Tayma Aramaic or early Nabataean?
This month, I spent two weeks teaching an introduction to Nabataean epigraphy at a small winter school set near the Nabataean ruins of Hegra, in northwest Arabia. One of my final tasks as part of the AlUla Inscriptions Corpus Analysis Project at Ghent University, this was an extraordinary experience and one that will feature in a number of blog posts to come. In this one, I want to focus on a question that came up in preparing my final class, on early Nabataean, centering on some fascinating Aramaic material from the next oasis over: Tayma.
Map from Healey (
1993; paywalled)
Under the editorship of Michael C.A. Macdonald and (for vol. 3) Muhammad al-Najem, a range of inscriptions from Tayma were published Open Access in 2023 and 2025. These include inscriptions in Akkadian, Taymanitic (the local, West Semitic language written in an Ancient North Arabian script), Imperial Aramaic,1 Nabataean Aramaic, and, in Macdonald’s view, a unique, local variety: Tayma Aramaic. The arguments for treating Tayma Aramaic as a distinct variety are not made explicit,2 but I think the reasoning is as follows:
The Tayma Aramaic script is for the most part more archaic than the Nabataean, so it cannot descend from it.The Tayma Aramaic script is occasionally more innovative than the Nabataean, so it cannot be ancestral to it either.Therefore, Tayma Aramaic is a distinct sister of Nabataean, like square script (and not one of its ancestors, like Imperial Aramaic, or one of its descendants, like Paleo-Arabic).By way of illustration, here is a gorgeous triply-inscribed funerary stela (or three or four separate stelae, I guess) in Imperial Aramaic, Tayma Aramaic, and Nabataean (from vol. 2: 100):
Imperial Aramaic:
npš g{z}{y}ʾh brt rgʿl ‘the funerary monument of G{Z}{Y}ʾH daughter of RGʿL’
Tayma Aramaic:
hy npš gzylh brt wʾlh ‘this is the funerary monument of GZYLH daughter of W
ʾLH’
Nabataean:
npš pṣyʾl brt ʿbydw byrḥ ʾlwl šnt XXIIII l-ḥrtt […]{w} rḥm ʿm-h ‘the funerary monument of PṢY
ʾL daughter of
ʿBYDW, in the month of Elul, year 24 of Aretas, [king of Nabatae]{a}, Lover of His People’
Macdonald’s commentary on the second inscription (vol. 2: 103) contains this very quoteworthy passage:
The importance of this inscription is its physical and chronological position between the Imperial Aramaic text (TA 10277 A) and the Nabataean one (TA 10277 C), a position which is reflected in the variety of its letter forms. The first two letters have their Imperial Aramaic forms, but the following letters are in a mixture of post-Imperial Aramaic shapes. Thus if one compares the first and last letters (h) one can see that the first is angular and the last is more rounded. More dramatically, comparison of the second letter with the eighth (y) shows that the stance of the letter has been reversed and that the short horizontal line half-way down the stem has disappeared. Comparison of the fifth letter (š) with the third letter in inscriptions A and C shows what looks like a steady, if purely theoretical, progression: from the Imperial Aramaic form (in A) to one local development (in B), and another (later) local development, the Nabataean form (in C). However, an indication of the complications of palaeographical ‘history’, is provided by the fourth letter (p) in TA 10277B. This letter, in a text which clearly predates the introduction of Nabataean into Taymāʾ, has a similar form to that at the end of line 2 in the Nabataean inscription of AD 203 found at Taymāʾ, see TM.N.004 in Macdonald – Al-Najem 2021 [and now also in vol. 3, BDS]. By contrast the form in the Nabataean inscription here (C, which is 188 years earlier than TM.N.004) is much closer in shape to the Imperial Aramaic p (the second letter in A); and there are other letter forms which will be discussed in the present author’s study of the development of the Aramaic script at Taymāʾ (Macdonald [in prep. a]).
This text shows that different forms of the same letter within the Aramaic alphabet could be held in the memories of scribes, and presumably readers, and used as they pleased to achieve various effects. Such a conclusion is not particularly startling but it shows the dangers of trying to use supposed palaeographical sequences to date inscriptions.
There’s a lot of very insightful observations here and I think the second paragraph is especially important (we’ll probably return to this point in my next post). But especially given the non-linear development of letter shapes Macdonald highlights here, I’m not fully convinced yet that Tayma Aramaic is definitely its own thing. In particular, I think it rather closely resembles the corpus of Nabataean inscriptions from the first (and maybe second) century BCE, before Nabataean had reached its classical, spidery shape.
The closest point of comparison, in my view, is in the Tayma Aramaic inscription TA 14285+14286+13651 (vol. 2: 117-118), which could actually be classified as Nabataean based on its contents:
bXI bʾb šnt ʿšr {w}šbʿ {l}mn{k}{w} mlk nbṭw ʾdyn qrb ʾlḥ{d/r}ym br ʿrgw {ʿ}lwʾ dnh lṣlm ʾlhʾ [l]ḥyy np{š}h wnpš […]{t}{h} {l}{ʿ}{l}{m} ‘On the 11th of Ab, year sevente{e}n {of} Mali{chus} king of the Nabataeans: then, ʾLḤ{D/R}YM son of ʿRGW presented this {
ʿ}LH to the god Salm, {for} his o{w}n life and the life of {his} […] {forever}.’
Macdonald (vol. 2: 118) relates this inscription to “the Nabataean king known in Classical sources as Malichus [II] … [who] reigned from AD 40–70 and so year 17 of his reign would be AD 56/57” (numeral “[II]” in original). But check out this inscription from Tell al-Shuqafiyya, Egypt:
dʾ rbʿtʾ dy ʿbd whbʾlh[y] br ʿbdʾlgʾ br ʾwšʾlhy ldwšrʾ ʾlhʾ dy bdpnʾ mṣryt šnt XVIII lmlkt qlptrw dy šnt XXIIIIII [l-] {m}nkw mlk nbṭw dy hy šnt II lʾṭlh byrḥ nysn‘This is the feasting couch which WHBʾLH[Y] son of ʿBDʾLGʾ son of ʾWŠʾLHY made for Dusares the god who is in Daphne, in Egyptian, year 18 of the reign of Cleopatra, which is year 26 of {Ma}lichus, king of the Nabataeans, which is year 2 of ʾṬLH, in the month of Nisan.’
The synchronism of Cleopatra’s year 18 and Malichus’ year 26 shows that the date corresponds to 34 BCE, ninety years earlier than the date Macdonald suggests for the inscription from Tayma (based on the overlap, we must be dealing with Cleopatra VII and Malichus I). Yet, although there are certainly some differences in particular letter shapes, the overall script type seems highly similar to me. For instance, both inscriptions have the typical round 𐢁 ʾ, the roofless 𐢃 b, the zig-zagging non-final 𐢍 y, back-looping 𐢙 ṣ, and the vertical but not (yet) connecting 𐢝 š (but in the Tayma inscription, note the broad-tailed 𐢘 p here too). Unless there is some archeological argument I’m missing, it seems preferable to date both inscriptions to roughly the same period and interpret the date in the one from Tayma as year 17 of Malichus I, 45 BCE. In terms of script, it would be odd to call an Egyptian inscription referring to a Nabataean king “Tayma Aramaic”, and in fact the editor on DiCoNab—none other than Michael Macdonald—opts for Nabataean.3
Several of the other Tayma Aramaic inscriptions reminded me of BCE Nabataean inscriptions. Compare TA 17431 (vol. 2: 118-120) to the Nabataean one above as well as CIS II 349 (from Petra, 69 BCE) below:
bII bnysn šnt II lšhrw mlk lḥyn ʾd{y}[n …
]mw wtbʿhn wbny ʿṣ{d/r}w rbʿtʾ dʾ […] ʾḥrhn lʿlm wʾbqy lbyt {ḥ}m{w} h[…]
t{l}t wlʾ yzbnn wlʾ yrhnn wlʾ yh[
wgrn …]
‘On the 2nd of Nisan in year 2 of ŠHRW king of Lihyan, th{e}[n …]MW and their followers and the sons of ʿṢ{D/R}W this feasting couch […] their descendants for ever and may he show mercy to the house of {Ḥ}M{W} […] th{r}ee and they may not sell nor give as a pledge nor re[nt out …]’
[… ṣ]{l}{m}ʾ zy rbʾl m{l}{k} {n}[b]ṭw […t] mlk nbṭw zy hqym lh […] br ḥymnny rbʾ wḥdth […] byrḥ kslw zy hw šmrʾ … XIIIIIIII lḥrtt mlkʾ‘[the st]{atu}e of Rabbel, ki{ng of the Na}[ba]taeans […as], king of the Nabataeans, that […] son of ḤYMNNY the elder erected for him, and […] renovated it in the month of Kislew, which is ŠMRʾ, […] 18 of King Aretas’
Or TM.TAr.003 (vol. 3: 40-41) to MPNab 3 (from Petra, completed 99 BCE):
ḥgrʾ dy qrb ʾḥbw{l}whw pny ḥṭmh lmnwh ʾlht ʾlhtʾ lḥyy npšh wn[p]š ʾḥrth lʿlm‘The vessel [?] which [unclear personal name] offered to Manawah, the goddess of
goddesses, for the preservation of his life and the li[f]e of his posterity for ever’
ʾlk ṣryḥyʾ w gbʾ zy ʿbd ʾṣlḥ br ʾṣlḥ dnh ṣryḥʾ dy ʿbd ʾṣlḥ br ʾṣlḥ ldwšrʾ ʾlh mn{k}tw ʿlḥyy ʿbdt mlk nbṭw br ḥrtt mlk nbṭw šnt I‘These are the halls and the well that ʾṢLḤ son of ʾṢLḤ made. This is the hall that ʾṢLḤ son of ʾṢLḤ made for Dusares, the god of MN{K}TW, for the life of Obodas, king of the Nabataeans, son of Aretas, king of the Nabataeans, year 1’
Others are just very close to Imperial Aramaic, which is also the case for the most archaic inscriptions mentioning Nabataean kings. Like an alleged Tayma Aramaic graffito, JSNab 334, and CNIK 1 (from Elusa, 2nd or early 1st c. BCE):
mšʿwdw mlk lḥyn ktb dnh‘MŠʿWDW, king of Lihyan, wrote this’
znh ʾtrʾ zy ʿbd {z}tyrw ʿl ḥywhy zy {ḥ}[r]tt mlk nbṭw‘This is the site which {Z}TYRW made for the life of {A}[re]tas, king of the Nabataeans’
Or, perhaps a better comparison, TM.TAr.004 (vol. 3: 41-44) and an Aramaic-Dadanitic bilingual from al-Ula, RCU.2023.150:
[… šn]{t} III mšʿwdw mlk lḥyn ʾdyn bnw ʿm{y}{w} {w}{ʿ}mrw […] {r}{b}ʿthn dʾ qrbn lrṣy ṣlmw ʾlhʾ lḥyy npš[…] w{ʾ}ḥrt-h{m} lḥgthn [w]{l}{r}pyhn w lb{k}{.}hm w ly{t}bh{n} w{r}ʿ{h}n wʾḥrthn lʿl{m}‘[… year] 3 of MŠʿWDW, king of Lihyan, then built ʿM{YW} {and} {ʿ}MRW […] this {feas}ting couch of theirs as an offering to please the god Salmu for the preservation of [their] life and [that of] the{ir} {p}osterity, for their festival [and] {for} their {hea}ling [?] and for their [?] and for the{ir} dwe{lli}ngs and the{ir} {neigh}bour and their descendants for ev{er}.’4[…]
zy ʿbd ʿmrw br š[lmw] wntnw b{r} zbynw wʾṣ[…] ʿmyrt ldwšrʾ ʾlh nb[ṭw] bywmt ḥrtt mlk nb[ṭw]‘[…] which ʿMRW, son of Š[LMW’] made with NTNW, so{n} of ZBYNW, and ʾṢ[…] ʿMYRT for Dusares, the god of the Nabataeans, in the days of Aretas, king of the Nabataeans’
In both these last comparisons, the Tayma Aramaic text is actually a bit more developed in both spelling and letter shapes than the text mentioning the Nabataeans, but mostly in ways that are shared with later Nabataean. (The exception, again, is the lengthened 𐢘 p).
In his commentary on TM.TAr.003, Macdonald writes: “The script of this text is remarkable for the variety of its letter forms and can be classed neither as Imperial Aramaic nor Nabataean, though it has some letter forms typical of each” (vol. 3: 40, emphasis mine). For the most part, this statement appears to me to be applicable to Tayma Aramaic as a whole. Given the apparent dating of the first entirely Tayma Aramaic inscription we considered to the middle of the first century BCE and the relative chronology where Tayma Aramaic precedes “Classical” Nabataean (see the triple funerary monument and occasional dating to kings of Lihyan, who ruled Tayma before the Nabataeans), can we just take Tayma Aramaic as a transitional, first-centuries-BCE script type between Imperial Aramaic and Nabataean? If so, that would fit nicely with the explicitly Nabataean inscriptions from this period, which similarly lack many typical features of mainstream Nabataean. I’m sympathetic to Macdonald’s implied objections against using the term Nabataean for inscriptions from Lihyanite Tayma (just like I wouldn’t want to call the ones from Egypt Tayma Aramaic).5 Maybe we could settle for Arabian Aramaic,6 or maybe Arabian square script and then explicitly contrast it with Jewish square script. Either way, I think we should include this material in discussions of the development from Imperial Aramaic to Nabataean proper, not sequester it in a niche of its own.
“The Nabataeans called, they want their letter shapes back”
At first, I was pretty confused by the Imperial Aramaic material from Tayma, which is mostly in a script that looks more archaic than the Imperial Aramaic I knew and therefore can’t directly descend from it. Thanks to Jérôme Norris for explaining that this is the usual (older) Imperial Aramaic lapidary script, which is indeed more archaic than the Imperial Aramaic cursive used in the papyri. This goes to show how readily you should accept my takes on Aramaic paleography. ↩︎In vol. 2, Macdonald refers to a forthcoming article in a journal I worked for for a little while, but which is now basically defunct. ↩︎The same goes for the other, longer-known inscription from Tell al-Shuqafiyya, which is fifteen to thirty years older (depending on which Ptolemy is meant) and written in a more archaic script than either of the inscriptions mentioning Malichus (three-legged ܐ ʾ [in the first half], roofed 𐢂 b, wavy 𐤑 ṣ, open 𐡒 q). ↩︎Translation modified more seriously than with the other incriptions. ↩︎Linguistically, it’s also interesting that at least the older Tayma Aramaic material lacks “wawation” and shifts *-at to *-ah in Arabic names, like tym and wʾlh for Nabataean tymw and wʾlt. This probably reflects a local difference in the Arabic dialect(s) the names were taken from, or maybe different spelling conventions. ↩︎Although that might be confusing, since the term has also been used for a hypothetical donor dialect of Aramaic loanwords into Old Arabic. ↩︎#Arabic #Aramaic #Nabataean