Laryngeal theory

Hypothesis that Proto-Indo-European had phonemes beyond those reconstructed through comparison

The laryngeal theory is a widely accepted hypothesis in the historical linguistics of the Indo-European languages positing that:

In its earlier form (see below), the theory proposed two sounds in PIE. Combined with a reconstructed *e or *o, the sounds produce vowel phonemes that would not otherwise be predicted by the rules of ablaut. The theory received considerable support after the decipherment of Hittite, which revealed it to be an Indo-European language.

Many Hittite words were shown to be derived from PIE, with a phoneme represented as corresponding to one of the hypothetical PIE sounds. Subsequent scholarship has established a set of rules by which an ever-increasing number of reflexes in daughter languages may be derived from PIE roots. The number of explanations thus achieved and the simplicity of the postulated system have both led to widespread acceptance of the theory.

In its most widely accepted version, the theory posits three phonemes in PIE: h₁, h₂, and h₃ (see below). Other daughter languages inherited the derived sounds, resulting from their merger with PIE short vowels and their subsequent loss.

The phonemes are now recognized as consonants, related to articulation in the general area of the larynx, where a consonantal gesture may affect vowel quality. They are regularly known as laryngeal, but the actual place of articulation for each consonant remains a matter of debate. (see below).

The laryngeals got their name because they were believed by Hermann Möller and Albert Cuny to have had a pharyngeal, epiglottal, or glottal place of articulation, involving a constriction near the larynx. While this is still possible, many linguists now think of "laryngeals", or some of them, as having been velar or uvular.

The evidence for their existence is mostly indirect, as will be shown below, but the theory serves as an elegant explanation for several properties of the PIE vowel system that made no sense until the theory, such as the "independent" schwas (as in *pəter- 'father'). Also, the hypothesis that PIE schwa was a consonant, not a vowel, provides an elegant explanation for some apparent exceptions to Brugmann's law in Indic languages.

The beginnings of the theory were proposed by Ferdinand de Saussure in 1879, in an article chiefly devoted to something else altogether (demonstrating that *a and *o were separate phonemes in PIE).

In the course of his analysis, Saussure proposed that what had then been reconstructed as long vowels *ā and *ō, alternating with *ǝ, was an ordinary type of PIE ablaut. That is, it was an alternation between e-grade and zero grade like in "regular" ablaut (further explanations below), but followed by a previously unidentified element. This "element" accounted for both the changed vowel colour and the lengthening (short *e becoming long *ā or *ō).

So, rather than reconstructing *ā, *ō and *ǝ as others had done before, Saussure proposed something like *eA alternating with *A and *eO with *O, where A and O represented the unidentified elements. Saussure called them simply "coefficients sonantiques", which was the term for what are now in English more usually called resonants; that is, the six elements present in PIE which can be either consonants (non-syllabic) or vowels (syllabic) depending on the sounds they are adjacent to: *y w r l m n.

These views were accepted by a few scholars, in particular Hermann Möller, who added important elements to the theory. Saussure's observations, however, did not achieve any general currency, as they were still too abstract and had little direct evidence to back them up.

This changed when Hittite was discovered and deciphered in the early 20th century. Hittite phonology included two sounds written with symbols from the Akkadian syllabary conventionally transcribed as , as in te-iḫ-ḫi "I put, am putting". This consonant did not appear to be related to any of the consonants then reconstructed for PIE, and various unsatisfactory proposals were made to explain this consonant in terms of the PIE consonant system as it had then been reconstructed.

It remained for Jerzy Kuryłowicz (ə indoeuropéen et ḫ hittite, 1927; Études indoeuropéennes I, 1935) to propose that these sounds lined up with Saussure's conjectures. He suggested that the unknown consonant of Hittite was, in fact, a direct reflex of the coefficients sonantiques that Saussure had proposed.

Their appearance explained some other matters as well: for example, why verb roots containing only a consonant and a vowel always have long vowels. For example, in *- "give", the new consonants allowed linguists to decompose this further into *deh₃. This not only accounted for the patterns of alternation more economically than before (by requiring fewer types of ablaut) but also brought the structure of these roots into line with the basic PIE pattern which required roots to begin and end with a consonant.

The lateness of the discovery of these sounds by Indo-Europeanists is largely because Hittite and the other Anatolian languages are the only Indo-European languages for which at least some are attested directly and consistently as consonantal sounds. Otherwise, their presence is to be inferred mostly through the effects they have on neighboring sounds, and on patterns of alternation that they participate in. When a laryngeal is attested directly, it is usually as a special type of vowel and not as a consonant, best exemplified in Greek where syllabic laryngeals (when they appeared next to only consonants) developed as such: *h₁ > e, *h₂ > a, and *h₃ > o.

There are many variations of the laryngeal theory. Some scholars, such as Oswald Szemerényi, reconstruct just one laryngeal. Some follow Jaan Puhvel's reconstruction of eight or more.[1]

Some scholars suggest the existence of a fourth consonant, *h₄, which differs from *h₂ in not being reflected as Anatolian [2][3] but being reflected, to the exclusion of all other laryngeals, as Albanian h when word-initial before an originally stressed vowel.[4]

E.g. PIE *h₄órǵʰiyeh₂ "testicle" yields Albanian herdhe "testicle" but Hittite arki- "testicle" whereas PIE *h₂ŕ̥tkos "bear" yields Alb. ari "bear" but Hittite hart(ag)ga- (=/hartka-/) "cultic official, bear-person".[5][3]

When there is an uncertainty whether the laryngeal is *h₂ or *h₄, the symbol *ha may be used.[3]

Another such theory, but much less generally accepted, is Winfred P. Lehmann's view, based on inconsistent reflexes in Hittite, that *h₁ was two separate sounds. (He assumed that one was a glottal stop and the other a glottal fricative.)

Some direct evidence for laryngeal consonants comes from Anatolian: PIE *a is a fairly rare sound, and in an uncommonly large number of good etymologies, it is word-initial. Thus PIE (traditional) *anti "in front of and facing" > Greek antí "against"; Latin ante "in front of, before"; Sanskrit ánti "near; in the presence of". But in Hittite there is a noun ḫants "front, face", with various derivatives (ḫantezzi "first", and so on), pointing to a PIE root-noun *h₂ent- "face" (of which *h₂enti would be the locative singular). (It does not necessarily follow that all reconstructed forms with initial *a should automatically be rewritten *h₂e.)

Similarly, the traditional PIE reconstruction for 'sheep' is *owi- (a y-stem, not an i-stem) whence Sanskrit ávi-, Latin ovis, Greek ὄϊς. But Luwian has ḫawi-, indicating instead a reconstruction *h₃ewis.

Considerable debate still surrounds the pronunciation of the laryngeals and various arguments have been given to pinpoint their exact place of articulation. Firstly the effect these sounds have had on adjacent phonemes is well documented. The evidence from Hittite and Uralic is sufficient to conclude that these sounds were "guttural" or pronounced rather back in the vocal tract. The same evidence is also consistent with the assumption that they were fricative sounds (as opposed to approximants or stops), an assumption that is strongly supported by the behaviour of laryngeals in consonant clusters.

Jens Elmegård Rasmussen (1983) suggested a consonantal realization for *h₁ as the voiceless glottal fricative [h] with a syllabic allophone [ə] (mid central unrounded vowel).[6] This is supported by the closeness of [ə] to [e] (with which it combines in Greek),[7] its failure (unlike *h₂ and *h₃) to create an auxiliary vowel in Greek and Tocharian when it occurs between a semivowel and a consonant,[8] and the typological likelihood of an [h] given the presence of aspirated consonants in PIE.[8]

It has been suggested by Beekes (1995) that *h₁ is a glottal stop [ʔ]. However, Winfred P. Lehmann instead theorized, based on inconsistent reflexes in Hittite, that there were two *h₁ sounds: a glottal stop [ʔ] and an h sound [h] as in English hat.

In 2004, Alwin Kloekhorst[9] argued that the Hieroglyphic Luwian sign no. 19 (𔐓, conventionally transcribed á) stood for /ʔa/ (distinct from /a/, sign no. 450: 𔗷 a) and represents the reflex of */h₁/; this would support the hypothesis that */h₁/, or at least some cases of it, was [ʔ]. Later, Kloekhorst (2006)[10] claimed that also Hittite preserves PIE *h₁ as a glottal stop [ʔ], visible in words like Hittite e-eš-zi 'he is' < PIE *h₁és-ti, where an extra initial vowel sign is used (so-called plene spelling). This hypothesis has met with serious criticism (e.g. Rieken (2010),[11] Melchert (2010)[12] and Weeden (2011).[13] Recently, however, Simon (2010)[14] has supported Kloekhorst's thesis by suggesting that plene spelling in Cuneiform Luwian can be explained in a similar way. Additionally, Simon's 2013 article revises the Hieroglyphic Luwian evidence and concludes that "although some details of Kloekhorst's arguments could not be maintained, his theory can be confirmed."[15]

An occasionally advanced idea that the laryngeals were dorsal fricatives corresponding directly to the three traditionally reconstructed series of dorsal stops ("palatal", velar, and labiovelar) suggests a further possibility, a palatal fricative [ç].

From what is known of such phonetic conditioning in contemporary languages, notably Semitic languages, *h₂ (the "a-colouring" laryngeal) could have been a pharyngeal fricative such as [ħ] and [ʕ]. Pharyngeal consonants (like the Arabic letter ح (ħ) as in Muħammad) often cause a-colouring in the Semitic languages.[16] Uvular fricatives, however, may also colour vowels, thus [χ] is also a noteworthy candidate. Weiss (2016) suggests that this was the case in Proto-Indo-European proper, and that a shift from uvular into pharyngeal [ħ] may have been a common innovation of the non-Anatolian languages (before the consonant's eventual loss).[17] Rasmussen (1983) suggested a consonantal realization for *h₂ as a voiceless velar fricative [x], with a syllabic allophone [ɐ], i.e. a near-open central vowel.[6]

Likewise it is generally assumed that *h₃ was rounded (labialized) due to its o-colouring effects. It is often taken to have been voiced based on the perfect form *pi-bh₃- from the root *peh₃ "drink". Rasmussen has chosen a consonantal realization for *h₃ as a voiced labialized velar fricative [ɣʷ], with a syllabic allophone [ɵ], i.e. a close-mid central rounded vowel.[6] Kümmel instead suggests [ʁ].[18]

The hypothetical existence of laryngeals in PIE finds support in the body of daughter language cognates which can be most efficiently explained through simple rules of development.

Unambiguous examples are confined to Anatolian languages. Words with Hittite (hh), Luwian h and Lycian x are explained as reflexes of PIE roots with h₂.[19]

Some Hittitologists have also proposed that "h₃" was preserved in Hittite as "ḫ", although only word initially and after a resonant. Kortlandt holds that "h₃" was preserved before all vowels except "*o". Similarly, Kloekhorst believes they were lost before resonants as well.

Reconstructed instances of *kw in Proto-Germanic have been explained as reflexes of PIE *h₃w (and possibly *h₂w), a process known as Cowgill's law. The proposal has been challenged but is defended by Don Ringe.[21]

In the Albanian language, a minority view proposes that some instances of word-initial h continue a laryngeal consonant.

Martin Kümmel has proposed that some initial [x] and [h] in contemporary Western Iranian languages, commonly thought to be prothetic, are instead direct survivals of *h₂, lost in epigraphic Old Persian but retained in "marginal dialects" ancestral among others to Modern Persian.[22]

In all other daughter languages, a comparison of the cognates can support only hypothetical intermediary sounds derived from PIE combinations of vowels and laryngeals. Some indirect reflexes are required to support the examples above where the existence of laryngeals is uncontested.

The proposals in this table account only for attested forms in daughter languages. Extensive scholarship has produced a large body of cognates which may be identified as reflexes of a small set of hypothetical intermediary sounds, including those in the table above. Individual sets of cognates are explicable by other hypotheses but the sheer bulk of data and the elegance of the laryngeal explanation have led to widespread acceptance in principle.

In the proposed Anatolian-language reflexes above, only some of the vowel sounds reflect PIE *e. In the daughter languages in general, many vowel sounds are not obvious reflexes.[23][24] The theory explains this as the result of

1 H-coloration. PIE *e is 'coloured' (i.e. its sound-value is changed) before or after h₂ and h₃, but not when next to h₁.

The results of H-coloration and H-loss are recognized in daughter-language reflexes such as those in the table below

Between three phonological contexts, Greek reflexes display a regular vowel pattern that is absent from the supposed cognates in other daughter languages.

Before the development of laryngeal theory, scholars compared Greek, Latin and Sanskrit (then considered earliest daughter languages) and concluded the existence in these contexts of a schwa (ə) vowel in PIE, the so-called schwa indogermanicum. The contexts are: 1. between consonants (short vowel); 2. word initial before a consonant (short vowel); 3. combined with a liquid or nasal consonant [r, l, m, n] (long vowel).

Laryngeal theory provides a more elegant general description than reconstructed schwa by assuming that the Greek vowels are derived through vowel colouring and H-loss from PIE h₁, h₂, h₃, constituting a so-called triple reflex.

An explanation is provided for the existence of three vowel reflexes in Greek corresponding to single reflexes in Latin and in Sanskrit
These presumed sonorant reflexes are completely distinct from those deemed to have developed from single phonemes.

The phonology of the sonorant examples in the previous table can only be explained by the presence of an adjacent phonemes in PIE. Assuming the phonemes to be a following h₁, h₂ or h₃ allows the same rules of vowel coloration and H-loss to apply to both PIE *e and PIE sonorants.

The hypothetical values for sounds with laryngeals after H-coloration and H-loss (such as seen above in the triple reflex) draw much of their support for the regularization they allow in ablaut patterns, specifically the uncontested patterns found in Greek.

In the following table, each row shows undisputed Greek cognates sharing the three ablaut grades of a root. The four sonorants and the two semi-vowel are represented as individual letters, other consonants as C and the vowel or its absence as (V).[26]

The reconstructed PIE e-grade and zero-grade of the above roots may be arranged as follows:

An extension of the table to PIE roots ending in presumed laryngeals allows many Greek cognates to follow a regular ablaut pattern.

The first row of the following table shows how uncontested cognates relate to reconstructed PIE stems with e-grade or zero-grade roots, followed by e-grade or zero-grade of the suffix –w-. The remaining rows show how the ablaut pattern of other cognates is preserved if the stems are presumed to include the suffixes h₁, h₂, h₃.[27]

In the preceding sections, forms in the daughter languages were explained as reflexes of laryngeals in PIE stems. Since these stems are judged to have contained only one vowel, the explanations involved H-loss either when a vowel preceded or when a vowel followed. However, the possibility of H-loss between two vowels is present when a stem combines with an inflexional suffix.

It has been proposed that PIE H-loss resulted in hiatus, which in turn was contracted to a vowel sound distinct from other long vowels by being disyllabic or of extra length.

A number of long vowels in Avestan were pronounced as two syllables, and some examples also exist in early Sanskrit, particularly in the Rig Veda. These can be explained as reflexes of contraction following a hiatus caused by the loss of intervocalic H in PIE.

The reconstructed phonology of Proto-Germanic (P-Gmc), the presumed ancestor of the Germanic languages, includes a long *ō phoneme, which is in turn the reflex of PIE ā. As outlined above Laryngeal theory has identified instances of PIE ā as reflexes of earlier *h₂e, *eh₂ or *aH before a consonant.

However, a distinct long P-Gmc *ō phoneme has been recognized with a different set of reflexes in daughter Germanic languages. The vowel length has been calculated by observing the effect of the shortening of final vowels in Gothic.[28]

Reflexes of trimoric or overlong *ô are found in the final syllable of nouns or verbs, and are thus associated with inflectional endings. Thus four P-Gmc sounds are proposed, shown here with Gothic and Old English reflexes:

Laryngeal theory preserves regularities in declensions and conjugations by explaining the trimoric sound as a reflex of H-loss between vowels followed by contraction. Thus

(Trimoric *ô is also reconstructed as word-final in contexts that are not explained by laryngeal theory.)

The reconstructed phonology of the Balto-Slavic languages posits two distinct long vowels in almost exact correspondence to bimoric and trimoric vowels in Proto-Germanic. The Balto-Slavic vowels are distinguished not by length but by intonation; long vowels with circumflex accent correspond to P-Gmc trimoric vowels. A significant proportion of long vowels with an acute accent (also described as with acute register) correspond to P-Gmc bimoric vowels. These correspondences have led to the suggestion that the split between them occurred in the last common ancestor of the two daughters.

It has been suggested that acute intonation was associated with glottalization, a suggestion supported by glottalized reflexes in Latvian. This could lend support to a theory that laryngeal consonants developed into glottal stops before their disappearance in Balto-Slavic and Proto-Germanic.[29]

A significant number of instances of voiceless aspirates in the Indo-Iranian languages may be explained as reflexes of PIE stop consonants immediately followed by laryngeals (*CH > *Cʰ)..

PIE resonants (sonorants) *r̥,*l̥,*m̥,*n̥ are predicted to become consonantal allophones *r, *l*, *m, *n* when immediately followed by a vowel. Using R to symbolize any resonant (sonorant) and V for any vowel, *R̥V>*RV. Instances in the daughter languages of a vocalic resonant immediately followed by a vowel (RV) are explained as reflexes of PIE *R̥HV with a laryngeal between the resonant and the vowel giving rise to a vocalic allophone. This original vocalic quality was preserved following H-loss.

Laryngeal theory has been used to explain the occurrence of a reconstructed sound change is known as Holtzmann's law or sharpening (German Verschärfung ) in North Germanic and East Germanic languages. The existing theory explains that PIE semivowels *y and *w were doubled to P-Gmc *-yy- and *-ww-, and that these in turn became -ddj-and -ggw-respectively in Gothic and -ggj- and -ggw- in early North Germanic languages. However, the existing theory had difficulty in predicting which instances of PIE semivowels led to sharpening and which instances failed to do so. The new explanation proposes that words exhibiting sharpening are derived from PIE words with laryngeals.

Many of these techniques rely on the laryngeal being preceded by a vowel, and so they are not readily applicable for word-initial laryngeals except in Greek and Armenian. However, occasionally languages have compounds in which a medial vowel is unexpectedly lengthened or otherwise shows the effect of the following laryngeal. This shows that the second word originally began with a laryngeal and that this laryngeal still existed at the time the compound was formed.

Further evidence of the laryngeals has been found in Uralic languages. While Proto-Uralic and PIE have not been demonstrated to be genetically related, some word correspondences between Uralic and Indo-European have been identified as likely borrowings from very early Indo-European dialects to early Uralic dialects. One example is the widespread word family including on the Uralic side e.g. Hungarian méz, Finnish and Estonian mesi, met(e)-, Mari мӱ /my/, Komi ма /ma/ 'honey', suggesting Proto-Uralic *mete; and on the Indo-European side, English mead, Greek methu 'wine', German Met 'honey wine', Slavic medъ and Sanskrit mádhu 'honey' etc.

There are several criteria to date such borrowings, the most reliable ones coming from historical phonology. For example, Finnic porsas, Erzya пурцос /purt͡sos/, Mokša пурьхц /pur̥ʲt͡s/ 'piglet' presuppose a common proto-form *porćas at an earlier stage of development. This is etymologized as a loanword from PIE *porḱ-, which gives Latin porcus 'hog', Slavic porsę 'pig', OE fearh (> Engl. farrow 'young pig'), Lithuanian par̃šas 'piglet, castrated boar'. Here loaning must have occurred predating the depalatalization of centum languages, and the later development into the Baltic *š reflected as Finn. h in borrowings, or Iranian *c medially reflected as Finn. t. If the PIE distinction between palatovelars and plain velars is reconstructed as one of velars and uvulars, then instead of the former condition also a lower limit can be set up for the loan, as postdating the satemization of *ḱ into a palatalized stop or affricate.

Work particularly associated with research of the scholar Jorma Koivulehto has identified several additions to the list of Finnic loanwords from an Indo-European source or sources whose particular interest is the apparent correlation of PIE laryngeals with three post-alveolar phonemes (or their later reflexes) in the Finnic forms. If so, this would point to great antiquity for the borrowings, since no attested Indo-European language neighbouring Uralic has consonants as reflexes of laryngeals. And it would bolster the idea that laryngeals were phonetically distinctly consonantal.

However, Koivulehto's theories are not universally accepted and have been sharply criticized (e. g. by Finno-Ugricist Eugene Helimski[30]) because many of the reconstructions involve a great deal of far-fetched hypotheses and the chronology is not in good agreement with the history of Bronze Age and Iron Age migrations in the Eastern Europe established by archaeologists and historians.

Three Uralic phonemes have been posited to reflect PIE laryngeals. In post-vocalic positions both the post-alveolar fricatives that ever existed in Uralic are represented: firstly a possibly velar one, theoretically reconstructed much as the PIE laryngeals (conventionally marked *x), in the very oldest borrowings and secondly a grooved one (*š as in shoe becoming modern Finnic h) in some younger ones. The velar plosive k is the third reflex and the only one found word-initially. In intervocalic position, the reflex k is probably younger than either of the two former ones.[citation needed] The fact that Finno-Ugric may have plosive reflexes for PIE laryngeals is to be expected under well documented Finnic phonological behaviour and does not mean much for tracing the phonetic value of PIE laryngeals (cf. Finnish kansa 'people' < PGmc *xansā 'company, troupe, party, crowd' (cf. German Hanse), Finnish kärsiä 'suffer, endure' < PGmc *xarđia- 'endure' (cf. E. hard), Finnish pyrkiä < PGmc. *wurk(i)ja- 'work, work for' etc.).

This list is not exhaustive, especially when one also considers several etymologies with laryngeal reflexes in Finno-Ugric languages other than Finnish. For most cases no other plausible etymology exists. While some single etymologies may be challenged, the case for this oldest stratum itself seems conclusive from the Uralic point of view, and corresponds well with all that is known about the dating of the other most ancient borrowings and contacts with Indo-European populations. Yet acceptance for this evidence is far from unanimous among Indo-European linguists, some even regard the hypothesis as controversial (see above).

Several linguists have posited a relationship between PIE and Semitic, almost right after the discovery of Hittite. Among these were Hermann Möller, though a few had argued that such a relationship existed long before the 20th century, like Richard Lepsius in 1836. The postulated correspondences between the IE laryngeals and that of Semitic assist in demonstrating their evident existence. Given here are a few lexical comparisons between the two respective proto-languages based on Blažek (2012), who discusses these correspondences in the context of a proposed relation between IE and Afroasiatic, the language family to which the Semitic languages belong:[31]

A feature of Proto-Indo-European morpheme structure was a system of vowel alternations termed ablaut ("alternate sound") by early German scholars and still generally known by that term (except in French, where the term apophonie is preferred). Several different such patterns have been discerned, but the commonest one, by a wide margin, is e/o/∅ alternation found in a majority of roots, in many verb and noun stems, and even in some affixes (the genitive singular ending, for example, is attested as *-es, *-os, and *-s). The different states are called ablaut grades; e-grade and o-grade are together "full grades", and the total absence of any vowel is "zero grade".

Thus the root *sed- "to sit (down)" (roots are traditionally cited in the e-grade, if they have one) has three different shapes: *sed-, *sod-, and *sd-. This kind of patterning is found throughout the PIE root inventory and is transparent:

In addition to the commonplace roots of consonant + vowel + consonant structure, there are also well-attested roots like *dhē- "put, place" and *- "give" (mentioned above): these end in a vowel, which is always long in the categories where roots like *sed- have full grades; and in those forms where zero grade would be expected, if before an affix beginning with a consonant, we find a short vowel, reconstructed as *ə, or schwa (more formally, schwa primum indogermanicum). An "independent schwa", like the one in PIE *pǝter- "father", can be identified by the distinctive cross-language correspondences of this vowel that are different from the other five short vowels. (Before an affix beginning with a vowel, there is no trace of a vowel in the root, as shown below.)

Whatever caused a short vowel to disappear entirely in roots like *sed-/*sod-/*sd-, it was a reasonable inference that a long vowel under the same conditions would not quite disappear, but would leave a sort of residue. This residue is reflected as i in Indic while dropping in Iranian; it gives variously e, a, o in Greek; it mostly falls together with the reflexes of PIE *a in the other languages (always bearing in mind that short vowels in non-initial syllables undergo various developments in Italic, Celtic, and Germanic):

Conventional wisdom lined up roots of the *sed- and *- types as follows:

But there are other patterns of "normal" roots, such as those ending with one of the six resonants (*y w r l m n), a class of sounds whose peculiarity in Proto-Indo-European is that they are both syllabic (vowels, in effect) and consonants, depending on what sounds are adjacent:

Saussure's insight was to align the long-vowel roots like *dō-, *stā- with roots like *bher-, rather than with roots of the *sed- sort. That is, treating "schwa" not as a residue of a long vowel but, like the *r of *bher-/*bhor-/*bhṛ-, an element that was present in the root in all grades, but which in full grade forms coalesced with an ordinary e/o root vowel to make a long vowel, with "colouring" (changed phonetics) of the e-grade into the bargain; the mystery element was seen by itself only in zero grade forms:

Saussure treated only two of these elements, corresponding to our *h₂ and *h₃. Later it was noticed that the explanatory power of the theory, as well as its elegance, were enhanced if a third element were added, our *h₁, which has the same lengthening and syllabifying properties as the other two but has no effect on the colour of adjacent vowels. Saussure did not suggest as to the phonetics of these elements; his term for them, "coefficients sonantiques", was not however a fudge, but merely the term in general use for glides, nasals, and liquids (i.e., the PIE resonants) as in roots like *bher-.

As mentioned above, in forms like *dwi-bhr-o- (etymon of Greek diphrós, above), the new "coefficients sonantiques" (unlike the six resonants) have no reflexes at all in any daughter language. Thus the compound *mṇs-dheH- "to 'fix thought', be devout, become rapt" forms a noun *mṇs-dhH-o- seen in Proto-Indo-Iranian *mazdha- whence Sanskrit medhá- /mēdha/ "sacrificial rite, holiness" (regular development as in sedur < *sazdur, above), Avestan mazda- "name (originally an epithet) of the greatest deity".

There is another kind of unproblematic root, in which obstruents flank a resonant. In the zero grade, unlike the case with roots of the *bher- type, the resonant is therefore always syllabic (being always between two consonants). An example would be *bhendh- "tie, bind":

This is all straightforward and such roots fit directly into the overall patterns. Less so are certain roots that seem sometimes to go like the *bher- type, and sometimes to be unlike anything else, with (for example) long syllabics in the zero grades while at times pointing to a two-vowel root structure. These roots are variously called "heavy bases", "dis(s)yllabic roots", and "seṭ roots" (the last being a term from Pāṇini's grammar. It will be explained below).

For example, the root "be born, arise" is given in the usual etymological dictionaries as follows:

The (A) forms occur when the root is followed by an affix beginning with a vowel; the (B) forms when the affix begins with a consonant. As mentioned, the full-grade (A) forms look just like the *bher- type, but the zero grades always and only have reflexes of syllabic resonants, just like the *bhendh- type; and unlike any other type, there is a second root vowel (always and only *ə) following the second consonant:

On the term "seṭ". The Pāṇinian term "seṭ" (that is, sa-i-ṭ) is literally "with an /i/". This refers to the fact that roots so designated, like jan- "be born", have an /i/ between the root and the suffix, as we've seen in Sanskrit jánitar-, jániman-, janitva (a gerund). Cf. such formations built to "aniṭ" ("without an /i/") roots, such as han- "slay": hántar- "slayer", hanman- "a slaying", hantva (gerund). In Pāṇini's analysis, this /i/ is a linking vowel, not properly a part of either the root or the suffix. It is simply that some roots are in effect in the list consisting of the roots that (as we would put it) "take an -i-".

But historians have the advantage here: the peculiarities of alternation, the "presence of /i/", and the fact that the only vowel allowed in second place in a root happens to be *ə, are all neatly explained once *ǵenə- and the like were understood to be properly *ǵenH-. That is, the patterns of alternation, from of Indo-European, were simply those of *bhendh-, with the additional detail that *H, unlike obstruents (stops and *s) would become a syllable between two consonants, hence the *ǵenə- shape in the Type (B) formations, above.

The startling reflexes of these roots in zero-grade before a consonant (in this case, Sanskrit ā, Greek , Latin , Lithuanian ìn) is explained by the lengthening of the (originally perfectly ordinary) syllabic resonant before the lost laryngeal, while the same laryngeal protects the syllabic status of the preceding resonant even before an affix beginning with a vowel: the archaic Vedic form jajanur cited above is structurally quite the same (*ǵe-ǵṇh₁-ṛ) as a form like *da-dṛś-ur "they saw" < *de-dṛḱ-ṛ.

Incidentally, redesigning the root as *ǵenH- has another consequence. Several of the Sanskrit forms cited above come from what look like o-grade root vowels in open syllables, but fail to lengthen to -ā- per Brugmann's law. All becomes clear when it is understood that in such forms as *ǵonH- before a vowel, the *o is not in fact in an open syllable. And in turn, that means that a form like jajāna "was born", which apparently does show the action of Brugmann's law, is a false witness: in the Sanskrit perfect tense, the whole class of seṭ roots, en masse, acquired the shape of the aniṭ 3sing. forms. (See Brugmann's law for further discussion.)

There are also roots ending in a stop followed by a laryngeal, as *pleth₂-/*pḷth₂- "spread, flatten", from which Sanskrit pṛthú- "broad" masc. (= Avestan pərəθu-), pṛthivī- fem., Greek platús (zero grade); Skt. prathimán- "wideness" (full grade), Greek platamṓn "flat stone". The laryngeal explains (a) the change of *t to *th in Proto-Indo-Iranian, (b) the correspondence between Greek -a-, Sanskrit -i- and no vowel in Avestan (Avestan pərəθwī "broad" fem. in two syllables vs Sanskrit pṛthivī- in three).

Stray laryngeals can be found in isolated or seemingly isolated forms; here the three-way Greek reflexes of syllabic *h₁, *h₂, *h₃ are particularly helpful, as seen below. (Comments on the forms follow.)

The Greek forms ánemos and árotron are particularly valuable because the verb roots in question are extinct in Greek as verbs. This means that there is no possibility of some sort of analogical interference, for example, happened in the case of Latin arātrum "plow", whose shape has been distorted by the verb arāre "to plow" (the exact cognate to the Greek form would have been *aretrum). It used to be standard to explain the root vowels of Greek thetós, statós, dotós "put, stood, given" as analogical. Most scholars nowadays probably take them as original, but in the case of "wind" and "plow", the argument can't even come up.

clarification and citation needed: consider "νέμω", seeing as "άνεμος" can be defined as that which is without "νομή"

Regarding Greek híeros, the pseudo-participle affix *-ro- is added directly to the verb root, so *ish₁-ro- > *isero- > *ihero- > híeros (with regular throwback of the aspiration to the beginning of the word), and Sanskrit iṣirá-. There seems to be no question of the existence of a root *eysH- "vigorously move/cause to move". If the thing began with a laryngeal, and most scholars would agree that it did, it would have to be *h₁-, specifically; and that's a problem. A root of the shape *h₁eysh₁- is not possible. Indo-European had no roots of the type *mem-, *tet-, *dhredh-, i.e., with two copies of the same consonant. But Greek attests an earlier (and rather more widely attested) form of the same meaning, híaros. If we reconstruct *h₁eysh₂-, all of our problems are solved in one stroke. The explanation for the híeros/híaros business has long been discussed, without much result; laryngeal theory now provides the opportunity for an explanation which did not exist before, namely the metathesis of the two laryngeals. It is still only a guess, but it is a much simpler and more elegant guess than the guesses available before.

The syllabic *h₂ in *ph₂ter- "father" might not be isolated. Certain evidence shows that the kinship affix seen in "mother, father" etc. might have been *-h₂ter- instead of *-ter-. The laryngeal syllabified after a consonant (thus Greek patḗr, Latin pater, Sanskrit pitár-; Greek thugátēr, Sanskrit duhitár- "daughter") but lengthened a preceding vowel (thus say Latin māter "mother", frāter "brother") — even when the "vowel" in question was a syllabic resonant, as in Sanskrit yātaras "husbands' wives" < *yṆt- < *yṇ-h₂ter-).

Like any other consonant, Laryngeals feature in the endings of verbs and nouns and derivational morphology, the only difference being the greater difficulty of telling what's going on. Indo-Iranian, for example, can retain forms that pretty clearly reflect a laryngeal, but there is no way of knowing which one.

The following is a rundown of laryngeals in Proto-Indo-European morphology.

Throughout its history, the laryngeal theory in its various forms has been subject to extensive criticism and revision.

The original argument of Saussure was not accepted by anyone in the Neogrammarian school, primarily based at the University of Leipzig, then reigning at the cutting-edge of Indo-European linguistics. Several of them attacked the Mémoire savagely. Osthoff's criticism was particularly virulent, often descending into personal invective.[32]

For the first half-century of its existence, the laryngeal theory was widely seen as ‘an eccentric fancy of outsiders’.[33] In Germany it was roundly rejected.[34] Among its early proponents were Hermann Möller, who extended Saussure's system with a third, non-colouring laryngeal, Albert Cuny, Holger Pedersen, and Karl Oštir. The fact that these scholars were engaged in highly speculative long-range linguistic comparison further contributed to its isolation.

Although the founding fathers were able to provide some indirect evidence of a lost consonantal element (for example, the origin of the Indo-Iranian voiceless aspirates in *CH sequences and the ablaut pattern of the so-called heavy bases, *CeRə- ~ *CR̥̄- in the traditional formulation[35]), the direct evidence so crucial for the Neogrammarian thinking was lacking. Saussure's structural considerations were foreign to the leading contemporary linguists.[32]

After Kuryłowicz's convincing demonstration[36] that the Hittite language preserved at least some of Saussure's coefficients sonantiques, the focus of the debate shifted. It was still unclear how many laryngeals are to be posited to account for the new facts and what effect they have had exactly. Kuryłowicz, after a while, settled on four laryngeals,[37] an approach further accepted by Sapir, Sturtevant, and through them much of American linguistics. The three-laryngeal system was defended, among others, by Walter Couvreur and Émile Benveniste. Many individual proposals were made, which assumed up to ten laryngeals (André Martinet). While some scholars, like Heinz Kronasser and Giuliano Bonfante, attempted to disregard Anatolian evidence altogether, the ‘minimal’ serious proposal (with roots in Pedersen's early ideas) was put forward by Hans Hendriksen, Louis Hammerich and later Ladislav Zgusta, who assumed a single /H/ phoneme without vowel-colouring effects.

By the 2000s, however, a widespread, though not unanimous, agreement was reached in the field on reconstructing Möller's three laryngeals.[38] One of the last major critics of this approach was Oswald Szemerényi, who subscribed to a theory similar to Zgusta's (Szemerényi 1996).

Today the laryngeal theory is almost universally accepted in this new standard form. Nevertheless, marginal attempts to undermine its bases are occasionally undertaken.[39][40]