Discover Magazine has a neat little article on irregular verbs that doesn't seem to be online:
Now for frustrated English-language students everywhere, there is good news: a formula that predicts hwen an irregular verb will convert to the regular form. However, they might have a long wait.Derived by Harvard University mathematician Erez Lieberman, the formula shows that verbs have their own "half-life", that is, the time it takes for half of the verbs in a particular group to become regular. The half-life is based on how popular a verb is: the more often it is used, the longer it will take to convert. For example, the verbs have and hold both have irregular forms in the past tense--had and held. Yet have is used roughly 100 times more often than hold. In line with that, we should expect held to become holded in nearly one-seventh the time (in about 5,400 years) it will take for had to become haved (in roughly 38,800 years).
To come up with the formula, Liebrman compiled a list of 177 Old English irregular verbs, many of which had regularized. He grouped them according to how often they popped up in modern English, which correlated with how long it took for the irregular forms to disappear.
The next irregular verbs likely to fall include infrequently used ones like slink.
I suppose in the beginning, all verbs were irregular. Which suggests that we're falling from an Edenic garden of plenty towards a monotonous, sanitized future in which all the verbs are exactly alike. Paging Rod Serling . . .
"a monotonous, sanitized future in which all the verbs are exactly alike"
Well, that assumes no new verbs enter the language. But they will. Primarily from Spanish these days, but I suspect a strong Chinese influence in the future.
Remember, English is the result of a Norman man at arms trying to get in bed with a Saxon barmaid.
IIRC, (from high school, decades ago) Latin has just one irregular verb. But Latin's been dead for 1500 years.
Why no comment on verbs going the other way? In my lifetime, the formerly regular verbs "dive" and "sneak" have become irregular ("dove" and "snuck" having replaced "dived" and "sneaked"), at least in casual speech. I don't see a future in which all verbs are alike. All forms of address, though, may become "Dude" fairly soon.
You've stumbled into an area in which I have some expertise, so I can address some of this.
First, the loss if irregular verbs doesn't mean that all verbs will become regular, or that all verbs were once irregular. Rather, the very process that creates regular verbs also leaves behind remnants, which become irregular verbs. An example:
Once upon a time, all verbs formed their past tense with a vowel alternation like we see in sit/sat. These were the "regular" verbs. The predecessor of the "-ed" was invented, probably to deal with foreign words or new coinings that didn't fit well into the usual system. These words were originally the exceptional ones, but as time went on the new -ed ending became more and more popular. Most verbs were switched to the new past tense; those that didn't were the most commonly used verbs, which suddenly found that they had become "irregular".
Then the process starts again. In Modern English we've already entered round 2, with words "said" and "had" that clearly go back to *sayed and *haved, but have become irregular as the rules for pronouncing that -ed ending have changed.
(There are other processes that create new irregulars, like contraction and changes in pronunciation, but this explains why we never reach a state of completely regular or completely irregular verbs.)
@wiredog: By one definition "esse" is the only irregular Latin verb. By another definition there are dozens, as the preterite of Latin verbs is often completely unpredictable from the present form.
One of the authors has posted a PDF of the original paper from the 11 October 2007 issue of Nature, for those who are interested: http://www.ped.fas.harvard.edu/people/faculty/publications_nowak/Nature07.pdf
One day to town there came a noun. He said,
"My purpose here is fixed, it be
A person, place, a thing you see
may be described by calling me.
This I do, and nothing more."
He stepped off the walk into the street,
he failed to look both ways,
and noun with bus did quickly meet.
"Alas," he mused, from by the curb,
"For it now it seems I have been verbed!
They will call my name as an act of fate
and noun no longer will they relate!"
A passing verb overheard the word
the noun did utter in passing, and said,
"My regularity
you have slain!
And take my name
in vain!
The lexicon shall hear of this."
That's right, folks. Original mouse poetry, complete with free-form rhyme. Get it while it's free.
In my lifetime, the formerly regular verbs "dive" and "sneak" have become irregular ("dove" and "snuck" having replaced "dived" and "sneaked")
Have you checked on the antiquity of "dove"? I suspect you're wrong. In general irregular verbs in English track their counterparts in Dutch almost exactly, and the irregular "dive/dove" corresponds to the irregular Dutch "duik/dook". (If it were regular in Dutch, it'd be "duik/duikte".)
Not sure about "snuck" vs. "sneaked".
I wonder what effect, if any, there will be from English becoming ever more the 'Lingua Franca' with a greater percentage of English speakers being non-native?
One might expect that this to increase the rate of regularization, as non-native speakers have more trouble picking up all the irregular forms.
What about plead --> pleaded/pled
That one always seems to be in stasis. Or maybe I'm just imagining things...
I'm a big fan of "snuck", despite my strong desire to make English easier to learn.
Proposal: make it mandatory to form past tense for all verbs (except "to be") as "did [present tense form]"
e.g. "I went to the store." --> "I did go to the store."
Advantages: already a valid, common grammatical form; eliminates many irregular verb forms; is parallel with the form used for questions ("did you go to the store?")
Disadvantages: Places a bizarre emphasis on the affirmation of the verb that sounds like pleading.
Both "dove" and "snuck" are comparatively recent, but not within any poster's lifetime. (The OED has cites for both dating back to the 19th century.)
What I want to know is whether the same process is at work converting "its" (possessive of "it") to "it's".
After reading this: "Discover Magazine has a neat little article on irregular verbs that doesn't seem to be online" I was looking for offline irregular verbs.
Japanese has, if I recall correctly, two irregular verbs. But they make up for it by having the world's craziest array of suffixes for counting
Japanese has, if I recall correctly, two irregular verbs
Japanese grammar does not lend itself well to western classification; verbs get conjugated as adjectives (especially in the past tense), certain adjectives are grammatically similar to nouns with the exception of one particle, and the closest approximation of "is" doesn't actually mean "is," or get used everywhere you'd normally want an "is."
But yes, the counting suffixes are clearly insane.
Linguistics is a rather unfortunate subject, in that not only do people who've not the slightest idea how language works feel free to pontificate at length on it, but even academic publications (Nature, f'rexample) publish linguistics articles that are misinformed or wrong.
The problem with the Lieberman article is twofold: a) lexicostatistics is almost completely discredited among actual historical linguists (Lieberman is a mathematician, not a linguist); b) it ignores how irregularity is reintroduced into the system by analogy (in many, though not all dialects, 'dive' and 'sneak' are now strong verbs) or sound change. (Although, English does have much less morphological irregularity than it once did, but then, it has much less inflectional morphology).
Irregular Japanese verbs? The least of your problems. Try irregular *noun counters*. All Japanese nouns are functionally "mass nouns" like water, furniture, sugar, roughage, i.e. nouns that cannot be pluralized without specifying a unit. ("pieces of furniture" rather than "furnitures")
The more mass nouns, the more special cases you have to learn, and each *irregular* counter adds to the pile-on.
This means that the Japanese word for motorcycle (ootobai) would be better translated as "motorcyclage" (compare to "roughage").
(My friends and I had a temporary phase in high school where we'd get kicks out of converting nouns into mass nouns, e.g. "look at all the tablage!" hahaha! "I hope there isn't much carrage on the roadage, or we'll be late!" hahaha; although obviously we didn't know the terminology for what we were doing linguistically.)
So you can't pluralize "motorcycle", no, that would be too easy. You have to refer to "two mechanized instances" (ni-dai, two + counter for mechanized objects) of motorcyclage (ootobai).
Rob, correct me if I'm making Japanese sound more bizarre than it really is.
For those who were wondering what Dan and I were talking about, Person has explained it. Some counters, from memory: Flat objects = mai, long thin objects = hon/pon/bon, years (duration) = nenkan, years (age) = sai, buildings = ken. I'm sure Google would reveal many that I have forgotten. And of course people are counted with "ri" for 2 or less, using Chinese numbers, but "nin" for 3 or more, Japanese numbers.
It will drive you bonkers. However, I'm not ready to say that it's more insane the the German requirement of inflecting adjectives differently depending on whether they are preceded by an inflected definite article.
Glad to see this story about the paper of a friend of mine on a blog that I read regularly. Erez is a brilliant and great guy... I'm sure we'll see more intriguing findings of his in the years to come.
Too bad the guy didn't tell us how much longer we need to hang on to the much abused objective case in English pronouns. I look forward to the day when I can finally think of the two sentences
- It seemed clear to me.
- It seemed clear to John and I.
as grammatically correct.
Death to the objective case!!
Re: Latin has just one irregular verb.
That's because Latin grammarians cheat and fail to class as irregularities the huge numbers of unpredictable perfect tense roots and past participle roots-- they just say that when you learn a Latin verb you need to learn three forms for each verb and then you know all of them. That's true for most verbs in English as well. In fact, the English irregularities often come from the same source as the Latin ones, the tendency in Indo-European languages to form alternate roots by varying the stressed vowel of the word.
I get the feeling that making a prediction of 5,000+ years vs. 30,000+ is the nod-nod, wink-wink equivalent of rejecting the predictive power of the model. I mean, 5,000 years is so preposterously meaningless when it comes to languages. Anyone has the article around? How long is "slink" going to take? Is it meaningful if it's 100 years? Or 20? Which makes me wonder about its explanatory power too. How well did it fit the data? Perhaps it's just the functional form that's wrong, but as some commentators suggest, it seems that the underlying idea, as neat and convincing as it first sounds (one of those "ah ha! of course!"), does not fit the "institutional" detail of the process as the target (what is irregular?) keeps moving.
I second Beth's comment quite enthusiastically, and I have a minor quibble with JSBangs', though it's mostly correct. The /Id/ ending in English that currently forms the regular past tense is only an innovation in its distribution from Proto-West Germanic; some kind of alveolar stop as an imperfective marker is present in German and Dutch as well (though now confined largely to very high register, literary German and Dutch). Of the top of my head, I think it goes back to Proto-Indo-European as well, having been an aspect marker alongside the ablauting processes that did most of the heavy lifting as far as past tense was concerned in Proto-Germanic, as JSBangs said.
Re: The predecessor of the "-ed" was invented, probably to deal with foreign words or new coinings that didn't fit well into the usual system.
Actually, it spread from the participle, though it was not universal there either (there are vowel-changing participles, some with -en, like "sung" and "frozen" too). The cognate ending exists in Latin with a "t" (laudatus, auditus, etc.).
JonF: If what you're saying about Latin is true, I'd have to agree. English, by that definition, also only has one irregular verb. See, if three forms define the verb, then yes, only one verb in English ("to be") isn't fully defined by infinitive + past + participle.
But isn't there more regularity in Latin in prepositions? One problem in English (and German, to my knowledge) is that the preposition you use to describe the relationship between a verb and a noun is non-obvious.
e.g. How do I know to say, "I'm dancing *to* 'Yellow Submarine'" instead of "I'm dancing *about/with/at/on* 'Yellow Submarine'"?
Doesn't Latin have this problem knocked out by use of cases?