by Patrick C. Ryan


N.B. When I completed this essay in 1998, I was not familiar with N. E. Collinge's The Laws of Indo-European, or the theory proposed by Theodor Siebs, which appears in the volume (1985:155-58), entitled Siebs' Law. Siebs published his proposal in the Zeitschrft für vergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen, as Anlautstudien (Berlin 1904: 37.277-324). Collinge characterizes Siebs' theory as a "word-initial alternation" for all plosives of the form:

#sk(h)- ~ #k(h)- ~ #g-
which I correct to:
#sk(h)- ~ #k(h)- ~ #g(h)-

but has the good grace to add that Siebs later explained that he considered that "the #g(h)- form is basic and the presibilantized variant is secondary". This is exactly the form of ‘my' theory, arrived at independently. Since I subscribe to the theory that all original PIE roots have the form *CVC, the unsibiliantized variant must be primary. I would go a step further than Siebs, and assert that occasionally the unsibilantized basis has not survived, and interlanguage-family comparisons can be made by assuming its onetime presence from analysis of forms beginning *sC-, where *C is any voiceless (aspirated or unaspirated) plosive. In view of the ‘laryngeals' which were not recognized when Siebs originally wrote, I would further formulate:

#sVC- ~ #HVC-

Several commentators have discussed and occasionally modified Sieb's theory to accommodate their own predispositions (Kurylowicz 1935; Illich-Svtych 1961; Szemerényi 1972).

My belated admiration goes to Theodor Siebs for his, in my opinion, insightful accomplishment.

Collinge concludes his discussion of this matter with: "On balance, there seem to be just enough sensible etymon-links to justify relating, e.g. sk(h) and g(h) etc. in Siebs' way."

e, a, o

in Pontic-Nostratic CA/E/O became Ca (with - or y/w) in earliest IE and Afrasian, allowing Ablaut in IE (and vowel-patterning in Afrasian) to be employed for grammatical purposes.

b, bh, p, p[h], w, d, dh, t, t[h], s, g, g[^], gh, g[^]h, k, k[^], k[h], k[^][h],
g[w], g[^][w], k[w], k[^][w], m, n, (n)g, (n)g[^], (n)k, (n)k[^],
r, l, l[^], L, H1,2,3,4

Note: *g[w]h, which has divergent reflexes in the IE daughter languages, is a result of an earlier combination of g[w] + H.

([h] indicates aspiration; [^] indicates a palatalization; [w] indicates velarization; L indicates a velar /l/;
l[^] indicates a palatal /l/; H indicates an unspecified "laryngeal", which can occur as
H1 (neutral), H2 (a-color), H3 (o-color), H4 (e-color))


When PIE causative *s(o:)-mobile, which is derived from PL S[H]O, ‘clan-member, that one,*he', is combined with an initial voiced stop or voiced aspirated stop, the voiced stop or aspirated stop is de-voiced: e.g. s + b(h)VC(V) becomes sp(h)VC(V);

  • when *s-mobile is combined with r-, a /-t-/ is inserted [s-t-r-] for euphony and ease of pronunciation: e.g. *2. streig-, "stiff", from s- + (*reig^-), "stretch out".)

  • when *s-mobile is combined with H-, apparently, the ‘laryngeal' (H) is occasionally lost. For example, it is possible that IE *swem-, ‘swim', is to be analyzed as s + **H2ewem from PL S[H]O + HHA-F[H]A-M[H]O, "cause to wander on the waters" (cf. also *swomb(h)o-s, "spongy, porous", from its characteristic of floating.

  • This s-causative occurs in Egyptian as s-, and in several Semitic (Afrasian) languages as s/š.

    There is, therefore, an interesting semantic relationship between many IE roots reconstructed with voiced stops and voiced aspirated stops
    apparently related forms with voiceless stops and (if you admit their existence) voiceless aspirated stops, which I believe derive from PL voiceless affricates, when preceded by *s-mobile.

    There are a number of interesting possible correlations:

    *1. bhel-, "shiny, white, split" and *2. (s)p(h)el-, "gleam, shimmer";

    *4. bhel-, "bloom" and *1. (s)p(h)el-, "split";

    *6. bhel-, "resound, roar" and *(s)p(h)el-, "speak loudly and emphatically";

    *bhend-, "sing, sound beautifully" and *sp[h]end-, "present a libation, promise";

    *1. bher-, "bring, carry" and *2. (s)p(h)er-, "strew, sow";

    *2. bher-, "well up, boil, ferment" and *1. (s)p(h)er-, "jerk, wriggle, shoot out of";

    *3. bher-, "score, split" and *4. (s)p(h)er-, "rip, shred";

    *7. bher-, "plait, weave" and *3. (s)p(h)er-, "turn, wind";

    *bhereg-, "hum, roar" and *(s)p(h)ereg-, "**speak (cf. MHG sprechen)";

    *1. bhereg[^]-, "break, crack" and *3. (s)p(h)ereg-, "jerk, jump, spray, burst";

    *4. del-, "drip" and *1. (s)tel-, "let flow, urinate";

    *5. del-, "long" and *2. stel-, "spread out";

    *3. dhen-, "strike, push" and *1. (s)t[h]en-, "thunder, rumble";

    *2. dher-, "hold, hold firmly, support" and *1. (s)t[h]er-, "rigid, be stiff, fixed object ";

    *5. dher-, "filth, defecate" and *8. (s)t[h]er-, "unclean liquid, manure";

    *geibh-, "buckled, humped" and *ske/e:ibh-, "crooked, limping";

    *1. gel-, "ball (up)" and *4. (s)kel-, "joint";

    *3. ger-, "turn, wind" and *3. (s)ker-, "turn, bend";

    *ghabh-, "grasp, take" and *sk[h]abh-, "support"; cf. also *ghabh(o)lo;

    In this process, palatalized and velarized dorsals are de-palatalized and de-velarized:

    *2. g[^]hel-, "cut" and *1. (s)k[h]el-, "cut";

    *2. g[^]her-, "scratch, score" and *4. (s)k[h]er-, "cut";

    *5. g[^]her-, "intestine" and *sker-(d)-, "defecate";

    *6. g[^]her-, "small" and *1. (s)k[h]er-, "shrivel up";

    *7. g[^]her-, "stiffen" and *1. (s)k[h]er-, "dried, frozen";

    *2. k[^]e/e:i-, "sharpen, whet" and *ske/e:i-, "cut, separate, part";

    *werg[^]h-, "press" and *swergh-, "care for";

    *g[w]er-u-, "pole, spear, stake, thorn" and 1. (s)ker-, "shrivel up, rumple, raw skin,
    scab, crust, dried out, emaciated, troubled"

    Of course, it was also added without any phonetic changes:

    *2. ak[^]-, "sharp, pointed, **high and *sak-, "sanctify, make a treaty";

    *3. aw-, "off, away from" and *saus[au]-, "cut";

    *10. aw(e)-, "blow" and *4. seu-, "boil, be strongly moved"

    *1. mai-, "hew, hew off" and *2. sme:i-, "carve, cut";

    *2. mai-, "bespot, dirty" and *smei:-, "smear, wipe over";

    *6. mel-, "dark, impure" and *2. smel-, "gray, dust-colored";

    *1. wei-, "turn, bend" and *swe/e:(i)-, "bend, swing";

    *we-n-g-, "be bowed" and *sweng-, "bend, swing";

    *6./8. wer-, "speak ceremoniously, speak/become aware of" and *1. swer-, "speak, orate";

    *7./10. wer-, "rip up, scratch/flow" and *4. swer-, "fester";

    The presence of *s-mobile for Pontic-Nostratic is attested by its ability to sporadically inhibit the change of N[H] and NE to IE /l/:

    *1. la:-, "move back and forth" but *sna:-, "wash";

    leubh-, "like, desire, dear" but *1. sneubh-, "engage, marry"

  • even when the *s-mobile has subsequently disappeared:

  • le/e:b(h)-, "hang down slackly, lip, *peel/hull" but *1. (**s)nebh-, "burst, unpeel itself"

    however *leu-, "hang down slackly, slack" and (s)leu-, "hang down slackly, slack"

    The *s-mobile, a very common phenomenon, has been characterized by IEists as imparting no particular meaning to the verbs it begins; a meaning that would be nearly undetectable is a simple causative nuance, particularly if the causative was allowed reflexive or passive use.

    The *s-mobile, so-called because it can be removed without apparent semantic change, should not be confused with another IE prefixed formative, su-, which is simply the IE element su/u:-, "well, good"`, i.e. "energetically . . ." or ". . . to completion".

    Interestingly, some Sumerian verb formations contain the prefixed element šu-2, "totality", which is probably an analogous device to the prefixing of IE su/u:-, "well, good", whether or not cognate:

    *H1ed-, "eat" and swa:d-, "sweet (="good-eating")" (*sowá?ed- -> *swá?d- -> swá:d-)

    the latest revision of this document can be found at

    Patrick C. Ryan * 9115 West 34th Street - Little Rock, AR 72204-4441 * (501)227-9947