Why don't we still use stone-age tools (or do we)?

While binge-watching a drama set a few hundred years ago in one of the world's empires, I realized that there was something intriguing about the mix of technology in use, which I think was a reasonably accurate rendition of what actually was in use at that time and place. This wasn't a meticulously-researched historical drama, so I'd expect some anachronisms, but the showrunners were constrained by using actual locations and a well-known and well-documented society, so I wouldn't expect to see the equivalent of a Roman senator on an e-bike.

On the one hand, there were ships at sail, intricately woven fabrics, expansive, multistory buildings with carefully crafted architectural details, and a highly-developed administrative state which, among other things, required people across a wide geographical range to carry personal ID. There was an extensive system of roads and detailed maps to show you how to get around them. There were cannon, rifles telescopes, steel farm implements and deadly-sharp swords.

On the other hand, buildings, including heavy-duty structures like jails and military fortifications, were built mainly of wood -- very well built, to be sure, but of wood just the same. Most people lived in small wood-and-stone huts. Some lived in very small dwellings made of sticks. Ladders were made from small logs bound together with rope. The roads were generally traveled on foot or horseback, or in wooden horse-drawn carts. Fighting was mostly hand-to-hand and the main projectile weapon was the bow and arrow.

In other words, while there was quite a bit of technology that would have been state-of-the-art at the time, many if not most of the objects most people dealt with day-to-day could have been made a thousand years before the time of the story, or even five thousand. Consider, for example, a clay cooking pot. While an archeologist could probably tell from the details roughly when and where a particular pot was made, from a purely technical point of view "clay cooking pot" narrows the time and place down to ... "probably not Antarctica" and "probably the last 20,000 years".

Impressive as the more modern-looking items were, few of them seemed essential. An elaborately-embroidered ceremonial robe is not purely decorative. It serves as a signal that the wearer has the resources, including human labor, at their disposal to make such things. Nonetheless, it's far from the only way to keep the wearer warm, and it's actively in the way of the wearer moving around easily (again a feature, not a bug, in context).

Relatively few technological developments are groundbreaking. An aluminum extension ladder is a better general-purpose ladder than a log-and-rope ladder. It's easier to carry. It's probably more weather-resistant and durable though, not knowing any better, I wouldn't want to underestimate how well a well-made wooden ladder would hold up. Likewise, it can probably bear more weight. Because it's easy to change the length of an extension ladder, it can be used in a wider variety of places.

Nonetheless, an aluminum extension ladder is still a ladder, and ladders have probably been around about as long as cooking pots. It's hard to tell for sure since a non-metal ladder is less likely to survive the millennia than a ceramic pot, but there is at least one surviving depiction of a ladder from 10,000 years ago. On a clay pot, of course.

An aluminum extension ladder is also a lot harder to make. While aluminum compounds, particularly alums, have been known for millennia, actually extracting aluminum requires electricity, and quite a bit of it. Aluminum has only been available in commercial quantities for a little more than a century. While we may think of aluminum foil and beverage cans, not so long ago aluminum was rare and expensive. It's probably not true that Napoleon III had a special set of aluminum silverware, but the aluminium statue of Anteros in Piccadilly circus, made in 1893 was the first of its kind, and so kind of a big deal.

Now that we can make aluminum ladders cheaply, they're everywhere. I have a couple in my garage. If someone gave me a carefully crafted wooden ladder, I'd keep it, because cool, and maybe even use it if it happened to come in handy. Like most of us, though, I have absolutely no reason to go out of my way to get one, particularly since it would almost certainly be much more expensive.

As far as technological developments go, I think aluminum ladders are more the rule than the exception. A modern freeway is in most respects a better road than the Via Appia, though not necessarily in durability. A modern car can go much farther and faster than a horse-drawn cart. A coat with a zipper is easier to get in and out of than one with buttons. It's somewhat easier to pay with a card or phone than it is to count coins and notes.

In general, a new development has to offer at least some advantage in order to catch on. That advantage doesn't have to be groundbreaking or even particularly great, however, and it doesn't even have to make sense on a larger scale. Plastic bags have their uses, but whether a new and cheaper way to make plastic bags catches on has little to do with whether we need more plastic bags, as long as we want some plastic bags.

Coming back to the title, in some sense we do still use stone-age tools every day. Ladders are still a thing. You can still buy rope at a hardware store and put it to the same kinds of uses it's been put to for as long as there's been rope. But in most cases, you'll be buying a stone-age tool made with modern technology. The rope you buy at a hardware store was almost certainly made by machine and very likely from a material that didn't exist a hundred years ago, much less several thousand.

I do keep thinking of getting a clay tagine pot, though.

What good is half a language?

True Wit is Nature to advantage dress'd
What oft was thought, but ne'er so well express'd
-- Alexander Pope

How did humans come to have language?

There is, to put it mildly, a lot we don't know about this.  Apart from the traditional explanations from various cultures, which are interesting in their own right, academic fields including evolutionary biology, cognitive science and linguistics have had various things to say about the question, so why shouldn't random bloggers?

In what follows, please remember that the title of this blog is Intermittent Conjecture.  I'm not an expert in any of those three fields, though I've had an amateur interest in all three for years and years.  Real research requires careful gathering of evidence and checking of sources, detailed knowledge of the existing literature, extensive review and in general lots of time and effort.  I can confidently state that none of those went into this post, and anything in here should be weighed accordingly.  Also, I'm not claiming any original insight.  Most likely, all the points here have already been made, and better made, by someone else already.

With that said ...

In order to talk about how humans came to have language, the first question to address is what does it mean to have language at all.  Language is so pervasive in human existence that it's surprisingly hard to step back and come up with an objective definition that captures the important features of language and doesn't directly or indirectly amount to "It's that thing people do when they talk (or sign, or write, or ...) in order to communicate information."

We want to delimit, at least roughly, something that includes all the ways we use language, but excludes other activities, including things that we sometimes call "language", but that we somehow know aren't "really" language, say body language, the language of flowers or, ideally, even computer languages, which deliberately share a number of features with human natural languages.

Since language is often considered something unique to humans, or even something that makes us human, it might be tempting to actively try to exclude various ways that other animals communicate, but it seems better to me just to try to pin down what we mean by human language and let the chips fall where they may when it comes to other species.

For me, some of the interesting features of language are

• It can communicate complex, arbitrary structures from one mind to another, however imperfectly.
• It is robust in the face of noise and imperfection (think of shouting in a loud music venue or talking with someone struggling with a second language).
• It tolerates ambiguity, meaning that (unlike in computer languages and other formal systems) ambiguity doesn't bring a conversation to a halt.  In some cases it's even a useful feature.
• Any given language provides multiple ways to express the same basic facts, each with its own particular connotations and emphasis.
• Different languages often express the same basic facts in very different ways.
• Related to these, language is fluid across time and populations.  Usage changes over time and varies across populations.
• It can be communicated by a variety of means, notably speech, signing and writing.
• From an evolutionary point of view, it has survival value.

I'd call these functional properties, meaning that they relate mainly to what language does without saying anything concrete about how it does it.  Structurally (from here on I'll tend to focus on spoken/written language, with the understanding that it's not the whole story),

• Language is linear.

That is, whatever the medium, words are produced and received one at a time, though there can be a number of "side channels" such as pitch and emphasis, facial expressions and hand gestures.

• The mapping between a word and its meaning is largely arbitrary (though you can generally trace a pretty elaborate history involving similar words with similar meanings).
• Vocabulary is extensible.

We can coin words for new concepts.  This is true only for certain kinds of words, but where it can happen it happens easily.

• Meaning is also extensible

We can apply existing words in new senses and again this happens easily.

• The forms used adjust to social conditions.

You speak differently with your peers after work than you would to your boss at work, or to your parents as a child, or to your prospective in-laws, and so forth

• The forms used adjust to particular needs of the conversation, for example which details you want to emphasize (or obscure).
• Some concepts seem to more tightly coupled to the structure of a particular language than others.

For example, when something happened or will happen in relation to when it is spoken of is generally part of the grammar, or marked by a small, closed set of words, or both.

• On the other hand, there is wide variety in exactly how such things are expressed.

Different languages emphasize different distinctions.  For example, some languages don't specially mark singular/plural, or past/present, though of course they can still express that there was more than one of something or that something happened yesterday rather than today.  Different languages use different devices to convey basic information like when something happened or what belongs to whom.

• Syntax, in the form of word order and inflection (changing the forms of words, as with changing dog to dogs or bark to barked or barking), collectively seem to matter in all languages, but the exact way in which they matter, and the degree to which each matters, seem to be unique to any given language.  Even closely related languages generally differ in the exact details.

There are plenty of other features that could each merit a separate post, such as honorifics (Mr. Hull) and diminutives (Davey), or how accent and vocabulary are such devastatingly effective in-group markers, or how metaphors work, or what determines when and how we choose to move words around to focus on a topic, or why some languages build up long words that equate to whole sentences of short words in other languages, or why in some languages directional words like to and of take on grammatical meaning, or why different languages break down verb tenses in different ways, or can use different words for numbers depending on what's being counted, and so on and so on ...

Many of these features of language have to do with the interplay between cognition -- how we think -- and language -- how we express thoughts.  The development of cognition must have been both a driver and a limiting factor in the development of language, but we are almost certainly still in the very early stages of understanding this relationship.

For example, languages generally seem to have a way of nesting one clause inside another, as in The fence that went around the house that was blue was red.  How would this arise?  In order to understand such a sentence, we need some way of setting aside The fence while we deal with that went around the house that was blue and then connecting was red back to the fence in order to understand that the fence is red and the house is blue.  To a compugeek, this means something like a stack, a data structure for storing and retrieving things such that the last thing stored is the first thing retrieved.

Cognitively, handling such a sentence is like veering off a path on some side trip and returning to pick up where you left off, or setting aside a task to handle some interruption and then returning to the original task.  Neither of these abilities is anywhere near unique to humans, so they must older than humanity, even though we are the only animals that we know of that seem to use them in communication.

These cognitive abilities are also completely separate from a large number of individual adaptations of our vocal apparatus, which do seem to be unique to us, notably fine control of breathing and of the position of the tongue and shape of the mouth.  While these adaptations are essential to our being able to speak as fluently as we do, they don't have anything to do with what kinds of sentences we can express, just how well we can do so using spoken words.  Sign languages get along perfectly well without them.

In other words, it's quite possible we were able to conceive of structures like "I saw that the lion that killed the wildebeest went around behind that hill over there" without being able to put them into words, and that ability only came along later.  There's certainly no shortage, even in modern humans, of things that are easy to think but hard to express (I'd give a few examples, but ...).  The question here, then, is not "How did we develop the ability to think in nested clauses?" but "How did we come to use the grammatical structures we now see in languages to communicate such thoughts?"

There's a lot to evolution, and it has to be right up there with quantum mechanics as far as scientific theories that are easy to oversimplify, draw unwarranted conclusions from, or to get outright wrong, so this next bit is even less precise than what I've already said.  For example, I'm completely skirting around major issues of population genetics -- how a gene (useful or not) spreads (or doesn't) in a population.

Let's try to consider vocabulary in an evolutionary context.  I pick vocabulary to start with because it's clearly distinct from grammar.  Indeed one of the useful features of a grammar is that you can plug an arbitrary set of words into it.  Conversely, one requirement for developing language as we know it is the ability to learn and use a large and expandable vocabulary.  Without that, and regardless of the grammatical apparatus, we do not account for the way people actually use language.

Suppose some animal has the ability to make one kind of call when a it spots particular predator and a different call for another predator, in such a way that is conspecifics (animals of the same species) can understand and react appropriately.  That's two calls (three if you count not making any call) and it's easy to see how that could be useful in not getting eaten.  Again, this is far from unique to us (see here, and search for "vervets" in the post, for example).

Now suppose some particular animal is born with the ability to make a third call for some other hazard, say a large branch falling (this is more than a bit contrived, but bear with me).  A large branch falls, the animal cries out ... and no one does anything.  The ability to make new calls isn't particularly useful without the ability to understand new calls.  But suppose that nobody did anything because they didn't know what the new call meant, but they were able to connect "that oddball over there made a funny noise" with "a big branch fell".  The next time a big branch falls and our three-call-making friend cries out, everyone looks out and scatters to safety.  Progress.

I'm more than a bit skeptical that the ability to make three calls rather than two would arise by a lucky mutation, but I think there are still two valid points here:

First, the ability to comprehend probably runs ahead of the ability to express, and certainly new ways to express are much less likely to catch on if no one understands what they mean.  Moreover, comprehension is useful in and of itself.  Whether or not my species is able to make calls that signal specific scenarios, being able to understand other species' calls is very useful (when a vervet makes a predator call, other species will take appropriate action as well), as is the ability to match up new calls with their meanings from context and examples.

In other words, the ability to understand a large vocabulary is liable to develop even without the ability to express a large vocabulary.  For a real-life example, at least some domestic dogs can understand many more human words than (as far as anyone can tell) they can produce distinct barks and similar sounds, and certainly more human words than they can themselves produce.

Second, this appears to be a very common pattern in evolution.  Abilities that are useful in one context (distinguishing the different calls of animals around you) become useful in other contexts (developing a system of specialized calls within your own species).  The general pattern is known as exaptation (or cooption, or formerly and more confusingly as pre-adaptation).

Let's suppose that the local population of some species can potentially understand, say, dozens of distinct calls (whether their own or those of other species), but its ability to produce distinct calls is limited.  If some individual comes along with the gift of being able to produce more distinct calls, then that will probably increase that individual's chances of surviving -- because its conspecifics will learn the new calls and so increase everyone's chance of survival -- and at least potentially its chances of reproducing, if only because there will be more potential mates around if fewer of them get eaten. 

If that particular individual fails to survive and reproduce, the conditions are still good for some other individual to come along with the ability to produce a bigger vocabulary, perhaps through some entirely different mechanism.  This in turn doesn't preclude some future individual from being born with the ability to produce a larger vocabulary through yet a third mechanism, or either of the original two.  If there are multiple mechanisms for doing something advantageous, the chances of it taking hold in the long run are better (I'm pretty sure, but I don't know if an actual biologist would agree.  Also, this isn't particular to vocabulary.).

If the community as a whole develops the tendency to find larger vocabularies attractive, so much the better, though the math starts to get hairy at this point.  Sexual selection is a pretty good way of driving traits to extremes -- think peacocks and male walruses -- so it's quite plausible that a species that starts to develop larger and larger vocabularies of calls could take this quite far, past the point of immediate usefulness.  You then have a population with a large vocabulary ready for an environment where it makes more of a difference.

In short, even some ability to produce distinct calls for different situations is useful, and it's no surprise many animals have it.  The ability to produce a large and expandable variety of distinct calls for different situations also looks useful, but also seems harder to evolve, considering that it's fairly rare.  Taking this a step further, we appear to be unique in our ability to produce and distinguish thousands of distinct vocabulary items, though as always there's quite a bit we still don't know about communication in other species.

It's clear that other animals can distinguish, and in some cases produce, non-trivial vocabularies, even if it's not particularly common.  How do you get from there to our as-far-as-we-know-unique abilities?  The usual answer for how complex traits evolve is "a piece at a time".

In order to find a (very hypothetical) evolutionary pathway from an extensible collection of specialized calls to what we call language today, we want to find a series of small steps that each add something useful to what's already there without requiring major restructuring.  Some of those, in no strict order except where logically necessary, might be:

• The ability to refer to a class of things without reference to a particular instance

This is one aspect of what one might call "abstract concepts".  As such, it doesn't require any new linguistic machinery beyond the ability to make and distinguish a reasonably large set of calls (which I'll call words from here on out), but it does require a cognitive shift.  The speaker has to be able to think of, say, wolf without referring to a particular wolf trying to sneak up.  The listener has to realize that someone saying "wolf" may not be referring to a wolf currently sneaking up on them. Instead, if the speaker is pointing to a set of tracks it might mean "a wolf went here", or if pointing in a particular direction, maybe "wolves come from over there".

This may seem completely natural to us, but it's not clear who, if anyone else besides us, can do this.   Lots of animals can distinguish different types of things, but being able to classify is different from being aware that classes exist.  An apple-sorting machine can sort big from small without understanding "big" or "small".  I say "it's not clear" because devising an experiment to tell if something does or doesn't understand some aspect of abstraction is difficult, in no small part because there's a lot of room for interpretation of the results.

[Re-reading this and taking Earl's comment into account, I think I've conflated two different kinds of abstraction.  Classifying "wolf" as opposed to "a wolf" is probably more basic than I have it here.  For example, a vervet will give the call for a particular kind of predator.  It doesn't have to develop a call for each individual animal, and a good thing, because that would only work for individuals that it had seen before.  A behavioralist would probably say this is all stimulus-response based on particular characteristics -- give the leopard call in response to the smell or sound of a leopard, and so on, and fair enough.  

Classification, then, would be more a matter of connecting particular attributes -- sound, smell, shape or whatever -- to physical things with those attributes.  The progression, as Earl suggests, would be leopard smell/sound --> leopard --> the leopard that just disappeared into the trees over there.  That is, it requires more cognitive machinery to be able to conceive of leopard as any animal that smells this way and/or makes this sound and/or has pointy ears and a tail or whatever, and it requires a different piece of machinery  -- a sort of object permanence -- to conceive of an absent leopard and connect that thing that was here but isn't any more to leopard and get that leopard that was here but isn't any more -- D.H 28 Oct. 2021]

• The ability to designate a quality such as "big" or "red" without reference to any particular thing with that quality.

This is similar to the previous item, but for adjectives rather than nouns.  From a language standpoint it's important because it implies that you can mix and match qualities and things (adjectives and nouns).  A tree can be big, a wolf can be big and a wolf can be gray without needing a separate notion of "big tree", "big wolf" and "gray wolf".  An adjective is a predicate that applies to something rather than standing alone as a noun does.

As I understand it, the widely-recognized stages of language development in humans are babbling, single words, two-word sentences and "all hell breaks loose".  A brain that can handle nouns and predicates is ready for two-word sentences consisting of a predicate and something it applies to.  This is a very significant step in communication and it appears to be quite rare, but linguistically it's nearly trivial.  A grammar to describe it has one rule and no recursion (rules that refer, directly or indirectly, to themselves).

As a practical matter, producing a two-word sentence means signifying a predicate and an object that it applies to (called an argument).  Understanding it means understanding the predicate, understanding the argument and, crucially, understanding that the predicate applies to the argument.  If you can distinguish predicates from objects, order doesn't even matter.  "Big wolf!" is just as good as "Wolf big!" or even a panicked sequence of "Wolf wolf big wolf big big wolf!" (which, to be fair, would require recursion to describe in a phrase-structure grammar).

From a functional point of view, the limiting factor to communicating such concepts is not grammar but the ability to form and understand the concepts in the first place.

Where do we go from predicate/argument sentences to something resembling what we now call language?  Some possible next steps might be

• Predicates with more than one argument.

The important part here is that you need a way to distinguish the arguments.  In wolf big, you know that big is the predicate and wolf is the argument and that's all you need, but in see rabbit wolf, where see is the predicate and rabbit and wolf are arguments, how do we tell if the wolf sees the rabbit or the rabbit sees the wolf?  There are two solutions, given that you're limited to putting words together in some particular order

Either the order of words matters, so see rabbit wolf means one thing and see wolf rabbit means the other, or there's a way of marking words according to what role they play, so for example see wolf-at rabbit means the rabbit sees the wolf and see wolf rabbit-at means the wolf sees the rabbit.  There are lots of possible variations, and the two approaches can be combined.  Actual languages do both, in a wide variety of ways.

From a linguistic point of view, word order and inflection (ways of marking words) are the elements of syntax, which (roughly speaking) provides structure on top of a raw stream of words.  Languages apply syntax in a number of ways, allowing us to put together complex sentences such as this one, but you need the same basic tools even for simple three-word sentences.  Turning that around, if you can solve the problem of distinguishing the meaning of a predicate and two arguments, you have a significant portion of the machinery needed for more complex sentences.

• Pronouns, that is, a way to designate a placeholder for something without saying exactly what that something is, and connect it with a specific meaning separately.

Cognitively, pronouns imply some form of memory beyond the scope of a simple sentence. Linguistically, their key property is that their meaning can be redefined on the fly.  A noun like wolf might refer to different specific wolves at different times, but it will always refer to some wolf.  A pronoun like it is much less restrained.  It could refer to any noun, depending on context.

Pronouns allow for more compact sentences, which is useful in itself since you don't have to repeat some long descriptive phrase every time you want to say something new about, say, the big red house across the street with the oak tree in the yard.  You can just say that house or just it if the context is clear enough.

More than this, though, by equating two things in separate sentences they allow linear sequences of words to describe non-linear structures, for example I see a wolf and it sees me.  By contrast, in I see a wolf and a wolf sees me it's not clear whether it's the same wolf and we don't necessarily have the circular structure of two things seeing each other.

• The ability to stack up arbitrarily many predicates: big dog, big red dog, big red hairy dog, etc.

I left this for last because it leads into a bit of a rabbit hole concerning the role of nesting and recursion in language.  I'm going to dig into that a bit here by way of arguing that some of the analytic tools commonly used in analyzing language may not be particularly relevant to its development.  Put another way, "how did language develop" is not the same question as "how did the structures we work with in analyzing language develop".

A common analysis of phrases like big red hairy dog uses a recursive set of rules like

a noun phrase can be a noun by itself, or
a noun phrase can be an adjective followed by a noun phrase

This is much simpler than a full definition of noun phrase in a real grammar, and it's not the only way to analyze noun phrases, but it shows the recursive pattern that's often used in such an analysis.  The second definition of noun phrase refers to noun phrase recursively.  The noun phrase on the right-hand side will be smaller, since it has one less adjective, so there's no infinite regress.  The example, big red hairy dog, breaks down to big modifying red hairy dog, which breaks down to red modifying hairy dog, which breaks down to hairy modifying dog, and dog is a noun phrase by itself.  In all there are four noun phrases, one by the first rule and three by the second.

On the other hand, if you can conceive of a dog being big, red and hairy at the same time, you can just as well express this with two-word sentences and a pronoun:  dog big. it red. it hairy.  The same construction could even make sense without the pronouns: dog big. red. hairy.  Here a listener might naturally assume that red and hairy have to apply to something, and the last thing we were talking about was a dog, so the dog must be red and hairy as well as big.

This is not particularly different from someone saying I saw the movie about the duck.  Didn't like it, where the second sentence clearly means I didn't like it and you could even just say Didn't like and still be clearly understood, even if Didn't like by itself sounds a bit odd.

From a grammatical standpoint (at least for a constituency grammar) these all seem quite different.  In big red hairy dog, there's presumed to be a nested structure of noun phrases.  In dog big.  it red. it hairy you have three sentences with a simple noun-verb structure and in dog big. red. hairy. you have one two-word sentence and two fragments that aren't even sentences.

However, from the point of view of "I have some notion of predicates and arguments, and multiple predicates can apply to the same argument, now how do I put that in words?", they seem pretty similar.  In all three cases you say the argument and the predicates that apply to it and the listener understands that the predicates apply to the argument because that's what predicates do.


I started this post with the idea of exploring how language as we now know it could develop from simpler pieces such as those we can see in other animals.  The title is a nod to the question of What good is half an eye? regarding the evolution of complex eyes such as we see in several lineages, including our own and (in a different form) in cephalopods.  In that case, it turns out that there are several intermediate forms which provide an advantage even though they're not what we would call fully-formed eyes, and it's not hard to trace a plausible pathway from basic light-sensitive "eye spots" to what we and many other animals have.

The case of language seems similar.  I think the key points are

• Cognition is crucial.  You can't express what you can't conceive of.
• The ability to understand almost certainly runs ahead of the ability to express.
• There are plausibly a number intermediate stages between simple calls and complex language (again, the account above is completely speculative and I don't claim to have identified the actual steps precisely or completely).
• Full grammar, in the sense of nested structures described by recursive rules, may not be a particularly crucial step.
• A purely grammatical analysis may even obscure the picture, both by failing to make distinctions (as with the jump from "wolf" meaning "this wolf right there" to it meaning "wolf" in the abstract) and by drawing distinctions that aren't particularly relevant (as with the various forms of big red hairy dog).

Do androids trip on electric acid?

Have a look at some of these images and take note of whatever adjectives come to mind.  If other people's responses are anything to go by, there's a good chance they include some or all of "surreal", "disturbing", "dreamlike", "nightmarish" or "trippy".  Particularly "trippy".

These aren't the first computer-generated images to inspire such descriptions.  Notably, fractal images have been described in psychedelic terms at least since the Mandelbrot set came to general attention, and the newer, three-dimensional varieties seem particularly evocative.  The neural-network generated images, however, are in a different league.  What's going on?

Real neural systems appear to be rife with feedback loops.  In experiments with in vitro neuron cultures -- nerve cells growing in dishes with electrodes attached here and there -- a system with no other input to deal with will amplify and filter whatever random noise there is (and there is always something) into a clear signal.  This would be a "signal" in the information theory sense of something that's highly unlikely to occur by chance, not a signal in the sense of something conveying a particular meaning.

This distinction between the two senses of "signal" is important.  Typically a signal in the information theory sense is also meaningful in some way.  That's more or less why they call it "information theory".  There are plenty of counterexamples, though.  For example:

• tinnitus (ringing of the ears), where the auditory system fills in a frequency that the ear itself isn't able to produce
• pareidolia, where one sees images of objects in random patterns, such as faces in clouds
• the gambler's fallacy, where one expects a random process to remember what it has already done and compensate for it ("I've lost the last three hands.  I'm bound to get good cards now.")

and so forth.  The common thread is that part of the brain is expecting to perceive something -- a sound, a face, a balanced pattern of "good" and "bad" outcomes -- and selectively processes otherwise meaningless input to produce that perception.

In the generated images, a neural network is first trained to recognize a particular kind of image -- buildings, eyes, trees, whatever -- and the input image is adjusted bit by bit to strengthen the signal to the recognizer.  The code doing the adjustment knows nothing about what the recognizer expects.  It just tries something, and if the recognizer gives it a stronger signal as a result, it keeps the adjustment.  If you start with random noise, you end up with the kind of images you were looking for.  If you start with non-random input, you get a weird mashup of what you had and what you were looking for.

Our brains almost certainly have this sort of feedback loop built in.  Real input often provides noisy and ambiguous signals.  Is that a predator behind those bushes, or just a fallen branch?  Up to a point it's safer to provide a false positive ("predator" when it's really a branch) than a false negative ("branch", when it's really a predator), so if a predator-recognizer feeds "yeah, that might be a four-legged furry thing with teeth" back to the visual system in order to strengthen the signal, survival chances should be better than with a brain that doesn't do that.  A difference in survival chances is exactly what natural selection needs to do its work.

At some point, though, too many false positives mean wasting energy, and probably courting other dangers, by jumping at every shadow.  Where that point is will vary depending on all sorts of things.  In practice, there will be a sliding scale from "too complacent" to "too paranoid", with no preset "right" amount of caution.  Given that chemistry is a vital part of the nervous system's operation, it's not surprising that various chemicals could move such settings.  If the change is in a useful direction, we call such chemicals "medicine".  Otherwise we call them "drugs".

In other words -- and I'm no expert here -- it seems plausible that we call the images trippy because they are trippy, in the sense that the neural networks that produced them are hallucinating in a manner similar to an actual brain hallucinating.  Clearly, there's more going on than that, but this is an interesting result.


When testing software, it's important to look at more than just the "happy" path.  If you're testing code that divides numbers, you should see what it does when you ask it to divide by zero.  If you're testing code that handles addresses and phone numbers, you should see what it does when you give it something that's not a phone number.  Maybe you should feed it some random gibberish (noting exactly what that gibberish was, for future reference), and see what happens.

Testing models of perception (or of anything else), seems similar.  It's nice if your neural network for recognizing trees can say that a picture of a tree is a picture of a tree.  It's good, maybe good enough for the task at hand, if it's also good at not calling telephone poles or corn stalks trees.  But if you're not just trying to recognize pictures, and you're actually trying to model how brains work in general, it's very interesting if your model shows the same kind of failure modes as an actual brain.  A neural network that can hallucinate convincingly might just be on to something.

The invisible oceans

Life as we know it needs water.  Two conclusions that don't follow from that premise: Life needs water, and where there is water, there must be life.  Nonetheless, in our present ignorance, the best bet for finding other life is to find water.  Liquid water, that is.

Over the past few years there has been a steady stream of discoveries of likely liquid water in our solar system.  Jupiter's moons Ganymede and Europa probably each have more than Earth does, even though both are considerably smaller than Earth.  Enceladus (a small moon of Saturn) probably contains significant amounts.  The planet asteroid dwarf planet Ceres also shows possible signs of an ocean.  Pluto and Charon might also.  We should know more about them in a few months as I write this.  In the case of Pluto and the moons, tidal flexing generates the heat that keeps the water liquid.  The case of Ceres is less clear, so from here on I'll restrict discussion to the moons.

What these masses of water all have in common, of course, is that they lie deep beneath the surface.  Tens of kilometers, at least.  Some, at least are also salty enough to conduct electricity well, as evidenced by their interaction with magnetic fields.  It's not clear what kind of pressure they are under, or what their range of temperatures is, except that they are likely sandwiched between layers of ice, or maybe between ice and rock, or maybe in alternating layers of various forms of ice (yes, there is more than one kind).  They will thus be literally ice cold in at least some regions, though they are probably considerably warmer in other places.  They certainly receive no sunlight.

Could life evolve under such conditions?  Who knows?  Personally I'd be reluctant to rule it out.  We've found life in all kinds of unlikely habitats on Earth, and in any case we just don't know that much about how life develops.  So suppose it did develop on one of these moons.  What would it be like, and what would be our chances of encountering it?


The first question is whether there is enough in these oceans to get microbial life going.  The short answer: no idea.  We don't know very much at all about the chemical composition of these oceans, though we can make some informed guesses, and again we don't know very much at all about how microbial life on Earth developed, though we can make some informed guesses about that, too.

It's pretty clear that there will be various impurities in the ocean water, so there might well be potential for some sort of self-replicating, information-carrying polymer, similar to RNA, to develop.  There is an outside source of heat from tidal flexing, and there will be temperature gradients as a result [perhaps as much as 40K over a few hundred kilometers], so the laws of thermodynamics don't rule anything out.  Let's assume that microbial life of some sort can develop.

The path from single-cell organisms to colonies of single-cell organisms to colonies of single-cell organisms with different forms (but the same genetics) to something we may as well call a multicellular organism is reasonably clear, although at least in the case of life on Earth it seems to have taken a good long time to develop.

Since we're totally speculating, let's assume there are multicellular organisms analogous to our own sea creatures, but possibly very different in form.  Again, in reality there might be nothing at all, or only single-celled organisms, or some sort of microbe without clearly defined cells at all, or who knows what.

What kinds of features might these critters have?

• Some sort of chemical sense analogous to smell or taste seems like a good bet.  It's almost a defining property of life that it will respond to chemical stimuli in some way or another.  Microbes do.  Animals of all sizes do.  Even plants and fungi do.  I wouldn't necessarily call a slime mold constituent detecting another's chemical signal "taste" or "smell" but ... some sort of chemical sense.
• For similar reasons, a temperature sense seems likely.
• Food webs, predator/prey relationships, mutualism, parasitism, commensalism, amensalism and any number of other relationships among organisms are pretty much inevitable when there is more than one kind of organism.
• Some way of shuffling genetic material as with sexual reproduction.  A source of variability beyond random mutation can be useful in surviving long-term in harsh, ever-changing environments.  Even microbes do this to at least some extent.
• Various ways of physically manipulating objects ... tentacles, pincers, pseudopods, maybe even something resembling hands
• Some sort of nervous system for communicating signals, including sensations from the senses, from one part of the body to another
• Sociality in at least some species.  We're assuming there are multicellular organisms, which is not so different from sociality at the cell level.  This is the next level: sociality among multicellular organisms.  Sociality requires some means of communication between organisms.  Chemical signals and sound seem like plausible candidates.

So far we have a world comprising organisms of various sizes and forms, some herding/schooling together, some hunting others, with the ability to grab things and move them around ... much like life as we know it, except quite likely totally different.

But then, we assumed many important premises based on experience with life as we know it, particularly cells, and we assumed that evolution would follow essentially the same rules as here.  The second assumption, at least, seems pretty reasonable, but who knows?

It's natural to think such an ocean would be dark and all its inhabitants blind, but maybe not.  Plenty of deep-sea creatures find it useful to have eyes and even to produce light.  Light-producing molecules are not that complex.  There are ones that require only carbon, hydrogen and oxygen, so we don't need to assume phosphorous or sulfur, just some source of carbon in the water, which is probably inevitable.

The evolutionary pathway to light and sight is not so clear, though.  On Earth there is a source of light independent of life and it's not surprising that eyes would evolve in the sunlit portion of the ocean, at least.  In naturally pitch-dark oceans, there would have to be some source of light, as a side-effect of something else, before light-detecting cells and organs become useful.

So let's assume it's dark.  Being visual animals, we might assume that that's a big deal, but maybe not.  There are plenty of other ways to get a good picture of the world without seeing it.  In particular, hearing will be important.

Sound carries well in water, and given that the whole world is constantly flexing, there ought to be at least some natural sources of sound.  Unlike the case of vision, it's almost inevitable that organisms that are reasonably large and able to move and to move things will end up making some noise doing so.  This all suggests it would useful to be able to hear.  If it's useful to hear, it becomes useful to be able to make sounds, both for communication and possibly for active sonar.

So we have a dark, noisy and probably fairly cold place teeming with organisms of various shapes and sizes, and schools of this and that all trying to eat and not be eaten.  Will we ever see it?  All we'd have to do to see it is fly a probe a billion kilometers or so and have it dig through kilometers of ice.  Keeping in mind that at the surface temperatures we're looking at, ice acts more like any other mineral than the softish stuff you can chew (to the horror of your dentist).

Since we're speculating, let's say that something we would recognize as an intelligent species develops.  Are they likely to come visit?

Humanity went into space driven by (among other things) the curiosity inspired by the sun, moon, stars and planets.  Even if you don't buy that, you have to admit that we could at least see the sun, moon, stars and planets without any technological help.  Our hypothetical ocean-dwellers could live indefinitely with no clue that there was a solar system out there, or anything else at all.  They would have no direct way of knowing that the parent planet, or even the surface of their own moon, existed.

Even the existence of gravity might be the subject of intense debate.  These moons are relatively small.  The surface gravity of Ganymede, for example, is about 0.1g.  If you're considerably beneath the surface the influence of gravity decreases since you also have matter above you, but that's probably not a major factor at the scales we're dealing with.  If our own oceans are any clue, most things will be near neutral buoyancy.

Put that together and you have a general tendency for most living things to float around in an ocean layer dozens of kilometers deep (thick?) and thousands of kilometers around.  Some inanimate things would tend to drift, very slowly, toward the inner surface (that is, sink).  Others would tend to drift, very slowly, toward the outer surface (that is, float).  This would not be nearly so easy to sort out as the general tendency of things to fall quickly to the ground on Earth.

Figuring out that the world is round would be a significant accomplishment.  The major cues the Greeks used -- ships sinking below the horizon, lunar eclipses, the position of the noontime sun at different latitudes -- would not be available.  The most obvious route left is to actually circumnavigate the world.  And figure out that you did it.

I'm very reluctant to say "such and such would be impossible because ...", but I think it's safe to say that a number of things we take for granted living on the solid surface of a planet with an atmosphere transparent in many wavelengths would be a lot harder in these worlds.  On the other hand, if an inhabitant ever did make it to space, they'd probably have a much better intuitive feel for it than we do.

There's one major factor I haven't mentioned here, that's been cited as a reason that no underwater species could ever develop technology.  Fire doesn't work.  Without fire, a host of things become much harder, if not impossible, including metal extraction and heat engines (steam, internal combustion, etc.).  Underwater rocketry also seems like a stretch, though jet propulsion is not a problem (the difference is that jet propulsion uses the medium one is traveling through, while a rocket creates its own exhaust).

From an earthbound perspective, knowing about our technologies, it's easy to say what familiar technologies probably wouldn't work in such a world, at least not the way they work here.  It's harder, though, to say what unfamiliar technologies could work.  This makes it tempting to say things like "There's no way that life in Ganymede's oceans could contact us.  They wouldn't even have fire."

But we've had, depending on how you count, at least thousands of years to figure technology out.  Depending on how things develop, particularly the transition from single-celled to multi-celled organisms, our hypothetical counterparts might have had tens of thousands, or millions of years.  Or no time at all since they haven't gotten far enough yet.

So what's at least possible in such a world?  Here are some wild, sketchily-informed guesses:

Mathematics: "Can mathematics develop?" seems very much the same as "Can intelligent life develop?", if only as a matter of definition.  If it can't handle at least some form of mathematical thought, can we really call it intelligent? Let's assume that these sea creatures have learned not only to count, but to "think abstractly", whatever that means.  While we're at it, let's assume something we would recognize as a language.

Writing: It's unlikely that anyone is going to whip out a ballpoint pen and write on the back of a napkin, so we need to define "writing" a bit more abstractly.  What we're really after is a permanent means of recording language in via discrete symbols, that is, in digital form.  This doesn't require much, for example, some sort of solid material that can be manipulated and will hold its form.  For example, it doesn't seem unlikely that there could be something resembling rope, which would enable something like the quipu.

Physics: If our critters have mathematical ability and curiosity, they will begin to notice and codify basic facts about their physical world.  Fluid dynamics would be an obvious subject of study, along with some aspects of thermodynamics, particularly at water/ice boundaries.  Electromagnetism, or at least magnetism, is not out of the question.

On the other hand, Newtonian mechanics might take quite a while.  We can ignore wind resistance and get reasonable answers for many problems.  Water resistance is a whole different matter, so it might take quite a while to develop a notion of inertia.  Modern physics -- particle physics, quantum mechanics, relativity, plasma physics, low-temperature physics, etc., requires something like modern technology, of which more in a bit.

Chemistry: This will certainly be tricky.  You can't just pour something in a beaker or dump a solid chemical into a liquid solvent -- at least not without producing the solid chemical in the first place.  But who knows?  If there is life, there are chemical reactions going on all the time naturally.  Perhaps someone learns that a particular gland-y thing from one creature does funny things when you stick a bone-y thing from another creature in it, or that you can use thus-and-such material to isolate a kind of water that acts unusually, and eventually this becomes a systematic body of knowledge.  Not out of the question, and not as much of a leap as astronomy would be, but still doesn't seem like it would be a strong suit.

Biology: Large parts of biology -- cell theory, for example -- require some sort of microscope, but other aspects just require careful observations of other living things.  Evolution is an interesting example.  It's not clear what kind of fossil record there might be, though non-living things would tend to float or sink, but evolution fundamentally requires figuring out that there's a family relationship among seemingly different living things, and realizing that the world is old.  That's certainly not out of the question.

Materials science: Metallurgy requires a supply of metal, which would likely be hard to come by, but one could learn a lot about the properties of the materials around -- tensile strength, hardness, elasticity and even, with more careful observation than we need in our environment, density, heat capacity and such.

Cities: There doesn't seem to be any fundamental reason there couldn't be something we might call  agriculture, and floating cities, at least, forming around it.  Suppose again that there's some sort of rope-like material, and suppose something edible likes to grow on it.  It might then be natural to build largish structures and garden them.  This in turn would provide a reason to stay close instead of wandering off, and along come civilization and its discontents.

Fire: At some point a technological species needs to figure out how to do things that don't happen easily in the natural world.  For us, you could argue it was extracting metals sometime in prehistory, or maybe harnessing steam, or maybe something in between.  For an undersea world, creating an environment where things could burn would be not only a significant achievement, but a gateway to what we might consider industrial technology.  If they can make fire, then it's hard to think what human technology would be out of reach, because to make fire, you need to recreate an environment not too different from ours.  At the very least you need a bubble of some sort of oxidizing atmosphere.  Once that happens, pretty much everything that seems implausible on the list above becomes possible.


So who knows?  There doesn't seem to be anything in principle keeping life on some extraterrestrial ocean from being able to get to us.  It just requires a long chain of unlikely events.  But then, life is full of those.  If you roll the dice billions of times and you get to keep lucky changes around for the next roll, the extremely unlikely can become quite plausible.  In this case, our chain of events looks something like

• A self-replicating molecule develops
• Microorganisms develop
• Multicellular organisms develop
• Some of them develop what we might call intelligence
• These develop a body of mathematical and scientific knowledge
• From that, they develop what we would recognize as technological tools, including
• Ways of storing energy and converting it to work
• Probably some way of creating sizable bubbles of gas, and ways of working inside them
• One way or another they find their way to the surface (which is much farther away than humanity has ever managed to dig into the Earth's crust)
• And then they get in contact with us, somehow.

On the other hand, we know we're coming their way (if they're there).  It's going to be quite some time before we can send a probe to their habitat, but there's one other way we could meet up: One of the reasons we believe that Ganymede and other worlds have water: cryovulcanism, that is, liquids coming to the surface through cracks in the crust (the "cryo" part is there because these liquids are a lot colder than the magma we're familiar with).

If water is coming up from the liquid interior of one of these moons, it might well carry something along with it.  With luck, it might be something like a tardigrade that can survive being dried out and frozen.  Or there might just be a fossil bed of life forms that might not have even made it to the surface alive.

Or, maybe, just maybe (and by "just maybe" I mean "almost certainly not, given all the things that would have to go right"), it might be a hardy explorer, wearing some sort of protective suit, struggling against the surface gravity that we would consider negligible, and wondering at the bizarre new world, and ... what's this thing coming at me from the ... what do you even call it without knowing "stars" and "empty space"?

Fermi and the revenge of the machines

Previously, I argued that the odds of our actually directly detecting an intelligent species from another star system are very low, partly because of the vast distances involved, but also because of timing.  However, there is another possibility besides detecting such a species on their home world: maybe they'll come to us.

"Or maybe we'll come to them", you might counter, and maybe, eventually.  I'm going to leave that aside for now because, while we have some interesting theoretical ideas of how we might reach another star system, at this point they're just that, and the more realistic ones involve moving maybe a ton of payload to a nearby star in somewhere around a human lifetime, and score around a seven on the scale of difficulty I gave earlier.

The amount of energy involved in doing even that is staggering.  By comparison, the New Horizons probe that's currently nearing the Pluto system as I write this was launched on one of the larger rockets ever produced, received a significant gravitational boost from Jupiter, and has been traveling for around ten years.  To get to Proxima Centauri in a small number of decades, you'd need to be traveling around 2000 times as fast.

The energy required goes as the square of the velocity (until you get to relativistic speeds, where it gets much, much worse), so we're looking at 4 million of the rocket that launched New Horizons.  Put another way, the kinetic energy of a one-ton object moving one-tenth the speed of light is comparable to

• The energy released by the largest nuclear bomb ever detonated
• The energy released by the eruption of Krakatoa
• The electricity consumed by Norway in a year

Actually accelerating such an object to that speed would take considerably more energy.

However, one of the main points of the previous post was that some significant portion of intelligent species would have arisen millions or hundreds of millions of years ago, and another was that there were likely a great many such species scattered throughout the galaxy, not to mention other galaxies.  Maybe billions in our galaxy alone, using not-totally-implausible guesses. Put that together, and there might be a large number of species which have had plenty of time to reach us, and plenty of time to develop the technology to do it.  The timing that undermined our prospects of spotting other life directly actually works in our favor now.

So let's say that some technological species arose ten million years ago and 50,000 light-years away.  If they'd built a craft capable of traveling 1/200 the speed of light, or about 100 times faster than New Horizons, that craft could have reached us by now.  And maybe even have stopped.

Or, it could have reached any of a few hundred billion other stars.  Even if it had launched, say, ten craft a year for a million years, each equipped with a means of navigating to a distant star and then slowing down, the odds would still be about 10,000 to one against one reaching us.


However, there's a scheme that's been floating around for quite a while that would dramatically increase the chances of something eventually getting to us: send self-replicating craft.  This is not something we could do right now but it's quite plausible that we could in the next century or so.  It doesn't seem too much to ask that some civilization, somewhere, with a million or more years of head start, could have done this.

It's an interesting scenario to contemplate.  Weight is critical at these speeds, so the craft will probably carry only the bare minimum it would need.  It would enter a new star system and find a solid planet or moon to set down on.  It would then start digging into the surface (assuming there's no one there to run across it and ask "Hey, what's this?") and gradually assemble raw materials.  From those it would assemble basic tools, use those to assemble more sophisticated tools, and eventually spacecraft parts, which it would then proceed to put together.  Some amount of time later it will have made a copy of itself, which would then take off for a nearby star system, while the original goes on building copies of itself ...


You'd want to be a little careful with this.  If you literally sent a copy to every nearby star system, that would include ones you've already visited (even the home world), and some of those would make their way back to where they started, and start making copies of themselves.  That is, you would get exponential growth, everywhere the probes visit.

The total number of craft might double every few centuries.  If every copy used a ton of raw materials, they would consume the mass of our Moon in about 70 generations of copying, or a few dozen millennia.  In a few million years, probes could eat every planet, planetoid and moon in the galaxy.  Or, at least, every unpopulated one.  Or, at least, the portion of the material of the unpopulated ones that was suited for making the craft.

The polite way to send such probes would be to remember which places you'd already visited and only send copies to new places.  In technical terms, this results in a breadth-first search of the galaxy.  Instead of exponential growth you get cubic, that is, a steadily expanding radius of exploration.  In our scenario of craft traveling 1/200 the speed of light (and not taking too long to build copies), this would cover the whole galaxy in around 20 million years, starting at the edge, or 10 million starting near the center.

In sum, if anywhere in the hundreds of billions of stars in the galaxy 20 million years ago someone had successfully launched a self-replicating interstellar probe, one copy (or more) could have made it to the solar system.  A heady thought, to say the least.


One question that always bothered me about this scenario was "why?"  Even if the various probes could relay information back to the home world (and this is totally handwaving how one would produce a signal strong enough), the finite speed of light becomes a problem.  It's going to take a thousand years to get information back from a system a thousand light years away, and 200,000 from one edge of the galaxy to the other.  Why bother?  I mean, who am I to say how long a lifespan, or attention span, an alien species might have, or what its motivations might be, but still ... it's hard to see the point.

However, you don't have to set out to cover the whole galaxy in order to end up doing it.  Suppose we just wanted to explore our neighborhood of a few thousand stars?  A self-replicating probe would be cheaper and easier than trying to send a probe to every star individually.  In which case, why stop?  If we're interested in data from a star 50 light-years away, why not 60, or 100?  The easiest approach is to just let the probes keep copying.

A bit unsettlingly, it's even easier not to bother screening out already visited systems, that is, letting exponential growth continue unchecked.

In a slightly less scary variant, the probes somehow send messages to each other ("heartbeats", effectively) and only send copies to stars that aren't sending anything.  This way if a probe goes dead, another will eventually take its place and the galaxy will remain completely covered indefinitely.  Among other things, this means that if you're actively trying to get rid of probes, that may turn out to be somewhat difficult -- it only encourages them.


Once we get into galaxy-spanning schemes, we're looking at hundreds of billions of copies.  There are bound to be a few mistakes here and there, and those mistakes may continue to propagate.  The ones that do will continue to make copies, and some of those copies will be imperfect.  "Mistake" here just means "different from the original".   Such a mistake might be innocuous, or it might result in a less worthy probe, or it might even result in an improvement.  One way or another, the population will evolve over time.

So we have something that can use energy in an organized way, reproduce and evolve.  We may as well call that "life".  It's most likely not the same life form as the one that first built it on the original home world, but that seems like a minor point.  It's quite possible that there is life spreading out through the galaxy, or already occupying every corner of it, even while its creators are still confined to a single world, or no longer around at all.  And that life might well have reached us, or even be here now.

This makes the question of "why haven't we noticed" a bit more interesting.

Language family trees, new and old

For the most part, change in language is gradual,  though with some exceptions.  People do develop new languages quickly, and not only with constructed languages like Esperanto or Klingon.  It can happen naturally in a couple of ways.

When populations with different languages are brought into close contact, it's common for people to work out a "pidgin", a sort of half-language with vocabulary taken here and there from either parent language, but with minimal syntax and grammar.  Or at least that's been a prominent theory.   There's been some revisiting of just how simple pidgins really are.

One way or another, though, the children who grow up hearing the pidgin end up speaking a "creole", which is a full-fledged language with its own grammar and syntax, and distinct from either of the parent languages.  Confusingly enough, these creoles are often referred to as pidgins.  Tok Pisin is a classic example.  It's been around long enough to develop its own dialects.

There are also sign languages that can be traced back to a small community that, as far as anyone can tell, invented its own language spontaneously since they needed to communicate and (obviously) couldn't use any of the spoken languages around.  These are interesting in that they say something about our innate ability to use language, even without being exposed to one.

Unlike pidgins and creoles, which hybridize existing languages, these sign languages are constructed from scratch and, to be clear, they have the same kinds of complex features that spoken languages have, including the use of arbitrary symbols for abstract concepts.  Like any other true sign language, these are not simple gestures and pantomime, and like other languages, sign or not, they too can develop variants and dialects over time.

There are probably other examples along similar lines.  People have been using languages for quite some time, and there have been quite a lot of people over that time.

Nonetheless, the vast majority of the world's thousands of languages trace their lineage back over thousands of years of gradual change, likely all the way back to the exodus from Africa some 50,000 years ago, and beyond.  Why should we think this?

The main reason to is that most of them bear a family resemblance to other languages.  That is, they share features in common with those other languages, and those features indicate a branching pattern of languages diverging repeatedly from earlier forms (but also intermixing and hybridizing, so we don't have a pure "tree structure").  In some cases we can directly trace the development of a language family, for example the Romance languages of Europe (French, Italian, Portuguese, Romansch, Romanian, Spanish and several others, all derived from Vulgar Latin).

By careful comparison of the various features of similar languages, and by clues in the structures of the languages themselves, we can form a fairly precise model of how the various languages developed from a common ancestor over time, in a way that matches up well with written records where they're available, but does not require them.

If we know for a fact that some language families developed over time from a common ancestor, and most of the others show the same kind of resemblances to each other as members of well-documented families, the simplest explanation is that that what looks like a family resemblance is a family resemblance.

This is not to say that we can assert with certainty that, say, Bantu, Swiss-German, Aleut and Japanese share a common ancestor.  Language changes quickly enough that the evidence becomes indistinct as we look at larger and larger families of languages.  The most thorough and successful reconstruction of an ancestral language, Proto-Indo-European, takes us back only five or six thousand years.

To go back further takes careful statistical analysis, and this is not without its problems and controversies.  However, there is no strong evidence that the languages listed above aren't related.  They all seem to follow the same general plan, albeit with markedly different details.  Again, the simplest assumption, until we know better, is that they all trace back to a small number of ancient languages, just as human ancestry traces back to a common ancestor (note that as always, "common ancestor" doesn't mean "first").

There's extinct, and then there's extinct

As a little follow-up to the previous post:

Both Manx and Old English have been extinct (I'll explain the careful phrasing in a bit).  We know exactly when Manx went extinct: with the death of Ned Maddrell on December 27th, 1974.   We can't say when Old English went extinct, and not only because it would have happened hundreds of years ago.

"When did Old English go extinct?" is the same sort of question as "When did Middle English arise?", though it's not the exact same question.  If we were to define some set of speakers as "the first speakers of Middle English" and some other as "the last speakers of Old English", then there was, pretty much necessarily, a period of time where both were alive.  Middle English was alive, by such a reckoning, but Old English wasn't dead yet.

In fact, it's plausible that some number of people could have been said to speak both, depending perhaps on the company and occasion.  As I argued before, what's really going on here is that there is really no clear line to be drawn in cases of continuous change.

On the other hand, when the number of speakers of a language dwindles, it's reasonable to speak of the language going extinct: when the last speaker speaks no more.

The situation may become somewhat better-defined if we consider features of languages.  It's plausible that there was a last time that someone used urum for "our" with a plural noun (in the appropriate grammatical case, etc.), or a first time (probably some time earlier and/or in some other place) that someone said oure in the like situation.

This at least clarifies some of the difficulties of the exercise.  First, features do not arise or disappear everywhere at once.  At some point, there would have been some people who said urum or oure, as the case may be and some who didn't.  At some later point there were fewer who said urum and more who said oure, and eventually, no one was saying urum anymore.  And even that may be an oversimplification.

Second, what set of features we call "Old English" and what set we call "Middle English" is largely arbitrary.  More realistically, as time went on, there were more people speaking English with what we would now regard as Middle English features and fewer speaking with what we would regard as Old English features.

Except that we still retain some features of Old English, for example, the words is, on, to and he from Ælfric of Eynsham in the previous post.  When we try to draw a line between Old and Middle English, we're really looking for features unique to each.  We could plausibly say that when there are, say, no longer any people who speak with more of Old English's unique features than Middle English's, Middle English has taken over.  When no one is using that particular set of features at all, we could say that Old English is extinct ... but then we should be careful not to choose any features for that set that do survive.

If this has a sort of unsatisfying feel to it, it's because the whole exercise is of limited use.  At the bottom of all this careful definition is a distinction without a difference.  The real story is that features come and go, with enough differences eventually accumulating that we start to feel that two different sets of features denote different languages.


Now, why did I say "have been extinct" above?  Manx is, it turns out, no longer dead.  There aren't many speakers, and perhaps none who speak exclusively Manx, or few that could be said to speak Manx as their first language, but people are speaking Manx again, with pride.  This is thanks not only to modern Manx speakers, but in part to recordings of Ned Maddrell and others, and to the efforts of the linguists who made those recordings and otherwise worked to preserve a record of the language.

Xeno and the history of English

Kids these days.  They don't talk like we did when I was a kid.  I'd give some examples, but they're probably already out of date.  You know what I mean, though.  It's one of the constants of life, the next generation doing things a bit differently from the last.  O tempora o mores.

When trying to make sense of the world's languages, dialects, accents, jargons and such, it's natural to look at differences among speakers, or among groups of speakers.  French speakers speak a different language from Italians.  New Yorkers, for the most part, speak differently from people in Clinch County, Georgia, or County Derry, Ireland.  It's easy, though, to neglect differences in a particular language over time, even though they can be just as significant.

Changes over time are in some ways similar to differences among contemporaries.  If someone were to start speaking, say, the English of Beowulf to a group of modern English speakers, they would likely hear it as just another foreign language.  At that level of remove, there's little chance that a modern speaker would think "That person sounds like they're trying to sound like someone from pre-Norman times, but they're definitely speaking English."

There is, however, one salient property of changes over time: Unlike differences between languages at any given time, they have to be small.  Even if a parent and child speak differently, they still need to understand each other.  Granted, this is generally more important to the older generation.  Kids throughout history have always been quick to invent their own vocabularies and otherwise make themselves more easily understood by each other than by their elders.  Even so, from a linguistic point of view they are speaking essentially the same language.

It takes quite a while for that gradual change to amount to what we would call a new language.  For example, here's a bit of a poem (They flee from me) that I studied in college.  I've modernized the spelling, but otherwise the words are the same

I have seen them gentle, tame, and meek,
That now are wild and do not remember
That sometime they put themself in danger
To take bread at my hand

This was written by Thomas Wyatt in 1535, almost 30 years before Shakespeare was born, but that wouldn't necessarily be your first guess, would it?

As we go back further, change becomes more apparent, but again gradually.  Here's a bit of Chaucer (The Former Age), again with modern spelling, from the late 1300s

No man yet knew the furrows of his land,
No man the fire out of the flint yet found,
Unkorven and ungrobbed lay the vine;
No man yet in the mortar spices ground

OK, a couple of unfamiliar words (unkorven may be translated as "unpruned" and ungrobbed as uncultivated), but then what do bodkin and fardels mean in Hamlet's soliloquy, written about 200 years later?  The word order is a bit weird, but just how do you parse the first lines of The Star-Spangled Banner, written about 400 years later?  Maybe that's just poetry for you.

I'm cheating a bit here by focusing on the written word, because there's ample reason to believe that English was pronounced considerably differently in the 1300s.  But then, contemporary speakers, depending on their dialect, will pronounce the same written words differently as well.  In the case of Chaucer, I'm also cherry-picking a bit.  Some passages of Chaucer sound considerably less familiar, and some contemporaries of Chaucer in other parts of England would be less familiar still.

Continuing our journey back in time, here's a passage from The Peterborough Chronicle, written in 1140:

Tha was England suythe todeled: some helden mid the king.  Some helden mid the empress.  For tha the king was in prison, tha wenden the earls to rich men that he never more should come out, sahtleden mid the empress, brought her into Oxford, iauen her the burgh.

This still seems something like English, but it's getting harder to decide what to clean up as just a matter of spelling, and what's just different.  For example, with is now mid, in line with the other Germanic languages.  Verbs have the -en ending when the subject is plural (as they do in Chaucer as well, though not in the passage I chose).  I could just as well have papered that over and written held instead of helden and, stretching a bit more, went instead of wenden.

That funny word iauen is really just gave: the letters u and v were not distinguished until later, and quite likely the i at the beginning is like the y in yellow, corresponding to an initial g in other dialects, so we get gaven, or plain gave when you lop off the -en.  That tha seems a bit like then and a bit like when, not unlike modern dialects that use what where others would use that.

It's worth noting that this was written with the Norman conquest of 1066 still in living memory, though only just.  Given that, I'm actually a bit surprised there aren't more French borrowings.  The only ones that stick out are prison and empress.

Go back not too much further, before the conquest, and we have something not too much different from the Peterborough Chronicle, but different enough from modern English that trying to smooth over the differences is a lost cause.  Here is Ælfric of Eynsham discussing the reflection in grammar of Christian theology.  First a straight transliteration (but with modern punctuation):

Oft ys seo halige þrinnys geswutelod on þisre bec, swa swa ys on þam worde þe God cwæð: 'Uton wyrcean mannan to ure anlicnisse'.  Mid þam þe he cwæð 'Uton wyrcean' is seo þrinnys gebicnod; mid þam þe he cwæð 'to ure anlicnisse' ys seo soðe annis geswutelod: he ne cwæð na menfealdlice, 'to urum anlicnissum', ac anfealdlice, 'to ure anlicnisse'.

Here's an attempt to smooth over the spelling differences and such, taking a few more liberties than in the last example:

Oft is the holy threeness geswutelod on this book, so so as the words that God quoth, 'Uton work man to our (an)likeness'.  With þam þe he quoth 'Uton work' is  the threeness gebicnod; with þam þe he quoth 'to our (an)likeness' is the sooth oneness geswutelod: he ne quoth na manifoldly (many-fold-ly) 'to our (an)likenesses', but one-fold-ly, 'to our (an)likeness'.

(That's sooth, as in soothsayer, meaning 'true', not the modern verb soothe).  And finally, glossing the words that don't seem directly related to modern ones, and updating a few that are,

Often is the holy trinity revealed in this book, just as the words that God said, "Let us make man to our likeness."  With that he said 'Let us make' is the trinity indicated.  With that he said 'to our likeness' is the true unity revealed: he neither said nor plural, 'to our likenesses', but singular, 'to our likeness'.

Even with the words glossed, the word order is a little funky, that "with that" is only approximate and doesn't parse easily, and the ne ... ne ... construct, still used by Chaucer and with traces here and there in Shakespeare, is gone now, the closest remnant being neither ... nor.  The forms of words have changed significantly (halige vs. holy, annis vs. oneness etc.).  There is more and different inflection in the original, both in nouns and pronouns (ure anlicnisse vs. urum anlicnissum, worde as the plural of word) and verbs (ge- ... -od for the past participle instead of just -ed today).

All in all, this is more a translation than a modernization.  Only after a third round of adjustments would Ælfric's English really look like modern English.

Even though the Peterborough Chronicle is in an East Midland dialect that is not a direct ancestor of Chaucer's, and Ælfric spoke a Wessex (West-Saxon) dialect which is, and the writers are writing in considerably different forms, it's not hard to see that these are examples of the same language, but a language which is, overall, changing gradually over time.

Thus even though we can point at plenty of texts and say "This is Middle English", or "This is Old English", there is no time in history we can point at and say "Middle English started here" or "Modern English started here".  As far as anyone at the time was concerned, they were speaking English.  No "Old" or "Middle" or "Modern" about it.

Even if you did pick some set of speakers and say, "the parents spoke Middle English, but the children spoke Modern English", you'd be hard pressed to tell the difference between the two.   There would be much more difference between our Modern English and theirs than between the parents and the children.

Gradual does not mean imperceptible, however.  We can trace the origins of newer words like blog or semiconductor with a fair degree of certainty.  We can remember words and constructions that people don't say much any more (at least in most places), like To whom did she give the book? instead of Who did she give the book to?  Going further back to old written sources, we can get a reasonable idea of when people stopped using -en as a regular plural marker for nouns, though of course that didn't happen all at once.  We can see individual changes happening, it's just that there are so many parts to a language that it takes quite a few -- many more than typically happen in a lifetime -- before we decide to call the result a new language.

Drawing a precise line between an older form of a language and a later form is largely arbitrary, like trying to pick a boundary between green and blue.  You can do it, but wherever you pick, there will be a reasonable argument for picking some slightly different boundary.  Even so, the language has changed and continues to.  You can quite rightly say "no one saw Modern English arise from Middle English," and yet it did.

This is all reminiscent of Xeno's arrow paradox: If an object in a definite place is at rest, and at any given moment a flying arrow is in a definite place, then, it would seem, a flying arrow must always be at rest, so how can it move at all?  Untangling this sort of thing rigorously is the wellspring of the mathematical field of analysis, but for most purposes it's enough to know that the arrow moves anyway, somehow unaware that it's not supposed to be able to do that.

Pan proiiciens

Humans have several unique qualities (and many more not-so-unique), but one that may not come readily to mind is that we throw things, and we throw them very well.  No other animal we know of could come remotely close to doing this, or this, or this.  Even an average human's throwing abilities are far beyond anything else in the natural world.  If I said that I'd seen someone throw a ball 50 meters, no one would think twice.  If I said that I'd seen a horse throw a ball 50 meters, the most likely response would probably be "No, you didn't", maybe followed by "How??"

If you think about it, it's actually somewhat surprising that humans would be unique when it comes to throwing things.  It's a very useful skill.  If you can bring down a small mammal with a well-aimed rock, you'll eat much better than if you have to run it down.  Even if you're a very good runner-downer, it still takes way less energy to throw a rock.  Throwing a rock can also get a high-hanging fruit out of a tree (if not easily).  Throwing is a good way to get something you're carrying up onto a high ledge that you can then get to with all your limbs free, and so forth.  You'd think something would have stumbled on it before our ancestors did.

On the other hand, a lot of body plans just aren't that well suited to throwing.  Birds have one set of limbs serving as landing gear and for locomotion when not airborne, and the other given over to wings.  That leaves the head and beak as all-purpose picker-uppers, severely limiting throwing potential.  Birds can throw things by picking them up and then heaving the head and letting go, but not very far or forcefully.

Fish and other sea creatures live in a viscous medium where throwing is not particularly useful.  Most mammals have all four limbs specialized for walking, running, jumping and so forth, leaving them in much the same spot as birds.  Lizards are also pretty well tied to the ground or other surface they're traveling on -- particularly the limbless ones, to say nothing of actual snakes (snakes are closely related to lizards, but there are also some proper lizards that lack limbs).

There's an obvious common thread here: To throw effectively you need a free limb, one not specialized to supporting your weight.  You need hands, as opposed to front feet.  You don't need to be fully bipedal, though that clearly ought to help, but you do need to be able to grab something, stand up on your hind limbs and let fly.

There aren't many animals in that category, but several primates are, by virtue of having hands (and feet, and tails) adapted to grabbing and swinging from tree branches instead of always walking on all fours.  Putting all this together, perhaps it's not so surprising that throwing would have only evolved recently, in a branch of the primate family tree, itself relatively recent.

It is a principle of evolution that behavior tends to change first, and then anatomy follows.  First certain fish started coming up out of the water, whether to escape from other fish who couldn't or to move from a shallow, drying-up pond to a deeper one, or for whatever reason. Later came Tiktaalik and its kin with pectoral fin bones (and other features) better adapted to life out of water.

Once the new niche was established on land, there was plenty of selection pressure to reshape the body for living on land, and eventually lose the older adaptations for water living entirely.  More strictly speaking, the new behavior of coming up on land meant that fish with more land-adapted bodies could have better survival chances.

In the case of primates throwing, it's not that other primates don't throw.  Some species of monkeys are well-known to throw excrement at others of their species, and chimps will throw things as a threat.  It doesn't seem to matter much to them what they throw, but it will generally be branches and occasionally rocks.  Chimps don't show signs of throwing with the same purposes we do, but it's significant that some sort of throwing behavior is established in our near relatives.  It's therefore plausible, though not certain, that it was also established in our common ancestors.

While chimps are known to hunt, and are known to throw things, and are known to be reasonably intelligent, they are not known to throw for purposes of hunting.  This is not completely shocking, considering what they hunt -- small monkeys -- and how they hunt them.  The strategy is to go in as a group, block off escape routes and send one member of the hunting party in for the kill.  Rock throwing would probably not help much in such a situation.  You'd have to encumber a hand carrying a rock up into the trees, and then you'd only have one rock, a bunch of branches in the way, and a monkey that would not be well-inclined to stay still in the path of a hurtling rock.

But imagine a tribe of ancestral chimps a couple of million years ago venturing out of the forest.  These wouldn't be exactly like today's chimps, of course, but today's chimps appear to resemble these ancestors much more closely than we do, so we can consider them ancestral chimps here.  In 1993, William Calvin laid out a hypothesis that a good way for such creatures to get food would be to stake out a water hole and go after the herds that gathered there to drink -- as several other predators do.

Naturally, herds of gazelles and such are adapted to dealing with predators around water holes.  The main strategy is to stampede away, which works well, except that any animal that trips or falls is likely to be trampled by the herd and left behind as an easy meal.  This actually works reasonably well for the herd as a whole (from an individual's view, if you have this behavior you're more likely to be one of the survivors than if you don't), and it works for the hunters as well.  It just doesn't work particularly well for the particular animal left behind.

Now suppose you hit upon a way to make a herd stampede, and at the same time make one of its members stumble and likely be trampled.  That would work even better for you, though again not so well for the unlucky victim.  Calvin's idea is that there's an easy way to do this, that those ancestral chimps could have done with the mental and physical abilities they had: Throw a reasonably-sized rock into the herd.  If it hits one member, that member will stumble, the herd will startle and, with a bit of luck, trample the stumbler.   This doesn't have to happen every time, just enough to make it noticeably easier to get food from the herds gathered at the water hole.

At that point, we're off to the races.  Any change, whether in behavior or anatomy,  that improves throwing force or accuracy, will make for better hunting and better-fed primates.  Richard Young cites paleontological evidence to support a claim that exactly this happened during the past couple million years, that the body, and the hand in particular, became better and better suited to throwing (and to clubbing, another area where we excel).  Doing these well requires a number of changes from the long-fingered, short-thumbed tree-branch-hooking hands of the rest of the chimp family.

Calvin, for his part, speculates that the famous Archeulean hand axe was actually an improved weapon to throw into a herd, as it would be more likely to cut into the prey's hide and cause it to instinctively collapse, making it more likely to be trampled.  As always, this doesn't imply that the hunters were thinking it through in that much detail.  It's enough that throwing a sharp rock works better than throwing one that isn't, and that some in the population had an innate proclivity for chipping away at rocks and so making them sharper.

It's been argued repeatedly that, if we weren't the ones doing the classifying, or weren't so subject to a certain prideful feeling of distinctness from the rest of nature,  we would be classified in the same genus as chimps and bonobos.  Either they would be Homo troglodytes and Homo paniscus, or we would be Pan sapiens.  Under the latter scheme, those savanna-dwelling rock throwers might best be called Pan proiiciens -- throwing Pan, the name Pan being taken from the Greek god of the forest, used to designate the forest-dwelling chimps and bonobos.  Except these particular Pan would be leaving the forest.

Behavior drives anatomical change, but anatomy limits behavior.  Just as only some fish had the right anatomy to even try moving on land, only some animals had the right kind of body to try throwing things.  Primates happened to be close enough, again likely because their hands had already adapted for something else besides walking, something that allowed for grabbing and flinging.  This is a common pattern in evolution.  It used to be called pre-adaptation, but that gives the impression that, for example, primate hands evolved for grabbing tree branches so that they could later evolve for throwing and clubbing.  That's not how it works, so the ungainly but unbiased term exaptation is preferred.

There is another interesting point that Calvin makes.  It's not enough to have the anatomy.  You need to be able to control it.  It takes around a tenth of a second for our bodies to carry out a conscious command.  To throw a projectile accurately [as we do now, as opposed to lobbing rocks into a herd -- D.H. Sep 2015], you need to time your movements within about a hundredth of a second.  Once you've decided to throw something, it's far, far too late to make adjustments as you go along.  You have to have the whole program ready to go ahead of time, adjusted for where your target is, or rather where it will be when the projectile reaches it.  Calvin speculates that this sort of plan-ahead was re-purposed into our control structures for things like language.

I'm not sure I quite buy this.  I doubt that throwing is the only thing in evolutionary history that requires this sort of plan-ahead.  Surely when a hawk dives for a mouse or a cheetah jumps for a gazelle it is doing the same sort of thing under similar constraints.  Nonetheless, it's an interesting idea, and plausible in a general sense, that the original behavior change of throwing things would have brought on a host of other changes, in both behavior and anatomy, that led to wholly new behaviors like speech and large-scale planning (for lack of a better term for the type of planning that we do and other animals don't).