Eric S. Raymond's Blog, page 17

Your identity is not your choice

There’s been a lot of public talk about “identity” lately, stimulated by high-profile cases of transsexuality (notably the athlete now named Caitlyn Jenner) and transracialism (Rachel Dolezal). It needs to be said: most of the talk, on all sides of these disputes, has been obvious nonsense – utter drivel that should not have survived five minutes of thought.

I thought we had reached the limit of absurdity with the flap over Rebecca Tuvel’s paper In Defense of Transracialism, about which it can only be said that while Tuvel seems marginally less insane than her attackers, everyone involved in that dispute has obviously been huffing unicorn farts for so long that oxygen no longer reaches their brains in appreciable quantities.

But that’s in a corner of academia where one rather expects postmodernism to have shut down rational thought. In its own way, the following statement in an exudation of mainstream journalism is much sillier, and has finally pushed me into writing on the topic. I quote it not because it’s a unique error but because it’s representative of a very common category mistake.

Thus should there be a weighty presumption against so blocking people, against subordinating them by substituting our judgments about their identity for their own.

This would seem to be a rather uncontroversial point, based on ordinary liberal arguments in favor of tolerance and respect for the dignity of others.

Ah, yes. So, what then would be amiss if I stood up in a public place and claimed to be the Queen of England? Who are you to substitute your judgment about my identity for my own?

There would actually be two different kinds of things wrong with this claim. One is that I can’t grant peerages – the people who administer the English honors system wouldn’t recognize my authority. The other is that the claim to be “Queen” (as opposed, to, say, “Prince-Consort”) implies an observably false claim that I am biologically female.

These criticisms imply a theory of “identity” that is actually coherent and useful. Here it is:

Your “identity” is a set of predictive claims you assert about yourself, mostly (though not entirely) about what kinds of transactions other people can expect to engage in with you.

As an example of an exception to “mostly”, the claim “I am white” implies that I sunburn easily. But usually, an “identity” claim implies the ability and willingness to meet behavioral expectations held by other people. For example, if I describe my “identity” as “male, American, computer programmer, libertarian” I am in effect making an offer that others can expect me to need to shave daily, salute the Stars and Stripes, sling code, and argue for the Non-Aggression Principle as an ethical fundamental.

Thus, identity claims can be false (not cashed out in observed behavior) or fraudulent (intended to deceive). You don’t get to choose your identity; you get to make an offer and it’s up to others whether or not to accept.

There was a very silly news story recently about “Claire”, a transsexual “girl” with a penis who complains that she is rejected by straight guys for ‘having male parts’. Er, how was “she” expecting anything different? By trying to get dates with heterosexual teenage boys using a female presentation, she was making an offer that there is about her person the sort of sexual parts said boys want to play with. Since “she” does not in fact have a vagina, this offer was fraudulent and there’s no wonder the boys rejected it.

More to the point, why is this “girl” treated as anything but a mental case? Leaving aside the entire question of how real transgenderism is as a neuropsychological phenomenon, “she” clearly suffers from a pretty serious disconnect with observable reality. In particular, those delusions about teenage boys…

I can anticipate several objections to this transactional account of identity. One is that is cruel and illiberal to reject an offer of “I claim identity X” if the person claiming feels that identity strongly enough. This is essentially the position of those journalists from The Hill.

To which I can only reply: you can feel an identity as a programmer as strongly as you want, but if you can’t either already sling code or are visibly working hard on repairing that deficiency, you simply don’t make the nut. Cruelty doesn’t enter into this; if I assent to your claim I assist your self-deceit, and if I repeat it I assist you in misleading or defrauding others.

It is pretty easy to see how this same analysis applies to “misgendering” people with the “wrong” pronouns. People who use the term “misgender” generally follow up with claims about the subject’s autonomy and feelings. Which is well enough, but such considerations do not justify being complicit in the deceit of others any more than they do with respect to “I am a programmer”.

A related objection is that I have stolen the concept of “identity” by transactionalizing it. That is, true “identity” is necessarily grounded not in public performance but private feelings – you are what you feel, and it’s somehow the responsibility of the rest of the world to keep up.

But…if I’m a delusional psychotic who feels I’m Napoleon, is it the world’s responsibility to keep up? If I, an overweight clumsy shortish white guy, feel that I’m a tall agile black guy under the skin, are you obligated to choose me to play basketball? Or, instead, are you justified in predicting that I can’t jump?

You can’t base “identity” on a person’s private self-beliefs and expect sane behavior to emerge any more than you can invite everyone to speak private languages and expect communication to happen.

Racial identity is fuzzier than gender identity becuse, leaving aside “white men can’t jump”, it’s at first sight more difficult to tie it to a performance claim. Also, people who are genetically interracial are far more common than physical intersexes. Although this may mean less than you think; it turns out that peoples’ self-ascribed race correlates very accurately with race-associated genetic markers.

Nevertheless, here’s a very simple performance claim that solves the problem: if you are a man or woman who claims racial identity X, and I do too, and we were to marry, can we expect our children to claim racial identity X and, without extraordinary attempts at deceit, be believed?

This test neatly disposes of Rachel Dolezal – it explains not just why most blacks think she’s a fraud but why she’s an actual fraud. To apply it, we don’t even have to adhere to an “essentialist” notion of what race is. But the test becomes stronger if we note that (see link above) a genetic essentialist notion of race is probably justified by the facts. Among other applications, genetic racial identity turns out to matter for medical diagnosticians in assessing vulnerability to various diseases – for example, if you are black but claim to be white, your doctor may seriously underweight the possibility that you have hypertension.

As a culture, we got to the crazy place we’re at now by privileging feelings over facts. The whole mess around “identity” is only one example of this. It’s time to say this plainly: people who privilege feelings over facts are not sane, and the facts always win in the end. Though, unfortunately, often not before the insanity has inflicted a great deal of unnecessary suffering.

You shall judge by the code alone

I support the open letter by Drupal developers protesting the attempted expulsion of Larry Garfield from the Drupal commmunity.

As a Drupal contributor who has never in any respect attempted to tie the project to his beliefs or lifestyle, Garfield deserves the right to be judged by his code alone. That is the hacker way; competence is all that matters, and no irrelevance like skin color or shape of genitals or political beliefs or odd lifestyle preference should be allowed to matter.

That I even need to say this in 2017 is something of a disgrace. The hacker culture already had judge-by-the-code-alone figured out forty years ago when I was a n00b; the only reason it needs to be said now is that there’s been a recent fashion for “social justice” witch hunting which, inevitably, has degenerated into the sort of utter fiasco the Drupal devs are now protesting.

Thomas Paine said it best: “He that would make his own liberty secure, must guard even his enemy from oppression; for if he violates this duty, he establishes a precedent that will reach to himself.”

It doesn’t matter how much you dislike Larry Garfield’s personal kinks. If you don’t defend him now, you may have nobody to defend you when some self-declared commissar of political and sexual correctness – or just a censorious project lead like Dries Buytaert – decides that you should be declared an unperson.

You shall judge by the code alone. That is the only social equilibrium that doesn’t degenerate into an ugly political bitchfest with expulsions controlled by whatever happens to be politically on top this week. It was the right community norm forty years ago, and remains so today.

The wreck of the Edmund Fitzgerald: the *science* version

My last G+ post reported this:

Something out there kills about one oceangoing ship a week.

It is probably freakishly large waves – well outside the ranges predicted by simple modeling of fluid dynamics and used to set required force-tolerance levels in ship design. Turns out these can be produced by nonlinear interactions in which one crest in a wave train steals energy from its neighbors.

Much more in the video.

So go watch the video – this BBC documentary from 2002 on Rogue Waves. It’s worth your time, and you’ll learn some interesting physics.

As I’m watching, I’m thinking that the really interesting word they’re not using is “soliton”. And then, doing some followup, I learn two things: the solutions to the nonlinear Schrödinger equation that describe rogue waves are labeled “Peregrine solitons”, despite not actually having the non-dissipative property of your classical soliton; and it is now believed that the S.S. Edmund Fitzgerald was probably wrecked by a rogue wave back in ’75.

In a weird way this made it kind of personal for me. I used to joke, back when people knew who he was, that Gordon Lightfoot and I have exactly the same four-note singing range. It is a fact that anything he wrote I can cover effectively; I’ve sung and played The Wreck of the Edmund Fitzgerald many times.

So, I’m texting my friend Phil Salkie (he who taught me to solder, and my reference for the Tinker archetype of hacker) about this, and we started filking. And here’s what eventually came out: Wreck of the Edmund Fitzgerald, the science! version:

The lads in the crew saw that soliton come through

It stove in the hatches and coamings

Her hull broached and tore, she was spillin’ out ore

That rogue put an end to her roamings.

Does anyone know where the Gaussian goes

When the sea heights go all superlinear?

A Schrödinger wave for a watery grave

It’ll drown both the saint and the sinner.

That is all.

From molly-guard to moggy-guard

In ancient lore, a molly-guard was a shield to prevent tripping of some Big Red Switch by clumsy or ignorant hands. Originally used of the plexiglass covers improvised for the BRS on an IBM 4341 after a programmer’s toddler daughter (named Molly) frobbed it twice in one day

The Great Beast of Malvern, the computer designed on this blog for performing repository surgery, sits to the left of my desk. This is Zola the cat sitting on it, as he sometimes does to hang out near one of his humans.

What you cannot quite see in that picture is the Power Switch of the Beast, located near the right front corner of the case top – alas, where an errant cat foot can land on it. Dilemma! I do not want to shoo away the Zola, for he is a wonderfully agreeable cat. On the other hand, it is deucedly inconvenient to have one’s machine randomly power-cycled while hacking.

Fortunately, I am a tool-using sophont and there is an elegant solution to this problem.

The solution is an ordinary refrigerator magnet. Being a magnet, it adheres to the ferrous-metal case top just firmly enough not to be easily pushed aside by a cat foot, but is readily moved by human fingers. Here it is in the ‘unsafe’ position:

And in the ‘safe’ position:

EDIT: Thanks to a brilliant comment by andyjpb. I have retitled this post and changed the following lexical item.

Thus, your word of the day: “moggy-guard”. Like a molly-guard, but for cats.

From molly-guard to zola-guard

In ancient lore, a molly-guard was a shield to prevent tripping of some Big Red Switch by clumsy or ignorant hands. Originally used of the plexiglass covers improvised for the BRS on an IBM 4341 after a programmer’s toddler daughter (named Molly) frobbed it twice in one day

The Great Beast of Malvern, the computer designed on this blog for performing repository surgery, sits to the left of my desk. This is Zola the cat sitting on it, as he sometimes does to hang out near one of his humans.

What you cannot quite see in that picture is the Power Switch of the Beast, located near the right front corner of the case top – alas, where an errant cat foot can land on it. Dilemma! I do not want to shoo away the Zola, for he is a wonderfully agreeable cat. On the other hand, it is deucedly inconvenient to have one’s machine randomly power-cycled while hacking.

Fortunately, I am a tool-using sophont and there is an elegant solution to this problem.

The solution is an ordinary refrigerator magnet. Being a magnet, it adheres to the ferrous-metal case top just firmly enough not to be easily pushed aside by a cat foot, but is readily moved by human fingers. Here it is in the ‘unsafe’ position:

And in the ‘safe’ position:

Thus, your word of the day: “zola-guard”. Like a molly-guard, but for cats.

PSA: “E-Shielder Security” and “CyberSec Buzz” are gangs of idiotic scum

This is a public service announcement: E-Shielder Security, describing itself as “leading importers and suppliers of high end electronic technology solution systems” is a gang of idiotic scum.

Yesterday they posted a Hacktivists on the rampage in 2017, which largely reproduced my Hacker Archetypes post.

They did so in obvious ignorance of who the hackers I was referring to actually are, going off on a tear about “hacktivists”. That term is, in general, a flare-lit clue that the term using it is either an idiot or a vandal trying to cloak destructive behavior in respectability – real hackers are proud of what they do, take responsibility for it, and don’t wear masks (with a limited exception for those under direct threat from totalitarian governments). In this case it was clearly idiocy.

Mere idiocy turned into something nastier. I left a comment on the post pointing out their error, something I had clear standing to do as the author of the article they were quoting.

The comment was suppressed. That was scummy behavior; thus “idiotic scum”.

Don’t do business with these clowns. Warn your friends. Propagate this widely, the clowns deserve some serious reputation damage.

Addendum: Title amended because the article may have originated at CyberSec Buzz, another ‘security’ blog run by drivelheads who obviously have no fscking idea what they’re talking about. It has been taken down where I originally found it.

Hacker Archetypes

There’s a book about martial arts called On the Warrior’s Path that tries to understand the differing psychologies of martial artists through the lens of half a dozen archetypes – Seeker, Ronin, Tribal Warrior, and others.

I have not yet read the book, but my friend and regular A&D commenter Susan Sons reports having found it very effective for motivating young and newbie martial artists. “It gave them their first glimpse of what they were trying to become,” she reports, “They both knuckled down not just in the obvious physical parts of training, but in the mental aspects, far more than they had before and far more than their age/experience peers.”

So, Susan had the idea that it might be a good idea to develop a parallel gallery of hacker archetypes to help motivate newbies. We brainstormed this on IRC for a while. One thing that had been blocking Susan is that, by her own report, she sucks at naming things. I, on the other hand, am pretty good at that; I was able to come up with names that helped the archetypes develop more definition.

We don’t think this is a complete set, and some of the names might change. But it’s enough of a start for some public brainstorming.

Also note: no hacker is only one of these, but in talking about a number of mutual friends we found it was always pretty easy to agree on both the friend’s dominant archetype and the secondary one that they display most after it. I think this is an indication that we are, even if imperfectly, zeroing in on real traits.

Here they are. Descriptions mostly Susan, names mostly me.

Algorithmicists: Very good at algorithms and sustained, intricate coding. Have mathematical intuition, and are one of the two types (with Architect) that have the highest tolerance for complexity. They like the idea of correctness proofs and think naturally in terms of invariants. They gravitate to compiler-writing and crypto. Often solitary with poor social skills; have a tendency to fail by excessive cleverness. Never let them manage anyone!

Tinkers: Hackers who are drawn to crossovers with the physical world – will design hardware as cheerfully as software. One of the two types (with Prankster) most likely to be lockpickers and locksmiths. Know practical electronics (including analog and RF), adept at reverse-engineering. When you can get them to pull their heads out of the details (which they may resist pretty hard) they make terrific whole-systems engineers.

Architects: The guys who are fascinated by, and very good at, blocking out architecture in complex systems. Kings of the productive refactor. Have an acute feel for design patterns and can see around corners in design space. Strong drive to simplify and partition; “It’s not done until it’s elegant.” The Architect failure mode is to lose sight of the ground. Architects don’t necessarily have communications skills; if they do, they can make worthy team leads.

Sharpshooters: Tenacious detail-obsessives who are most comfortable with a bottom-up view of code and like rifle-shooting bugs more than almost anything else. In past times they would have been happy writing assembler. Polar opposite of the Architect, very productive when paired with one (and vice-versa). Not a good bet for managing anything.

JOATs: The great strengths of the jack-of-all-trades are adaptability, fast uptake of new ideas, and mental flexibility. The JOAT doesn’t do any one thing better than the other types, but can do a bit of everything – including people and social engineering. The JOAT failure mode is to try to do everything themselves. A JOAT is more likely than other types to make an excellent team lead, as long as he remains aware enough to delegate deep technical decisions to others.

Pranksters: Their natural bent is adversarial – they’re great at thinking up ways to disrupt and subvert systems (or just put them to unexpected and hilarious uses). They gravitate to infosec and test engineering. The really good ones can social-engineer people more ruthlessly and effectively than any of the other types.

Castellans: Supreme control freaks who get their power from focusing on what they’re responsible for and knowing it inside out. Castellans memorize manuals; they love language-lawyering, process automation, and vacuuming up domain-specific knowledge about whatever they’re working on. Old-school sysadmins are often castellans: “That will never happen on my system” is their promise (and of course Pranksters love to prove them wrong).

Translators: The type that bridges between human and machine: tends to excel at UI/UX development, documentation, policy and supply-chain stuff, requirements analysis, user training, and so on. Highly social, less hard-core technical than others, but in a way that helps them help other hackers understand how non-hackers see and interact with technology. Some of them make good project managers, but like JOATs they need to understand their technical limitations and mostly leave the hard decisions to types that naturally swim in deeper technical waters. Of all the types, Translators are the least likely to self-identify as hackers even if they are intimate with the culture and working within it.

What archetypes, if any, are we missing? Are there places where the boundaries need adjusting?

(Oh, and me? Mostly Architect with a side of Algorithmicist and a touch of JOAT.)

Final note: This post is being edited as we collect more feedback. “Translators” weren’t in the first version at all.

Four modes of creole formation

A ‘pidgin’ is a language formed by contact between speakers of different languages. A ‘creole’ is what happens when a pidgin becomes a birth language for children raised where a pidgin is spoken. Pidgins are simple languages, stripped to the running gears, Often creoles re-complexify in later generations, retaining grammar mostly from one parent language and vocabulary mostly from the other.

My interest in the historical linguistics of pidgins and creoles began a very long time ago when I noticed that pidgins, wherever they arise, are usually morphologically a lot like English – analytic (positional) grammar with few inflections, SVO order oftener than can be accounted for by the fact that English is often one of the parent languages. Why should this be?

Nicholas Ostler’s excellent Empires of the Word deepened the question by proposing that analytic SVO grammar is the common factor in languages like English, Chinese and Malay that have been very successful at spreading from their original homelands. In his account, that is because this class of language has the lowest complexity barrier to acquisition for adult speakers.

That would explain pidgins all right – they look like they do because they’re invented by adults as the simplest possible way to establish communication. And English, with similar traits, is a non-pidgin that has spread like crazy because it combines the prestige of the Anglosphere with being exceptionally easy for native speakers of other languages to learn.

Er, but why is English like that in the first place?

A few years back I tripped over – and was instantly fascinated by – the notion that English is best understood historically as an old creole. Most educated people know the ha-ha-only-serious line about English being the result of attempts by Norman men-at-arms to pick up Saxon barmaids; this refers to pidgin formation between late Anglo-Saxon and old French after 1066, with later Middle English viewed as the succeeding creole.

But there may have been an even more important creolization 150 years previously when Vikings conquered the region of central England known as the Danelaw and their West Norse collided with midlands Anglo-Saxon of the time. Midlands Anglo-Saxon could be considered to have been replaced by an Anglo-Norse creole with a simplified version of old Anglo-Saxon grammar and a lot of Norse vocabulary (including items as basic as most of the pronouns that English has used since).

So, double creolization with the grammar getting simplified at each stage. But wait! There’s more! We don’t know for sure because the Anglo-Saxons who invaded Britain after the Roman collapse in 410 weren’t literate, but it is quite possible that there was an even earlier creolization following on their domination of the Celtic natives.

So, running this forward in time, the oldest Anglo-Saxon collides with Celtic languages; the resulting creole becomes middle Anglo-Saxon. Which then collides with Old Norse; the creole from this becomes late Anglo-Saxon. That is what collides with Old French and, badda boom badda bing, Middle English.

At each stage the language grammar gets progressively more stripped to its running gears – more analytic, more SVO, more pidgin-like – only to partly re-complexify (as creoles do) and pick up loads of vocabulary items along the way.

Unfortunately we have only a badly foreshortened view of all this because the Anglo-Saxon manuscripts are mostly from the late period. The only creolization process we can more or less track is the latest of the three; the previous two have to be inferred from, for example, the fact that late Anglo-Saxon had already adopted the pronouns of West Norse.

The other problem with this theory is actually a problem with linguists. Ever since philologists started reconstructing the history of the Indo-European language group in the 1800s, linguists have loved nice tidy evolutionary tree structures. Crosslinks that mess up that picture, like sprachbunds or creole formations, are not loved. There is still – at least it seems to me, as an amateur but careful observer – a tendency in academia to want to banish pidgin and creole formation to the periphery, denying that these can be central features of the history of “great” languages.

Part of this comes from historical connotation; pidgin and creole formation were first noticed in the record of the Age of Exploration, when Europeans were having a high old time rambling all over the globe trading with, warring on, enslaving, and having sex with various kinds of dusky-skinned natives. Pidgin/creole formation has still got an associative whiff about it of dandies keeping octoroon mistresses that’s a bit unseemly.

OK, so what if we try looking past that? What happens if, instead of admitting there’s been language hybridization and applying the label “creole” only when forced and the contact event was recent, we think of creolization as a language-formation process that is historically normal under certain circumstances, and go looking for examples and recurring patterns?

Now, much of the rest of this is mostly me being speculative. And I’m not a trained historical linguist with a PhD union card, so its unlikely any of the pros will care much what I think. But I see some fascinating vistas opening up.

For one thing, there are recurring patterns. I can see four: trade creoles, conquest creoles, camp creoles, and court creoles. The lines among these are not perfectly sharp, and sometimes a creole will be repurposed after it forms, but they illustrate four basic modes of formation.

A trade creole is a language evolved from a trade pidgin. There are numerous examples in the South Pacific, of which Tok Pisin in Papua New Guinea is among the best known; these are recently formed within the last 200 years and linguists do apply the label “creole” to them.

On the other hand, Swahili seems to be an old trade creole, resulting from contact between coastal Bantu languages in West Africa and Arab traders and beginning to form not long after 600CE. Linguists normally don’t label it as a creole, but it has kept the pattern of simplified and regularized grammar relative to its root languages. One marker of this is that unlike most of the area’s Bantu languages, Swahili has no tonal system – a feature contact pidgins invariably drop.

The Indonesian archipelago is rife with dozens of trade creoles formed by contact among different Austronesian languages and others (including Chinese and Sanskrit). National Indonesian and Malay are themselves well understood to be old trade creoles, though in the normal way of such things the “creole” label is seldom applied.

A conquest creole is a case like Middle English where an incoming military elite forms a contact pidgin with the natives that displaces the native “pure” language. I have previously noted that this seems to fit what happened to middle Anglo-Saxon in the Danelaw beginning 150 years earlier. If there was a still earlier creolization event mixing Anglo-Saxon and Celtic, Middle Anglo-Sazion too would have been a conquest creole.

A thing to look for in conquest creoles is the formation of a creole continuum in which the language of the invaders is the acrolect, a relatively less modified version of the indigenous language is the basilect, and individuals routinely code-switch from lower to higher forms and vice-versa without recognizing that this involves not just a change in vocabulary but substantial morphological shifts as well. I don’t think you get this in trade creoles, where social relationships along the contact front are more horizontal.

Another example, helpfully demonstrating multiple creolizations in a language’s back story in case we are tempted to think of English as unique, is Maltese – grammar from Arabic conquerors, vocabulary from Romance-speaking subjects. Recently (last 150 years) “old” Maltese has been largely replaced by an Anglo-Maltese creole.

Next, the camp creole. This is a creole formed from a pidgin invented as a military command language in a multilingual empire.

The best known of these is Hindi/Urdu, the latter name for which literally means “camp language” (it’s related to Mongol “ordu” and the derived English word “horde”). It originated as a military pidgin formed by contact among a largish group of related North Indian languages around the 7th century CE (so, about as old as Swahili).

I tripped over a minority theory of the origin of modern German a while back. The usual story about this is that it’s the language of Luther’s Bible, but apparently some experts think it is more properly viewed as having originated more recently as a military pidgin in Frederick the Great’s Prussia. At the time there were different so-called “dialects” of German that were quite mutually unintelligible, so this theory makes functional sense.

My source even proposed that this is why modern German tends to verb-final order in sentences – as a way of making command verbs more prominent.

The label “creole” is, as you are probably expecting by now, not generally applied to either Hindi or Modern German. In the case of Hindi, though, it fits the generally accepted interpretation of the language’s history. And I’m betting the military-pidgin account of modern German hasn’t gotten quite the attention it deserves.

Our fourth variety is the court creole. This is like a camp creole, but instead of arising from a military pidgin it develops among the ruling elite in the capital of a multi-lingual empire. Because once your polity gets past a certain size you need to recruit administrators, servants, concubines and whatnot from places where the language isn’t yours, and they then have the same contact problem as a military camp.

There are a couple of solutions to this problem that don’t involve spinning up a new language. You might be able to actually impose your language on the natives; the Romans were pretty effective at this. You might be able to adapt a camp creole that your armies already speak; this is how Hindustani became a court language with a literary tradition. But if all else fails, your capital is going to grow its own creole because it has to.

The type example I had in mind for “court creole” when I began writing was Mandarin Chinese, but there was the problem that it’s tonal, a feature normally lost during pidginization. On the other hand, Wikipedia says straight up that Mandarin arose during the Ming dynasty “as a practical measure, to circumvent the mutual unintelligibility of the varieties of Chinese”, which certainly sounds like my court-creole notion in action.

The standard account describes Mandarin as a “koine”, which is elsewhere defined as a contact language arising among mutually intelligible varieties – and thus without undergoing drastic complexity reduction through a pidginization phase. That would remove the mystery about the retention of tonality.

But something is off here. We actually know that mutual unintelligibility was a problem at the time; one Emperor put a complaint on record that he couldn’t understand the speech of certain provincial officials, and founded language academies to attack the problem. This doesn’t really sound like the conventional koine-formation story.

We might be running up against an edge case for the “koine” and “creole” categories where it’s difficult to know which applies. Or linguists might be exhibiting their usual flinch about using the label “creole” outside of a dandies-and-octoroons situation. It’s difficult to know and I am certainly too ignorant of historical Chinese linguistics to justify a strong opinion.

What I think we can propose is that the specialists ought to take a fresh look at the period sources and see if they can detect any traces of what looks like pidginization (with its characteristic loss of grammatical complexity) in the period when Mandarin was forming.

What we can say without dispute is that like English and Malay (but, to be fair, unlike Swahili) modern Mandarin has retained a lot of pidgin-like traits commonly found in creoles – SVO analytic grammar, simple phonology, a spoken form easily acquired by adults from other language groups (even as its written form is infamously difficult).

Which brings us to the end of my speculation. I wish I had an unambiguous example of a court creole to lay down, but just this brief survey should make clear that creolization events have been both more common and far more important than you’d think from the linguistics textbooks.

Even the oldest attested human language may have been a creole. There are structural and lexical indications in Sumerian that it may have fused from a couple of rather dissimilar languages spoken still earlier in the Fertile Crescent!

So, hey, academic linguists, stop being such prudes about language hybridization, eh? It’s limiting your vision.

Odlyzko-Tilly-Raymond scaling

I’ve been ill with influenza and bronchitis for the last week. Maybe this needs to happen more often, because I had a small but fundamental insight into network scaling theory a few minutes ago.

I’m posting it here because I think my blog regulars cast a wide enough net to tell me if I’ve merely rediscovered a thing in the existing literature or, in fact, nobody quite got here before.

Back in the naive days of the dot-com boom people used to talk about Metcalfe’s Law: the value of a node of n networks scales as O(n**2). This heuristic drove a lot of the early excitement about Internet-related stocks.

But then the boom busted, and in 2006 along come Odlyzko and Tilly to explain that. What they said is: empirically, networks actually seem to top out at O(n log n) value, and this is superlinear but way lower than O(n**2), and thus dot-com boom fall down go bust.

You can read the Odlyzko/Tilly paper here. It’s good; it’s lucidly written and deserves its status as a seminal classic. The explanation of O(n log n) that the authors give is that in a world where not all connections have equal value, people build only the connections with the best cost-benefit ratio, and due to an effect called the “gravity law” the value of traffic between any two nodes falls off superlinearly with distance. This produces a substantial disincentive to build long-distance links, leading to a network of clusters of clusters with O(n log n) link density and value scaling.

After Odzylko/Tilly, complexity theorists looked at real-world networks and found that they frequently evolve towards a topology that is self-scaling or fractal – clusters of clusters at any scale you examine. Circulatory systems in the body, neural networks in the brain, road and rail networks in human cities, the Internet itself – over and over, we find self-scaling nets anywhere evolution is trying to solve optimal-routing problems.

So here is my small stone to add to the Odlyzko/Tilly edifice: their assumption in 2006 was stronger than it needed to be. You still get selective pressure towards an O(n log n) self-scaling network even if the cost of connections still varies but the value of all potential connections is equal, not variable. The only assumptions you need are much simpler ones: that the owner of each node has a finite budget for connection-building small in relation to the cost of providing links to all nodes, and that network hops have a nonzero cost.

To see why, we need to recognize a new concept of the “access cost” of a node. The value of a node is by hypothesis constant: the access cost is the sum over all other nodes of any cost metric over a path to each node – distance, hop count, whatever.

in this scenario, each node owner wants to find the best links to the network, but the valuation minimizes access costs . Under this assumption, everyone is still trying to solve an optimal routing problem, so you still get self-scaling topology and O(n log n) statistics.

That’s it. Same result as Odlyzko/Tilly but with weaker assumptions. Put their case and my case together and you have this:

The value of a network of n nodes will rise as O(n log n) under the following assumptions: (a) node hops have variable costs, and (b) each node owner has a small budget.

Is this in the literature anywhere? If not, I guess it’s worth my time to do a more formal writeup.

How to act like you’re bright

This blog post is brought to you by a recent bad experience I had watching a 5-minute clip from Big Bang Theory on the recommendation of a friend who thought I might find it amusing.

Bleagh. This is supposed to be a show about geniuses? It’s not. It’s a show about a dimwit’s idea of what bright people are like. The slowest person in my peer group could out-think and out-create any of these sad-sack imitations of “smart” on any day of the week.

These actors are not bright, and don’t know how to fake it on screen. It occurred to me that I have seen this pulled off occasionally; the example that leaps to mind was Jennifer Love-Hewitt playing a bright scientist opposite Jackie Chan in Tuxedo (2003). She did a good enough job that I was later quite surprised at how relatively free of the ravages of intelligence she sounds in propria persona.

Ms. Love-Hewitt must have been at least smart enough to know that she should emulate the mannerisms of very bright people, and then set about doing it. After thinking about this, I thought it would be entertaining (and possibly useful) to compile some actionable advice for actors finding themselves in a similar situation.

Here goes a list of bright-person behavior signals which, while not universal, are very common…

Bright people have very precise diction and tend to self-assimilate to educated speech norms even if their formal education is minimal. Enunciate as crisply as you can. If the character is designed to have a regional or lower-class accent, dial it down a little. [UPDATE: I may have overgeneralized a bit here. Strongly true of STEM geeks, but maybe not as reliably of other kinds of brights.]

Bright people concentrate. Their casual attention to a task or person is as intense as most peoples’ full attention. So fixate on those targets – not to the point of being glassy-eyed about it, but to the point where stillness and attention dominate your body language.

Bright people spend a larger fraction of their time in an ‘on’ state of mental alertness or conscious thought than non-brights do. This has consequences in visual saccades that are easy to see – with a little practice, you can grade people by intelligence in bank or movie-theater lines by watching eye movements. Look for relatively little time spent in a defocused, half-asleep state – or, conversely, lots of time when the eyes are tracking or making motions indicative of either imaginative activity or memory retrieval. Thus, when you play a bright person, always be looking at something.

Do not fall into the robot trap. Bright people are not emotionless, not at all. They do tend to be more introspective and more controlled, which makes their emotional signaling less obvious. A good way to approach this mental stance is to behave like someone who is seeing wry, dry humor in everything.

The most common minor failure mode of bright people in dramatic situations is that they’ll have visible difficulty tolerating stupid behavior by those around them. The thing to get is that this is not egotism and shouldn’t be pushed out that way unless the character is an asshole by design; playing it not as assholery but as weary exasperation is usually truer.

Bright people move differently. This one is complicated. If they’re naturally physically graceful, the always-on/full-attention trait amplifies that a lot. If they’re physically clumsy, they may still exhibit startling if confined physical competence in trained skills – typing, playing a musical instrument, martial arts, whatever. What ties this together is that they’re good at all kinds of learning, including learning to use whatever physical ability they have efficiently.