Rebracketing (also known as juncture loss, junctural metanalysis, false splitting, misdivision, or refactorization) is a common process in historical linguistics where a word originally derived from one source is broken down or bracketed into a different set of factors. It is a form of folk etymology, where the new factors may appear meaningful (e.g., hamburger taken to mean a burger with ham), or may seem to be the result of valid morphological processes.
Rebracketing often focuses on highly probable word boundaries: "a noodle" might become "an oodle", since "an oodle" sounds just as grammatically correct as "a noodle", and likewise "an eagle" might become "a neagle", but "the bowl" would not become "th ebowl" and "a kite" would not become "ak ite".
Technically, bracketing is the process of breaking an utterance into its constituent parts. The term is akin to parsing for larger sentences, but is normally restricted to morphological processes at the sublexical level, i.e. within the particular word or lexeme. For example, the word uneventful is conventionally bracketed as [un+[event+ful]], and the bracketing [[un+event]+ful] leads to completely different semantics. Re-bracketing is the process of seeing the same word as a different morphological decomposition, especially where the new etymology becomes the conventional norm.
The name false splitting in particular is often reserved for the case where two words mix together but still remain two words (as in the "noodle" and "eagle" examples above). The name juncture loss may be specially deployed to refer to the case of an article and a noun fusing (such as if "the jar" were to become "(the) thejar", or if "an apple" were to become "(an) anapple").
Re-bracketing is part of the process of language change, and often operates together with sound changes that facilitate the new etymology.
Rebracketing is a common mechanism for new word formation. For example, the English word adder derives from the Old English næddre, snake, re-bracketed from "a nædder" to "an adder" (c. 14th c.); the word "nedder" for snake is still present in some Northern English dialects. Similarly, "nickname" is a refactorization of "an ekename" (1303, ekename=additional,little name).
Some common name forms are also rebracketings, e.g. Ned or Neddy may have risen from generations of children hearing "mine Ed" as "my Ned" (mīn is the Middle English form of the first person possessive pronoun, and the my form was also emerging around the same time). Similarly "mine Ellie" --> "my Nellie".
As another example, alone has its etymology in all+one (cognate to German allein). It was subsequently rebracketed as a+lone (akin to aflutter, afire), so the second part seemed likely to be a word, "lone".
Similar processes may also add a syllable on occasion, e.g. humble pie, is derived from the umble pie, where umble referred to the inner parts of a deer, and an umble pie was a less palatable meat. Clearly, the etymology "humble pie" seemed to fit. Umble is long gone, but this phrase continues.
* The word hamburger had its origins in a form of ground meat dish popular in Hamburg, Germany (where it is still called Tartar steak; it was apparently popular with the Mongol armies). A possible bracketing for the original may be [[ham+burg]+er], but after it became popular in America, it was soon factorized as [ham+burger] (helped by the fact that ham is a form of meat). This led to the independent suffix -burger: chickenburger, fishburger, etc. Note that in the original etymology, burg was town and burger was a resident, or something related to the town; after refactorization it becomes a chunk of meat for a sandwich.
* The English helico•pter (from Greek heliko- and pterōn) has been rebracketed to modern heli•copter (as in jetcopter, heliport).
* cybern•etics: (from Greek kubernan and -ētēs) has become modern cyber•netics (as in cyberspace).
* prosthodontics (= false teeth) is from prosth(o)- + Greek odont-; odont- = "tooth", and prostho- arose by misdivision of "prosthetic", which was treated as supposed stem prosth- and suffix -etic, but actually came from Greek pros = "in front of" and thē- (the root of the verb tithēmi = "I place").
An example from Persian is the word shatranj (chess), which is derived from the Sanskrit chaturanga (2nd c. BCE), and after losing the "u" to syncope, becomes chatrang in Middle Persian (6th c. CE). Today it is sometimes factorized as shat (hundred) + ranj (worry / mood), or "a hundred worries" - which appears quite a plausible etymology.
As demonstrated in the examples above, the primary reason of juncture loss in English is the confusion between "a" and "an". In Medieval script, words were often written so close together that for some Middle English scholars it was hard to tell where one began and another ended. The results include the following words in English:
* adder: Middle English a naddre ("a snake") taken for an addre.
* aitchbone: Middle English a nachebon ("a buttock bone") taken for an hach boon.
* apple pie order: English a nappes-pliées (meaning "neatly folden linen" in French) taken for an apple pie (this is also an example of transposition).
* apron: Middle English a napron taken for an apron.
* auger: Middle English a nauger taken for an auger.
* eyas: Middle English a niyas taken for an eias.
* humble pie: Middle English a numble taken for an umble (ultimately from Latin lumbulus, this is also an example of homorganicness).
* lone: Middle English al one (all one) taken for a-lone.
* newt: Middle English an eute (cognate with eft) taken for a neute.
* nickname: Middle English an eke name ("an additional name") taken for a neke name.
* the nonce: Middle English, for old English þen ānes (the one [occasion]).
* omelette: 17C English from French la lemelle ("omelette") taken for l'alemelle; ultimately from Latin lamella ("blade"), perhaps because of the thin shape of the omelette (SOED).
* ought: Middle English a nought ("a nothing") taken for an ought. Ultimately distinct from Old English naught ("nothing"), of complex and convergent etymology, from na ("not") and wight ("living thing, man"), but cf. aught ("anything", "worthy", etc.), itself ultimately from aye ("ever") and wight (SOED).
* tother: Old English (now dialectal) t[he] other, taken for t-other.
* umpire: Middle English a noumpere taken for an oumpere.
In French similar confusion arose between "le/la" and "l'-" as well as "de" and "d'-".
* French démonomancie ("demonomancy") taken for d'émonomancie ("of emonomancy").citation needed
* Old French lonce ("lynx") taken for l'once, thus giving rise to once (hence English: ounce), now more often applied to the snow leopard.
* Old French une norenge ("an orange") taken for une orenge.
* boutique from Greek-derived Latin apoteca, a change common to most Romance languages (e.g. Italian bottega, Spanish bodega), a putative proto-Romance l'aboteca or l'abodega taken for la + lexeme.
In Arabic the confusion is generally with non-Arabic words beginning in "al-" (al is Arabic for "the").
* Greek Alexandreia (Alexandria) taken for al Exandreia (and thus Al-Iskandariyah; this is also an example of metathesis).
* Greek Alexandretta taken for al Exandretta (and thus Iskenderun; this too is an example of metathesis).
* alligator from Spanish el lagarto ("the lizard").
* alone from all one.
* atone from at one.
From Arabic "al"
Perhaps the largest form of this sense of juncture loss in English comes from the Arabic al (mentioned above):
Spanish
* Arabic al-faṣfaṣa in Spanish as alfalfa, alfalfa.
* Arabic al-kharrūba in Spanish as algorroba, carob.
* Arabic al-hilāl in Spanish as alfiler, pin.
* Arabic al-hurj in Spanish as alforja, saddlebag.
* Arabic al-qāḍī in Spanish as alcalde, alcalde.
* Arabic al-qā’id in Spanish as alcaide, commander.
* Arabic al-qaṣr in Spanish as alcázar, alcazar.
* Arabic al-qubba in Spanish as alcoba, alcove.
* Arabic al-‘uṣāra in Spanish as alizari, madder root.
* Arabic al-rub in Spanish as arroba, a unit of measure.
* Arabic al-zahr ("the dice") in Spanish as azar, "randomness", and in English as "hazard"
Portuguese
* Arabic al-bakūra in Portuguese as albacor, albacore.
* Arabic al-ġaṭṭās in Portuguese as alcatraz, albatross.
Medieval Latin
* Arabic al-’anbīq in Medieval Latin as alembicus, alembic.
* Arabic al-dabarān in Medieval Latin as Aldebaran, Aldebaran.
* Arabic al-ḥinnā’ in Medieval Latin as alchanna, henna.
* Arabic al-‘iḍāda in Medieval Latin as alidada, sighting rod.
* Arabic al-jabr in Medieval Latin as algebra, algebra.
* Arabic al-Khwarizmi in Medieval Latin as algorismus, algorism.
* Arabic al-kīmiyā’ in Medieval Latin as alchymia, alchemy.
* Arabic al-kuḥl in Medieval Latin as alcohol, powdered antimony.
* Arabic al-qily in Medieval Latin as alkali, alkali.
* Arabic al-qur’ān in Medieval Latin as alcorānum, Koran.
Other
* Arabic al-ġūl in English as Algol.
* Arabic al-majisti in French as almageste, almagest.
* Arabic al-minbar in Medieval Hebrew as ’almēmār, bema.