A Curious Miscellany of Items Philosophical, Historical, and Literary

Manus haec inimica tyrannis.

Wednesday, July 27, 2016

Our Robots, Ourselves

Anyone who has been following the US election spectacle — and really, the media have made it impossible not to — will by now realize that the insurgencies of Donald Trump and Bernie Sanders are a collective cry for help from the middle and working class. People are struggling, and the usual ways of doing things have not helped them one iota. Hollowed-out factory towns, main streets full of boarded up shops, and parents whose lives now revolve around heroin rather than around jobs that pay a living wage. These are the legacies of several decades of politics as usual.

Despite their quite radical ideological differences, there is an interesting and significant overlap between them on this issue: both acknowledge (or at least claim to acknowledge) the growing numbers of formerly prosperous Americans who have fallen through the cracks. I never in my lifetime thought I’d hear the presidential nominee for the Republican Party call for ripping up trade deals and penalizing companies that offshore jobs. That kind of Bolshevism is not the usual stuff from which Republican presidential hopefuls are made. And to find that these positions of Trump’s tally with those of Sanders, well, we live in interesting times.

However, even within these broad areas of agreement between Trump and Sanders, there are more subtle differences. Each of them blames the problems facing the working class on different enemies. Regarding trade deals, Bernie Sanders views the Trans-Pacific Partnership (TPP) as the devil’s right hand, while Trump views the North American Free Trade Agreement (NAFTA) as the worst trade deal in the history of the United States. Sanders blames offshoring for job losses, while Trumpian nativism tends to focus on the role of (illegal?) immigrants in supposedly taking jobs away from hardworking American-born citizens.

I read an article recently that made me think they were both missing what is — or is about to become — the biggest enemy facing the middle class, namely robots. The article was about how workers in China are being replaced by robots. The Chinese wage differential that was boosting their economy to such heights is no longer a comparative advantage. Soon, it seems, there won’t be anywhere left in the world where humans are cheap enough to hire. Embedded in the article were some statistics on the purchasing of industrial robots around the world that made for chilling reading. For the benefit of readers, I have charted the numbers and provided year-over-year percentage changes. Here is the picture worldwide:



These numbers are just for the past few years. Note particularly the large jump from 2015 to 2016. Now, here is the picture for North America (the US, Canada, and Mexico):



“Roboticization” is not as pronounced in North America, at least as judged by year-over-year increases. But I posit that this is simply an indicator of a mature economy that has been roboticizing for decades now. On the other hand, the chart shows a steady increase, without the dip in year-over-year increase that occurred in the worldwide numbers between 2014 and 2015. And the surge in the red line between 2015 and 2016 is still very pronounced in North America.

Now, what do these numbers have to tell us about the human impact of roboticization? The article mentions a kitchen utensils factory in Foshan, China, which replaced 256 workers with nine robots. If, therefore, we were to assume that each new industrial robot represents 28 jobs lost, it would mean that some 1,232,000 North American workers are poised to lose their jobs in 2016. Of course, many of these jobs will be Mexican, which won’t cause many Trumpists to shed tears. But still…

It might be the case that the robots in the Chinese example are extraordinarily efficient. For the sake of argument, let's instead assume that on average each new North American robot only replaces five workers — an admittedly arbitrary number. In that case, “only” 220,000 workers will lose their jobs to robots in 2016. But look again at that red line, at that year-over-year increase in robot purchases. If that increase only remains where it is, another 276,540 people will lose their jobs in 2017. The damage really begins to add up. And this is not a new process. It has been going on for awhile now. Is it possible that robots have something to do with all the empty factories, boarded up shops, abandoned homes?

So far, to my knowledge, this issue hasn’t made it into the campaign speeches of either Sanders or Trump (or anyone else, for that matter). They instead blame job losses on illegal “aliens”. Or on bad trade deals. Or on offshoring. Or, in Sanders’ more vague language, on “Wall Street”. Any of these might be a contributing factor to some extent. But look just look at at those numbers above. At some point in the not-too-distant future politicians will have no choice but take notice, as roboticization advances and becomes as plain as the nose on one’s face.

It may happen sooner rather than later if roboticization moves up the income ladder and starts gobbling up white collar jobs. We are on the cusp of a brave new world of robots, artificial intelligence, and big data, a world in which a chatbot can already outperform lawyers in overturning parking tickets.

Imagine it: a world without lawyers. Maybe they can replace our politicians too. I for one welcome our new robot overlords.


It raises some philosophical questions. If we are all fated to be replaced by robots and chatbots, what do we say about a species that makes itself obsolete? Or, if we find a way to structure our economy such that "benevolent" robots have merely freed us all from drudgery to do more pleasurable things, what do we say about a species that only lives for pleasure? What is the place of work within humankind's moral economy?

Monday, July 4, 2016

Piers Plowman and the Cock’s Egg

I have a friend who works at the same university I do, and last fall we decided to spend our lunch hours reading Milton’s Paradise Lost. Key to this experience was that we would take turns reading the text aloud and discussing it as our lights led us. It was an enjoyable and profitable experience. So much so, in fact, that once we were finished we decided to tackle William Blake’s Milton: A Poem.

That experiment didn’t work out nearly as well. We right away decided that we needed to have a look at some secondary literature. Problem is, the secondary literature seemed to be as opaque as the Blake text itself, in some cases more so. After three wasted lunch hours, we gave up on Blake. Certain things about the poem — including its tangled syntax and occasional lack of agreement between subject, verb, and object — led us to the conclusion that Blake (poor soul) was clearly mentally ill. The only explanation we could come up with for this literary mess was that he was schizophrenic. This diagnosis was more charitable than deeming him an outright fraud. Pity we could not extend this charity to the myriad academics who have made their careers in Blake studies, including, I’m sorry to say, my countryman Northrop Frye.

So setting aside Blake, we picked up Piers Plowman by William Langland. It’s rather heavy going. Compared to Chaucer’s Middle English, Langland’s is idiosyncratic, much further away from modern English, and the story is heavily allegorical. And although one should not expect standardized spelling in medieval English, Langland’s (or his scribe’s) is all over the place.

The edition we chose (the Norton Critical Edition) also does not help. It is a parallel text, with Middle English on the left page and modern English on the right. This is good, because unlike with The Canterbury Tales, I find I need a translation to refer to as I struggle through the original. Indeed, my friend and I have taken to reading a couple of pages of the translation aloud, and then going back and reading the corresponding Middle English aloud. However, in transcribing the text, the editors have chosen not to use the “yogh” or ʒ character, which would have been pronounced somewhere between a hard g and a y. Instead, they replaced every instance of it with either a g or a y — usually the latter. What this means is that a word like “ʒift” gets transcribed as “yift” when it should clearly be “gift”. It makes parsing a difficult text unnecessarily harder.

The translation could be better. At one extreme, the translators often leave untranslated and unglossed words that require explication. At the other extreme, they often kill the metre and alliteration by providing a bad or awkward translation for a word that is perfectly intelligible in the original — they obtrusively translate where translation is unnecessary. Although, the translation is laid out in verse, it is not what one could call poetic. In addition, they sometimes provide English translations for Langland’s Latin quotations in their places in the verse, while at other times they relegate the translation to a footnote, while at still other times they provide no translation at all.

But perhaps worst of all, they sometimes provide a translation that is simply wrong. An interesting example occurs at Passus VI, ll. 284-285, where Piers describes the simplicity of his diet. Here is the translation, followed by the original Middle English (italics mine):

      “And yet I say, by my soul, I have no salt bacon
      Nor any hen’s egg, by Christ, to make ham and eggs,”

      “And yet I sey, by my soule, I have no salt bacoun
      Ne no kokeneye, bi Cryst, coloppes forto maken,”

Once you realize that kokeneye is “cockeneye” or “cock’s egg”, you immediately see that the translation makes little sense. First, by default, all chicken eggs are hen’s eggs, so there is nothing to be added by describing them as such. Second, at the risk of stating the obvious, a rooster is not a hen. Hence, a rooster’s egg is not a hen’s egg. Hence, the translation is bad in a quite literal sense.

But that is not all. Rendering “cockney” as “hen’s egg” also misses much nuance. First, let us agree on one simple fact: roosters do not lay eggs. As such, a cock’s egg is an imaginary entity, like a unicorn or pixie dust. Thus, in a sense, Piers could be understood as saying, “Not only do I not have an egg. I don’t even have an imaginary egg.” (i.e. he has less than nothing). That is conceivably the sense a reader might get if the translators had simply translated kokeneye as “cock’s egg” or “rooster’s egg”. At the very least, some of Langland’s earthy colloquialism would have been preserved.

Langland’s Piers is trying to convey a notion of scarcity. For him, food beyond the very basic is hard to come by. It is worth noticing that there is a related figure of speech still used today to convey the notion of scarcity: we sometimes describe a thing as being “scarce as hen’s teeth.” Just as roosters don’t lay eggs, neither do hens have teeth. In both cases, to describe something as a cock’s egg or a hen’s tooth is to describe a nothingness. By contrast, to use the term “hen’s egg”, as the Piers Plowman translators do, is to refer to something that is extremely common, thus undermining what I take to be the author’s intention.

Interestingly, Langland’s is the earliest citation for “cockney” in the Oxford English Dictionary. It is a word with an interesting history. Its primary use was to refer to an egg that was small and misshapen. This may have been the way Langland intended the word, if you wish not to accept my more figurative glossing of a cock’s egg as a fictional entity used to describe a nothingness.

However, shortly after Langland’s time, “cockney” came to refer to an overly coddled child, and, by extension, to a “milksop”. It could also refer to someone with overly refined and delicate tastes. This is the sense intended by Robert Burton in The Anatomy of Melancholy (1628), when he writes of people who are “overprecise, Cockney-like, and curious in their observation of meats.”

By 1520 “cockney” was being used to refer to town-dwellers, presumably because they were perceived as being soft and effeminate. I posit that there is another reason that the term came to be applied to townies: in The Passions of the Minde in Generall (1604), Thomas Wright remarks that “Sundry of our rurall gentlemen are as well acquainted with the civill dealing, conversing, and practise of citties, as many Kockneis with the manuring of lands, and affayres of the countrey.” In other words, it wasn’t just that townsfolk were effeminate; they were also ignorant — ignorant of agriculture, husbandry, and rural pursuits. They were the kind of people that wouldn’t know the difference between a rooster and a hen, or whether his eggs came from the one or the other.

My theory here is supported by a couple of lines from “The Reeve's Tale” in Chaucer's Canterbury Tales, written just a decade or two after Langland wrote Piers Plowman:

      “Whan this iape is told another day
      I sal ben halden a daf a Cokenay”

As daft or clueless as a Cockney.

Funny enough, the Norton Critical Edition translators of Piers Plowman were not the first to trip over this word. In a footnote, the editor of an 18th-century edition of Chaucer’s Poetical Works (Edinburgh, 1782) glosses Chaucer’s use of “cokenay” in the above passage as follows: “That this is a term of contempt, borrowed originally from the kitchen, is very probable. A cook, in the base Latinity, was called coquinator and coquinarius, from either of which cokenay might easily be derived.” He then cites our very passage from Piers Plowman and says that “it seems to signify a cook”. This explanation obviously makes no sense whatsoever as applied to the Piers passage. He then goes on to suggest that it might also have something to do with Cockaigne, the mythical land of idleness and luxury. Again, rather a stretch — although one might see a plausible path leading from Cockaigne to an idle and pampered Londoner.

In short, “cockney” is a word with a very rich web of overlapping significations. I’m not completely sure what Langland’s precise intention was in using it, though I’ve given you my theory. But what I do know is that he didn’t intend it to mean “hen’s egg”.