Laudator Temporis Acti

There is nothing elementary about the history of the word elementary, or rather its Latin root elementum. The Oxford Etymologist, Anatoly Liberman, has recently discussed elementum three times, in his Monthly Gleanings for August 2007, October 2007, and November 2007.

Elementum first appears in Lucretius, as a translation of Greek στοιχεῖον (stoicheion), both of which can mean a letter of the alphabet. Liberman rejects the idea that elementum comes from the sequence of letters in the middle of the alphabet, L-M-N. Before reading Liberman, this is the derivation I always thought was correct. I may have learned it in school, or from the confident assertion in Allen and Greenough's New Latin Grammar, p. 145, note: "So elementum is a development from L-M-N-A, l-m-n's (letters of the alphabet), changed to elementa along with other nouns in -men."

In his commentary on Horace, Satires, 3rd ed. (Leipzig: Herbig, 1859), pp. 10-11 (on 1.1.26), L.F. Heindorf stated:

Someone acutely derived the etymologically inexplicable word elementa from the sequence of letters L,M,N, as we say the A,B,C's.

Scharfsinnig leitete jemand das etymologisch nicht erklärbare Wort elementa aus der Zusammenstellung der Buchstaben l, m, n her, wie wir sagen das A,B,C.

But he does not say who that "someone" was. Ernst Sittig, "Abecedarium und Elementum," in Satura. Früchte aus dem antiken Welt, Otto Weinreich zum 13. März 1951 dargebracht (Baden-Baden, 1952), pp. 131-138, apparently tried to defend the derivation from L-M-N, but this book is unavailable to me.

Quintilian (1.1.26, tr. H.E. Butler) says:

I quite approve on the other hand of a practice which has been devised to stimulate children to learn by giving them ivory letters [eburneas etiam litterarum formas] to play with, as I do of anything else that may be ways to delight the very young, the sight, handling and naming of which is a pleasure.

Relying on this passage from Quintilian, Hermann Diels, Elementum: Eine Vorarbeit zum griechischen und lateinischen Thesaurus (Leipzig: B.G. Teubner, 1899), derived elementum from *elepantum, which he supposed was a Latin form of Greek ἐλέφας (elephas = elephant, ivory), in the same way that the Romans said Tarentum for Greek Τάρας (Taras), Agrigentum for Ἀκράγας (Akragas), and cilibantum for κιλλίβας (killibas).

Liberman favors a connection between elementa and alimenta:

Elementa (plural) seems to have been an individual coinage prompted by alo ~ olo "nourish," a pun on something "growing" (progressing), as the sequence of letters does in an alphabet.

On this supposed connection see also F.A. Trendelenburg, Elementa logices Aristoteleae, 6th ed. (Berlin: Bethge, 1868), pp. 50-52.

Here are some other discussions of the question which I haven't looked at:

• Leo Meyer, "Elementum," Beiträge zur Kunde der indogermanischen Sprachen 1 (1877) 86-107
  
• S. Reiter, Zur Etymologie von elementum (Prague, 1900)
  
• W. Vollgraff, "Elementum," Mnemosyne 2 (1949) 89-115
  
• Alfred Adam, "Die sprachliche Herkunft des Wortes Elementum," Novum Testamentum 6 (1963) 229-232

I also haven't been able to consult the standard Latin etymological dictionaries (Ernout-Meillet and Walde-Hoffman).

By the way, Sherlock Holmes never said, "Elementary, my dear Watson, elementary." The phrase apparently first occurs in P. G. Wodehouse, Psmith Journalist (1915).