I have found that Early Arabic poetry is in some ways even more difficult to translate than its Medieval Arabic successor, for totally opposite, yet parallel, reasons.
In the case of Medieval and Modern Arabic, just like Medieval and (Early) Modern Hebrew, an awareness on the part of the modern scholar (and intended audience) of the literary tradition, with all its weight and complexity, allows a plethora of meanings to be injected into, and extracted from, even the briefest fragments- such that even a single hydronym like "the (river) ˁAqīq" or a single personal name like "Laylā" in a poem by ˁUmar ibn Al-Fāriḍ can be incredibly suggestive. Similar phenomena, on a much smaller scale, lie behind the phantasmagoria of associations which might suface upon encountering a phrase like "as if on Juliet's balcony", or the reason why we smile upon hearing a phrase like "lead us not into temptation, for we can find it ourselves". For a translation to embody the context of its own allusions is, of course, impossible. Whether it is the poem that fails by requiring the reader to "import" the allusive framework, or the reader who fails for not being able to do so is something of an academic question. In traditions as insular or intertextual as Arabic, Persian or Hebrew, or Classical Chinese for that matter, it is probably a meaningless question anyway.
In any case, with Ancient Arabic literature (roughly from the late 5th to the late 7th century AD), also paralleling the case of its Ancient (biblical) Hebrew counterpart, the problem is one of deficit, rather than surplus. Here, one is working with the very root of the poetic tradition as it survives, yet not the root of the tradition as such. The tradition itself as we have it did not spring into being suddenly, but is the product of a long process of development whereby formulae, symbols, conventions etc. slowly took shape in a period of time now shrouded in the mists of Arab autochthony. Almost all of that process is sadly lost to us. We therefore may have little idea what a given symbol even meant to its original 6th or 7th century audience, if symbol it even be. Often all we know is what later audiences did with it, and how they developed it. Sometimes we simply have no idea what a given word even really meant.
For example, when the poet Al-Shanfarā uses the term Umm Qasṭal "mother of (Qasṭal)" as an epithet (presumably) for "War", we have no clue what "qasṭal" means. The word's later meaning of "dust" (and therefore war being "dust's mother", in reference to the dust of war) was apparently derived from its occurrence in Al-Shanfarā's poem, presumably because (or perhaps with the result that) the word's earlier meaning was lost, (along with the original meaning of the name Al-Shanfarā, it would appear.)
Yet a further complicating factor is that this poetry was transmitted orally for a matter of centuries before being set down in writing, with all the issues oral transmission entails, including the very serious question of whether it even makes sense to speak of an "original" text or of "original" meanings or even of an "original" author at all, let alone embark on a tortuous journey to discover same, anymore than a modern Classicist should go about wondering what the "ur-text" of the Iliad was like (when it never existed) or wonder what it would be like to meet Homer (when neither did he.)
Suzanne Stetkevych in her study "Pre-Islamic Poetry and the Poetics of Ritual" makes the case that the association of various early poems with various poets' names is thematic or semantic, and not historical. Certainly the biographies of the poets as they have been handed down to us appear to be essentially folkloric material derived from the poems.
Even so, I would argue that, unlike the Homeric poems, Early Arabic verse displays what might be called "an individual voice". The fact that some poems were composed at a discernible point in time or even for distinct individuals further suggests that the concept of an original author is not entirely meaningless here. In this, Early Arabic poems have more in common with Hesiod than with Homer. The latter is simply a figura around which a body of literature coagulated. The former is the name of an actual historical individual, whose personality and style is discernible through some (though not all) of the works attributed to him- works which, admittedly, underwent much transformation in the process of being handed down over time so that we most likely haven't anything like his ur-text. Particularly the mukhaḍramūn or "co-Islamic" poets who lived during the birth of Islam seem to be real people in real time, even if many of the akhbār or poetic biographies attributed to them are clearly legends, with discernible literary personalities. Note: I use the term "Early Arabic" as a nice catch-all to combine both pre-Islamic (jāhilī) and "co-Islamic" (mukhaḍram) literary and linguistic phenomena. Anyway, in literate terms, one may speak of Labīd or al-Khansā' as author, and of the oral transmission as collective editor.
A historical, diachronic perspective is, furthermore, essential to scholarly appreciation (though not the artistic enjoyment) of these poems- and it therefore becomes necessary to situate these poems at something approaching a particular point in time. What did a word mean at the time of a verse's composition? If a line is a late accretion, what were the semantics and pragmatics governing word-use a the time of addition? These questions require answers if one is to understand the world of Early Arabic poetry. The ways in which (pre-literate) Early Arabic differed from later (literary) "Classical" stages of the language, to say nothing of the modern (standard) form of the language which students (including, nowadays, myself) are most familiar with, are considerable. There are a number of early constructions that later changed their meanings or died out e.g. laˁalla meaning "so that" and not "perhaps"; qad with the imperfect meaning "indeed"; lam with the jussive as a present negative; kāna meaning "he is" etc.. Moreover, and perhaps more seriously, there has been a significant change in vocabulary.
For example, in Labīd's most famous lament for his step-brother, the phrase šaqiyyun bi-l-maˁīšati qāniˁ, when read as if it were modern literary Arabic (the stylistically odd separation of a participle from its noun by a qualifying adverbial phrase notwithstanding) might give a meaning roughly translatable as "a vile wretch who is content to live the way he does." Unless one is familiar with older layers of the language, one would be tempted to read qāniˁu as the present participle of the verb qaniˁa/yaqnaˁu "to be content" governed by the preposition bi- "with" which modifies maˁīša "livelihood, way of life." But qāniˁ here is more likely the present participle of the (now defunct) verb qanaˁa/yaqnaˁu "to beg, ask for alms" (which has been replaced in modern literary Arabic with the verb šaħaða), with a meaning close to that found in Qur'an 22:36 aṭˁimū l-qāniˁa "feed the beggar". Bi- (alongside li- which survived with this usage in later Arabic) could be used in Early Arabic to introduce a verb's object when the accusative was undesirable for some reason (in this case a metrical one.) Thus a far more likely, and indeed more contextually intelligible meaning of the same phrase is "one who is unlucky and (reduced to) begging for his sustenance." But perhaps bi- is intended to mean "in" here, and thus one could read "one who is unlucky in his livelihood and remains poor." Which of these is more likely? How can one tell?
Speaking of which, the verb qaniˁa/yaqnaˁu "to be content" did also exist in Early Arabic. So is a polysemantic pun intended here, or is only the primary denotative meaning of "begging" relevant? Again the same questions surface: Which is more likely? How can one tell?
The answer to the question "how can one tell" is often simply "we can't." Leaving aside the relatively meager corpus of inscriptional Old Arabic (which contains, to my knowledge, just one poem of three lines), our sources for early Arabic are limited to the Qur'an on the one hand, and on the other to the early poetry itself. Sometimes we simply lack the information, contextual, lexical, historical or otherwise, to know what exactly was meant. More seriously still, we sometimes simply lack enough evidence, which we do fortunately have in the above case, to figure out what kinds of questions should even be posed.
Even the classical commentators, to whom many a student such as yours truly might be tempted to turn in their darkest lexical hour, though often goldmines of information, can often turn out to be worse than unhelpful. The commentators are often guessing, or relying upon the educated guesses of lexicographers. The coming of Islam had occasioned a massive shift in vocabulary, as an Arabian society dominated by nomadic (=bedouin) elements and a kaleidoscopic intermix of Jewish, Christian and Polytheistic communities, in which contact with non-native Arabic speakers must have been relatively limited, gave way to the new township-centered colonialism of a vast caliphal empire in which Arabian "Pagans" were either killed off or converted, and where new populations were assimilated wholesale making native Arabic speakers, for a considerable while, an elite minority in many areas. In such a context, language change, even in the prestigious "literary" register of the language, was inevitable, extensive and swift.
The numerous anecdotes about the bedouin being consulted on problems of vocabulary show that the difficulties caused by the shift were formidable and well-recognized and that considerable effort was made to seek out sources that were thought to shed light on the vast and increasingly obscure lexicon covering the minutiae of nomadic life in the desert. The meanings of many words are permanently lost.
For example, when the poet Abu Dhu'ayb refers to a she-donkey as being jadūd, we basically have no idea what the word could possibly mean in that context- the poem's classical commentators appear to have just been guessing based on the word's context.
I try to deal with Early Arabic poetry in a way that is true to the literate, Islamic tradition (for the urban Islamic milieu is the filter through which all of this literature comes down to us- and many late Early Arabic poets such as Labīd were indeed Muslims) while at the same time keeping in mind that this poetry is aural and oral, and that Islam, even for the 7th century poets who converted to it, was a new phenomenon, one that had yet to inscribe itself in the language. Therefore, it must be translated in a fashion which gives primacy to the mnemonic attributes of the original. This means, of course, "formalism" in some sense, either syntactic, metrical or phonetic. Verse is seldom "free" in pre-literate or (as was the case with the ancient Bedu) proto-literate societies. In translating Arabic verse, as with the Persian, assonance, rather than full rhyme, seems to provide a convenient and worthy substitute for the latter.
At times times when I come up against a word whose meaning is, for all intents and purposes, lost to us there is often no choice but to use the commentator's best guesses. If the context affords some clue that the commentators seem to have ignored or missed, I firmly believe the translator should feel at liberty to go with that, as I did when translating this poem by an anonymous woman. This uncertainty, however, should serve as a reminder that translations of Early Arabic poetry, more so than translations of poetry from some other languages and periods, are renditions of verse whose milieu and even language are, and at this point in time must be, only partially understood.