Like every other cool person out there, I am of course fascinated by the minutiae of ancient languages. Especially the parts of these languages which still lay at least partly outside the grasp of our full comprehension can draw me into a scholarly book or article and leave me surprised that the sun is already gone when I finally look up – though living in the far north arguably makes this a bit less extraordinary… Today I want to share some of this enthusiasm with you by discussing a linguistic phenomenon that is still debated among contemporary scholars and of which some even say that it isn’t even real – the R-stem in the ancient Akkadian language.

We therefore once again return to Akkadian. Not unlike me, this language is unfathomably old. It is one of the tongues that was spoken throughout what is now Iraq and parts of Syria in the three millennia before our common era, and it functioned as a shared language for diplomacy in ancient West-Asia and eastern Mediterranean of the late Bronze Age.¹ But where in previous blogs we were mainly interested in the cuneiform script with which Akkadian was primarily written and curious words in its lexicon that could tell us more about antiquity, we now venture into its grammar! So first, I am going to introduce – very briefly! – the system of stems in Akkadian. Then we are going to discuss the skeptical account of Bert Kouwenberg and his three criteria for the existence of a hypothetical R-stem. Subsequently, we will ascertain whether John Huehnergard’s examples of this stem in the surviving Akkadian text corpus fulfill these criteria. So let’s do something exciting today: let’s (dis)prove the existence of a mysterious grammatical form in a long deceased language!

This blog is also available in Dutch.

Grammatical Stems in Akkadian

Like other members in the Semitic language family, the primary meaning of words in Akkadian is conveyed by the roots of words. These are combinations of mostly three, but sometimes two or four, consonants. And by adding prefixes, infixes, suffixes respectively before, between or after these consonants, as well through several sets of alternating vowels, you can create nouns, verbs and all the other words you need to impress people at parties. Let us take us take a look at the root which carries meanings associated with stealing stuff as a perfectly innocent example. Well, not literally innocent, of course… But merely talking about thieves and stealing is, to my knowledge at least, legally allowed. This root consists of the consonants Š R Q and provides us with many words. The following are merely an illustration of the pertinent vocabulary for ancient criminal activities:²

ŠaRāQum = to steal
ŠaRRāQītum = thief (♀)³
ŠaRRāQum = thief (♂)
iŠRiQ = he/she/it stole
iŠaRRiQ = he/she/it steals/will steal
niŠRiQ = we stole
niŠaRRiQ = we steal/will steal

We immediately notice that the difference between having stolen something and being busy with stealing or planning to steal something is – at least for this specific verb – a doubling of the second consonant of the root and the adding of a vowel. But this is not the only way in which one can convey meaning by systematically changing what happens before, between, after, and with the consonants that make up the root of a verb. And here we meet the idea of different grammatical stems. The verbal forms we encountered above are part of the most basic conjugation of Akkadian verbs, which is called the Grund– or G-stem – ‘Grund’ meaning ‘basic’ in German. But there are other grammatical stems that are derived from this G-stem, which each add their own layer of meaning to a verbal form. Most prominent in this regard are the D-, Š-, and N-stems. Within these stems the signifiers which indicate, for example, the time or subject of a verbal form can differ.⁴

The D-stem primarily entails the doubling – get it? – of the second consonant and, like the other derived stems, its function depends on the meaning of the G-stem of the verb in question. By and large, the D-stem creates a factitive form for a number of adjectival verbs and transitivizes certain active-intransitive verbs or creates a causative form for them. In other words: these verbs now mean making something happen, often to something/someone. The root DMQ, for example, has meanings connected with be(com)ing good. So the D-stem of the verb damāqum, that being dummuqum, is used to for making something or someone good. And if we take the verb for ‘to disappear’, ḫalāqum, we find that the D-stem, that being ḫulluqum, means ‘to cause to disappear’. So far, so easy right?

Then there is the Š-stem which sadly breaks the naming pattern we had established thus far, as it adds the consonant ‘š’ to verbal forms. This grammatical form indicates primarily two things. For a few adjectival verbs, the Š-stem steals the D-stem’s thunder and denotes making something happen. To wit, the verb marāṣum, ‘to be(come) sick’, has a morbid Š-stem, šumruṣum, which means ‘to make sick’. But mostly, the Š-stem – like the D-stem – denotes causing something for certain active-intransitive verbs. In fact, it is considered the primary means to create a causative version of many G-stem verbs. To give another morbid example: maqātum means ‘to fall’, while the Š-stem šumqutum means ‘to cause to fall’. Additionally, the grammatical form that is the Š-stem can make active-transitive verbs doubly transitive! These verbs will now take two accusative objects, one pertaining to the action associated with the original G-stem and one to the causing thereof.

Finally there is the N-stam which – you guessed it! – adds the consonant ‘n’ to the verbal form. Though you do not see the ‘n’ often, because it almost always assimilates. Simply put, the ‘n’ takes the shape of other consonants when these are placed to their right. The N-stem mainly indicates one of two things. First, it can give an active-transitive verb a passive meaning. So while the verbal form in the G-stem iṣabbat means ‘he will seize’, iṣṣabbat – the added ‘n’ has assimilated to the ‘ṣ’ on its right – means ‘he will be seized’. And, second, it can make a verb reflexive. This is as simple as it sound. If you have a verb saḫārum, meaning ‘to turn’, then nasḫurum means ‘to turn oneself’.

To make things even more easier and understandable, there are also variations on these stems which mainly add other infixes into the proverbial mix. Like the Dt-stem, which adds the consonant ‘t’ to the verbal form of a D-stem, or the Gtn-stem, which adds the syllable ‘tan’ to the verbal form of a G-stem. And these extra additions naturally impact the meaning of the verbal form that is thus constructed. Is yet another stem among this wealth of grammatical versatility and myriads of verbal forms, that being our hypothetical R-stem, therefore really such an outlandish idea?

Kouwenberg’s Criteria for the Existence of an R-Stem

One of the most systematic and comprehensive books on the Akkadian verb ever – just shy of 700 pages and with rather small letters! – is Bert Kouwenberg’s magnum opus The Akkadian Verb and its Semitic Background. This is unironically one of my favorite language books of all time. And lucky for us, in the fifteenth chapter Kouwenberg discusses the possibility for an Akkadian verbal form wherein a syllable is reduplicated – because that is where ‘r’ in the term R-stem comes from.⁵

Kouwenberg posits that there are three prerequisites to determine whether a grammatical phenomenon like a hypothetical R-stam does exist. There need to be a consistent form, an identifiable function, and a certain productivity. In other words: there should be a recognizable R-stem, that has a specific role, and which is actually consistently used in our ancient sources. Alas, when it comes to R-stems, we find relatively little regularity when it comes to its forms and its use. In addition, there does not seems to be a specific impact on meaning, which characterized the other derived stems that we encountered above. With such little material to work with, there is also a very real chance that the verbal forms which could be R-stems are merely the result of scribal errors.

Sadly for us, Kouwenberg therefore only sees his three prerequisites with a potential Dtr-stem. That is, a Dt-stem as I introduced above but with a reduplicated second consonant and additional vowel. The function of this verbal form, according to him, is indicating reciprocity or pluractionality. An example of the former is the verbal form uktaṣaṣarū, derived from kaṣārum, where you can clearly see the titular reduplication and which means ‘they assembled together’.⁶ And even here, I am shocked to say, we may not deal with a proper R-stem. In an impressively substantiated article, David Testen has proposed that the curious verbal form which is often called the Dtr-stem, is actually an artifact of how Dt-stems were written earlier in the history of the East-Semitic branch of the Semitic language family that Akkadian is a member of. We are not dealing here with a fairly arbitrary reduplication, Testen argues, but the legacy of an earlier mediopassive element within our historically attested Akkadian. There is no Dtr-stem, only Dt-stems with a mediopassive element⁷

How then can it be that some of the greatest philological minds – and more humbly, yours truly – still adhere to the existence of an R-stem? From his own perspective, Kouwenberg points to two interlaced explanations.⁸ In the first place, there is the possibility of confusing reduplication with gemination. Gemination takes place when a consonant is doubled or lengthened, while proper reduplication entails repeating a syllable – so at least, when it comes to Akkadian, a consonant and a vowel. Furthermore, we should not confound reduplication within the root – the aforementioned group of consonant which primarily carry meaning in Akkadian – with a verbal form that requires reduplication. Simply put, when such a group of consonants, especially when there is four of them, include two of the same consonants, it does not follow that the subsequent conjugations have a reduplicated part that points to a novel R-stem. It is just the same letter! Is there any reason left, you may wonder, to believe in the existence of an R-stem that goes beyond a very specific Dt(r)-Stem?

Huehnergard’s Substantiation of the R-stem

There are clues that an R-stem may indeed have been part of the Akkadian language – even if our surviving textual evidence is meager. And indications for the existence of an R-stem can be found in much of the relevant literature of the last decades.⁹ The most accessible argument is made in John Huehnergard’s A Grammar of Akkadian, which is an excellent and elaborate tool to independently learn the language, as well as one of the books I would take with me to a deserted island.¹⁰ He sees two kinds of verbal forms that possibly point to the existence of an R-stem: those that reduplicate the syllable with the third consonant of the root and those that do the same with the second.

The former variation, reduplicating the syllable with the third consonant, seems the most promising, according to Huehnergard. There are fairly clear examples and these occur a number of times. We can also propose a function: the intensification of the meaning associated with the root. Like the dramatic verbal form šaḫururum, which means ‘to be(come) completely paralyzed with fear’. It is rather fitting then, as Huehnergard notices, that these specific hypothetical R-stem verbal forms are regularly used to describe divine and other numinous qualities. This kind of R-stems is likely too few and far in between to fit Kouwenberg’s criterion of productivity, though. (And I presume he would not agree with all of Huehnergard’s examples) But we can – I think – tentatively establish a recognizable form and function that would comply with the other two criteria.

Ironically, Huehnergard is more hesitant when it comes to the reduplication of the syllable with the second consonant. The examples we can rely on are even fewer and are more readily explained as errors. As such, he thinks most of them were not acceptable grammatical forms at the time. With the exception of – hold on to your seat! – the Dtr-stem, that is. Which he endowes with roughly the same meaning as Kouwenberg did. We therefore see many of the same considerations on this subject by both scholars.

The underlying problem in trying to prove the existence of an R-stem, however, is that by the Old-Babylonian period, roughly the first half of the second millennium BCE, this verbal form was seemingly no longer used as such. It is therefore mostly absent from the majority of our surviving texts. What is more, those verbal forms which could have been R-stems, where reanalyzed and employed by ancient writers as being a combination of the D- and Š-stems – the wonderfully named ŠD-stem. Though far from me to criticize those who could actually speak Akkadian, Huehnergard plausible suggests with multiple relevant examples that this was probably a wrong interpretation by the ancients. Furthermore, as Akkadian was mostly written with a syllabic script – that being cuneiform writing – and a hypothetical R-stem would involve reduplicating a syllable, the chance that a scribe made an error is, as both Kouwenberg and Huehnergard highlight, not that far-fetched.

Conclusion: Lost to Time?

If you’d ask me – which you shouldn’t, as I am just an enthusiastic amateur who writes blogs on the worldwide web – I tend to think Huehnergard has the right of it. Though we may never know for certain, because of a fundamental and inevitable problem. Despite the hundreds of thousands of clay tablets that have been excavated, we still lack enough material to be entirely sure. Just think of all the spoken words in the millennia long existence of Akkadian that we miss out on! One thing that I am rather firm on, however, is that we should be careful when dismissing the proposed verbal forms that may be proof of the existence of a hypothetical R-stem as being mere ancient mistakes.

Because it is a dangerous business to try and discern mistakes in ancient texts that were written in languages which we as yet do not entirely understand. A brilliant and mind-expanding book which elaborates on this subject is Principles of Akkadian Textual Criticism by Martin Worthington. Mistakes are indeed present in our ancient corpora. Especially because many of the surviving texts, like those from the aforementioned Old-Babylonian Period, were school exercises and copies made by example or by ear. And the physical circumstances in which copyists worked may also not have been conductive to attentive labor, as some of them eloquently complain in their own words.¹¹ Through such settings and practices, not all these texts are as fit to deliver grammatical perfection to posterity.¹² But, as Worthington also warns, some things we now consider mistakes may receive an exonerating explanation in the future.¹³ It is therefore pertinent to keep approaching mysterious differences between our verbal forms with an open mind and consider all possibilities. And that’s where the R-stem might still be useful.

And this is so fascinating about linguistics and philology. In almost any scholarly debate within these disciplines, all sides are worth studying and all the minds involved have something to teach us. Next week we are going to again encounter some educational scholarly disagreements when we are looking for yet another forgotten god: the ancient Anatolian deity Telipinu.

Please enable JavaScript in your browser to complete this form.

Subscribe and never miss a blog!

Submit

References

1. John Huehnergard, A Grammar of Akkadian (Winona Lake: Eisenbrauns, 2005), p. xxi-xxvi; Rebecca Hasselbach-Andee, “Akkadian”, in: Rebecca Hasselbach-Andee (ed.), A Companion to Ancient Near Eastern Languages (Hoboken: John Wiley & Sons, 2020), p. 129-131.
2. Huehnergard, A Grammar of Akkadian, p. 522.
3. Note that the form ‘šarrāqītum’ has not been attested in the Old-Babylonian dialect, which is what I use for most of my examples throughout this blog. Only a later form without the final -m is known: šarrāqītu, see: Jeremy A. Black et al (eds.), A Concise Dictionary of Akkadian (Wiesbaden: Harrassowitz, 2000), p. 361; Erica Reiner et al (eds.), The Chicago Assyrian Dictionary – Vol 17: Š (Part I) (Chicago: The Oriental Institute, 2004), p. 69-70. I added the final -m, also called a mimation, to make the examples appear uniformly Old-Babylonian and so avoid confusion.
4. Of course, these derived stems have more functions than I have time for today and those may someday merit their own blog. For more details, see: Bert J.C. Kouwenberg, The Akkadian Verb and its Semitic Background (Winona Lake: Eisenbrauns, 2011), p. 245-246, 271-277, 294-299, 327-333. For these introductory remarks and the following overview – including the examples given – see: Huehnergard, A Grammar of Akkadian, p. 252-258, 297, 299-300. 358, 361-362.
5. Kouwenberg, The Akkadian Verb and its Semitic Background, p. 438-439.
6. Ibidem, p. 440-441.
7. David Testen, “Internal Reduplication among the T-Derived Verbal Stems of Akkadian”, Journal of the American Oriental Society 2021, 141 (2), p. 355, 362-363.
8. Kouwenberg, The Akkadian Verb and its Semitic Background, p. 438-439.
9. A few influential examples will suffice, see: Burkhart Kienast, “Weiteres zum R-Stamm des Akkadischen”. Journal of Cuneiform Studies 1961 15 (2), p. 59-61; Jean-Marie Durand & Dominique Charpin, “Nouveaux Exemples de «R Stems»”, Nouvelles Assyriologiques Brèves et Utilitaires 1988, 17 (1), p. 11-14; Brigitte R.M. Groneberg, “Reduplications of Consonants and “R”-stems”, Revue d’Assyriologie et d’Archéologie Orientale 1989, 83 (1), p. 27–34.
10. Huehnergard, A Grammar of Akkadian, p. 363-365.
11. Worthington, Principles of Akkadian Textual Criticism, p. 6, 8, 17, note 62.
12. Martin Worthington, Principles of Akkadian Textual Criticism (Berlin: Walter de Gruyter, 2012), p. 20, note 73. Paul Delnero, “Pre-Verbal /N/: Function, Distribution, and Stability”, in: Jarle Ebeling & Graham Cunningham (eds.), Analyzing Literary Sumerian: Corpus-Based Approaches (London: Equinox, 2007), p. 110-111.
13. Worthington, Principles of Akkadian Textual Criticism, p. 8,19, 162, 309.