Fortson Chapter 3 Laryngeals

• "Laryngeals" are considered to be fricatives produced in the velar region, but we have inadequate phonetic knowledge to say exactly what sound they were. Suggestions include glottal, uvular, velar, etc. sounds. There simply is no direct evidence for precisely what sound they were.
  

• In 1879 Fernand de Saussure suggested that such phonemes existed, based on his analysis of Sanskrit.
• What he noticed was that in 2 verb patterns, there was a similarity that could be easily explained by hypothesizing one no-longer-existent consonant that had left a different trace in each of the two patterns.
  
• Only a few people believed his theory, among them Möller and Cuny: they expanded the number of laryngeals to 3!
• THEN, in 1927, Jerzy Kury łowicz discovered that Hittite, recently deciphered, had a sound transliterated as ḫ exactly where the laryngeal hypothesis suggested it should (it is said to be pharyngeal in Hittite and uvular in Luwian!). We will see more details about this in Chapter 4.
• Side note: Glottalic theory is an alternative explanation of the IE stop inventory: currently, it suffers from a lack of any direct evidence: in that way, it is similar to the situation of laryngeal theory before Hittite was deciphered. As Fortson points out, glottalic theory suffers from other problems.
  

• Laryngeals are consonant sounds that sometimes play the role of vowels.
• There are usually 3 of these laryngeals: *h₁, *h₂, and *h₃, (also written as *H₁, *H₂, and *H₃ or sometimes also as *ə₁, *ə₂, and *ə₃)

• 3.18 Coloring occurred when a laryngeal was next to a short e:
  
• *h1e didn't color but stayed *h1e;
  
• *h2e colored to a,
  
• while *h3e colored to o.
• 3.19When next to vowels, laryngeals were lost (they disappeared)
• Except in Hittite, where *h2 and (most likely) *h3 remained as a sound that has been transliterated as ḫ. But *h1 was lost in Hittite too.
  
• Avestan and Vedic Sanskrit have reflexes of laryngeals that are glottal stops between vowels or hiatuses. Also, taking into account the lost laryngeals makes some of the metrical schemes in Sanskrit work, which is evidence that the formulaic system preserved a trace of them even after the sounds had changed.
• 3.20 When a sound is "lost," it often leaves some trace. In the case of the loss of laryngeals, there is 'compensatory lengthening' of a neighboring vowel. See examples in 3.20.