In the Bronze Age, two incompatible alphabetic writing systems developed independently in Canaan—to the North, Ugarit’s alphabet had an interface derived from syllabic cuneiform and, to the South, miners in the Sinai used a hieroglyphic ideogram interface. From the outset, both of these regional scripts relied upon omission, where readers had to fill in gaps by supplying missing vowels. Also from the outset, both symbolic systems used individual graphemes to represent individual consonants. Both systems had their strengths, similar to the communications format war of the 1980s between Sony (Betamax) and JVC (VHS). If that reference seems obscure, you might compare these parallel but incompatible alphabetic technologies to the choice of Blu-ray versus HD-DVD, or Chrome versus Safari—although both performed well, only one format thrived. The cuneiform version had the benefit of deriving from a common international cuneiform system, using common clay scribal media and tools, and having the advantage of disambiguating three specific vowel combinations with the consonant ˀalp (ˀa, ˀi, ˀu). This Ugaritic version would play the part of Betamax in our analogy, a version with a bit more precision but perhaps not as accessible.
The ideographic version had the benefit of borrowing certain known monoconsonantal Egyptian graphemes (“ox” or “house”), using common papyrus scribal media and tools, and applied them to Canaanite dialects using the acrophonic principle, where the value of the initial consonant symbolized the consonant itself (𐤀 for ˀalp “ox” and 𐤁 for bayt “house”). In our simile, Proto-Sinaitic would parallel the VHS, proving more accessible and eventually gaining a larger following. This type of alphabetic symbolism pertains directly to our analysis of abbreviations, or initialisms, where the initial consonant symbolizes the entire word. Note how, when learning the alphabet, students still use posters and flashcards that employ this acrophonic principle:
Neither of the two competing systems had spread very widely before the catastrophic system-wide collapse at the end of the Bronze Age ca. 1200 (or 1177 BCE). The Phoenician city-states along the Levantine coast, however, had been early adapters of the same type of Canaanite system found in the Sinai. Their version of the alphabetic format, and the derivative format adopted by the Aramaeans, became the dominant design during the Iron Age. Versions of the Phoenician alphabet spread from the Levant to the Aegean and even to the far West of the Mediterranean, while the Aramaic adaptation spread throughout the Assyrian and then Persian empires (א for alāp “ox” and ב for bayit “house”). These Iron Age alphabets ushered in a transformative advancement in communication technology, on par with the Gutenberg press. Again, if that might seem obscure,or if even the telegraph might seem obscure, then think of the transformation in communications brought on by smart phones and the internet as analogous to the alphabetic advancement.
As mentioned in passing above, the Phoenician design relied on omission of vowels. This economy of signs required readers to fill in the gaps, yet their alphabet still functioned well within the triconsonantal root system used by the Semitic language families. Forming verbs and nouns with such radicals renders vowels nonessential for expression, since they can be readily supplied by context. Just as today, when a fluent speaker of Arabic or Hebrew can navigate public signage lacking vowel markers, so too in antiquity those who used this alphabetic system could supply the vocalization (NB archaic Hebrew used Phoenician letter forms, both for epigraphic monuments and daily record keeping). In later eras, however, some vowel marking did prove necessary for disambiguation (i.e. matres lectionis, Masoretic diacritical marks, etc.). When pronunciation began to diverge over time (due to chronological distance) or between regions (due to geographical distance), vowel marking could preserve or recover a presumed original pronunciation, an artificial method to permanently fill in the gaps.
When Greek city states adopted this dominant design, they borrowed the Phoenician alphabet wholesale, with acrophony but without comprehension. For example, the utility of the initial phonetic value for A (alpha) and Β (beta) still pertained, but their phonemic value did not—i.e. the terms became meaningless labels, no longer signifying “ox” and “house.” Eventually the direction of the text also varied, becoming retrograde rather than sinistrograde. Indo-European language families had a greater need to indicate vowels for comprehension, not being based upon on triconsonantal radical system. Fortunately, they had a lesser need to mark as many consonants, so those symbols could transfer to vowels. For example, the weak glottal stop represented by ˀalp in Phoenician could mark instead the following vowel. This Indo-European accommodation of vowels, a move made out of necessity, has been marked by some as the true origin of the alphabet, relegating the non-vocalic Semitic systems to the category of ˀabjad. But this modern term was contrived to mark vocalic disambiguation as the key characteristic of an alphabet, magnifying the significance of the Greek adaptation. We have little interest in polemical claims to priority or superiority. and employ only the term “alphabet” to describe these divergent forms of the dominant design.
Beside omission of vowels from individual syllables, at the order of phrase or sentence structure, certain votive inscriptions were so common and so formulaic that the system could omit words or entire clauses and yet readers could still reasonably fill in gaps. Votive inscriptions make up most of the Phoenician-Punic epigraphic corpus, with brief inscriptions from open-air precincts dedicated to Baˁl comprising an astonishing majority. In fact, more than 90% of all inscriptions that survive, in both the Phoenician block script and in the Western Punic/Neo-Punic cursive script, originate from just one sanctuary, dedicated to Baˁl and his consort Tinnit in Carthage. These terse and rigidly formulaic texts might start with a prepositional phrase of dedication (“to the deity”), a predicate nominative in the main clause (“this is a votive”), a relative clause naming the dedicant (“which PN vowed”), perhaps a temporal clause (“in the magistracy of PN”), and finally a causal blessing (“because the deity heard my voice and blessed me”), but they need not fall in this order and we rarely find all of these phrases together. Often elided were both the temporal clause and the final blessing, but even the main noun phrase could be omitted with the expression still understood. The reader could fill in those gaps because something had to be offered, on some occasion, when some blessing was received. In the main clause, the predicate verb “to be” was always elided, an omission that many language families allow if not expect. When the dedication did actually appear as the predicate nominative, it commonly and poetically repeated in the verb of the following clause—i.e. NDR ˀŠ NDR “[this is] the votive which he vowed,” MTNT ˀŠ YTN “[this is] the gift which he gave.” Whenever the pleonastic main clause would be omitted, a denominative form of the verb from the following relative clause could be understood to fill in the gap—i.e. “[the votive] which he vowed.” Rigid formulaic syntax allowed omission of the very core of the sentence, the main clause, yet the full ritual fomulae could still be expressed.
Nevertheless, the efficiency gained and the time saved might come at the cost of clarity. Brevity achieved through omission could lead to misunderstanding, particularly among interlocuters unfamiliar with the syntax of the formulaic system:
Efficiency would only function properly with shared understanding, within a group, thus excluding foreigners ignorant of the language, or even those of a different age, social class, or level of education. The in-group could understand the entire formulaic expression with very little information, even when reduced to individual terms—the name and ritual action of the dedicant (NDR ḤNˀ “[that which] Ḥanno vowed” CIS I.4968 – see image below), or just the name alone (ḤNˀ “[that which] Hanno [vowed/sacrificed]” CIS I.4980, 5220, 5514) or ritual act/artefact alone (NDR “[that which someone] vow[ed]” CIS I.5631, or ZBḤ “[that which someone] sacrifice[d]” CIS I.3807). We lack epigraphic examples that preserve only the name of the divinity, but we do have non-alphabetic symbols—e.g. solar disc, lunar crescent, “sign of Tanit”, orant hand, beytl, etc.—used as icons to convey the presence of the divine. These intact and well-crafted stelae, with only individual words and symbols inscribed, must convey a message similar to the full inscriptions placed adjacent in the same archaeological/social context.
So, too, intact and well-crafted stelae with individual alphabetic symbols—one letter alone or a few widely spaced—still carried the full message, if only a reader could fill in the gaps. With all of the previous as prelude, we have reached the focus of our essay—how Phoenicians used individual letters to symbolically augment or supplant epigraphic ritual commemorations. The use of individual symbols increases the opportunity for misunderstanding, again trading clarity for efficiency. Perhaps more than those in other eras, we recognize the necessity for efficient expression when limited to 140/280-character tweets. This necessity has generated a wide range of new expressions using individual graphemes, whether abbreviations or contractions, acronyms or initialisms:
Across the changing landscape of Twitter, rapid innovation when forming abbreviations quickly excludes not only non-native speakers but also older generations. In our case, chronological and social distance makes recovery of the ritual meaning behind isolated Phoenician letter forms a serious challenge.
We have preferred to use the term initialism here, because we assume that each letter was expressed individually (FBI, CIA, irl), rather than using the term acronym under the assumption that letters would form new words (yolo, fomo, potus). In extreme cases of ellipsis, the initialisms (or acrophony) might reduce the entire ritual to just a few symbols, possibly representing a divine name and epithet (e.g. B Ḥ for “Baˁl Ḥammon” CIS I.4781) or representing paired divine names (e.g. B T for “Baˁl [and] Tinnit” CIS I.436, 4782, 4784):
Beside pairs, a lone alphabetic symbol might provide the initial letter of a proper name (e.g. B — CIS I.3808), either of the deity (“Baˁl”) or of the dedicant (“Baˁl hanno,” “Baˁladdir,” etc.). Even with such extreme ellipsis, the entire ritual of dedication, votive action, occasion, and blessing must still be understood within that single letter form. Besides proper names, we do have rare cases of possible abbreviation when marking the occasion, not in isolation but introducing the temporal clause (ˁT R “at the time of the magistracy of…” CIS I.132, cf. Mactar 32-36). Here the reˀš might stand for some abstraction (from √rbb – to become “numerous” or “great,” cf. Latin magis > magister). In our prior examples with letters appearing individually, we presume that these initials marked personal or divine names but, in this unique example, we presume that the rigid formulae, themselves ripe for abbreviation, allowed for reduction of the temporal clause (cf. BC or AD). We presume because, with such extreme ellipsis, we can only hope for a plausible reading rather than a secure one. Even in antiquity, the meaning of such initials was likely unknown outside of a limited circle.
In situations where we have more than one disconnected letter symbol, rather than initialism early editors inclined toward abbreviation or contraction, in the same way that certain titles or proper names might be abbreviated with first and final letters (Dr., Mr., Wm.). In fact J.-B. Chabot, one of the editors for the Corpus inscriptionum Semiticarum, consistently and ingeniously presumed that pairs of alphabetic symbols represented first-and-final abbreviations (e.g. ˀ S for “ˀ[Eshmunˁama]s” CIS I.914, 3815; ˁ T for“ˁ[Abdmelquar]t” CIS I.2125; ˁ R for “ˁ[Akba]r” CIS I. 4777; Ḥ L for “Ḥ[anniba]l” CIS I.3006), despite that fact that in each case these letters were separated from each other by non-alphabetic icons—e.g. caduceus, vase, etc.—and were normally set above full votive formulae. In other cases multiple letters, might be uninterrupted but integrated within the iconography, e.g. set aside between the legs of a ram (ˁ T, perhaps again“ˁ[Abdmelquar]t” CIS I.3683), below or within a “sign of Tanit” ( ˀ M Q L Y, CIS I.3683; B ˀ G T, CIS I.3812), or on the base supporting a “bottle idol.” (Ḥ B N ˁ, CIS I.3821). Rather than proposing ingenious interpretations for these letter clusters, we grant only that they must have held deep symbolic meaning to their engravers and their in-group.
For our final examples of this type of symbolic meaning, we turn to Cirta (or Constantine, modern el-Hofra). There inscribed stelae were discovered ex situ in a secondary deposit, buried in a fossa, but they must have once stood in an open-air sanctuary dedicated to Baˁl, similar to one in Carthage. As the dominant design of alphabetic script spread across the Mediterranean, the Phoenician and Greek systems diverged but then converged again here in Cirta by the 2nd century BCE. As an example of this convergence, we find rare examples of Greek letters preserving vocalized Phoenician, with votive dedications to ΒΑΛ ΑΜΟΥΝ and his consort ΘΙΝΙΘ (el-Hofra 01 GR) or ΘΕΝΕΙΘ (03 GR). As in Carthage, we find examples of Phoenician initialisms separated by icons and set outside of cartouches holding votive dedications (e.g. B for “Baˁl,” el-Hofra170 PUN). But we also find examples of convergence, with a Phoenician initial centered below the Greek dedication and set within the cartouche (ˀ possibly for “ˀ[Adon]” 06 GR) or with Greek initials set outside the cartouche and centered above the Phoenician dedication (Α Ν, perhaps as a first-and-final abbreviation for the epithet Α[μου]ν, 139 PUN – see above). In both instances the engraver, writing in one language, used alphabetic characters from another, symbols that carried meaning similar to an icon.
In one intriguing case, however, we have a letter serving as both grapheme and icon – both an anthropomorphic alpha and a “sign of Tanit”. We also have convergence of Greek and Phoenician not only in the transcription of the divine name but also in the directionality of its script, but this does require a series of cognitive leaps:
First, the anthropomorphic alpha differs from the clearer “sign of Tanit” icon set directly above the cartouche, even though the alpha form does resemble icons from Carthage. Second, interpretatio graeca at the Cirta precinct either translated Baˁl Ḥammon as Cronus (el-Hofra 03 GR) or simply transcribed the divine name as Bal Amoun (01 GR), although in this instance the interpetatio evokes only an epithet. Third, the bidirectionality of the text must move outward from the central symbol. Note that each of these leaps is tenuous. For comparison, the stylized initialism of the logo for Nine Inch Nails can be read both sinistrograde, as in Phoenician (ИIИ), and retrograde, as in Greek (NIN), with the stance of the alphabetic symbols indicating directionality. In our case (15 GR), however, without any change in stance we must still presume to read outward from the center icon, twice, simultaneously quasi-sinistrograde (ν[υομ]Α for Phoenician Ḥammon) and quasi-retrograde (for Greek Α[μου]ν, cf. 139 PUN). If all three conditions hold—central alpha/icon, epithet in place of divine name, unique bidirectionality—then this ingenious reading shows symbolic representation of both the deity and his consort. In any case, this final inscriptional text demonstrates alphabetic convergence and the power of individual alphabetic signs to bring meaning to ritual.
~ Brien Garnand, Visiting Research Fellow, NINO Leiden University