from Historical Linguistics, an introduction, pg 250
by Lyle Cambell.

Uralic is the ancestor language of the Finn-Ugor and Samoyedic languages and is derived into stone age times. The application of the comparative method to languages of the Uralic family reveals a proto-language with the following numbers;

dual *-kè(-) (later becomes plural for Hungarian, két=two (things)) but also in most Uralic languages for two.
plural *-t and *-j (also in Altaic)
singular *ø.

Direct object of finite verbs were marked by the "accusative" case ( *-m),
but the objects of imperative verb bore no accusative marker. Case and definiteness were related; the genetive and accusative cases implied definiteness, while indefinite nouns took no marking. That is in form they were not distinct from the nominative case. The "genetive" ( -*n) case marked not only the possessor but also served to signal an adjective attribute before its head noun.

Proto Uralic verb tenses included:

past *-j
past perfect *-mè [ Hung meg- ]
present *-pA (A denotes vowel harmony with the attached root)
past *s'A [Hung s'>t ]
negative verb *e- [Hung á, el- ]

Sentences minimally had a nominal subject and a predicate (verb or nominal); the subject could be signaled by personal pronominal suffixes attached to the predicate. The predicate agreed with its subject in person and number. There was no other agreement. The predicate of embedded clauses was in form a verbal noun, where personal possessive pronominal suffixes were used to signal its subject. The role of the embedded clauses was in form a verbal noun, where personal possessive pronominal suffixes were used to signal its subject. The role of the embedded clause in the overall sentence was shown by case markings on the verbal noun ( a nominalization) which was the core of the embedded clause.

Proto Uralic had no overt conjunctions or relative pronouns; embedded verbal nouns, nominalizations, were the only means of showing subordination. In brief, the application of the comparative method to the reconstruction of Proto-Uralic morphosyntax has proven quite successful and this case shows that at least in some instances, we are capable of syntactic reconstruction (Jahunen 1982; Cambel 1990).

------end-------