Numerals are a class of lexemes whose basic function is to express a quantitative value. From the grammatical perspective, it is a highly heterogeneous group, and its scope may vary depending on the defining criteria applied. In semantically-motivated grammatical classifications, a large portion of lexemes classified as numerals from a purely formal perspective have the properties of adjectives (e.g. pierwszy, drugi, kilkukrotny), nouns (e.g. setka, milion), adverbs (samotrzeć, dwojako), pro-forms (jeden in such contexts as Jeden pan o ciebie pytał.) or particles (jeden in contexts such as Ty durniu jeden!). On the other hand, more recent approaches based on inflectional and syntactic criteria have significantly restricted the scope of this class.
The development of numerals as a separate grammatical category is a process that has been shared to a large extent across all Slavic languages. In Slavic languages, we may easily distinguish between two groups of numerals with distinct morphological-syntactic properties. From the syntactic perspective, in Proto-Slavic the expressions referring to values 1–4 behaved as adjectives, while the expressions referring to values 5–10 were nouns (Siuciak 2008: 12). It is believed that this differentiation is motivated by the limits of human perception, and the capacity to subitize, or, simply put, to quickly and effortlessly determine the number of elements in a specific set. The human capacity to subitize decreases substantially for sets of five and more elements – to determine the number of elements in such sets one needs to be able to count or, in other words, use a mental process that requires one to rely on the abstract concept of a number. There is more on this topic in: Ifrah 1990; Rutkowski 2003.
Knowledge of the development of words used to express quantitative values is reflected in the earliest descriptions of grammar of the Polish language. In early 19th century studies, the expressions now classified as numerals were described as a subcategory of names. For example, Onufry Kopczyński distinguishes a category of numeral names (in Polish imiona liczbowe, e.g. dwa, trzy, cztery, pięć, sto, tysiąc). Nevertheless, the author did not regard them as a separate grammatical category. What is more, taking into account the diversity of the criteria on the basis of which the remaining groups of names were distinguished, not to mention their formal and semantic diversity, one can hardly speak of any coherent classification in this case. On the basis of their formal properties, Kopczyński classifies jeden, jedna, jedno, dwaj, dwie, dwa, trzej, trzy, czterej, cztery as adjectives, while – according to him – pięć, sześć, siedem “and all the subsequent ones” are neuter nouns (Kopczyński 1917: 68–69).
Both the older and more recent proposals concerning the internal organisation of the class of numerals have their source in the classifications of this part of speech, which appeared in the grammars of the Polish language as early as in the 19th century and originate from the grammar of Latin. Antoni Małecki defines this class in the following way:
Numerals (nomina numeralia) are words used to designate quantity, or, to put it differently, the number of persons or things – for instance to specify if there are two, three or more of them, or maybe only one? – or if the thing is the first one in an order, the second, the third, or takes yet further place in a sequence, etc. (Małecki 1863: 117).
Małecki divides all numerals into definite and indefinite ones. The former refer to a specific number (the author included in this category such expressions as dwa, trzy, sto, tysiąc, drugi, setny, sześcioraki), while the latter refer to an approximate number (e.g. kilka, kilkoro, kilkanaście, kilkadziesiąt, kilka set, mało, wiele, ile, tyle, siła, ćma). Alongside this division, the researcher introduced the classification into cardinal numerals (numerals proper, cardinalia), which “are the clearest manifestation of the essence of a numeral and […] all the others derive more or less from them” (e.g. jeden, dwie, czterej, jedenaście, dwadzieścia, sto, tysiąc) (Małecki 1863: 117). In this group, the author distinguishes a subgroup of collective numerals, which he also refers to as diverse, “since they are usually used when we refer to things or
persons of different types or of different nature” (e.g. dwoje, oboje, obojgo, troje, trojgo, dziesięcioro, dwanaścioro, trzydzieści czworo) (Małecki 1863: 18). Apart from cardinal numerals, the scholar distinguished ordinal numerals (ordinalia) which “express a place in an order occupied by a thing or a person” (e.g. drugi, wtory, trzeci, dziesiąty) (Małecki 1863: 127) and multiplicative numerals (multiplicativa) which “express some kind of plurality or numerousness of parts, manners, properties, kinds etc. which make part of a thing” (e.g. dwoisty,
troisty, pojedyńczy, jednaki, dwojaki, trojaki) (Małecki 1863: 128) and are then used to derive multiplicative numeral adverbs (e.g. podwójnie, potrójnie, dwoiście, troiście, dwojako, trojako, kilkorako, wielorako, w dwójnasób, w trójnasób, w czwórnasób). What is more, Małecki noted that, in the Polish language, distributive numerals (distributiva) and adverbia numeralia described in the grammars of Latin are only expressed through analytical structures, but with the latter being used to derive adjectives that do not have their Latin lexical counterparts (e.g. jednokrotny, dwukrotny, kilkokrotny, tylokrotny).
In subsequent grammars, this classification continued to be developed, becoming increasingly detailed in the process. For instance, in Henryk Gaert ner’s Gramatyka współczesnego języka polskiego, the classification of numerals is much more hierarchical. The author divided this class of lexemes along two axes – between specific (definite) and general (indefinite) numerals on the one hand, while on the other he distinguished between “the kinds of numerals on the basis of the manner in which they restrict the scope of nouns” (Gaertner 1938: 159). The two main categories are quantitative and ordinal numerals (e.g. pierwszy, drugi, trzeci). The former are subdivided into specimen- and kind-specific (in Polish osobnicze and gatunkowe, e.g. jednaki, dwojaki, trojaki, kilkoraki), with the specimen-specific category being further split into combined and tandem numerals (in Polish: łączne and posobne, e.g. jednokrotny, dwukrotny, trzykrotny, kilkakrotny). The combined numerals are divided into fractional (e.g. pół, dwie trzecie) and mixed (e.g. półtora, półtrzecia, półczwarta), as well as integral numerals that could be further split into cardinal (e.g. jeden, dwa, trzy, kilka) and collective ones (e.g. dwoje, troje, kilkoro).
The most widespread classification of numerals, and which owes its popularity to being the prevailing textbook model, consists of seven groups: cardinal (trzy, pięć), collective (troje, pięcioro), ordinal (trzeci, piąty), fractional (pół, ćwierć), multiple (trzykrotny, pięciokrotny), manifold (trojaki, pięcioraki), multiplicative (potrójny). Furthermore, quantitative assessments may be also conveyed through expressions that – in this approach – are treated as numeral adverbs (dużo, pełno, masę, trochę, mało, nieco, odrobinę) and numeral pro-forms, subdivided into four groups: deictic (tyle, tyloraki, wszystek), indefinite (kilka, ileś), interrogative (ile?), negative (żaden) (Klemensiewicz 1981: 58–60).
In pre-war and some post-war grammars (e.g. Klemensiewicz 1981; Bąk 1986; Strutyński 1997), the scope of the class of numerals is determined by semantics. In more recent studies, we have seen some attempts at defining this class on the basis of formal criteria. In classifications by Zygmunt Saloni and Marek Świdziński (Saloni 1974; Saloni 1977; Saloni, Świdziński 1985), based on the inflectional criteria, the class of numerals only includes the lexemes traditionally described as cardinal and collective numerals (all the remaining lexemes meet the criteria to be assigned to other grammatical classes). Saloni and Świdziński define numerals as lexemes that, while having the inflectional category of case and gender, do not have the inflectional category of number (Saloni 1977: 147; Saloni, Świdziński 1985: 95). The defining criteria assumed by these authors exclude such expressions as tysiąc, milion or miliard from the class of numerals, as they do have the inflectional category of number, without having the inflectional category of gender (dwa tysiące/miliony/miliardy etc.). In consequence, in accordance with the classification proposed by the authors, such words are nouns. Nevertheless,
the syntactic relations formed by these expressions prove that the inflectional criteria are insufficient to distinguish the grammatical class of numerals. They occur in contexts typical of numerals, and when a phrase containing a numeral fills the position of the subject, they may impose on the verb the singular number and the neuter gender. In consequence, not only such sentences as Tysiąc kobiet zastrajkował., Tysiące kobiet zastrajkowały., but also Tysiąc kobiet zastrajkowało., Tysiące kobiet zastrajkowało. are acceptable. From the perspective of syntax, the words tysiąc and tysiące in the first pair of sentences are nouns, while in the second they act as numerals. A similar classification problem concerns the expression jeden, inflected both for number and gender (cf. Widzę jednego chłopca., Widzę jednych chłopców.,
Widzę jedne dziewczynki.). In the classification by Saloni and Świdziński, it meets the defining criteria of an adjective, but is also described as a pro-form (cf. e.g. Linde-Usiekniewicz, Rutkowski 2003: 128). The exclusion of the lexeme jeden from the class of numerals would be consistent with the postulate put forward by Andrzej Bogusławski, who highlighted the fact that jeden does not bind with such quantifiers as wszystkie, niektóre (Bogusławski 1966: 50–51). Yet another solution to this problem is to classify the expression as being representative of two different lexemes – an inflected (variable) and an uninflected (invariable) one
(cf. e.g. Laskowski 1999: 345). The typical contexts in which the word is usually interpreted as a numeral include: “1. Sequences of numerals, in which it loses the paradigm of gender and case and is always represented by the first person nominative masculine, e.g. dwadzieścia
jeden mężczyzn […] 2. In numeral syntactic structures such as: jeden i pół […] 3. In mathematical operations and in series of calculations […]” (Mieczkowska 1994: 11). What is more, the most recent studies highlight the expressive function of jeden (e.g. Ty durniu jeden!) (Żabowska 2022; Kosek 2020).
However, when formal criteria are adopted as a basis for the classification of lexemes into parts of speech, compound expressions that could be treated as numerals in semantically-motivated approaches (e.g. sto dwadzieścia jeden, dwie trzecie, na pęczki, po piąte) must be excluded from this class.
In the classification proposed by Saloni and Świdziński, cardinal and collective numerals are not divided into two separate subclasses. The authors argue that the group of expressions traditionally classified as collective numerals should simply be treated as an inflectional variant of the general class of numerals. They do highlight, however, the specific distribution of collective numerals, set out by the five classes of nouns: 1. certain neuter animate nouns (dziecko, kurczę, szczenię), 2. pluralia tantum nouns whose referents do not have the dual structure (drzwi, skrzypce), 3. pluralia tantum nouns that refer to a group of people (rodzice, państwo, generałostwo), 4. masculine personal nouns (or pronouns) referring to sets of people of different genders, 5. other nouns in strongly phrasalized (dwoje oczu, dwoje uszu) or archaic combinations (dziesięcioro przykazań, obojga płci) (Saloni, Świdziński 1985: 179).
Other scholars, such as Roman Laskowski, represent a different view. Laskowski distinguished the class of numerals on the basis of syntactic criteria. In this approach, the class also consists of cardinal and collective numerals which – according to Laskowski – “form
two syntactically, morphologically and semantically distinct subclasses” (Laskowski 1999: 63). Laskowski further distinguishes a third subclass – partitive numerals which “inform about the size of a portion of an uncountable object (e.g. a substance, as in trochę wody, wiele czasu) or the size of a portion of a set of countable objects (cf. e.g. dużo osób, tyle ludzi)” (Laskowski 1999: 342). Laskowski defines the class of numerals by pointing out that “in a nominal group, they perform the function of a dependent […] remaining in a relation of mutual grammatical determination with the head: the numeral determines the case and the number of the noun (pronoun) syntactically combined with it, while the gender of the numeral is determined by the noun” (Laskowski 1999: 63).
Older grammars dedicate relatively more room to inflection and the syntactic relations between cardinal numerals and nouns, taking account of the relevant historical changes affecting this group. This is not only a consequence of the fact that the group has changed rather dynamically, but also a product of its internal formal diversity, typical of all Slavic languages. In contemporary Polish, cardinal numerals referring to the values from the range 1–4 have different morphological and syntactic properties than numerals referring to the values from 5 above. At the syntactic level, this difference is manifested in the types of phrases that these expressions form with nouns and verbs. Numerals from the first group (apart from personal masculine numerals dwóch, trzech and czterech) agree the values of case and gender
with the noun with which they form a phrase (cf. Dwaj chłopcy zjedli owoce., Trzy dziewczynki poszły spać., Cztery krzesła stały przy stole., Jan pomógł trzem kobietom.). The situation of numerals from the second group and the personal masculine numerals dwóch, trzech, czterech is more complex. In their case, the type of syntaxctic relation that they form with a noun depends on the position in the sentence and on the case imposed on the numeral by the verb. The gender of numerals depends on the gender of the noun. However, whenever a numeral is in the dative case in the position of the subject or an object, it imposes the form of the dative on the right-handed noun combined with it. If a numeral fills the position of an object, but is not in the dative form, it combines with a noun in agreement (cf. Widzę pięć dziewczynek., but Śpiewam pięciu dziewczynkom.). Additionally, if a phrase with a numeral fills the position of the subject, the numeral imposes the singular number and the neuter gender on the verb (e.g. Dwóch chłopców zjadło owoce., Pięciu chłopców zjadło owoce., Sześć dziewczynek poszło spać., Dziesięć krzeseł stało przy stole.). The problems of syntax and the inflection of numerals from diachronic and synchronic perspectives have been discussed in detail in a number of studies (Klemensiewicz 1930; Mieczkowska 1994, 1995; Laskowski 1999; Siuciak 2008; Derwojedowa 2011; Słoboda 2012).