Robert asked, whether there a general term for the shift we see when

'a nadder' becomes 'an adder,'
'a naranja' becomes an orange,'
'a napron,' becomes 'an apron,' or
'an ewt' becomes 'a newt'.

And tsuwm answered, "This sort of 'false division' is known linguistically as metanalysis." That was a new word to me, and a little research proved to be interesting. Results below.

Clearly metanalysis has been used that way, but has it caught on enough to be considered to be the word for this linguistic change?

Among the one-look dictionaries, only MW (plus one of the odd-word sites) lists that word with that meaning. Even MW doesn't seem convinced, for when it gives the etymologies of words that that have undergone such a change, such as "apron", it calls the change a 'faulty separationfaulty separation', but not a 'metanalysis'. On-line Etymology also call it a 'faulty separation'; Compact Oxford calls it a 'wrong division'; and AHD, Encarta and Chalmers all avoid using an specific term.

What about usage elsewhere? Googling metanalysis gives a lot of misleading hits, involving another unrelated meaning of that word as 'a statistical technique where all data from all available studies of something are combined'. But a search for only the linguistic usage (by googling metanalysis norange OR napron OR noranja OR nadder) you'll find only 36 hits, far fewer than even 'epicaricacy' gets! Admittedly you wouldn't expect many hits on so narrow a subject, but over 900 sites talk about norange OR napron OR noranja OR nadder in the sense of words (not place names), and very few of them call their evolution 'metanalysis'. Nor does any other term seem particularly oft-used.

My conclusion is that no word has yet caught on for the meaning Robert describes. So the meaning still awaits an accepted name. My own preference would be 'faulty separation', as in MW, for that term, unlike 'metanalysis', wouldn't force you to run to a dictionary, and it would be unambiguous.

My conclusion is that no word has yet caught on for the meaning Robert describes.

You are, of course, free to call metanalysis whatever you like. tsuwm was just pointing out what the technical term in historical linguistics is.

The Oed Online gives this definition of metanalysis:

quote:

---

The reinterpretation of the form of a word resulting in the creation of a new word; esp. the changing of the boundaries between words or morphological units (see quots. 1966, 1972).

---

It gives citations from 1914 to 1989. Apparently the word was coined by Otto Jespersen in 1914.

I found this on the Catholic Information Network, of all places:

quote:

---

Date: Sat, 26 May 2001 00:07:52 -0400
From: "H. R. Stockert" <the.avatar_at_att.net>
Subject: Today's word: metanalysis

The history of ourselves and our nation lies in the (sometimes) strange origins of how words work their way in to our common language.

Today's Word:

Metanalysis

Metanalysis is a technical term in linguistics that was coined in 1914 by Otto Jespersen, a Danish linguist who was one of the great grammarians of English.

Metanalysis can be broadly defined as 'a reinterpretation of the division between words or syntactic units', as the Oxford English Dictionary does. A usual example in English is the shift of the letter n based on a misunderstanding of the previous indefinite article a/an. So--and I'll be oversimplifying some of the changes here--the word adder was originally nadder, but the phrase "a nadder" was reeinterpreted as "an adder." Likewise, apron was originally napron.

In the other direction, newt 'type of salamander' was originally ewt, but "an ewt" was taken as "a newt." And nickname was originally an ekename (eke is an archaic term for 'also').

Metanalysis is also used to refer to syntactic interpretations. Jespersen gave the following example: "It is good for a man/not to touch a woman," which could be apprehended as "It is good/for a man not to touch a woman."

Sometimes metanalysis is given an especially broad range, where it is applied to any process of counter-etymological regularization--as when the plural of Walkman becomes Walkmans instead of Walkmen, or when the nonce past tense of sing-song is sing-songed instead of sang-song. But it's likely that by the time you encounter the word used in such a restricted way, you'll already be a linguist, and won't need my help anymore.

---

Tinman

This message has been edited. Last edited by: tinman,

coined in 1914 by Otto Jespersen

"I have ventured to coin the word 'metanalysis' for the phenomenon frequent in all languages that words or word-groups are by a new generation analyzed differently from the analysis of a former age. Each child has to find out for himself in hearing the connected speech of other people, where one word ends, and the next one begins, or what belongs to the kernel and what is the ending of a word, etc. In most cases he will arrive at the same analysis as the former generation, but now and then he will put the boundaries in another place than formerly, and the new analysis may become general. Familiar instances are a nadder which through metanalysis becomes an adder, North Thriding which becomes North Riding, sur + ound which becomes surround and is felt as if from round, vegetar-ian which is felt as veget-arian and gives rise to fruitarian and even nutarian. I shall give elsewhere a detailed classification of various kinds of metanalysis."

[Otto Jespersen Modern English Grammar vol. II, ch. 5.61 Numerical Metanalysis]

Thanks for tracking down that down, Tinman. It is interesting that two of the best historical grammars of present day English were written by foreigners: Otto Jespersen and Hendrik Poutsma. Jespersen's Language and Philosophy of Grammar are still in print and/or easily obtained. There great reads.

I feel vindicated!

jheem quotes Jespersen himself as giving this example of 'metanalysis': "vegetar-ian which is felt as veget-arian and gives rise to fruitarian and even nutarian.. This, of course, is not the same shifting of a letter or sound within a phrase, as Robert mentioned; instead, the existing word is broken apart in a new way, and then one resulting part is attached to a different base. Jespersen defines 'metanalysis' as change from a new division. His definition includes but is not limited to the sort of change that Robert mentioned.

Similarly from Grover Hudson, Essential Introductory Linguistics: "... a process of ordinary language change in which the stream of speech is analyzed in a novel way, known as metanalysis. Wrong cutting is a subtype of this, in which a word boundary is mislocated." Later, he says, "Metanalysis was noted as the process involved in backformation. Learners and even mature users of a language analyze a word, phrase or sentence in a new way, and then start using the parts of that analysis. They may do this as a means of being clever, or because of imperfect learning. One of his examples is very like the Jespersen example above: alcoholic, treated as if the olic were a suffix, yields chocoholic and workaholic.

Some authorities use the term differently. Bill Bryson seems confused at page 63 of The Mother Tongue. He defines 'metanalysis' as the process of detaching an initial n from an noun; attaching that letter is distinguished as "a similar process". But his examples of 'metanalysis' cover both cases. Here is his text in full:

Other words underwent changes, particularly those beginning with n, where there was a tendency for this letter to drift away from the word and attach itself to the preceding indifinite article. The processes is called metanalysis. Thus a napron became an apron, a nauger became an auger, and an ekename became (over time) a nickname. By a similar process [emph. added], the nickames Ned, Nell, and Nan are thought to be corruptions of "mine Edward," "mine Ellen," and "mine Ann."

PS: All this discussion of what 'metanalysis' means is, of course, seperate from the question of whether Jespersen's coinage has become an accepted word for the meaning.

This message has been edited. Last edited by: <wordnerd>,

The general term I use is 'reanalysis', and more specifically 'resegmentation'. I find this latter is not in my dictionary of linguistics and there are almost no web hits for this sense, so perhaps I made it up. The word 'metanalysis' was new to me, so it's not in common use in my reading.

This message has been edited. Last edited by: aput,

I feel vindicated!

Glad for you, but Jespersen specifically mentions a nadder where it crosses word boundaries. Another example from Tinman above was from syntax. The term that I learned in historical linguistics is metanalysis, and that's what I'll continue to use. I'd think that using the term faulty separation will entail an explanation to people who are not familiar with metanalysis anyway. If you concentrate on morphemes being reanalysed rather than words, some of the difficulty goes away. Some morphemes are free standing like words, others are bound to other morphemes like affixes and clitics. I like metanalysis (or perhaps morphemic resegmentation or reanalysis) because they are neutral as to the final outcome, whereas faulty separation implies some error.

Amazon allows you to search the texts of its books, to the extent publishers have allowed. To save other researchers the trouble of repeating work done, here's a list created that way of usages of the term metanalysis, eliminating non-linguistics usages. You can click to the actual text using that word. Great tool!

1. Bill Bryson, The Mother Tongue
2. David Crystal, The Cambridge Encyclopedia of the English Language
3. C. T. Onions, Robert D. Eagleson, A Shakespeare Glossary
4. Tom McArthur, The Oxford Companion to the English Language
5. Barbara Fennell, A History of English: A Socioloinguistic Approach
6. Grover Hudson, Essential Introductory Linguistics
7. R. L. Trask, Larry Trask, Historical Linguistics
8. Andrew Spencer, Arnold M. Zwicky, The Handbook of Morphology
9. Roger Lass, Old English : A Historical Linguistic Companion
10. Tom McArthur, The Oxford Guide to World English
11. Stuart C. Poole, An Introduction To Linguistics
12. John C. Wells, Accents of English 1: An Introduction
13. Kenneth Muir, Shakespeare Survey: Volume 23, Shakespeare's Language
14. Michiel de Vaan, The Avestan Vowels
15. Raymond Hickey, Motives for Language Change
16. John Edward Terrell (Editor), Archaeology, Language, and History: Essays on Culture and Ethnicity
17. Bhadriraju Krishnamurti, The Dravidian Languages (Cambridge Language Surveys)

Aput, as to your term 'reanalysis': as I recall, at least one of these texts uses 'reanalysis' in the broad sense suggested above for 'metanalysis'; it confines the later word to the narrow sense of the nadder/adder shift.

This message has been edited. Last edited by: <wordnerd>,

And, of course, there are other definitions for meta-analysis.

Meta-analysis (and I vaguely recall mentioning it here before) is gathering all available studies on a subject and re-analyzing the data together. In nursing, especially, we tend to do small studies and a met-analysis is an excellent way to find what the data really means.

That's interesting, I own Spencer and Zwicky and can check around it. They only use 'metanalysis' once, in a sentence mentioning Jespersen and newt, but around it they're calling it reinterpetation. (And it's not in the index.)

Larry Trask writes 'nice examples of reanalysis involving nothing more than the movement of a morpheme boundary, a type of change impressively called metanalysis', in his typically sharp-tongued way implying that it's an unnecessary word.

Wells does use it once, talking about the sentences 'Justice was metered out', 'The church is foundered upon a rock' in accents such as Australian where the mistakes can arise by metanalysis: [mi:t@d] = [mi:t@ + d] rather than [mi:t] + [@d].

McArthur (4) simply defines it as the term for it, with Jespersen's name and example, but doesn't use it. Crystal likewise defines it in a list between metalanguage and metaphor but doesn't use it.

McArthur (10) uses the term as one of a list applying to Strine, but doesn't give an example; however, one obvious one is Strine's author Afferbeck Lauder.

Poole uses it once and also defines it at the back. Onions uses it once. So it seems to be a known term for the newt phenomenon, but not actually in common currency. The re- terms like reanalysis and reinterpretation have the advantage of being transparent in meaning.

Aha! Here it is in Ciardi! He calls it nunnation or nunation, but proposes a new term.

quote:

---

noncing/to nonce More commonly labeled nunnation/nunation after Arabic nun, the letter n. But Arabic nunnation is the addition of a declensional n at the end of certaiin words, whereas Eng. (and Fr.) noncing occurs at the beginnings of nouns, the n shifting on or off by some undefinable force in the sequential illogic called idiom. In these notes, therefore, I have taken for the nonce to be the eponym of the process and have used the forms nonce word, nonced, noncing as perferable to nunnation, nunnated. [Ciardi gives examples shifting in either direction.]

---

Unfortunately, I can't find anyone else using nunnation this way. Even OED is very vague about it, listing 1. the Arabic process, and 2. a "similar process" in English.

Unfortunately in Britain nonce has another distinctly unsavoury meaning. It is a prison slang term for a sexual offender, especially a child molester.

That quote shufitz gives doesn't make it clear that 'for the nonce' is itself an example of metanalysis. It comes from 'for then once' (or the Middle English thereof), where 'then' is an inflected form of 'the'.

Old English had a fully inflected 'the' just like German (der, den, des, dem etc.), traces of which survive in a couple of surnames by metanalysis: Noakes from 'at then oaks', Rea from 'at ther ea' (ea = river).

The usual spelling of the Arabic term is 'nunation', and the original Arabic is tanwin (or tanween). The indefinitive article is a suffix [n], and instead of being written with the consonant letter for [n] (nun), it's indicated by doubling the case-ending vowel.

aput: 'for the nonce' is itself an example of metanalysis. It comes from 'for then once' (or the Middle English thereof), where 'then' is an inflected form of 'the'.

Not certain of this, but "dressed to the nines" may have been something similar. That is, originally "to the eynes" (to the eyes).