As probably the only person in the world to have been awarded a university degree thanks to the puca, it behoves me to attempt to correct some very slovenly scholarship in the New Oxford Dictionary of English.
Under the word "Puck" (p. 1499) we are told it is another name for "Robin Goodfellow . . . a mischievous or evil spirit . . . origin, Old English puca; it is unclear whether the word is of Celtic or Germanic origin."
Celtic origin? Impossible. Celtic lost its initial P in prehistoric times (viz. The Irish Language by Mairtin O Murchu, 1985). This is illustrated widely in modern Irish; compare Latin piscis with Irish iasc; Latin pater with Irish athair; Latin porcus with Irish orc, English port, Spanish puerco, etc. Modern Irish has many words with an initial P, due to the influence of Church Latin, which came to us with the missionaries of the fifth century, and later via English and other contemporary tongues, particularly French. But, in Old Irish, there are very, very few - and all a bit of a mystery (e.g. portan, a crab). Among them is peta (modern Irish, peata). According to the New Oxford Dictionary of English (p. 1388), the word "pet" originated in the late 16th century but is "of unknown origin".
The Irish Times columnist Diarmaid O Muirithe, in his latest book, A Dictionary of Anglo-Irish (Four Courts Press, Dublin, 1996), under the word peata, writes: "The English pet is from the Old Irish peta, al. petta, a tame or domesticated animal. See Revue Celtique, XLIV."
In Chambers Dictionary (Edinburgh, 1980), however, the Celtic origin is denied emphatically, but with no authority quoted. I wrote to Chambers seeking knowledge, but received no reply, no even an acknowledgement, alas.
However, the word appears in Old Irish (viz. Dictionary of the Irish Language, compact edition, Dublin, 1983), not arriving in English until the 16th century. O Muirithe appears to be right.
As for puca, the word came to Ireland with the Normans and we appear to have strong evidence that in this Marstrander was right, if we examine the puca tradition as defender in the castles and big houses of the Pale (e.g. Slane and Malahide).
Probably the oldest role of the puca is that of helper.
In Irish folklore, as recorded, and available in the archives of Roinn Bhealoideas Eireann, UCD, we have many examples of the puca as helper. In some of them, when the puca has acted to prevent a late night wanderer from getting into trouble with the slua si (fairies), on being asked as to who he is he replies, most unusually for such entities: "Is mise an puca."
In Northern European countries, particularly in Denmark, where this entity is known as Puge or Nisse, once again the dominant role is that of helper. A "mischievous or evil spirit"? There are some examples of both in folklore, certainly, but these are not the dominant roles.
Germanic origin? Again, impossible. As pointed out in the last century by the German linguists Jakob and Wilhelm Grim, Teutonic also lost its initial P a very long time ago. The oldest surviving example of a long prose work in Teutonic is set to be found in the Old English story Beowulf. As the Grimms pointed out, this work contains not even one example of a word beginning with P. I checked this for myself a few years ago. The work is found in a 10th-century manuscript but appears to be much older. Puca does appear in Old English but in material later than the 10th century (viz. Napier, Old English Glosses).
If I may be permitted an educated guess as to the origin of the word, there are two possibilities: that it is a pre-Indo-European survival (viz. Wolfgang Meid, 1984; J. P. Mallory, 1985; etc.); or that it reached northern Europe via Greek Christian missionaries working among the Slavs (the word exists in classical Greek; and in Russian, the initial P is replaced by B).
If ever you have pondered the origin of the traditional advice to be careful when dealing with Ps and Qs, you may be curious as to what the New Oxford Dictionary of English has to say about it. In page 1329 (s.v. P) we find this: ". . . mind ones Ps and Qs, see Mind." No apostrophe between that "e" and that "s"; language growth? In page 1176, it is as, well, normal: "Mind one's Ps and Qs." Turn to page 1513, and we return to "ones' instead of and "one's".
Is there a new, official attitude to the apostrophe? I turn to page 78 (s.v. apostrophe), to be told that this consists of a punctuation mark ('), used to indicate either possession . . . or the omission of letters or numbers. Fair enough; No change. There is no reference to what used to be known popularly in some linguistic circles as "the greengrocer's apostrophe", the habit of creating plurals merely by adding an "apostrophe s".
Let us now return to the letter P (p. 1328), to be instructed that the plural of P may be either Ps or . . . wait for it . . . P's. Now, no doubt, all the English-speaking greengrocers may relax and get on with it.
The reflexive verb "to avail oneself of something" has not been, well, liberalised, reduced to "avail of something", as we see so often in the public prints and hear so often from RTE, never mind the semi-literate politicians.
However, the fashionable spelling "anymore" (instead of "any more"), though not quite completely approved, does get recognition. Of course. Doesn't it rhyme with Avonmore, Tullamore, Moneymore?
Enough said, written, God help us all!