Updated on November 13, 2016.
The ABCs of Linear-Script Decipherment
Any decipherment that focuses on one of the three linear scripts (A,B, or C*), at the exclusion of the others, overlooks the mutual influence among these scripts and misses the opportunity to view the greater picture behind the development of both the Greek and the Japanese languages. While the addition of Linear C to this list certainly adds greater complexity to linear-script decipherment, it also lends a poetic completion to the endeavor.
* “Linear C” is sometimes used as a reference to the Arcado-Cypriot script of ancient Cyprus. I use the term, Linear C, not only to encompass all pre-alphabetic scripts of Cyprus but to also acknowledge the relationship of this body with linear scripts A and B.
Linear A or Linear B to Ancient Greek
Note: Due to the complexity of translation to ancient Greek, this key is in process.
The translation from the linear scripts to ancient Greek is not as neat as that to Japanese (see below). Those who attempt translation must be mindful of multiple factors, which may not be limited to the following:
- The five vowels of the linear scripts had increased to seven vowels by the eighth century BCE.
- Each of the many Greek dialects (e.g. Aeolic, Doric, and Ionic) had its characteristic consonant and vowel (vocalic) changes.
- While some consonant/vowel (CV) sign groups had weakened to complex signs (e.g. /pte/ for /pe-te/), others had arbitrarily weakened by the eighth century BCE, in Homeric prose.
- Some consonants have aspirated counterparts.
- A shift from an unvoiced to a voiced consonant typically indicates a late Greek development in LinB.
Consequently, translation comprises a hierarchy of expected phonetic values for each sign. These hierarchies are undergoing constant refinement so may be subjected to change.
Note that I do not espouse “progressive spelling”, as promoted by Roger Woodard et al. This practice inserts consonants where none exists in LinB to agree with Greek words (e.g. pe-ma to σπέρμα (sperma) “seed”), on the premise that the “borrowed” phonetic system was inadequate to represent the language. On the contrary, my extensive research demonstrates that proto-Greek, as embodied in LinB, was a moraic language that did not initially recognize consonant clusters, which primarily developed as vowels weakened to the point of elision (see #3 above).
A second reason for the development of consonant clusters is found in metathesis. I believe that, in some dialects, /kV/, /pV/, and /tV/ were pronounced /ksV/, /psV/, and /tsV/. For the first two values, compare ξ (ksi) and ψ (psi). For /tsV/ values, I take my cue from the LinA /tu/ to Japanese /tsu/ and the Greek /ts/ and /tz/ values, which are evident not in ancient but in modern Greek. Compare ancient Greek ῥητίνη (hretine) “resin” to modern Greek ῥητίνη or ρετσίνα (hretsina) “resinated wine”. In ancient Greek, however, LinB consonants such as /tV/ (possibly pronounced /tsV in some dialects) may have experienced phonetic metathesis (e.g. ka-ra-ti-ri-jo to χρηστήριο(ν) (khresterion) “an oracle”).
Essentially, I believe that the early Greeks were more than competent to represent their language; conversely, our task is to develop the competence to reveal the language’s secrets.
- /a/ to
- (1) /η/ [eta or long e] (Ionic influence),
- (2) /ε/ [epsilon or short e], and
- (3) /α/ [alpha or a] (Doric influence).
- /a2/ or /i.ja/ to /ια/
- /a3/ or /a.i/ to /αι/
Notes: AB *08 A suggests H (eta) with the inclusion of a vertical stroke.
- /e/ to
- (1) /α/ [alpha or a]
- (2) /ε/ [epsilon or short e], and
- (3) /η/ [eta or long e],
Notes: Variations of AB *38 E suggest A (alpha) with one and two horizontal strokes. However, in diphthongs, such as e-u, /e/ shifts to /ε/.
Note for /a/ and /e/: With some exceptions, these vowel hierarchies may be applied to the consonant+vowel (CV) signs.
Notes: Richard Vallance Janke has astutely noted that AB *010 U suggests the obsolete F (digamma) . (When he sees this reference, he will provide me with the link.) In modern Greek, /ευ/ is pronounced /ef/ or /ev/, so, given the morphology of AB *010, it is quite probable that ancient Greek was pronounced in a similar manner.
- /ja/ to /ια/ (cf. Italian /ja/ to /ia/, esp. in proper names)
- /je/ to /ιε/
- /jo/ to /ἰο/
- final /jo/ or /Ci-jo/ to /ἰο(ν)/ (ion) or /ἰο(ς)/ or
- final /jo/ to /ζω/ (zo)
Notes: For final /jo/ or /Ci-jo/, /ἰο(ν)/ and /ἰο(ς)/ may indicate an adjective, a noun, or a genitive noun; and /ζω/ indicates a verb. An understanding of context will help to determine the word’s role. Moreover, I believe that, at this early stage, the first-person aspect was inherent in the verb but not formalized until much later; LinB appears to be written in third person (s/he or it). Following are some examples. For an instance of both noun and verb in the same tablet, see KN Db 1232.
- a.i-ni-jo | αἰνίζω (ainizo) | to allow, to recommend; to approve, to praise
- a-re-i-jo | Ἡραῖο(ς) (Heraios) | of Hera
- a-ri-jo | ἠρίο(ν) (erion) | one of two; the other
- da-mi-jo | δήμιο(ς) (demios) | of the people; the public
- di-pi-si-jo | δίψιο(ς) (dipsios) | parched, thirsty
- ka-ra-ti-ri-jo | χρηστήριο(ν) khresterion | an oracle
- ka-ra-ti-ri-jo | χρηστήριο(ς) khresterios | oracular, prophetic
- o-mi-ri-jo | ὁμηρίζω (omerizo) | (1) to imitate Homer, prob. from (2) to indulge in unnatural lust (see Homer’s Daedalus)
- /ka/ to
- (1) /χη/
- (2) /χε/
- (3) /χα/
- /ke/ to
- (1) /χα/
- (2) /χε/
- (3) /χη/
- ki/ to
- (1) /χι/
- (2) /χη/
- (3) /χε/
- /ko/ to
- (1) /χο/
- (2) /χω/
- (3) /κο/
- (4) /κω/
- (5) /γο/
- /ku/ to
- (1) /χυ/
Notes: All /kV/ values may shift to the Ionic /κV/. Some words may be represented by one or both values. Compare ki-to to χιτώ(ν) and κιθώ(ν) “a garment worn next to the skin, a tunic”. Also, /kV/ values may shift to /σχ/ (skh). I believe that this is an example, not of “progressive” spelling, but of metathesis for medial /k/ values that were pronounced /ks/ (embodied in ξ, pron. ksi). Cf. ξ with the morphologies of A *081 KU > B * 081 KU.
- /pa/ to /φα/ [The sound of φ may be described as both a soft /p/ and a soft /f/: midway between the /p/ in “parent” and the /ph/ in “phase”.]
- /pa2/ [see Q values].
Notes: Compare the shift of LinA /pa/ to Japanese /fu/. Exceptions may include foreign toponyms that retain their pronunciations.
- /pa2/ or /qa/ to
- /πV/, /φV/
- /κV/, /χV/
Notes: AB *016 has alternately been assigned /pa2/ and /qa/. While philologists seem to have settled upon /qa/, the Greeks were not so quick to discard the one for the other. The alternation of κ and π is, perhaps, confounded in k͡p, a labio-velar consonant cluster that may be more accurately described as a velolabial, because it is articulated first at the soft palate (velum) and then at the lips. A telling example is found in to-ro-pa2 / to-ro-qa: compare τρόπα (tropa) and τροπή (tropā) “a turn, a turning” with τροχή (trokhā) “a course, a running”. Compare, further, τροχή with “Turkey”, the ancient site of numerous stadiums and of the hippodrome at Constantinople.
Consequently, it is quite plausible that the value of AB *016 is k͡p. While it can certainly be argued that one example of alternation hardly proves the rule, this example does compel further examination of AB *016, which, in many LB words, assumes one or the other consonantal value. Moreover, k͡p may be compared to the ancient-Greek koppa, which is morphologically similar to AB *016. Consequently, I have resolved the phonetic dichotomy by reassigning the value of AB *016 as QPA (koppa), so that to-ro-pa2 / to-ro-qa may be transcribed as to-ro-qpa.
Notes: LinA and LinB are characterized by the inclusion of /rV/ values and the absence of /lV/ values, whereas LinC is characterized by the inclusion of both /rV/ and /lV/ values. Nevertheless, scholars attempt to reconcile the earlier absence of /lV/, and the later presence of both values in classical Greek, with the presumption of a shift from /rV/ to /lV/ in many LinB words. With limited exceptions, I largely disagree, because I believe that the phonetic split between /r/ and /l/ did not occur until the emergence of the Arcado-Cypriot dialect (ca. 1200 BCE), which is believed to have descended from Mycenaean Greek, as founded on LinB; certainly, the emergence of LinC coincides with the presumed decline of LinB.
On the contrary, it is most likely that the early Greeks did not recognize /l/ as a separate value prior to its inclusion in various Cypriot syllabaries, in which the sign for /lo/ was equivalent to the earlier sign for /ro/. This dichotomy may perhaps be explained in terms of the Japanese syllabary, which is similarly marked by its absence of /lV/ values. It is significant that, as one native Japanese speaker explains, the pronunciation of Japanese /r/ is actually much closer to that for /l/.
- /ta/ to
- (1) /τη/
- (2) /τε/
- (3) /τα/
- (4) /θV/ or /θ/ (transcribed as /te/)
Notes: /ta/ to /θV/ or /θ/ suggests that the latter were originally pronounced /τη/; note the aspirated role of /η/ (eta), which equals English H/h. Cf. Greek H (capital eta). Note, also that /τη/ will typically shift to /θV/ when the vowel is stressed and to /θ/ when the vowel is unstressed. This is not a hard rule, but it is consistent enough to provide a starting point.
- /te/ to
- (1) /τα/
- (2) /τε/
- (3) /τη/
- /wa/ to /α/ [alpha or a]
- initial /we/ to /ε/ or /
- final /we/ to /ς/
- /wi/ to /ι/
- /wo/ to /ω/
Notes: W values typically drop the glide /w/ before the vowel. The main exception is found in the values for /we/, which resembles the word-final /ς/. When /we/ is in final position, it acts as the /-ς/ stem. There are many LinB word pairs, in which one word features the final /we/ (e.g. a-ta-ro / a-ta-ro-we; e-ri-nu / e-ri-nu-we). These suggest that the inclusion of a final consonant without a vowel nucleus was either a later development or a contemporaneous development in an existing or an emerging dialect. As for /we/ in the initial and the medial positions, the tentative conclusion is that /we/ shifts to /ε/.
Note that the ubiquitous final /we/ is replaced by final /se/ in LinC. While these values are not necessarily interchangeable, this predominant value in LinC may underscore rather than negate my theory that LinB final /we/ shifts to /ς/.
Linear A or Linear B to Japanese
This list arises from consistent shift patterns between LinA and Japanese. Nevertheless, while many words defy these patterns, diligent research continues to reveal phonetic values that resolve the anomalies of those words.
Note that the Japanese vowel order — /a/, /i/, /u/, /e/, and /o/ — appears to correspond to the order of frequency in LinA [Packard 1974: 95, 99, 113], in which /a/, /i/, and /u/ show strong relative representation and in which /e/ and /o/ show noticeably weak representation.
Blue entries denote the shifts of LinB values to Japanese.
- /di/ to /ji/
- initial /du/ to /zu/ (tentative)
- medial and final /du/ to /dzu/ or /zu/
- initial /ja/ to /ha/; medial and final /ja/ to /ya/
- initial /je/ to /he/ (tentative)
- /ka/, /qa/, and /qe/ to /ka/
- /pa/ to /fu/ (cf. shift from LinB /pa/ to Greek /φα/ (pha))
- /ra2/ to /ria/; initial and medial /ra2/ to /ra/ before glides /ja/ and /wa/ and before pure vowels such as /u/ (tentative)
- final /re-u/ to /ru/
- final /te-u/ to /tu/ to /tsu/
- /si/ to /shi/
- /ti/ to /chi/
- /tu/ to /tsu/ (cf. Greek consonantal clusters /ts-/ and /tz-/)
- /wi/ to /i/
- /wo/ to /o/ (tentative)
Exceptions to the phonetic key include foreign toponyms that retain their pronunciations.
Linear C to Ancient Greek
While some complex LinB values may be carried over in some words, the values of LinC seem less likely to change in ancient Greek. Fo example, vocalic values are generally retained as written. Nevertheless, my scholarship of LinC is still young, so this conclusion is likely to be revised.
Note that lexemes or words found in series (i.e. in compound words and undivided sentences) may include asterisks in prefix or suffix position, so that, for example, ti-o-se* indicates a lexeme or a word at the beginning of a series, *ti-o-se* indicates a lexeme or a word in the middle of a series, and *ti-o-se indicates a lexeme or a word at the end of a series.
- /a/ to
- (1) /α/ [alpha or a] (Doric influence),
- (2) /η/ [eta or long e] (Ionic influence).
- (3) /ε/ [epsilon or short e], and
Notes: While many LinC words demonstrate the Doric / Ionic duality that is found in LinB, /a/ is generally retained as written, which may indicate a strong Doric influence. Nevertheless, this hierarchy is still being tested.
- final /se/ to /σε/, /σσε/, /σι(ς)/, /ξι(ς)/, or /ς/ (e.g. a-mu-se to ἄμυσσε or ἄμυξις, e-ko-to-se to ἔκτοσε or ἐκτός, and *ti-ma-se to τίμασις)
Notes: The ubiquitous final /we/ is replaced by final /se/ in LinC. While these values are not necessarily interchangeable, the predominant /se/ in LinC may underscore rather than negate my theory that LinB final /we/ shifts to /ς/.