![[]](conflag_25gif.gif)
Jörg Rhiemeier's Conlang Pages
Jörg Rhiemeier's Conlang Pages
Indo-European is the largest language family of the world and the most dominant family in Europe. It is therefore hardly surprising that many diachronic conlangers choose to build their conlang(s) as members of the family. Besides languages that belong to one of the existing branches such as Germanic, Celtic, Romance or Slavic, languages derived from Proto-Indo-European (PIE), often called pielangs in the community, are popular. This page is meant as a guide for those who endeavour to create such languages.
Building a diachronic conlang on Proto-Indo-European involves some challenges, as PIE is not an easy language - but a very interesting one, which certainly is part of its attraction. First of all, you should familiarize yourself with it, which requires knowledge where to get the necessary information. Fortunately, there are good resources available on the Web for free. For phonology and grammar as well as culture and archaeology, Wikipedia actually is quite good. For lexicon, this spreadsheet, compiled by Sam McCabe, is the best available source known to me; it can be downloaded for offline use here. If you prefer printed matter, the best books are Indo-European Language and Culture by Benjamin Fortson IV for phonology, grammar and some culture, as well as overviews of the individual branches, and The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World by James P. Mallory and Douglas Q. Adams for lexicon and also culture and mythology.
Among the first decisions to make is whether you accept the glottalic theory which was proposed, independently, by T. Gamkrelidze (Georgia) and V. Ivanov (Russia), and by P. Hopper (USA), in the early 1970s (see Gamkrelidze - Ivanov 1995), and states that the set of stops traditionally reconstructed as voiced unaspirated (*b, *d etc.) actually were glottalized (ejective). This idea did not fare very well in Indo-European studies as the evidence adduced to it has been considered insufficient, but it is legitimate to explore it in a conlang, and some conlangers have done so. (I formerly did myself, but later decided not to use the glottalic theory.)
What were the phonetic values of the elusive PIE laryngeals? Short answer: Nobody knows. Long answer: We can only make some guesses. First, it is probably misleading to consider *h1 "e-colouring". Rather, it is non-colouring. Like all other consonants except *h2 and *h3, it does not affect vowel qualities. It may have been almost everything, but it is quite likely that it was simply [h]. The other two laryngels have the effect to change a preceding *e into a back vowel. Hence, it seems as if *h2 carried the feature [+back], and *h3 also the feature [+round].
Some Indo-European conlangers (inluding formerly me) maintain the idea that the three laryngeals were just the fricative members of the three velar series, matching their places of articulation. Thus: *h1 = *x́, *h2 = *x, *h3 = *xw. Fair. But there are some objections against this:
These problems led me to abandon this neat idea. I now think that *h1 was just [h] (see above), *h2 was something farther back than [x], i.e. a uvular or pharyngeal fricative or approximant, and *h3 the same but labialized (and possibly voiced). Yet, in your conlangs, you are free to pursue other ideas.
The pre-ablaut stage of PIE (see below on the origin of ablaut) seems to have had a very small vowel inventory, but it may be the case that the velar series preserve old vowel features [+front] (the palatovelars) and [+round] (the labiovelars); apparently, there were no front rounded vowels when this happened. That there is no palatalized laryngeal may be because they were further back than velars, and therefore had retracted adjacent front vowels (try to say [χe] without either fronting the uvular or backing the vowel, and you get the picture), which also explains why *h2 is so common - it has absorbed the missing palatalized laryngeal by this effect.
Once again, the short answer is: We do not know. There are several different theories in circulation, and you better just take the ablaut grades as given: there is no reliable rule that predicts which grade appears where. It is no sin to simplify a bit - most IE languages have done so. My personal idea is that the pre-ablaut stage of PIE, with a vowel inventory of **a, **i and **u, had a penultimate accent which was overridden by "strong" (perhaps long) vowels attracting the accent; yet, there were other factors at work, too. The vowel **a yielded the ablaut series *e ~ *o ~ *0, while **i gave *ei ~ *oi ~ *i, and **u resulted in *eu ~ *ou ~ *u. Yet, there are many instances where these rules do not predict the correct form, so there must have been more going on.
The PIE accent is closely associated with ablaut, and equally unpredictable - which is why so many IE languages have abandoned it in favour of something more predictable, such as the initial accent in Germanic and Celtic, or the accent system of Latin which is based on the weight of the penultimate syllable (if that syllable is heavy, it bears the accent, if not, the antipenultimate syllable is accented). There is nothing wrong with simplifying it in your conlangs as well.
PIE nouns are not very difficult, most forms are well-reconstructed. Most PIE nouns involve ablaut, though this is levelled out in most daughter languages. There is some incertainty concerning the dative and ablative plural, and especially with the dual number. Otherwise, at least the non-Anatolian languages show much agreement, and the PIE forms can be reconstruted with great certainty, and can be found in the handbooks. The Anatolian languages, however, disagree with the rest of IE in some points, so Early PIE may have had only four or five cases (nominative, genitive, dative, accusative and perhaps a vocative) and two numbers (singular and plural) as well as only two genders (animate and inanimate).
Adjectives work just like nouns, except that the gender is not fixed but agrees with the noun they modify. Most IE languages also have comparative and superlative inflections; the suffixes vary among the branches (the most common forms are *-yos- for comparative and *-isto- for superlative), and Anatolian languages express these notions differently. There is thus not one true way to handle this in your conlang.
The PIE pronouns are well-reconstructed, though there are two different relative pronoun roots, namely *kwi- (which is also the interrogative pronoun) and *yo-, with unknown difference in meaning; perhaps *yo- is the older of the two as a relative pronoun, while *kwi- was just an interrogative pronoun. Possibly, the latter had a restrictive force while the former did not.
The PIE verb is complex, and sources such as Mallory & Adams give only roots because the verb formations often differ between teh branches of the family. In Late PIE, each verb had up to three different stems, all formed from the same root but not predictable. These are the present stem from which the present and the imperfect (imperfective past) were formed, the aorist stem of the aorist (perfective past), and the perfect stem whose meaning was similar to the English present perfect. There are more than a dozen different ways of forming the present stem, and four different ways of forming the aorist stem; only the perfect stem is always formed the same way. These formations had to be learned for every verb, and it is perhaps more convenient to consider them three different dictionary entries.
Because once you do this, things become easier - what remains are quite regular inflections. Things are easier in Anatolian, where all forms of a verb (not that many - there is just a present, a preterite, and an imperative) are formed from a single stem. But many IE languages have simplified the Late PIE three-stem verb inflection to a large degree, and many modern IE languages thus form all verb forms from a single stem (usually a continuation of the present stem).
PIE appears to have been a head-final language in which modifiers (such as adjectives) precede the noun and the verb comes last in the clause. But because of the rich inflection of nouns, adjectives, pronouns and verbs, it allowed great freedom of word order, and the daughter languages developed in different ways - almost everything is possible.
© 2025 Jörg
Rhiemeier
Last update: 2025-12-19