The Evolution of the Indo-European Languages

Dr. C. George Boeree

Similarities and Differences among the Indo-European Languages

First, let's look at phonetic similarities and differences among some of the oldest representatives of the main branches of the Indo-European languages.  The letters in red are ones that differ from the original Proto-Indo-European:

PIE Sanskrit OCS Lithuanian Armenian Greek Latin OIR Gothic Hittite Tokharian*
P P P P H, -W P P --- F, -B P P
B B B B P B B B, M, -W- P P P
Bh Bh B B B, -W- Ph F-. B B, M, -W- B P P
T T T T T T T T. -Th TH. -D T T, C
D D D D T D D D, -Dh T T T, ---, S'
Dh Dh D D D Th F, D, B D, -Dh D T T, C
S' S S' S K K K H, -G K K, S'
J Z Z' C G G G, -Gh K K K, S'
H Z Z' J, Z Kh G G, -Gh G K K, S'
K K, G K, C', C K K' K K K, -Kh H, -G K K, S'
G G, J G, Z', DZ G K G G G, -Gh K K K, S'
Gh Gh, H
G, Z', DZ G G, J Kh H, G G, -Gh G K K, S'
Kw K, C K, C', C K K' K, P, T Kw, K Kw, K Hw Kw Kw, K
Gw G, J G, Z', DZ G K B, D, G Gw, G, W B, M, -W- Kw, K Kw K, Kw
Ghw Gh, H
G, Z', DZ G G, J Ph, Th, Kh Gw,G,W,F Gw Gw, G, W Kw K, Kw
S S, S' S S S, H, --- S, H, --- S, R S, --- S S S
L R, L R, L L L L L L L L L, Ly
M M M M M M M B, M, -W- M M M
N N N N N N N N N N N, Ny
Y Y Y Y --- Y, --- Y ---, Y Y Z, H, --- Y
W V, W V V G, V W, H, --- W F, ---, W W W W, Y
PIE Sanskrit OCS Lithuanian Armenian Greek Latin OIR Gothic Hittite Tokharian*

(Chart based on Christopher Gwinn's "Indo European Phonology,"

* Note: The column on Proto-Tokharian was added by myself. Although it begins as a Centum language, most consonants were later palatalized before front vowels. This happened with some other Centum languages as well.

While S, R, L, M, N, Y, and W do not change that often, the plosives often change dramatically.

One basic set of differences is that between the Satem languages - Sanskrit, Old Church Slavonic (OCS), Lithuanian, and Armenian - and the Centum languages - Greek, Latin, Old Irish (OIR), Gothic, Hittite, and Tokharian.

In Centum languages, PIE palatalized velars k', g', and g'h merge with k, g, and gh, while bilabialized velars kw, gw, and ghw retain their identity (up to a point!). In Satem languages, PIE k' becomes s or ʃ, g' becomes z or ʒ, g'h becomes z or ʒ, while kw, gw, and ghw merge with k, g, and gh, with some additional palatization occuring later in their development. Notice how Armenian loses the voiced-unvoiced distinction among the plosives, and Hittite and Tokharian lose that as well as the distinction between plain and aspirated plosives.

The Indo-European languages of the Tarim Basin in far western China known as Tokharian have some Celtic sound qualities - leading some to suggest that a branch of early proto-Celts wandered all the way to China.  Another scenario is a little more sensible:  The ancestors of The Tokharians left the eastern side of the main PIE area at about the same time as the Celts and other western Centum languages left the western side, perhaps a little earlier, and so prior to the same changes in the Balto-Slavic and Indo-Iranian dialects that the Celts missed.

In Greek , PIE bh, dh, and gh become phi, theta, and chi, i.e. first become unvoiced aspirated stops and eventually unvoiced fricatives f, θ, and χ.  Note that in Latin, bh, dh, and gh become f, f, and h, respectively.  I suspect that the progression was to ph, th, kh, then to phi, theta, chi, and finally to f, f, and h.  In most other languages, bh, dh, and gh become b, d, and g (or z/ʒ, if satem), respectively.  Also, Greek collapses PIE kw, gw, and gwh somewhat indiscriminately into k/p/t, b/d/g, and phi/theta/chi.

In Gothic we see some of the most dramatic (although systematic) changes:  PIE p, t, k, and kw become f, θ, h, and ʍ, while b, d, g, and gw become p, t, k, and kw.

In OIR, PIE b, gw, bh, and m all become b or m or w (depending on context).  Note also that all plosives become aspirated at the ends of words, and eventually become fricatives.

Note:  Languages related to Latin, as well as ones related to OIR, sometimes use p in place of kw.  This is another commonality, then, between the Italic family and the Celtic family.

What groupings might we tentatively make on the basis of these comparisons?

1.  Satem - Sanskrit, OCS, Lithuanian, and Armenian -  vs Centum - Greek, Latin, OIR, Gothic, Hittite, and Tokharian;

2.  Hittite and Tokharian are significantly different from all the other languages, and likely represent an early divergence;

3.  Sanskrit, OCS, and Lithuanian share many commonalities in the Satem group;

4.  Armenian is an outlier in the Satem group;

5.  Greek, OIR, and Latin share many commonalities in the Centum group;

6.  Gothic is an outlier in the Centum group.

Putting it into outline form:

I.  Satem
        A.  Sanskrit-OCS-Lithuanian
                1.  OCS-Lithuanian
                2.  Sanskrit
        B.  Armenian
II.  Centum
        A.  Greek-OIR-Latin
                1.  OIR-Latin
                2.  Greek
        B.  Gothic
III. "Odd" Centum
        A.  Hittite
        B. Tokharian


Unfortunately for the elegance of the preceding, other approaches give us other results.  If we look at certain grammatical constructions, for example, we find the following:

For the genitive singular of nouns, we have Greek, Armenian, and Indo-Iranian (Sanskrit) using -osyo, Slavic (OCS) and Baltic (Lithuanian) using -ó, Baltic and Germanic (Gothic) using -eso, and Celtic (OIR) and Italic (Latin) using -í.

For indirect and dative cases, we find a similar pattern:  Greek, Armenian and Indo-Iranian using -bhi, Slavic, Baltic, Germanic, and Tokharian use -m . Hittite uses -n. Celtic and Italic using -bhos (dative singular)

The Celtic-Italic link is fortified by such constructions as the comparison in -samo (vs -tero, -isto) and medium voice in -r (vs -oi, -moi).

The Greek-Armenian-Indo-Iranian link is fortified by the fact that all three have an athematic and a thematic aorist.

Hittite differs in many ways from the others, although showing more similarity to Celtic-Italic than to other groups.

(Following Cyril Babaev's "Indo-European Proto-Dialects."

So now we have a chart that looks like this:

I.  Greek-Armenian-Indo-Iranian
II.  Slavic-Baltic-Germanic
III.  Celtic-Italic

So, while Celtic-Italic remain independent, Greek and Germanic have jumped the Centum ship, you might say, to join the two Satem sub-groups - or more likely, the reverse: left the central dialect area (Slavic-Baltic and Armenian-Indo-Iranian) prior to the Satem innovation, but well after Celtic-Italic left, and even further after the Hittite and Tokharian left.


And finally, there is the issue of common words.  Analyses seem to show that Hittite and Tokharian split from the mainstream earliest.  Later, the Celtic and Italic languages split off, followed by Greek and proto-Armenian.  Albanian also splits off somewhere in the same time frame.  Finally, we have the Indo-Iranian and Germanic languages splitting off from the central mass of Indo-European, which eventually becomes Balto-Slavic.

Lexicon analysis generally supports the morphemic hypotheses, and is not too contrary to the phonetic hypotheses.

A sticking point is the place of Germanic:  It seems to be both clearly tied to Baltic, yet phonetically quite distinct.  It is, in fact, quite distinct from all its relations!  One theory is that its similarity to Baltic is in part due to close proximity and continued contact, and that its phonetic distinctiveness is due to a pre-Germanic substrate of megalithic people sometimes referred to as "the Folk."

It is possible that pre-Indo-European neolithic people in France and England likewise influenced the development of the Celtic languages as well - another "Folk" perhaps.  The neolithic people of the Iberian peninsula and southwestern France - including the ancestors of the modern Basques - do not appear to have had too significant an effect on the Celtic of that area, nor on the later Latin.

Other languages have had more obvious contact with other language substrates.  The Hittites and other Anatolians clearly came into an area already well-populated with non-Indo-Europeans, which no doubt hastened the rather dramatic phonetic and morphemic simplifications characteristic of those languages.

The Mycenaean Greeks may have been influenced by an Anatolian substrate in Greece and the islands (the Pelasgians?), further complicated by a second wave of Greeks (Dorians).

The Armenians appear to be closest to the Indo-Iranians in phonetic and morphemic structure, and were possibly influenced by centuries of contact with them as well as the Hittites, Semites, and Caucasians.

An Evolutionary Time-line of the Indo-European Languages

c. 5000 bc.

Homeland:  The Danube River valley (Wallachia and Hungary).  Farming learned from the people of Asia Minor.  Cultivation of native rye and oats and domestication of native pigs, geese, and cattle begins.  Strong tribal sociey develops. This is the hypothesis first presented by Igor M. Diakonov. Note: Most linguists follow Marija Gimbates' Kurgan theory. A smaller group follow Colin Renfrew's Anatolian theory. I am admittedly only an amateur, but I prefer Diakonov's Balkan theory, which is also a part of Renfrew's extended theory.

There are many reasons for choosing the Danube River valley:  Farming is possible, although the land is less than desirable to more powerful tribes from the south; the flora and fauna of the valley, as well as for other natural features such as hills and rivers, are represented by the oldest words we can reconstruct; it includes the natural ranges of wild horses which, when later domesticated, would become the Indo-European's "ace card";  the area is central to the eventual expanse of the Indo-Europeans, with due allowance for the more rapid expanse commonplace over steppe-lands;  the area is also in close proximity to some of the most conservative recent representatives of the family.

The most compelling reason is the presence of the Danubian culture, with its linear incised pottery, at this same time.  The culture spreads soon after in exactly the directions that would account for the spread of PIE.

There are, of course, many other possibilities.  The most common suggestion is the steppes north of the Black Sea, for many similar reasons.  I believe that the strong tribal social structure suggests that the Indo-Europeans were farmers before they were pastoralists.  It is highly unlikely that they went straight from steppe hunter-gatherers to sophisticated pastoralists in one step.

c. 4000 bc.
Proto-Anatolians move east to the northern Caucasus.  They would be profoundly influenced by the advanced cultures of Asia Minor and beyond.

Proto-Tokharians  move east into the Ukraine.  These people are the most likely originators of the horse culture.  There is also plenty of evidence of ox-drawn wagons with disk wheels in the western steppes.

A western dialect emerges on the upper Danube and beyond.  The enclosed steppe of the Hungarian Plain is an ideal position to blend farming with a horse culture.

c. 3000 bc.
Copper working, learned from the people of Asia Minor, begins in Thrace and the Danube valley and reaches Germany by 3000 bc.

Domestication of the horse spreads from the Ukraine.  Within a thousand years, horsemanship spreads from the Ukraine throughout the Indo-European area, even into Scandinavia.  It is the steppe inhabitants who change most dramatically into true pastoral societies.  In the more wooded areas of Europe, horse ownership begins to differentiate a warrior nobility from commoners.  Of course, use of the horse spreads to the non-IE societies of the Middle East as well.

The disk-wheel wagon has spread from Russia across Europe to Holland.

The Proto-Anatolians move from the Caucasus to Asia Minor.

The Proto-Tokharians continue east to the steppes, towards the Tarim Basin in northwestern China.  They may be the people known to the Chinese as the Yüeh-chi, and may have been the core of the Kushan Empire of the first century AD.

The Proto-Celts separate from the rest of the western dialect and expand west into southern Germany and France, where they develop the Michelsburg culture and begin to strongly pressure the pre-PIE people, likely including the ancestors of the Basques and Aquitanians.  The remaining western dialect tribes edge into the modern Slovenia-Croatia area as well as northern Germany.

The main body of Indo-Europeans expands into Thrace, the Ukraine, Bohemia, and Poland, and begins to differentiate into a northern dialect (Bohemia, Poland, and Hungary, represented by the Funnel Beaker culture) and a southern dialect (Wallachia, Thrace, and Ukraine, continuing the Danubian culture). The original inhabitants north and east of the Carpathians, likely speakers of Uralic languages, are pushed further north and east.

c. 2500 bc.
Bronze working develops throughout Indo-European area.
The Proto-Italics, who speak a western dialect, move west and south from the Slovenia area into Italy. There they would encounter well-established pre-PIE people, possibly the ancestors of the Etruscans and Rhaetians.

The Proto-Illyrians, speaking a western dialect (perhaps), move south from the northern Croatia area into Illyria (the Dalmatian coast).

One branch of the southern dialect - Proto-Hellenic - moves south into Macedonia, Greece, and the Aegean islands, absorbing much of the Pelasgian people and culture.  By 1500 bc, the southern-most tribes would establish the Mycenaean culture.

The Proto-Germanics move into Scandinavia. Odd aspects of Proto-Germanic may be due to interaction with northern Celtic tribes, Baltic tribes, and possibly to the presence of native speakers of Uralic languages in Scandinavia.

The remaining body of Indo-Europeans (the Baltic, Poland, Bohemia, the Hungarian Plain, Wallachia, Thrace, the Ukraine and the neighboring steppes) - both northeast and southeast dialects - undergoes the Satem phonetic changes.

c. 2000 bc.
The horse-drawn, two-wheeled chariot, with spoked wheels, is developed in the western steppes, and spreads quickly to the Balkans as well as the Middle East.

A branch of the southeastern Satem dialect - Proto-Indo-Iranian - expands from Ukraine and the steppes into Afghanistan, Iran, and into India.  One tribe - the Mittani - goes as far west as northern Mesopotamia. The well-established cultures influence the newcomers greatly, but the Proto-Indo-Iranians of the steppes maintain their language.

The main body of the southestern Satem dialect expands into the Ukraine to become the Cimmerians, leaving the Dacians in the original homeland.  I suspect that the Dacians and Thracians spoke a Cimmerian-like dialect. These people would develop the steppe version of the Battle Ax culture.

The main body of the northeastern Satem dialect - Proto-Balto-Slavic - expands north from Poland into Belarus and the Baltic coast.  With the Germans, they would develop the northern version of the Battle Ax culture.

The Celts expand further into western Europe and, in a retrograde move, back into Hungary.  A powerful society, they pressure the original peoples of western Europe, as well as their own relations to the east.  They develop the Bell-beaker culture and, later, the Urnfield culture.

Anatolians (most notably the Hittites) establish themselves in Asia Minor, where they become a major power.  Their languages are profoundly affected by neighboring non-IE languages.

A second wave of Hellenics (Doric Greeks) moves into Greece from Macedonia.

c. 1500 bc.
Proto-Phrygians - possibly a branch of the Cimmerians - move from Thrace across the Bosporus to northwestern Asia Minor.  The Phrygians would move into the power gap left by the collapse of the Hittite Empire around 1200 bc.

Proto-Armenians - possibly another branch of the Cimmerians - move into Asia Minor, probably by means of the Bosporus.  In the next 1000 years, they spread over much of northern Asia Minor, but are eventually pressured into the Lake Van region.

Albanian may be the sole survivor of the Illyrian languages, its many variant features due to long contact with a variety of neighbors.  Or it could be the lone descendent of a Dacian dialect that later moved into the Albanian region.  It is very difficult to tell!  So little of Dacian, Thracian, Illyrian, etc. is left to us.

The western Celts expand into western Iberia and the British Isles, where they absorb most of the prior inhabitants.  The original inhabitants of Spain survive well into the Roman era, while the original inhabitants of southwest France survive to the present as the Basques.

The Balto-Slavics differentiate into Baltic and Slavic.  Both begin to expand east- and northward, at the expense of the hunter-gatherer Finno-Ugric people.

The Indo-Iranians differentiate into Indic and Iranian. The Indic group rapidly expands across northern India as far as Magatha.  The Iranians split into powerful tribes, notably the Persians and the Medes, by the 800's bc.  The Iranians remaining in the steppes would come to be known as the Scyths and Sakas.  The powerful Scyths expand westward at the expense of the remaining Cimmerians.

c. 1000 bc.
Iron working begins in the Balkans by 1000 bc.  It reaches Britain by 800 bc.

The "Age of Empires" begins in earnest.

Major resources:  I. M. Diakonov, On the Original Home of the Speakers of Indo-European, Journal of Indo-European Studies (1985, vol. 13, pp 92 to 174) and J. P. Mallory's In Search of the Indo-Europeans (1989).  Maps based in part on Colin McEvedy's The Penguin Atlas of Ancient History (1967).

Return to Homepage