D]iffering from you in the essay, I’m not sure if I agree that math and logic are outside natural language, but rather have developed formalized symbols that represent a healthy chunk of natural language.

    I think you are going too far here. ‘Natural language’ is a term with an accepted meaning. Taking your line would involving changing the meaning of the term, in effect undermining a useful (and entirely harmless) concept.

    It’s funny but I really improved in math after I realized that all the symbols were shorthand for statements about an action, relationship, or object. You can therefore translate equations back into (what I consider) “natural language.”

    As I say, you seem to want to redefine a perfectly good concept. That said, your ideas on learning maths sound like they might be very helpful to others (bearing in mind of course that no single approach suits everybody).

    And if you couldn’t I’m not sure what meaning they would have.

    Logic can (to a point), but I doubt that much mathematics can be translated into natural language – or anything like it. Doesn’t maths have its own (I am tempted to say intrinsic) meaning? It seems to me that practitioners learn a set of concepts and techniques which are sui generis: the concepts only arise in the context of number theory, say, or whatever the area of maths involved and can only really be understood in that (or a related mathematical) context.

    As far as what science (and math and logic) tackles, I think it is about adding greater precision and finer detail to our accounts/narratives.

    I am inclined to think that the phrase “adding greater precision” significantly understates what, say, the mathematical apparatus of physics or any branch of pure mathematics does.

    Like

  30. Labnut

    … Of course we are the superior beings in respect to many creatures and already do some of this as appropriate.

    Only this doesn’t quite work (as a satisfactory response to the original question) because the animals we study don’t have science, broader culture, language, etc..

    Our science makes progress. Will this kind of progress involve *radical* changes in social/moral understanding? This is the sort of question I’m getting at. (And I’m suggesting: *maybe* not.)

    (Dbholmes’s examples (magician, toy cars, etc.) catch the spirit of what I am trying to say – or contemplate.)

    Like

  31. Hi Mark, well I would want to argue that some portions of math (surely basic functions and geometry) can be translated to natural language. I believe these grew out of dealing with quantities and spatial relations in the world.

    But maybe I am pushing this idea of “translation” too much, and (even if possible) what that means for whether math itself lies within the realm of “natural language”.

    So I’ll take back my claim and go with your account which is definitely more straightforward.

    “I am inclined to think that the phrase “adding greater precision” significantly understates what, say, the mathematical apparatus of physics or any branch of pure mathematics does.”

    Not sure what you mean here. Could you explain?

    It might be that I meant more than you thought I did. For example, by adding precision and detail that entails better models and so predictive tools. And depending on what particular field it may mean a broader (zoom out) or more magnified (zoom in) picture of the world.

    Like

  32. labnut

    Let me more directly address your suggestion of what would interest advanced intelligences. You suggest that the best way to answer the question is to imagine ourselves to be those intelligences. But isn’t there a degree of begging the question here?

    “… our diversity means that we have a wide range of scientific interests. We would therefore deploy scientific teams from the full range of our scientific endeavour to study the matter.”

    Your understanding of “the full range of scientific endeavour” is (I think appropriately) broad (includes historians of science). But one of the points I was trying to make in my previous comment was that our science is still a work in progress. Imagine a civilization where the fundamental work was basically complete: in other words they know virtually all there is to know about how the world works, came to be, etc..

    My suggestion is that this would mean that science would no longer be a core endeavour. Sure, they may see our peculiar customs etc. as grist to their historical/anthropological mill, document and classify them and so on. This sort of science never ends. But what gives science its status and importance in our world is that it represents a quest for fundamental understanding.

    If that fundamental understanding is achieved we find ourselves in a very different place. Some of us may be looking at the face of God. Others listening to the music of Bach or Gershwin or to punk rock; or going to the opera; or reading, or dancing or playing old-fashioned or virtual reality games. Or (as dbholmes suggests) adding to our collection of matchbox toys!

    Or perhaps spying and eavesdropping – just for amusement – on the hapless inhabitants of a planet like ours.

    When you think about it, you could see the reading of fiction or the watching of films as an approximation to this. There’s something (healthily) voyeuristic about it, I think.

    Like

  33. dbholmes

    I would want to argue that some portions of math (surely basic functions and geometry) can be translated to natural language. I believe these grew out of dealing with quantities and spatial relations in the world.

    Fair enough.

    But maybe I am pushing this idea of “translation” too much, and (even if possible) what that means for whether math itself lies within the realm of “natural language”.

    It’s the latter point I’m most concerned about. (I’m not denying that ‘translations’ of various sorts can be useful/enlightening.)

    “I am inclined to think that the phrase “adding greater precision” significantly understates what, say, the mathematical apparatus of physics or any branch of pure mathematics does.”

    Not sure what you mean here. Could you explain?

    It might be that I meant more than you thought I did. For example, by adding precision and detail that entails better models and so predictive tools. And depending on what particular field it may mean a broader (zoom out) or more magnified (zoom in) picture of the world.

    This raises all sorts of deep and difficult questions. But I’m tempted just to say that, to the extent that you did mean more than I thought you did (in the way you explain), to that extent you did *not* understate math’s power and importance. 🙂

    Liked by 1 person

  34. Mark,
    You suggest that the best way to answer the question is to imagine ourselves to be those intelligences. But isn’t there a degree of begging the question here?

    First off, this is a fascinating line of thought. The only guideline that we have on the behaviour of another intelligent species is to examine the behaviour of a known intelligent species. Since the only example is us we must employ introspection. Reason should indicate how rational species might behave under certain broad conditions, even if that seems question begging. Otherwise we have to accept that our rationality is fundamentally flawed and then all bets are off.

    Imagine a civilization where the fundamental work was basically complete: in other words they know virtually all there is to know about how the world works, came to be, etc..

    Is it even possible to ‘know all there is to know about how the world works, came to be, etc‘? Is scientific knowledge finite? I find it hard to grant your premise.

    Fundamental scientific knowledge is codified in the laws of nature. Now if you subscribe to the descriptive concept of the laws of nature(rather than the prescriptive) then scientific knowledge can be finite and discoverable for the simple reason that the number of kinds of particles and their interactions is finite(though still very large).

    But if the prescriptive view of the laws of nature holds then the source of the laws of nature is wholly unknown and we can say nothing about their extent or finite nature. All our experience indicates they are invariant in time and space. But this might just be an artefact of our small patch of observable space. We just don’t know, but all of science precedes on the assumption of invariance. If they are not invariant the problem of knowability becomes even worse.

    We have no way of determining whether the descriptive or prescriptive view is the correct one. It seems to me that the philosopher’s metaphysical worldview predetermines his answer. From my point of view the descriptive view is the appropriate approach at higher levels of examination where randomness and complexity confound causal explanation. At lower levels causal explanation is possible because the influence of randomness is minimised(or can be controlled for). And where strict causal relations hold the laws of nature are prescriptive. You might dispute this last statement but would find it very hard to make your case. And, by the way, if you believe in strict determinism, as many of a certain metaphysical predisposition do, then you must accept the prescriptive view of the laws of nature.

    Is the reverse line of reasoning true? If I believe in the prescriptive view of the laws of nature am I required to believe in strict determinism?

    OK, you might rebut my arguments thus: we can only investigate prescriptive laws of nature where they observably manifest themselves by interaction with particles and fields. This is a finite set and thus the observable laws of nature are finite, therefore, in principle, knowable by your very superior intelligent species.

    Like

  35. My lovable, intelligent and devoted dogs lack three fundamental abilities. They have no episodic memories, they cannot imagine a future and they lack language. The first two deficits arguably stem from the lack of language. Now here’s the thing, my dogs are wholly unaware of their deficits and so a vast range of knowledge is hidden from them. They don’t know of their deficits and they do not know of the existence of the hidden knowledge. Nor can they ever know this.

    Now if this is true of my dogs then I must admit the same is possibly true of myself, that fundamentally I am unable to conceive of or perceive certain truths about our universe. Just as my dogs can never discover their deficits so might we never be able to discover our deficits. We might have the cognitive equivalent of colour blindness.

    How might we know that we suffer from cognitive colour blindness? I maintain that there would be indications of this. The laws of nature would have baffling gaps and inconsistencies that we simply could not resolve. Our description of the universe would always be incomplete, inconsistent or contradictory. And this is exactly what advances in science seems to be finding. Certain things do not make sense and they might never make sense if we suffer from cognitive colour blindness. But I digress. A better answer in my next comment.

    Like

  36. Mark,
    “Imagine a civilization where the fundamental work was basically complete: in other words they know virtually all there is to know about how the world works, came to be, etc..”

    I am going to answer your question in another way by appealing to the simulation hypothesis. Imagine that your superior species has constructed a simulated world that we inhabit. Now why would they do that? In a deterministic world their intelligence and computing power allows them to know what all possible outcomes are. They are in effect Laplace’s Demon. If you know the answers why do a simulation? Why even bother to look at one?

    Now let’s imagine also that Laplace’s Demon is also creative and that creativity is its main raison d’etre. That poses a real problem because creativity is pointless when you can, with perfect clarity, know all possible outcomes in all your simulations. There is one way out of this dilemma, if we imagine that Laplace’s Demon can give his simulated creatures free will. He necessarily possesses free will and has found a way to impart free will to his simulated creatures.

    That changes the game completely since he no longer can foresee the outcomes. Now he can watch and enjoy our creative activity. Laplace’s Demon, in this account, is essentially creative, and realises his creative urge through the unforeseeable, free-willing, creative activity of his simulated creatures.

    Now we know why Laplace’s Demon had to create his simulations. They are necessary for the fulfillment of his creative urge.

    You might ask why Laplace’s Demon should be necessarily creative? Because without a creative urge he would be one very lonely Demon in a boring, endlessly unchanging world. What would be the point of that?

    On this account then, the creative activity of our species would be the main source of interest to the superior species(Laplace’s Demon?). And now Mark, you will be pleased to note that I have wholly conceded your point 🙂 But I think I have produced a better rationale 🙂 🙂

    This line of reasoning also constitutes an interesting test of the hypothesis that we are in a simulation. Given a deterministic universe, if we possess free will then we inhabit a simulation whereas if we do not possess free will we do not inhabit a simulation.

    Like

  37. labnut

    I am having trouble understanding your point of view as expressed in these final comments, partly because you bring in philosophical topics like determinism (not a topic on which I have strong or developed opinions).

    Again, the way you talk about the laws of nature puzzles me.

    “… you might rebut my arguments thus: we can only investigate prescriptive laws of nature where they observably manifest themselves by interaction with particles and fields.”

    I don’t see so-called laws of nature “interacting” with anything.

    “Just as my dogs can never discover their deficits so might we never be able to discover our deficits. We might have the cognitive equivalent of colour blindness.”

    I think our being language-using creatures puts us into a different category. Advanced intelligences could learn our language and say at least *something* about what they knew and we didn’t: we can’t do this with dogs.

    Your last comment seems to be trying to update certain traditional theological ideas. Again, much of this is based around traditional philosophical concepts which I don’t really have a lot to say about – other than thinking that the traditional dichotomies may not be all that useful and that employing other conceptual frameworks might be more productive.

    “This line of reasoning also constitutes an interesting test of the hypothesis that we are in a simulation. Given a deterministic universe, if we possess free will then we inhabit a simulation whereas if we do not possess free will we do not inhabit a simulation.”

    This is way too speculative for me! (Also, wouldn’t predictability be a simpler and less problematic concept for you to use in this context than “free will”?)

    Like

  38. Mark,
    I don’t see so-called laws of nature “interacting” with anything.

    Here are some quotes from Sean Carroll. I quote him because he is an articulate, philosophically informed physicist.

    a bunch of atoms and particles, mindlessly obeying the laws of physics,

    electrons and nuclei act according to the laws of physics

    The world consists of things, which obey rules. A simple idea, but not an obvious one, and it carries profound consequences.

    just as positions and momenta obey Hamilton’s equations, the wave function obeys the Schrödinger equation,

    certainly obeys local determinism

    there will still be some underlying equations that are rigidly obeyed

    all that matters is that there are laws. If the atoms and particles that make up human beings obey those laws, there is no free will in this strong sense

    The evolution of states in quantum mechanics works just like it does in classical mechanics; it obeys a deterministic rule–the Schrödinger equation–that allows us to predict the future and past of any specific state with perfect fidelity. And that’s all there is to it.

    If particles and fields obey the laws of physics(as Sean Carroll repeatedly says in his writings) that is certainly a form of interaction. Of course the mechanism of that interaction is completely unknown and constitutes the greatest mystery of all time.

    From dictionary.com

    Interaction: noun
    1.
    reciprocal action, effect, or influence.
    2.
    Physics.
    the direct effect that one kind of particle has on another, in particular, in inducing the emission or absorption of one particle by another.
    the mathematical expression that specifies the nature and strength of this effect.

    You might argue that the interaction is between two particles or between a particle and field. That is quite true(in most cases), but what constitutes this interaction? Why does a particle interact with a field, or another particle? Somehow the one influences the other. All we know is that these interactions behave accordingly to exact, mathematically described laws of nature and that is what I mean by saying that the particles interact with he laws of nature. That is because the laws of nature determine how the interactions are carried out, or as Sean Carroll repeatedly says, the particles and fields obey the laws of nature.

    Like

  39. Mark,
    Your last comment seems to be trying to update certain traditional theological ideas

    Or perhaps I am trying to introduce some new thinking. Labelling a line of thought does not constitute an answer to that line of thought.

    other than thinking that the traditional dichotomies may not be all that useful and that employing other conceptual frameworks might be more productive.

    I encourage that approach but we must do more than just wish for more productive conceptual frameworks. Again you have labelled a line of thought(incorrectly, I think) but that does not constitute an answer.

    Like

  40. Mark,
    “I don’t see so-called laws of nature “interacting” with anything.”

    Ah, I understand your objection. Interaction is reciprocal. Particles and fields cannot influence the laws of nature therefore the effect is not reciprocal and technically not an interaction. The influence is one directional, the laws of nature influencing the particles and fields. So I have used the word ‘interaction’ loosely, but even so, despite your technical objection(which I fully grant) you should understand what I mean.

    Like

  41. labnut

    “I encourage that approach but we must do more than just wish for more productive conceptual frameworks. Again you have labelled a line of thought (incorrectly, I think) but that does not constitute an answer.”

    I am not pretending to be giving full answers to your comments and speculations, just responding in the sort of (usually) piecemeal way one does in comment threads. If I label something, I am doing that to give you an idea of how I am seeing it.

    Essentially what I was saying was that you seem to be working with some concepts which I recognize as traditional philosophical (and philosophico-theological) concepts which I prefer not to work with and so have little to say about. (I did give one positive suggestion: about unpredictability being perhaps a simpler and more appropriate concept than free will to make a particular argument you were making but I wouldn’t want to press this point.)

    You understood my point about interaction. But I also have some doubts about seeing the laws of nature as if they exist independently of the natural processes which exemplify them (I prefer a word like ‘exemplify’ to ‘obey’). I am happy to say they *exist* – but not independently, if that makes sense. Or better: Nature operates in a law-like way.

    Like

  42. Hi Mark,

    As usual I largely agreed with your post. But let me raise a couple of quibbles.

    I think you overstate the difference between mathematical and “natural” language (which I take to mean everyday language). Some mathematical terms are themselves used in everyday language, such as numbers, arithmetic operators and geometric shapes. I think that counting, arithmetic and geometry evolved gradually out of less mathematical-looking ways of speaking. They weren’t a completely new type of language invented ab initio. Of course, mathematicians have enormously extended the language, enabling us to express certain relationships more precisely, and to express relationships that weren’t previously expressible (or imagined) at all. Even in everyday discourse we extend the language with new concepts when we need them. The mathematical language extensions used by scientists are broadly a similar sort of thing.

    (P.S. On reading more comments, I see that dbholmes has made a similar point.)

    In comments, you wrote:

    I agree with you, however, that semantics is basic to natural language – that semantics is built in from the start. (Whereas formal systems are defined essentially in syntactic terms.)

    This is a bit of a hobby horse of mine. I think the terms semantics and syntax are much misused in such contexts. Any language in use has both syntax and semantics, even if they’re not explicitly stated. Roughly speaking, the syntax is the symbolic structure of the language (the ways that the symbols can be put together). The semantics is the meaning of the symbols. If a language is in use then its symbols have meaning.

    When people talk of “formal systems” I’m not sure whether they are including systems like arithmetic, which can be formally defined, but can and have been used without any formal definition. In such systems, the meanings of the symbols is fixed by their use. If subsequently the system is formalised, then the formal rules must correspond to the prior usage. If a a new system is formally stipulated, then the formal definition fixes the use and meaning of the symbols. Roughly-speaking it’s the transformational rules and the axioms that define the semantics of the system.

    If you’re interested in more by me on this subject, please see the appendices of a post I wrote about John Searle’s syntax-and-semantics argument against Strong AI:
    https://barbedsextant.wordpress.com/2015/10/14/searles-argument-from-syntax-and-semantics/

    Like

  43. richardwein

    It seems that you and dbholmes are pushing in one direction and Neil Rickert in another. You quoted my response to Neil in which I agreed that semantics is basic to language. I think we can safely speculate that language began with simple signifiers and fairly primitive syntax. Language learning also starts with learning the meaning (or use-in-context) of its simpler elements. The syntax employed at the early stages of learning is nowhere near the complexity of standard syntax. (The main point of syntax is to serve semantic ends: I agree that syntax and semantics are tied together.)

    I also said that I saw formal systems as being defined essentially in syntactic terms. What I meant is that *what makes them formal* is that they can operate mechanically (without any reference to semantic concepts). But they have, as you are rightly insisting, a dual (syntactic/semantic) aspect.

    As you suggest, ordinary arithmetic (like ordinary language) has meanings built in from the start, and formal arithmetic inherits these meanings. But formal arithmetic can also operate in a purely rule-based way – temporarily blind, as it were, to any interpretation.

    I’ll have a closer look at those notes of yours. (Am not expecting to find anything there with which I seriously disagree.)

    Like