Average vs. cumulative outrage

There’s a ceiling on how much outrage I can feel at any given moment, much like there’s a limit on how much I will consider paying for a set of dishes, even if that set contains 10,000 bone china plates.

Over the past week I’ve seen, as you have, a string of successive and increasingly shocking affronts to human decency from the President and his advisors, which should have added up to at least a few instances of my head literally hitting the ceiling.

The thing is, outrage doesn’t seem to add up in this way, and rather than each new bit of news causing me to hit the ceiling, it has simply been added to the lively simmering crock pot of intense disappointment I’ve had in my head for the last few weeks.  I get the feeling most of my liberal friends feel somewhat like this, and are just motivated enough to tweet, complain on facebook or maybe send the ACLU some money via Paypal.  I’m not criticizing this, just remarking that the emotional state of many liberals is less:


and more:


This isn’t a political post at all, actually.  It’s just something that I noticed about the way I feel about Trump and the news that might point to something interesting about how people think in general.

People, it shouldn’t need to be pointed out, are poor intuitive statisticians.  This much is obvious, and provable by trying to explain any statistic to anyone, ever.  As a species we seem to have a counting module that can think about quantifiable things only as “one” “some” or “a lot”.  What is interesting is that this tendency applies to things that don’t even seem quantifiable, like feelings of outrage or indignation.  I can be outraged at one thing, outraged at some things, or outraged in general, but my subjective experience during the last of these isn’t all that much more intense than the first.  I have an upper boundary on how much of any particular emotion I can feel, and more input that would push me in that direction simply escapes and is lost as outrage radiation.

On the other hand, any countervailing information that I get cancels out far more outrage than it should.  If I hear that Trump’s son might have a disability, or I see people making fun of Melania for speaking learner’s English, I can pretty quickly forget about the last 5 terrible things that Trump himself did.  The bad things he does and the good (or at least not bad) things about him are both quantified in my brain as “some stuff” and weighted surprisingly equally.  Whenever I am made to recall at least one redeeming thing about him, my outrage drops down from its ceiling to “some outrage”, until some fresh news item (or just remembering the last one) pushes it back up.

It’s as if my outrage is averaged out rather than summed.  Rather than adding up travesty after travesty to get to 10,000 travesty points before subtracting 100 because he seems to love his children, all the travesties are normalized to within a narrow range and then have positive things I know about him subtracted.

The principle in action here seems to be a variation on the principle outlined (as many great principles are) in Daniel Kahneman’s Thinking, Fast and Slow, that the value of any given set is usually thought of as the average of its components rather than their sum.  That is, a set of 10 like-new dishes is priced higher than a set of 12 like-new dishes and 3 broken ones.  The broken ones seem to taint the set as a whole, bringing down the value of the entire package, although the package still contains at least as many like-new dishes as the alternative.  Ergo, as long as any number of things are conceived as a set rather than taken individually, their value is likely to be considered as a mean rather than a sum.  I’m surprised and a bit disappointed (in addition to fearful of what this could mean for how we think of Trump during his administration) in light of the fact that this principle seems to apply just as well to our feelings about a person’s set of actions as to our valuation of sets of dishes.

(I had a class in which I demoed this principle improvisationally on pieces of paper, handing each student a different description of a set of dishes and asking them to price it.  The principle was proven in real time, to the surprise of several managerial types in attendance.)

Things language teachers know #1 – impermanence

A truth you’re exposed to pretty early on in your training in SLA is that correctness is a matter of making your utterances target-like rather than meeting some objective standard.  This is because those who insist on a purer, correcter version of their language that nobody happens to speak are, well, incorrect – the rules of languages are defined by the people who use them.  Target-like means similar to the community that you want to be a part of, and if that community changes its mind, then what is “correct” changes too.  We EFL/ESL teachers help our students to be members of new language communities, not learn objective facts.  What appears objective and true about the rules of English only lasts as long as English speakers’ belief in or maintenance of them.  That’s right, languages and Tinkerbell have something rather crucial in common.

I understand the need to feel like you’re part of something permanent, or are playing by rules that are not just someone else’s opinion or the product of consensus.  If it bothers you, it’s best not to think about it, but impermanence is the law to which all the things we experience are temporary exceptions.  The groups at various levels of abstraction that you consider yourself part of, the morality you espouse and sometimes observe, and the language you speak are all are.. well, you know who said it best.

The dust and the wind are also impermanent.
Read More »

Machine translation requires human-like AI

Some of you alive in the 90s might remember an episode of Star Trek: TNG that is held as an example of the philosophy of language showing up in popular culture.  In this episode, the voyagers of the Starship Enterprise arrive on a planet where the inhabitants speak a highly allegorical language, using phrases about mythic or historical figures a la”Shaka, when the walls fell” to convey messages such as “oops” or “I see your point”.  As a result of these literal translations, the Enterprise’s crew members are forced to decipher what the dense metaphors mean contextually rather than in their normal English idiom as the universal translators usually supply.  Universal translators, as you can probably guess, are supposed to work with any language on the first encounter with that language or even with the species using it, and as far as I know this is the only episode where this particulary difficulty arises.


The problem is, if a universal translator can’t work with the very (infeasibly, as the article above points out) allegorical language spoken in that episode, it shouldn’t work with any language.  Even very closely related human languages use vastly different grammar and vocabulary to express greetings, thanks, obligation, and anything else under the Sol System’s sun.  To know that the verb phrase “thank you” is a show of gratitute in English (not a command, as verb phrases in isolation generally are), while an adverbial like “doumo” serves that purpose in Japanese, a universal translator would need to be a mind-reader before it was a translator, as there is no way to ferret out the fact that “doumo” and “thank you serve the same purpose from first principles or even from the grammar of that language (which universal translators don’t always have access to; they work on every language even on the first try). Moreover, it would need to do this mind-reading on species whose physiology it has never encountered before, meaning it would need to determine where the locus of that species’ cognition is, make intelligent predictions about how the patterns of (presumably) chemical synapse firings correlate to intentions, and map those intentions onto speech acts as they occur in real time.  The prerequisite technology for a universal translator is much larger than mere substitution and reordering of words, and approaches impossible, even by sci-fi standards.

In our world, people often discuss non-sci-fi machine translation like Google Translate as if it also were a scaling problem of existing technology, as if adding more of the same gears and cogs we already have would result in perfect language-to-language recoding.  In essence, people think the incremental improvement of current machine translation technology can save us from the years-long process of mastering new languages ourselves.  This post, with its oddly long prologue, is meant to argue that perfect machine translation would require a project of enormously grander scale than the visible inputs and outputs of textual language, and like the universal translators in Star Trek, would have a project of imposing complexity as a prerequisite, one whose implications would go far beyond mere translation.  In the case of machine translation that prerequisitive is a complete human-like artificial intelligence.

Read More »

Free will and vocabulary size

Criminals commit crimes of their own volition.  Society certainly plays a role in creating the conditions that produce crime, but ultimately the criminal is responsible for his or her own actions and should be held accountable as a free individual.

Words have identifiable meanings at their core.  People may use them incorrectly, or their meanings may be corrupted over time, but in order to understand them, use them, and teach them you need to know their real definitions independent of context.

I think both of these statements are wrong, but they have something interesting in common, which is a perceived need to regard certain things in isolation for some purposes when in reality they depend on other things for almost all their characteristics.  I intend here to draw a connection between what I studied in college (criminology) and some issues that haunt my current field of applied linguistics.  The connection that I will draw is one of an ultimately false atomized view of both human agents and words in language, and then I will show that the realization that atomized view is false nonetheless has very limited implications due to institutional restraints on how we can deal with human agency and how we can teach and test vocabulary.


That started off much more formally than I intended.  I guess the issues I’m going to bring up are fairly high-minded, but keep in mind I’m used to dealing with these topics in academic writing.  I will try to address the least educated of my readers.  No, not you.  The other guy.  You know the one.

Read More »

Strain Theory, Eigo and Eikaiwa

Criminology, being more or less a specialized subset of sociology focused on how societies treat deviance, offers many lessons applicable to education.  Classrooms and student bodies are types of little societies after all, and it stands to reason that they would have their own versions of deviance, criminality, and sanctions; these being necessary parts of any society.

Merton’s Strain theory was one of the most memorable theories from my undergrad years (nominally spent studying social ecology, an interdisciplinary major of which criminology was a component).  Strain theory, as I remember it without referring back to Wikipedia, attempts to categorize deviant (non-mainstream) behavior in terms of acceptance or rejection of mainstream goals and means.

In general terms, under strain theory one can accept or reject mainstream goals (e.g., getting into college) and mainstream means (e.g., studying hard) independently of each other – you can dismiss college as the goal of education while being a fierce autodidact, or hold college admission as a goal while gaming the system by cheating on your SATs, or (for some reason) cheat on your SATs while not intending to go to college.  Which behaviors fall under the different categories below, of course, depend on the means and goals particular to that culture and/or society.

The canonical example of strain theory has financial success as the goal and employment as the means.  One could accept the goal of getting rich while robbing banks to get there, and this would be Innovation.  One could also give up on getting rich while still clocking in every day, which would be Ritualism.  Or one could spend all one’s time feeding the ducks at the park, in Retreatism.

OK, I had to check Wikipedia to make sure I had these labels right.

I’ve made two new versions of the above graph describing cultures of English learning in Japan, in which the goals and means for learning English differ.  The cultures corresponding to these graphs are captured under the terms eigo and eikaiwa.  If you are unfamiliar with these terms, chances are you have never taught in Japan.  The two are dichotomized quite strongly, and as we shall see, mainstream values in one are often stigmatized in the other.  Briefly, eigo is closely aligned with mandatory education and eikaiwa with English as a means of international (mostly verbal) communication.  For more detailed treatments of these different cultures of English in Japan, check out my MA thesis and its list of references.

Also see Diane Hawley Nagatomo’s new book Identity, Gender and Teaching English in Japan for a comprehensive and practical review of how these two ideologies of English affect individual teachers in Japan.

Read More »