|Abstract: ||The work constitutes a monographic perspective of a complex process of numeral grammaticalization in Polish, between the 16th and the beginning of the 20h century. The process was conditioned by many structural changes and a general Slavonic tendency to differentiate number determinants at the semantic, inflectional and syntactic level from other parts of speech. It is cardinal numbers and collective numerals that, according to the latest theories, belong to the grammatical category in the work in question.
The basic phenomena taking place in the process of shaping the grammatical category of number are first of all stabilization of numerals in the quantifying function and formal inclusion of such lexemes as kilka, wide, tyle into this class, and second, the development of a separate declension of cardinal numbers and a homogeneous paradigm of collective numerals, third, the establishment of syntactic relations between a number and a quantified noun, and fourth, stabilization of relations between a nominal phrase, number and predicate. The majority of the processes mentioned is compensatory and unifying in character however the reducing phenomena do exist. What is most important here is the loss of the category of number which became the most basic determinant of numerals among other parts of speech. In the case of numbers, the very process began as early as in the pre-Polish period and lasted for many centuries, whereas collective numerals underwent this innovation in a different way, i.e. these grammatical changes were connected with the loss of many additional semantic functions of an attributive type. Lexemes, from 2 to 900, continued a declension inherited from pra-Slavonic language, according to 11 separate morphological types (2, 3—4, 5—10, 11—19, 20, 30—40,
50—90, 100, 200, 300—400, 500—900). At the turn of the 15th and 16th centuries unifying processes began. They were connected with vital changes of the Polish language system. One of the most important changes involves the loss of a dual number typical of the majority of the Slavonic languages. The substitution of a tripartite opposition singularity : evenness : multiplicity with a clearer dychotomic distinction into singularity and
plurality allowed for a formal closeness not only of the lexeme two, and neighbouring
units such as three and four (including indefinite several), but also of 12, 20, 200 based
on the stem two, and close 13, 14, 30, 40, 300, 400.
The second most important for an inflectional and syntactic consolidation of numerals was the process morphologising the plural masculine gender conducted between 17th and 18th centuries. This fundamental change concerning all nominal speech parts made all the gender-neutral lexemes create the possibility of expressing a new opposition. The very process was marked by a series of analogies which finally led to the A = G syncretism of masculine and transition of the accusative opposition of pięć, trzydzieści, czterysta H pięciu, trzydziestu, czterystu to the nominative case. The fact that numerical lexemes reflect a complex arithmetic structure, constituting a variety of numerical combinations, enabled a mutual inflectional relationship of
particular units. As a result of these centuries-old influences, a unified paradigm of cardinal numbers 5—900 was created, with the -u ending typical of this category, covering almost all cases (apart from a nominative-accusative of masculine ones and instrumental ending of -oma). A multifunctional morpheme -u became the basic determinant of numerical declension among the remaining speech parts. Numbers 2—4 remained
their inflectional identity, though, the processes in question were also evident among 2, 3 and 4 paradigms different in the beginning. A grammaticalization of cardinal numerals also consisted in a partial unification of
syntactic behaviours. A hereditary division stayed long in the Polish language according to which lexemes 2 and 4 as adjectives combined with quantified nouns on the basis of congruency whereas the units from 5 and above being formal nouns or composites were given a primary role in the nominal group. The influence of syntax typical of 2—4 lexemes in dependent cases was responsible for the advantage of congruency in noun combinations with numbers from 5 and above foregrounding the noun in the nominal group.
The full unification of syntactic behaviours in the case of numbers did not happen because the units from 5 and above remained their formal primacy in nominative and accusative cases respectively.
Two groups of cardinal numerals ( 2—4 and 5—900) remained their formal identity in connections with verbs. Although 2—4 lexemes as an attribute did not influence the form of predicate connoted by the noun, the numbers from 5 and above formed an integral group with a quantified noun in the function of the subject influencing the verbal part as a whole. At the beginning of the 16th century a competition between the predicate
in singularis neutrum and a semantically justified plural predicate can be observed, the advantage being given to the former one, although the full normalization took place as late as in the 19th century.
A syntactic pattern of behaviour similar to the numbers from 5 and above was to be find also in the case of collective numerals which, as the primary elements of nominal phrases, connote a predicate in singularis neutrum. On the basis of the historio-linguistic analysis conducted, one can draw a conclusion that essential analogies within syntactic behaviours were evident between cardinal numbers and collective numerals which,
finally, led to the unification (apart from 2—4 cardinal numbers). A closer look at the Polish numerals from a diachronic perspective allows for understanding why this category is currently so complicated, controversial (especially in the syntactic description) and outlining different research points of view. It is not a feature
of the Polish language, but the characteristic of the majority of Slavonic languages in which the numeral as a separate speech part developed late whereas the process itself, according to many researchers, was not completed yet.|