Any discussion of the grammar and phonology of Proto-Indo-European (PIE) would be incomplete without reference to Ablaut (German for 'vowel gradation'). Ablaut signals grammatical differences by altering the quantity or quality of the root vowel. Remnants of Indo-European (IE) Ablaut are visible in English in verbs such as 'ride', which forms its preterite as 'rode' and its past-participle as 'ridden'. The consonants in the root remain the same, but the vowels change.
The reconstructed vowel system for PIE usually contains three vowels: */e/, */o/, and */a/ (though there are some scholars who question the validity of reconstructing PIE */a/). However, only */e/ and */o/ seem to have participated in IE Ablaut. Those particular roots which have /e/ as their vowel are said to be in the 'e-grade', and those roots with /o/ are in the 'o-grade' of the root. There was, in addition to the e- and o-grades, a so-called 'zero-grade' of the root in which there was no vowel. An example of these three grades of the root *sed- 'sit' is still evident in English, e.g. sit < *sed- (e-grade), sat < *sod- (o-grade) and in the word nest < *ni-sd- (zero-grade).
It is possible, however, that the connections between the grade of vowel in a root and its grammatical function were not intact in the proto-language. Rather, one can view Ablaut as the result of particular phonological conditions (namely accent) which came to be associated with particular grammatical distinctions. Accented root-vowels became the e- and o-grades of the root (the full-grades), and unaccented roots became the zero-grade of the root. This is best compared to the differences evident between English 'do not' (in which 'not' receives more stress than 'do') and 'don't' (in which 'do' received more stress than the negator, which originally had contained a vowel that has since disappeared).
Accent of the vowel in Proto-Indo-Europen was most likely tonal, i.e., it was related to the pitch of the vowel rather than to the stress as it is in English. Furthermore, PIE had in all likelihood two different pitches: high and low (relative to one another). This would provide a good explanation as to how the basics of the IE Ablaut system could have arisen. If we assume that all roots originally had the same vowel, */e/, then its transformation into the other grades would have been the result of its accentuation within the word. Roots containing vowels with high tone became the e-grade, and those with low tone the o-grade. This is due to the fact that the vowel /e/ tends to have a relatively higher second formant frequency in comparison to /o/. The lowering of the pitch of the vowel /e/ was most likely reinterpreted as if the vowel were /o/. Elements of these accent/ablaut associations are still evident in reflexes of PIE *bheidh- 'persuade' in Greek: peíth-omai (e-grade, accented), pé-poith-a (o-grade, following accent), and e-pith-óme:n (zero-grade, unaccented).
As such, though, these differences in accentuation are merely phonologically conditioned phenomena. What is key in the development of a system of Ablaut is that these phonological conditions be associated with grammatical changes. Although the grammaticalization of Ablaut varies from daughter-language to daughter-language, the process was probably similar in each case. In English, for example, one can see the on-going productivity of the Ablaut series found in 'drink, drank, drunk' as applied to a verb which previously did not participate in Ablaut. Thus the past tense and participle of 'ring' can be said to be 'rang' and 'rung', even though there is no etymological basis for this. Dissassociated from their original phonological conditions, these vocalic alternations are able to take on an independent existence to mark grammatical features.