The proto-indo-european language develops in Central Asia
The Indo-European proto-language was spoken in a homeland South-East of the Black Sea by a collection of semi-nomadic clans and pastoral tribes which more or less could understand each other (8th - 6th Millennium B.C.). As such, we might better conceive of Proto Indo-European as a group of related dialects which evolved from one branch of the proposed primitive Nostratic parent language macro-family. Scholars have determined the location based on an extensive reconstructed vocabulary of Proto Indo-European, and the habitat it describes. This was an adaptive language, one with which they sang, joked, loved, lamented and prayed. Linguistic evidence indicates that they prayed to *Deiwos ( = the God of Light ). These Proto Indo-European dialects were either of the peripheral tribes, or of the central tribes. Innovations which occurred within the central tribal dialects might not be reflected in some of the peripheral dialects. Migrations due to climate shifts further differentiated the dialects, as various groups dispersed to seek opportunities. Outside influences on the peripheral dialects might not be felt by the central dialects. Thus, it is difficult to say what "Proto Indo-European" was like if we do not accept the diversity of that proto language and it's speakers. The migration from the mountains of Eastern Turkey (approx. 7th - 4th Millennium B.C.) to the steppes of Russia separated the Anatolian group from linguistic contact with the semi-nomadic groups that migrated on to settle the expanse in the North beyond the Caucasus mountain range. The "Epoch" of Proto Indo-European in Southern Russia was comparatively short, and characterized by dissolution of tribal groups dispersing and eventually migrating to distant lands. The spread of Y-chromosome DNA haplogroup R1a1 is associated with the spread of the Indo-European languages, too. The vocabulary of these groups adapted to new environments, but enough "core" terms survived to enable the reconstruction of those terms.
The proto-language dialect of each of these migrating tribes evolved differently from the other dialects. Some dialects seldom lost prolonged contact with each other, as in the case of Baltic and Slavic. This explains the similarities between those two groups, both of which have many archaic features still worthy of serious study. Also, some Proto-Indo-European tribes (dialects) maintained tribal alliances (linguistic contact) up to their various distinct Proto-Stages, as in the example of Proto-Baltic/ -Slavic/ -Germanic. Certain lexical correspondences and innovations (as in the words for "thousand", or "wax") bear this out.
The Proto Indo-European language slowly evolved from a dialect of Nostratic, primitive at first, but expressive. With time, it's speakers innovated new ways to render it more precise and effective. One innovation lead to another, and eventually the everyday speech of these people resembled something somewhat similiar to the reconstructed synthetic proto-language theorized by scholars. It is doubtful that it ever possessed the elegant complexity often set forth about it. That complexity was nonetheless accomplished later, in both humble and renown languages, all derived from Proto Indo-European. Such are the languages as Sanskrit, or Lithuanian.
The transition from active (fientive) to the later, and more complex, declensional system was accomplished using various "quasi-paradigmatic" adverbial forms in the dialects.
The four cases of West-Baltic (Prussian & Sudovian) declension are not an innovation but an archaic feature uniting West Baltic with Germanic and Greek. Only nominative, genitive, dative and accusative forms have constant intercrossing functions in various Indo-European languages, while forms used for the instrumental or locative cases (traditionally declared to be "Common Indo-European"), have related functions. Such intercrossing elements were used for semi-paradigmatic adverbial forms, differently paradigmatized in the various Indo-European languages.
"The traditional academic construct of a seven case declensional system for Proto Indo-European is as synthetic as it is theoretically convenient." (Jeannette DeBusk Cox)
The differentiation between each dialect became more pronounced as time went by. Those dialect tribes that remained in closer contact later resembled each other more, as in the case of Baltic, Indo-Iranian, Germanic, and Slavic (each with their somewhat similiar grammatical innovations). In the case of Baltic, with many of it's supra-archaic qualities, a clearer window into past developements is possible to determine how such innovations took place. Thus, Baltic Studies will continue to enrich and redefine Proto-Indo-European Studies, now and far into the future.