Proto-Indo-European Language Info Contents

Since Proto-Indo-European has been reconstructed, there is no problem with spelling it regularly. A problem does arise, however, from the reconstructions having been made by linguists for their own uses -- PIE words are written phonemically rather than phonetically. That is, they are written so as to reflect their linguistic structure, not the way they're pronounced.

For example, "to be" is H1es-. To form one of the versions of the imperative, this root is put into the zero-grade (which means dumping the vowel), and -dhi is added, resulting in H1s-dhi. The hyphen is put in to show the separation of the pieces. However, there is a rule of PIE phonology that says that voiced consonants cause voicing in preceding consonants. Also, it's likely that the H1 would change to a schwa in front of a consonant. This means that H1s-dhi is actually pronounced <əzdhi>. Because most of the material in PIE on this site is intended for speaking I've decided to write the words phonetically. This may cause a bit of confusion for those used to the traditional system, but I think it will make the words easier to read in ritual performances. (I wrote these pages before I decided that the laryngeal would stay as a schwa; in the current version, an initial laryngeal followed by a consonant disappears. I'll be fixing that eventually.)

Consonants
Consonants in Proto-Indo-European have the same value as in English, with the following exceptions: bh, dh, gh, kh - These are aspirated consonants. They are pronounced like the English "b," "d," "g," and "k," followed by a short puff of air. In English, these consonants are always aspirated at the beginning of a word. Thus "pit" has an aspirated "p." Compare the sound to the one in "spit," and will see a difference. In some languages, including English, it makes no difference whether a consonant is aspirated or not. In others, however, like Sanskrit or Proto-Indo-European, a word beginning with an aspirated consonant, and one begining with the same consonant, but unaspirated, have different meanings.

g - Always hard, as in "get."

ǵ - A palatalized "g." If you say the sound "g," you will see that it is done at the back of the mouth, by the entrance to the throat, the vellum, which is why a "g" is called a velar sound. A "ǵ, however, is pronounced at the palate, the roof of your mouth. It is the sound made by the first "g" in "gewgaw" (as in "gimcracks and gewgaws"), except that in that word there is a bit of a glide after the "g," a little bit of a "y." In PIE "ǵ,"there isn't that glide. See the description of "ḱ" for a more common English word that has the unvoiced version of this. In older printed material you will sometimes see this represented with a circumflex rather than an accent.

gʷ - A labialized "g," a "labio-velar." Round your lips as if you were going to say "o," but say "g." A voiced form of "kʷ.

ḱ - A palatalized "k." Just like the "ǵ," only voiced. Compare the initial sound in "call" with that in "cute." The "cute" version of "c" is a palatal one. In "cute" this is followed by an off-glide sound, like a "y;" in the Proto-Indo-European "ḱ" there isn't that glide. In older printed material you will sometimes see this represented with a circumflex rather than an accent.

kʷ - A labialized "k," a "labio-velar." Pronounced just like a labial-velar "g" (i.e., "gʷ"), only unvoiced. Round your lips as if you were going to say "o," but instead say "k." This is actually how the English "qu" is pronounced, despite your English teacher telling you it was a k followed by a w..

qʷ - A voiced velar fricative (the same sound as "x," only voiced), a sort of gargling noise, which has been labialized. The "q" is usually represented by the Greek letter "gamma," but"q" is more browser friendly. Without the labialization, it is pronounced similar to the way some Parisians pronounce the "r" in "Paris." I've used this to indicate the sound I think was the PIE laryngeal H3. (In earlier versions of this site and in Deep Ancestors I maintained that this was simply , but I've now changed my mind. I'll be changing all the PIE q's on this site to "qʷ;" if you stumble across one that hasn't been changed yet, just assume that it should be.)

r - There is no way of knowing if the Proto-Indo-European "r" was trilled, rolled, or pronounced like the English "r." It is a matter of personal taste, then. I like to give it a bit of a trill, just because I think it sounds pretty.

x - The final sound in German "Bach;" a voiceless velar fricative. I've used this for the PIE laryngeal "H2."

ʔ - A glottal stop. Like the sound between the "t" and "l" in "little" when spoken in a Cockney accent.

There are some details that we don't know. For instance, was the pronounced with the tongue touching the teeth as in English, or the alveolar ridge as in German, or even somewhere else, like interdental? We don't know. So pick a version and stick with it; you may be right, or you may be pronouncing your PIE with an accent. Vowels
The short vowels have what is called their "Italian values." In English terms:

Father
Weight
Machine
Rote
Lute

It's important to know, however, that English long "a," which is what we call the sound in "weight," is actually a diphthong, with an off-glide at the end. This is one of the things that has to be beaten out of English-speakers when they learn other languages, in which the sound is a single one. There's also a touch of at the end of English "o," so be careful with that one.

Long vowels are just that - the short vowels held a little longer, and with an accent.

Some linguists think that the PIE accent was one of pitch rather then stress, but that's debatable. Since a stress accent has a pitch component (compare the pitch of the second syllables of "desert" and "dessert"), and pitch accents have stress components, I don't think it's a big deal for our purposes.

Diphthongs
ai - Like the English long "i."

ei - Like the English long "a" (technically , since in most languages that's the letter with which the sound is written) which, unlike that in Spanish or most other languages, has a touch of off-glide at the end; that is, it's not actually a single sound. As in "day."

eu - We don't have this sound in English. As a dipthong, it's made by quickly switching from the to the .

oi - Like the Yiddish "Oy!"

Semi-vowels
In PIE there are cases where vowels ablaut. That means that they change according to how a word is used. We have that in English in some of our old irregular verbs: sing, sang, sung. In PIE, ablaut can be even more radical; a vowel can disappear completely, forming what is called the "zero-grade." When this happens, a following consonant can become a vowel, and these consonants are therefore called "semi-vowels." We have these in English too, although they generally aren't recognized. "Button," for example, is usually said to be pronounced , but it really isn't; there isn't a schwa there. Instead, the "n" is serving as a vowel. The same thing is true of "butter," "bottle," and, to a lesser degree, "bottom." PIE has the same semi-vowels, with the addition of a vowel version of "w."

Semi-vowels are spelled with a dot under them: ḷ, ṃ, ṇ. Linguists write the "w," whether as a consonant or a vowel, as a "u" with an upside down short vowel sign (a breve) underneath: u&815;. (I can't find how to write this in html; if anyone knows how to do it, I would love to know.) This is to indicate that it can be either a or a , depending on where it is found, and to distinguish it from the superscript "w" used to indicate labialization for , , and . I've just used "w" for the consonant version of this and "u" for the vowel. As a vowel, it's pronounced like the "oo" in "book."

A few words about laryngeals

(You don't have to know anything in this section to be able to pronounce the words in the rituals. It's just there to explain how I got from the linguists' phonemic representations of laryngeals to my phonetic ones.)

PIE had sounds called "laryngeals." The name comes from an early belief that they were pronounced in the larynx, but now that is known not to be true. Unfortunately, there isn't any consensus on how they were. This is because although they had affects on neighboring s, accents, and voicing, they disappeared at an early point in PIE. When I wrote the ritual words on this site and in my book, I believed that they had disappeared before the end of PIE, but after having had their affects, and I wrote the words accordingly. I have now decided, however, that they hadn't completely disappeared by that point. I therefore need to make some modifications in the forms of the PIE words. I haven't done this yet, so if you want to use the rituals as written you will just have to go with the forms I've presented.

There are generally believed to have been three laryngeals in PIE. (Some linguists use a different number of them, but three is the most common.) Because we aren't sure how they were pronounced, they are generally written as H1, H2, and H3. (Sometimes a lower-case "h" is used, and sometimes the numbers aren't subscripted.) When next to an , H2 changes it to an and H3 changes it to an . When it's found next to a voiceless consonant, H3 causes voicing. H1 doesn't make any changes to the .

There are numerous theories of how these were pronounced. Since I intend these rituals to be performed, I've had to choose a pronunciation for each laryngeal. I've gone with what seems to be the most common views.
H1 is assumed to be either an or a glottal stop (<ʔ>). This is most important when the laryngeal is initial. Originally I had used , but I've decided that <ʔ> was more likely. So if you see a word starting with "h," you can assume that I haven't gotten around to changing it, and replace the with <ʔ>.
H2 is assumed to most likely have been a voiceless velar fricative, which is the sound found at the end of the German "Bach" and the Scottish "loch." I've spelled this with the standard IPA symbol "x."
H3 is a little trickier, and there isn't much agreement on it. The candidates are , <&gamma>, and <&gammaʷ>. As I said earlier, on this site and in my book I used "q" instead of &gamma because it's easier to type and read so I'll start using it now. I don't think that it could have been because it causes voicing, and since isn't voiced, there's no reason why it would do that. It also causes rounding, though, since it changes the flat to the round . Since this is most easily caused by a labialized consonant, the voiced and labialized <&gammaʷ> seems the most likely. That's what I've used then (although spelling it "qʷ"). If you see a "q" all by itself on these pages it's because I haven't gotten around to fixing it; just labialize the sound and you'll be fine.

These are the ways that I've assumed laryngeals were pronounced either between vowels (VHV) or initially before a vowel (HV), with their having had their effects on the following vowel (e.g., > ). When they are either between consonants (CHC) or initial before a consonant (HC), however, they've manifested themselves as a schwa, ə. (Thus > <əs>. And finally, when they are preceded by a vowel and the two are either in-between consonants or at the end of a word I've allowed the laryngeal to have its affect, but then disappear with compensatory lengthening. (Compensatory lengthening is a phenomenon common in languages, whereby when a sound disappears the neighboring sound lengthens, as if to make up for it.) So , and > . So that's how I've made PIE possible to be pronounced. If you have any comments or suggestions, especially if I haven't made things clear, just contact me, and I'll see what I can do, and possibly explain things better here.

Contents