In an awesome paper [1] published in 2007, it was shown that irregular English verbs have been dying out (i.e., "regularized") at an incredibly precise and measurable rate. The paper shows "how the rate of regularization depends on the frequency of word usage. The half-life of an irregular verb scales as the square root of its usage frequency: a verb that is 100 times less frequent regularizes 10 times as fast."
I recently felt that effect in an old norse class here in Germany.
I had to translate brast into German.
Bresta follows the 3rd class of strong verbs (old norse still has more or less rules for the strong verbs):
bresta, brast, (brustum,) brostit - to burst, bursted, burst (archaic: to burst, brast, bursten)
The same verb also exists in German:
bersten, barst, geborsten
However it's not that common today and the past tense is also extremely uncommon in speach. This led me inflect it weak (regular).
It is indeed always interesting to see that it is not only the common words but also those little subleties, which are not noticable at the first glance, have survived in the different languages.
As Steven Pinker points out in this very interesting article, the category of irregular verbs in English will surely shrink over time, as old, little used verbs become regularized over time, but new verbs formed in English are never formed as irregular verbs. For all that, though, "many of the irregulars can sleep securely, for they have two things on their side. One is their sheer frequency in the language. The ten commonest verbs in English (be, have, do, say, make, go, take, come, see, and get) are all irregular, and about 70% of the time we use a verb, it is an irregular verb. And children have a wondrous capacity for memorizing words; they pick up a new one every two hours, accumulating 60,000 by high school. Eighty irregulars are common enough that children use them before they learn to read, and I predict they will stay in the language indefinitely." Cool.
> the category of irregular verbs in English will surely shrink over time
I just don't think that's true.
English, like all other languages, is in an unbroken line of continuous speech from the first group of humans to use speech at all. (Or if you don't follow the theory of a single origin of language, an origin set of language users.) Where you place the borders between what you call "English" or "Middle English" or "unnamed branch of Old High Saxon spoken by a particularly small group of villages" is fuzzy, but nobody ever decided "What I was speaking 10 minutes ago was Language A but starting now this is called Language B and they're totally different in every way".
So we've been developing English for a very long time. If "little used verbs become regularized over time, but new verbs formed [are] never formed as irregular verbs", then why do we have still have irregular verbs at all? Why wouldn't they have been wiped out thousands of years ago?
In the most obvious cases, English hasn't changed at a uniform rate. It's experienced gradual splits, mergings, conquerings, being alternatingly a vulgar and prestige dialect, immigration, emmigration, wars, trade explosions, and regular old influence of other languages nearby.
The claim you're quoting seems to be that if a language is left to its own devices that it will gradually approach regularity. I definitely disagree that point, and there are somewhat well-understood methods for these changes it occur even in an isolated language (for instance, vowel shifts that affect some words more than others, after which words that used to follow the same rule no longer do). But let's set that aside. Even if we ignore the normal linguistic processes that can increase irregularity in an isolated language, what makes us think that the "artificial" events like wars or interactions with other languages will decrease? Why would those things stop?
You make several false assumptions or gross misunderstandings of how language changes.
The reason irregular verbs originally existed is because English is a Germanic language, meaning, that a thousand years ago, what we now call irregular verbs were actually regular verbs. There was a logical ordering and well understood way to modify the stem of a verb in order to agree with the subject and tense. At the time, irregular verbs changed their stem, but in a predictable and universal manner.
This all changed with the conquest of the Norman French. After the slow introduction of French and Latin into English, the verb forms we are now familiar with entered the language. In time, all verbs of Germanic origin were used the same way they had been used in Old English, while verbs brought in by the French used the forms we are familiar with today.
The reason it makes sense to see language regularization is because we have grown to expect the French form of verb agreement rather than the Germanic one. The Germanic system is more complex.
In short, the reason that "new verbs formed [are] never formed as irregular verbs ..." makes sense, is that the change from irregular Germanic verbs to regular French verbs is only about 800 years old.
Sorry but this is totally wrong. The weak inflection in english with a dental suffix is an proto-germanic invention shared with all germanic languages.
The reason why there are so many weak verbs is that the rules for strong verbs complicated over time so that the rules weren't obvious anymore and that the verbs were seen just as
irregular (there is thought to have been just two verb classes originally in PIE http://en.wikipedia.org/wiki/Indo-European_ablaut, but there are already 7 of them in the old germanic languages http://en.wikipedia.org/wiki/Germanic_strong_verb#Strong_ver...).
Therefore new verbs couldn't be integrated as strong verbs by analogy.
> You make several false assumptions or gross misunderstandings of how language changes.
That's a little over the top. I only made a few claims, none specific to English:
Claim #1: There a processes that create and remove irregularities in a language. Some of these can occur in isolation, some are interactions with other languages.
Claim #2: At the very least, the ones resulting from interaction will continue to occur in the future. That is, we still have wars and immigration.
Claim #3: Implicit to my argument is my opinion that the creations occur approximately at the rate of the removals.
Certainly #3 there is the hardest to justify (and I only offer "they aren't gone yet" as my evidence). A much easier and not too different claim is "we won't reach 100% regularity" and if you like you can pretend that I argued that instead as it's not too different a claim. But "gross misunderstanding" seems unfair.
So which of these is a gross misunderstanding of language changes?
Presumably, new irregular verbs used to appear, but no longer do. Groups of speakers are less isolated, we have a standardized written language, and demarcations between languages are clearer.
It's (supposedly) pronunciation changes that cause irregular words, right? You start with a regular rule, pronunciation of all words change so that the rule no longer works, all the uncommon words quickly regularize on a new rule, and what's left is irregular.
It's been a while since my last linguistics class, but that seems like the natural result of words retaining their pronunciation roughly according to how often they're used. That sounds like a continuous process we should expect to slow as pronunciation change slows (due to recorded media), but not necessarily to stop at some cut-off point.
> Presumably, new irregular verbs used to appear, but no longer do
Why not? Have now we invented all of the words we need, but hadn't yet invented "seen" or "got" in 1200 AD? Then why would it stop now, instead of in 1200 AD?
> we have a standardized written language
But we've had written language for much longer than English has been around
At one time, the "irregular" verbs were regular: English (or rather Proto-Germanic, the common ancestor language of English, German, Swedish, etc.) had two classes of regular verbs, strong and weak, and newly invented verbs could be added to either class. However, the strong class over time became somewhat obsolete and irregular, and is now relatively closed to adding new members (though not completely: sneak/snuck, dive/dove and shit/shat are relatively new additions -- certainly long post-dating the irregularization).
Written language has become more prominent, recorded language has been introduced, and more of us experience some formal instruction in what the rules are. Those seem all like reasons new words are more likely to conform to the rules.
>So we've been developing English for a very long time. If "little used verbs become regularized over time, but new verbs formed [are] never formed as irregular verbs", then why do we have still have irregular verbs at all? Why wouldn't they have been wiped out thousands of years ago?
we haven't been developing English for thousands of years. Who knows what will happen then! We'll be speaking in python no doubt.
The low frequency of use of many of the irregular verbs in English and their restriction to activities mostly not talked about at home ensures, as Pinker correctly points out, that they are not part of the "unbroken line of continuous speech" that you write about. My ancestors trace back through the paternal line to English speakers for as far back in history as you can go and still find speakers of English. And the last few generations of my family have included college-educated English majors who were very fussy about proper rules of English grammar. But none of that passed on to me in living speech the "correct" (irregular, or "strong") forms of some of the irregular verbs in English. The attested historical process, as Pinker correctly writes, is that the category of "strong" verbs in English and in other Germanic languages has shrunk over time.
For onlookers, I should list here the English irregular verbs. All of you will see for yourselves how rarely used some of them are, and how unlikely it is that a child would learn them from home conversation. Others, of course, the ones we all know, are the ones Pinker thinks will last indefinitely. The ones with alternate forms shown are likely to become regular sooner, I think.
Instead of posting a lengthy and yet incomplete list, perhaps next time you could oblige us with a shorter list of examples, a hyperlink to the full list, or both.
> As Steven Pinker points out in this very interesting article, the category of irregular verbs in English will surely shrink over time, as old, little used verbs become regularized over time, but new verbs formed in English are never formed as irregular verbs.
At least, until we start pronouncing "plated", "played", and "placed" so differently that kids wonder why they're all spelled the same :)
Choosing between more/er and most/est has always seemed quite random to me for two syllable adjectives. The problem with English is that there ALWAYS seems to be an exception to the rule.
If I may take this unique opportunity, does anybody know of a command line tool (say "past") such as wn [1] that could output something like the following?
$ past go
go went gone
I'm talking about a stand-alone program that would work offline. Thanks!
Your comment seems irrelevant - the article never claimed that english has the most (or even a lot of) irregular verbs. If you have some linguistic insights regarding irregular verbs in Polish I would be curious to hear them (I really mean it)
Your comment seems irrelevant - the grandparent never claimed that polish has the most (only many) irregular verbs. If you have some pedantic insights regarding relevant comments on Hacker News I would be curious to hear them (I really mean it)
> Not only is the irregular class losing members by emigration, it is not gaining new ones by immigration. When new verbs enter English via onomatopoeia (to ding, to ping), borrowings from other languages (deride and succumb from Latin), and conversions from nouns (fly out), the regular rule has first dibs on them. The language ends up with dinged, pinged, derided, succumbed, and flied out, not dang, pang, derode, succame, or flew out.
I'd love a quick straw poll. Who says flied out?
I've only ever heard flew out - I've never heard anyone say "I flied out yesterday."
Honestly "pang" (past tense) and "pung" (past participle) don't sound terrible to me, in analogy to sing/ring. Native English speaker from USA, if that matters.
The real question is how rule based natural language systems can evolve. Otherwise they are subject to becoming moribund.
Of course that's not simply learning the latest usages on the cutting edge of human language evolution. Rather, natural language AI evolves concurrently with the technology people use to communicate, which technology (heavily?) influences their language use.
I'd go so far as to suggest the influence of technology on language is more important than historical rules of verbiage and exceptions.
Long live the irregulars! Having said that, I would surmise that the past tense of a hockey stick is a tree. I just couldn't say which kind, I'm not an expert.
The thing with the most commonest verbs "come" and "go" is they're the same verb but with different directionalities, just as "come" and "came" is the same but with different tenses. Ditto "bring" and "take". So the form irregularity sits not only on the tense but also the direction of movement.
i've had a pet theory that irregularity helps memorization. the thought is that when something is deducible rather than arbitrary, you remember the rule, not the result of the rule. when its arbitrary, you just remember it. does anyone know of anything to suggest this is true or false?
Only barely related, but I do know that the speed limit in the parking lot at my local mall is 18 MPH. Not 15 or 20, but 18. And of course the only reason I know that is because they choosed an arbitrary value and not a standard "slow parking lot" value.
The fact that the 10 commonest verbs are all irregular seems to agree with your theory,
What's to say that "payed" won't end up winning in the end? I mean, it's a hell of a lot more logical, pay->payed, than pay->paid.
The really amusing part about loose vs. lose is that both have a negative connotation, so the similarity in sense lends a dissonance to the "attuned eye" not unlike two musical notes slightly off pitch played interchangeably.
If you want to control language evolution, perhaps you'd prefer the existence of something like the Académie Française and invented words like "courriel"?
Also possibly like geegled, from gögled, or goegled. The old way was for broad vowels to become slender, for either pluralisation of nouns or for thingumming of verbs. A, O, and U would be broad vowels, and I and E slender, and the whole bucket of dipthongs follow somehow.
Or should it be that the whole bucket of dipthongs follows? I, personally, support metonymic shifts.
[1] "Quantifying the evolutionary dynamics of language" by Erez Lieberman, Jean-Baptiste Michel, [...], and Martin A. Nowak http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2460562/