¹ This construction, and others involving expressions of naming, I treat at length in a monograph now in process of preparation.

² Occurrences in Latin, if they really exist, are certainly imitations of Greek. For instance, two may be cited in Vergil, Aen. 3.613–4 sum patria ex Ithaca, comes infelicis Ulixi, nomen Achaemenides, and 12.514–5 maestum mittit Oniten, nomen Echionium. In both the reading nomine also exists, but nomen is probably preferable as the lectio difficilior. If correct, it is assuredly a borrowing from Greek and not an inheritance from Indo-European.

³ The use of the possessive with the word for ‘name’ seems redundant to us. The question whether there is a possible connection of the Hittite laman-šet (or ŠUM (-an)-šet as it is often written) with the Akkadian šum-šu is a difficult one. At all events, the usage is not unknown to Indo-European languages, for it recurs in Old Irish.

⁴ For Hittite, see e.g. Johannes Friedrich, Hethitisches Elementarbuch 1.69–70 (Heidelberg, 1940–6).

⁵ Cf. my discussion of partitive apposition in Homer, TAPA 85.197–239.

⁶ Quite as much so as physical parts of his body or as his mind or soul, the types of word mainly met in the construction in Homer. It is from the use of partitive apposition with words of this sort that some scholars, notably Brugmann and Delbrück, have derived the usual type of accusative of specification met so frequently in Greek as to be named accusative graecus; I offer fresh arguments in their favor, TAPA 85.239–89. In Sanskrit, in which partitive apposition does not seem to have been a feature, the accusative of specification, used mainly of parts of the body, did not develop; nāma is viewed as simply a kind of adverbial accusative.

⁷ One example has been cited for Avestan, Yt. 13.120, but this is doubtful and probablyspurious. Two examples are met in Old Persian, A²Hc 14–5 and A³Pa 18–20, but these are very late, and belong to a period when the Old Persian case system seems to have unaccountably gone to pieces. In Greek, the earliest example known to me of ónoma used with a noun not in the nominative or accusative is as late as the Christian era, Plutarch, Solon 12.4 . There seems little doubt that ónomais an accusative of specification in Xenophon, Anab. 1.2.23, but this cannot be proved.

⁸ In Sanskrit, the development may have been forwarded also by the common use of the accusative as the adverbial case. Cf. above, fn. 6.

⁹ Hittite laman, Indo-Iranian nāma, Greek ónoma, Latin nōmen, etc.

¹⁰ Greek also has a feminine word for ‘name’, epíklēsis, which Homer uses only in the form epíklēsin. This of course can be accusative only, but it does not disprove my thesis, for in Homer it occurs exclusively with another accusative denoting the name's owner. The earliest example of its combination with a noun in a different case is in Homeric Hymn 2.207–8 (or 1.385–6) ‘then they all pray to the lord as the Telphusian by name’. By the time that this was written, clearly the development of the accusative of specification was complete, at least for epíklēsin; but, as already indicated (fn. 7), there is no similarly cogent proof for ónoma until centuries later. Is this mere chance, or does it suggest a lingering on (of course unrecognized by the speakers themselves) of what I believe to have been the original inherited construction in the use of the original word for ‘name‘?

¹¹ Also in the Baltic languages, Lettish and Lithuanian, ‘name’ is masculine; but there it comes from a different root.

¹² In Gothic and the Scandinavian languages the word remains neuter.

¹³ It would not, however, have any bearing on the second part of my thesis, namely, my belief that it was the very ambiguity of the nominative-accusative form that led to its ultimately being viewed as an accusative of specification and not as an appositive. According to this portion of my hypothesis, the use of an accusative to express the notion ‘in name’ should not be capable of development in a language-group in which the nominative and accusative of the word for ‘name’ are not identical. As a matter of fact, ‘name’ did not so develop in West Germanic ; but this argumentum e silentio has not complete cogency, for it did not so develop in North or East Germanic or in Latin, either. A more persuasive argument is that the development should be expected in a language which at an early stage uses the nominative and accusative in partitive apposition to a considerable degree; this is certainly true of Greek more than of any other Indo-European language known to me, and the development of the accusative of specification was certainly more pronounced there than in any other language (cf. fn. 6).

¹⁴ That it does not occur in Homer may be merely the result of metrical necessity.

¹⁵ So too did the dative compete with the accusative in words denoting parts of the body. Cf. e.g. Il. 3.194 and 227 .

¹⁶ Some take toúnoma (i.e. tò ónoma) as a nominative, ‘the name was Joseph’ (which is in line with Brugmann's theory of the origin of the construction, on which see below, fn. 40); but this I think would require a genitive autoû or a dative autôi indicating the possessor of the name.

¹⁷ Contrast the dative in the parallel passage, Lk. 23.50–1 Wulfila, who apparently interprets toúnoma as nominative, supplies a genitive in his rendering, þizuh namo Iosef. That he either did not recognize the Greek form as an accusative of specification, or, if he did so recognize it, deliberately departed from it with unusual freedom, would seem to imply that the accusative of specification 'in name' had no place within his linguistic ken. He did, however, in two instances (Jn. 11.44 and Eph. 6.14), where he found it in the original, use the accusative of specification for parts of the body.

¹⁸ I have already pointed out that the accusative in such an instance, if it exists at all, is a deliberate Hellenism. See fn. 2.

¹⁹ Thus toúnoma in Mt. 27.57 (quoted just above) is rendered in the Vulgate by nomine exactly as is onómati in Lk. 23.50 (quoted in fn. 17) and elsewhere; the two passages in Latin run respectively homo dives ab Arimathaea, nomine Ioseph, and vir nomine Ioseph. On the other hand the accusative of specification for parts of the body in Jn. 11.44 and Eph. 6.14 is retained by St. Jerome as it is by Wulfila (on the latter, see fn. 17).

²⁰ As it happens, only four passages exemplifying this are preserved in the Gothic version: Mk. 5.22; Lk. 1.5, 5.27, 16.20.

²¹ In studying Old English usage as exemplified in renderings of the Vulgate, I have examined the three sets of glosses conveniently collected in Walter W. Skeat's The Holy Gospels (Cambridge, 1871–87): for Northumbrian, the Lindisfarne MS (also called the Durham Book), now in the British Museum; for Old Mercian (Mt. only), the Rushworth MS, now in the Bodleian Library at Oxford; for West Saxon (often miscalled ‘Anglo-Saxon‘), the Corpus MS, now in the library of Corpus Christi College at Cambridge.

²² On can govern the accusative as well as the dative, and naman can be an accusative as well as a dative; but I think we can safely assume that we have a dative here.

²³ It is rather interesting to note that all instances of on naman are confined to the West Saxon version of Luke, and that this version never renders the Latin nomine in any other way except in 5.27, which is a special case, as it departs from the original Latin (publicanum, nomine Levi), running publicanum, he wæs oþrum naman Leui gehaten ‘a publican, he was called Levi by his other name‘ (emphasis mine).

²⁴ Genuine hypotaxis seems to me impossible, since Old English has no true relative. But sometimes the demonstrative pronoun or the particle þe seems to approach relative force.

²⁵ Antoine Meillet, in Introduction à l'étude comparative des langues indo-européennes 345 (Paris, 1934), calls 'nom' as used in Sanskrit, Old Persian, and Greek a 'nominatif apposé'; but he does not give details, and his use elsewhere (e.g. ib. 359 and 360) of the terms apposé and apposition shows that he employs them in a much broader sense than an English reader would expect. Jean Humbert, who quotes his phrase 'un nominatif apposé' in Syntaxe grecque 251 (Paris, 1945), indicates clearly that he is taking ónoma as a predicate noun, not as what we would call an appositive.

²⁶ Instead of nāman we would write nōmn̥ today.

²⁷ Our English idiom likewise demands use of the verb ‘give’, but the Old English scyppan means rather ‘make, create’. Similarly Latin in examples of the sort uses not only do but also facio; it likewise uses indo, which in its literal sense means ‘put into’. Hittite dai- and Greek títhēmi, regularly employed in the corresponding construction, also have the general meaning ‘place, set’. In this connection it is of interest to note that both of them, and also Latin facio and indo, are all cognates etymologically. So far as the meaning goes, we may also compare the verbs used in the Latin and Germanic versions of the Biblical passage cited to illustrate the construction.

²⁸ Or theîo. The reading is disputed ; but this has no bearing on the point under discussion.

²⁹ This is not a perfect parallel for the other examples; but the Latin passages that in other ways are closer to the Hittite and the Greek involve attraction of the name into the case of the name's owner, as in Plautus, Men. 77 iuventus nomen fecit Peniculo mihi, Capt. 69 iuventus nomen indidit Scorto mihi. This attraction seems to me to furnish fresh evidence of the intimate relationship between the man and his name. I am puzzled by Gray's comment (IF 11.309) on the contrast between Aul. 164 and Capt. 69: 'Meines Erachtens geht dieser Wechsel auf eine ursprüngliche Auffassung von nomen in appositioneller Bedeutung zurück.' This would make sense if he were positing, as I am, a close connection between the man and his name; but since he views the word for 'name' as in apposition with the nomen proprium, it is astonishing that he finds support for this view in the attraction of the nomen proprium in case away from the word for ‘name’.

³⁰ We would expect Paitru (accusative). Is the foreign name here treated as an indeclinable, as apparently in the Latin version? Since both translations, Latin and Gothic, from the original Greek alike use the peculiar nominative form, should we suppose that possibly the Greek passage once read Pétros?

³¹ Cf. fn. 30.

³² In this clause the translation departs completely from the syntax of the original. Cf. the West Saxon treatment of Lk. 5.27 (above, fn. 23).

³³ To be sure, the word order may be thought to support him, since the appositive seems likely to follow rather than to precede the noun with which it is in apposition; but the practice in regard to this is too variable to furnish safe guidance. As it happens, in all the parallels for Beowulf 78 offered by me directly above in which both the proper noun and the word for ‘name’ occur—in other words, all but Od. 19.403–4 and the West Saxon version of Mk. 3.16 (for which see fn. 32)—the word for ‘name’ comes first; and—what is more to the point—this is also true of a number of Gray's own examples (all cited by him IF 11.308–10), namely RV 6.66.1, Yt. 13.79, Od. 19.183 and 24.306, Lk. 1.31 and 2.21, and Plautus, Aul. 164 and Capt. 69 (on which see above, fn. 29).

³⁴ Of course this rendering of mine is impossible as Latin (cf. fnn. 2 and 18). And we shall see that Gray's interpretation is likewise impossible as Old English.

³⁵ I trust I need not indicate my awareness that all the translations offered here are literal, without regard for English idiom.

³⁶ Modern English, on the lips of the semiliterate, sometimes quaintly violates this natural and normal rule. The substandard speaker, in narrating some personal experience, will frequently say, ‘and so I met this man’ or something of the sort, with reference to a man who has not previously entered into his tale.

³⁷ The usual idiom in Beowulf is ‘he is named X’, as in 102, 262–3, 373, 2602 (corresponding to the active ‘they name him X’, as in 364 and 1354). We have already noted the tendency in the Old English glosses in general to substitute the construction ‘a man—he is named X’ for the Latin ‘a man, X in name’; and we have also (fn. 32) observed the substitution in a particular West Saxon gloss (Mk. 3.16) of the corresponding active ‘he named him X’ for the Latin ‘he gave him the name X’. But if we do have in Beowulf the construction ‘the name is X’, we would rather expect the owner of the name to be represented not by a dative but by a genitive, on the basis of the use of the possessive adjective in 343 Bēowulf is min nama ‘Beowulf is my name’. Similarly in this type of construction the West Saxon glosses invariably use a genitive or a possessive adjective. On the other hand the Anglian versions (Northumbrian and Mercian) almost invariably use a dative. Hence Beowulf 343 corresponds to Mk. 5.9 Legio mihi nomen est in the West Saxon version, min ('meurn') nama is Legio, as opposed to the Northumbrian version, Here to noma me ('mihi') is; but Beowulf 1457 corresponds to Jn. 18.10 erat autem nomen servo Malchus in the Northumbrian version uses ... noma ðæm esne ('illi servo') Malchus, as opposed to the West Saxon version, þæs þeowan ('illius servi') nama wæs Malchus. However, this does not need to arouse our suspicions. There is plenty of dialect mixture in Beowulf; the consensus of opinion seems to be that the poem as we have it is primarily in West Saxon, but that in origin it was Anglian. It may be worth noting that the form mēce (instead of m¸) in the very same line (1457) is Anglian, not West Saxon. Furthermore, even though the West Saxon version of the Gospels lacks the construction with the dative, we do find it elsewhere in works regularly regarded as West Saxon, e.g. Genesis 1106, Andreas 1322, Elene 418 and 437.

³⁸ The meaning of this word is far from certain. Various suggestions have been made, among them ‘spokesman’, ‘major-domo’, ‘jester’.

³⁹ Or at least supposedly valuable. The loan is usually viewed as an act of courtesy and magnanimity on Unferth's part, which is surprising after his earlier malicious attack on Beowulf (506–28). But when Beowulf is actually engaged in combat with the horrible she-monster, the sword fails him (1522–8), and only a miracle saves him (1553–62), as he afterwards reports (1655–64). He expressly says that Hrunting is good (1660), and in returning it to Unferth thanks and praises him (1807–12) ; yet I cannot help wondering whether there may not be an implication of some sort of trick on Unferth's part, which Beowulf is too chivalrous to recognize.

⁴⁰ And two other eminent scholars, Foy and Brugmann, the great Brugmann himself, perpetuated his error, for Foy copied the line from Gray (IF 12.178) and Brugmann copied it from Foy (IF 27.144), in both cases with Gray's impossible translation of the first part, ‘war ihm ein Heftschwert’, reproduced, but in both cases with Gray's also impossible translation of the second part, ‘ “Hrunting” sein Name’, changed to the equally impossible though quite different translation ‘Hrunting (war) der Name’. Foy (12.172–8) attacks Gray's entire paper in a polemic which to me, despite my general disagreement with Gray, seems too violent, although Brugmann (27.145) terms it in great part thoroughly justified; yet neither one of them tested the accuracy of Gray's interpretation of a passage which formed an important basis for his main argument, and, later, for theirs. Foy (12.178), directly after citing as 'eine andre Quelle für den Gebrauch von ai. nāma usw. im Sinne von “mit Namen” ' 'solche Sätze wie Beow. 1457' (as if there were other instances of the sort as interpreted by Gray and Foy!), translates the line as I have reported above; and then, after expressing the view that such investigations as Gray's into the origin of the construction are of no use to linguistic science, concludes with unconscious irony: 'Begnügen wir uns damit, die Verhältnisse der Einzelsprachen genau kennen zu lernen' ! And Brugmann, who is endeavoring to prove (IF 27.143–6) that the proper noun and the word for ‘name’ were originally subject and predicate of an interpolated independent clause, seizes upon Gray's citation, or rather Foy's altered translation of it, as an apt exemplification of his theory, and repeats this translation word for word (144). 'Passend verweist Gray auf Beowulf 1457,' says he (144), without knowing either how Gray actually explained the passage, or how the passage really ought to be explained. — Error dies hard; this particular one has been alive for over half a century. Gray perpetrated it in 1900; Foy and Brugmann perpetrated it in 1901 and 1910 respectively. As recently as 1954, Helmut Humbach, copying Gray or else some other copyist of Gray rather than Foy and Brugmann, in an article entitled Kompositum und Parenthese (Münchener Studien zur Sprachwissenschaft 5.90–9), wrote as follows in regard to 'die Namenparenthese' (91): ‘Gut bekannt und oft zitiert sind ae. waes þaēm haeftmēce Hrunting nama Boew. 1459 [sic] “es war ihm ein Heftschwert, Hrunting der Name” und āsīd rājā Nolo nāma “es war ein König, Nala der Name”.‘ Nala yes, Hrunting no!

⁴¹ It seems appropriate to stress realization of this fact in the issue of Language which is dedicated to an eminent lexicographer.