▪ There seems to have been, in earlier research on
minǧal, a kind of “filter bubble” suggesting that the semantic distance between ‘sickle’ and other values realized in the Ar root √NǦL was too big to be explained by derivation; therefore, the idea that it could be a borrowing was readily accepted, and Sum, Eg, Copt and Grk etyma have been proposed. Apart from the semantic distance just mentioned, the adherents of an extra-Sem origin also point to the scarcity of *NGL items throughout Sem – there are clear cognates only in Hbr and Aram, while the relation to some Gz lexemes is obscure and far from reliable. At the same time, a weak point is the fact that there
are some of those *NGL lexemes, even in SSem, and quite many of them in Ar itself, the semantics of which can hardly be explained as deriving from ‘sickle’ alone. Moreover, there may also be extra-Sem cognates (see below).
▪ The weakest of the borrowing hypotheses are probably those that assume a
Copt < Grk or an
Eg origin, for the simple reason that none of them accounts a) for the shift of meaning from ‘pickaxe, hoe, mattock’ (Eg, Copt, Grk) to ‘sickle’ (Ar, Hbr, Aram, Akk), and b) for the existence of the extra-Ar ‘sickles’ (Hbr, Aram, Akk) and the many Ar and Sem *NGL items, with a large variety of meanings, suggesting a deep temporal dimension that would have allowed for the development of such a diversity. While Copt
mankʸale ~
mančale may indeed be a loan from Grk
makélē ~
mákella, it is hardly convincing, in the light of the pre-Islamic attestations of Ar
minǧal as well as its Can parallels, that the Copt word should be the source of Ar
minǧal – the word existed in Ar before the Islamic expansion to Egypt and intense Copt-Ar contacts. Corriente’s idea that both the Ar and the Grk item might go back to a common Eg (i.e., pre-Copt) etymon can sound more plausible at the first instance. On a closer look, however, it turns out that while Eg, as the language of experienced farmers, of course has a number of words for ‘sickle’ (
ȝzḫ,
ḫȝb ‘Sichel’,
mȝ ‘sichelförmige Holzstange’ –
TLÆ) as well as for ‘pickaxe, hoe, mattock’ (‘Breitblatthacke’ –
TLÆ) or ‘chisel’ (
mǧȝ.t,
ḫnr –
TLÆ), none of these qualify as possible ancestors of Grk
makélē ~ mákella or Ar
minǧal.
1
Therefore, a more likely scenario here is: Eg was not involved at all; the Grk word is a loan from Sem (Hbr
maggāl, Aram
maggǝlā), and Copt
mankʸale ~
mančale a borrowing from Grk, perh. later influenced by Ar phonology (explaining Copt
‑nc‑/‑ng‑ instead of Grk
‑k‑ < Aram
‑gg‑).
2
The semantic difference between Sem ‘sickle’ and Grk/Copt ‘pickaxe, hoe, mattock’ could be due to a loan at a time when the Sem etymon still also signified an tool used to till the ground. It is not attested as such in any Sem language; but Ar knows the vb.
naǧala also with the meaning ‘to till the ground’ (see below), so that the Sem n.instr. formed from a NGL vb., at the time of its borrowing into Grk, may indeed have signified something like a ‘pickaxe, hoe, mattock’. As Beekes2009 observes, the variant
mákella of Grk
makélē points to an early borrowing (pre-Grk *
‑alʸa), and the cognate Arm
markeł ‘mattock’ could indicate that Grk
makélē and Arm
markeł are »from a common source« (which we think could well have been a Sem language, e.g., Aram). The Sem origin may also help to explain the formal and semantic similarity between Grk
makélē ~ mákella and Grk
díkella ‘mattock, two-pronged how’ (
dí‑ ‘two’ +
kellō ‘to drive on, run
a ship to land, put to shore, into harbour’) as the result of the cross that has been assumed but for which, until now, »a convincing explanation has not yet been found« (Beekes2009: ibid.).
3
▪ In contrast to the Copt < Grk or the Eg etymologies, discussed in the preceding paragraph, both the
Sum and the
AfrAs > Sem connection have a much higher degree of plausibility – if perh. only in combination with each other. In marking Akk
niggallu ~
ningallu ‘sickle’ as a word »of foreign origin«, vonSoden seems to be reluctant to assume a specifically Sum source; irrespective of this, however, it is clear that he thinks the word is not genuine Akk.
CAD explicitly identifies the source as Sum (though without naming the Sum word itself; would that be
níŋ‑ŋál(‑la) ‘sickle’, as given by Halloran_3.0? Semantics not confirmed by PSD, which renders Sum
níŋ‑ŋál(‑la) as ‘possessions’! – Rolland’s »hypothèse personelle« that Akk
niggallu ~
ningallu is from Sum
engar ‘farmer’ sounds slightly far-fetched and is phonologically problematic). A borrowing from Sum into Akk would make sense also from the cultural-historical point of view: it sounds only natural that the Sem nomads who immigrated into Mesopotamia from the west integrated important agricultural terminology from the language of the experienced Sum farmers into their own idiom. However, as mentioned above, an exclusively Sum etymology can neither explain the large semantic variety within the Sem root NGL nor the AfrAs parallels – it is impossible to make all this dependent on only one initial Sum > Akk borrowing. Although Zimmern1914 does not mention Akk as a source of any WSem NGL item, the semantic and phonological similarity between Akk
niggallu ~
ningallu and Hbr
maggāl, Aram
maggəlā and Ar
minǧal is so high that some kind of relation can be taken as a given. However, how exactly the items are related among each other, is difficult or impossible to decide. The borrowing may have happened independently for each of the recipient languages (i.e., Akk > Hbr, Akk > Aram, and Akk > Ar, separately) or first into Hbr or Aram and from there into Ar. In the first case,
‑nǧ‑ in Ar
minǧal would be directly from Akk
‑ng‑, in the second, it would be the result of dissimilation of Hbr or Aram
‑gg‑ to Ar
‑nǧ‑. In all these cases, initial Akk
ni‑ would have become
ma‑/
mi‑, probably to make the loanword conform to familiar noun patterns (like the Ar
miC₁C₂aC₃ pattern for nomina instrumenti). But even if one assumes a Sum > Akk > WSem borrowing, the semantic variety within WSem NGL as well as the extra-Sem (Chad) parallels remain to be explained. For
EtymArab, the most convincing solution to this problem is the assumption of a root √NGL in Sem that predates the borrowing from Akk into WSem so that the borrowed word was interpreted
as if from the already existing root √NGL (Militarev2002 suggested Sem *
mi‑/
ma‑ngal‑). This would explain not only the existence of the Chad parallels but also the replacement of initial Akk
n‑ with WSem
m‑. It may also account for the difference in semantics between the Copt and Grk words (‘pickaxe, hoe, mattock’) and the WSem ones (‘sickle’): one could assume that Grk
magélē ~
mágella was the etymon of Copt
mancela but was itself a loan from Hbr
maggāl or Aram
maggəlā at a time when the WSem words still meant something like ‘pickaxe, hoe, mattock’, but that this value later changed to become ‘sickle’ when WSem came into contact with the Akk
niggallu ~
ningallu ‘sickle’. In this scenario, Hbr, Aram, and Ar may have lost their original value, which, however, at some time, may indeed still have existed: for Ar, for instance, one can easily assume the existence of a n.instr. *
minǧal from an Ar vb.
naǧala in the sense (now obsolete, but attested for earlier times) of ‘to till the ground’ (see
†[v
15] in root entry ↗√NǦL), lit., to make a
naǧl, i.e., an ‘opening in the earth to plant s.th.’.
minǧal is not attested in that meaning, but
naǧala and
naǧl are.
4
Moreover, both are probably the result of a semantic shift (owing to the Neolithic revolution and the introduction of agriculture?) from a still earlier
†‘to split, pierce (s.th.,
bi‑ with a spear)’ (attested as such in Ar, too, and thought to be the Ar vb.’s most elementary value by Fraenkel1886). – The question that remains to be solved in this theory is the existence of the Chadic parallels meaning ‘sickle’, ‘knife’ or ‘sword’. If ‘sickle’ is not the original meaning in Sem and if both Sem and Chad were from a common AfrAs source, then the Chad items shouldn’t mean ‘sickle’ but *‘instrument to pierce’ or, more specifically, *‘tool used to till the ground’. So, are the Chad parallels perh. no genuine cognates but borrowings from Sem? Or the results of a crossing of a borrowing with earlier semantics, similar to the changes that the Sem words underwent after the borrowing from Sum?