Vous êtes ici : Accueil / Langue(s) / Linguistique / Generic reference: the exceptional status of human nouns

Generic reference: the exceptional status of human nouns

Par Ismaël Zaïdi : Doctorant en linguistique - Université Grenoble Alpes
Publié par Marion Coste le 25/09/2023
This article examines the particular status of generic references (statements that assign a characteristic to a class or subclass) that concern human subjects. While such statements are often described as interchangeable, our study shows that each form of generic reference has social implications when human subjects are under study. Between non-acceptability, the expression of stereotypes and a need for context which is usually absent in studies, human nouns highlight more than ever the relationship between grammar and semantics within generic references.

Article réalisé dans le cadre du diplôme de l'ENS.


Generic reference is used to apply a property to a class or subclass: birds fly or ducks lay eggs express a generalization over the class of birds and ducks, regardless of "how many members of the kind or category have the property" (Leslie and Lerner 2016). Generic reference can also apply to sentences: John likes a cigar after dinner is generic, although the subject itself is a specific individual. Generic statements are essential for one's knowledge of the world, starting from the very early childhood, in order to make sense of the child's surroundings, and extending throughout our whole life. A sentence such as sharks attack bathers, uttered by figures of authority (parents and teachers, for instance), will consequently allow children to swim away when seeing a shark. Through such utterances, "children can gain information about the people and events they are observing" (Cimpian and Markman 2011: 471). This is also true of generic statements about humans: they are useful in our attempt to ascribe specific characteristics to various groups, from workers or people belonging to a given social class to bodies of people sharing the same nationality. As such, generic statements about humans may exist regardless of whether or not they are true of a majority of the class.

There exist two types of generic reference: the first is "reference to a kind" (Krifka et al. 1995: 2): dodos have died out, in which the predicate have died out applies to the entire class, with no possible exception; the kind itself prevails over the members it comprises. The second is "characterizing generalizations" (Radden 2009: 275), illustrated by birds fly: penguins, which cannot fly but belong to the kind birds, do not jeopardise the truth of the sentence. These generalizations thus license exceptions and merely consist in ascribing a characteristic to a class. Both the noun phrase (a noun and its constituents, with the noun as the head, also known as NP) and the predicate must be taken into account since, considered on its own, an NP might be generic (Germans speak good English), specific (the Italian over there cooking dinner), or non-specific: "If an American moves to Denmark and reduces his electricity consumption by half his electric bill will double" (Blog 2012); "an American" does not designate a particular American but a virtual one, who exists in the situation defined by the predicate and the subordinator IF. It is not generic either as the reference designates a potential American to which the predicate would be applied, and not the entire class.

Generic reference with regard to human nouns implies studying characterizing generalizations rather than reference to a kind: statements relating to a class of people very often admit exceptions as they are prone to stereotypes, defined as "an individual's set of beliefs about the characteristics or attributes of a group" (Judd and Park 1993: 100, my emphasis). This is why linguists have taken an interest in characterizing generalizations in their study of generic reference about human nouns: as entrenched as beliefs might be, they are the expression of one's mindset and stereotypical statements cannot be said to apply to a group with no counterexamples. There only exist a few statements that do not license exceptions for humans, such as Armenians are humans, in a context where, at the beginning of the twentieth century, a speaker would openly rise up against the Armenian genocide.

In terms of grammar, genericity is commonly associated to three forms: definite NPs (the tiger), bare plurals (students frequently engage in voluntary activities), which are also definite (just as one would say les étudiants rather than des étudiants in French), and, for most linguists at least, indefinite NPs (a lion). The aim of this study is to give an overview of the intricacies and specificities of each form in generalizations over humans, enlightening their social implications as well as furthering the grammatical scope of generic reference. In other words, I will seek to point out what makes generic reference to humans stand out among other groups.

1. Interchangeability: establishing the unique grammatical status of generalizations over humans

At first glance, uttering a simple sentence that, for example, ascribes the property of having a mane to a lion does not meet with strong grammatical or social restrictions: lions have manes, a lion has a mane, the lion has a mane all seem to be synonymous as they denote the same process. Nevertheless, from a strictly grammatical point of view, they do differ:

  • Lions have manes: with bare plural generic reference, the members of the class are   perceived as a group, with no instance more salient than the other.
  • A lion has a mane: with indefinite singular generic reference, a member of the class is extracted (core value of the indefinite article) from the class as a whole and stands for it as a prototypical instance; the predicate has can be applied to any individual member of the class.
  • The lion has a mane: with definite singular generic reference, the lion is seen as a kind whose members are specimens: "singular generic NPs with the definite article the (like the owl) pick out a group individual which do[es] not allow any reference to its members (Borik and Espinal 2012: 126). The label "a group individual" forecasts the challenges in making such generalizations coincide with those over humans since the kind prevails over the members of the class: this amounts to seeing a human body irrespective of its members.

Certainly all three sentences are not perceived as fundamentally divergent — they are rather interchangeable, in that one can be replaced by another, depending on the context. Thus, indefinite singular generic reference can only occur when the predicate applies to one element at a time (the predicate is then said to be distributive, as in a lion has a mane). Definite singular generic reference cannot be used with "[collective] predicates that require a count interpretation" (Borik and Espinal ibid.: 127) since, as we have just seen, it does not refer to the members of the class: *the lion is numerous is not grammatically correct.

The specificity of definite singular generic reference with human nouns is that it often comes across as offensive, in particular in the case of nationality nouns (just as in French, in which such an NP as l'Indien bespeaks a form of racism, with the classification of the Indians as a type, which is all the more inacceptable as their history is marked by a colonial violence that erased consideration for the members of the class and grouped them under one shapeless mass). Thus "the American thinks, talks, and acts under the influence of high national and Christian precepts" (Myrdal 1962) is semantically inappropriate today, for a prototypical behaviour is predicated of a human group. As grammatically correct as they may be, such statements can readily be interpreted as the expression of the speaker's contempt for the class:

Humans have, in addition to lower-order properties, higher-order properties […]. As a result humans tend to be individualistic and vary from person to person and hence can hardly be generalized about. The use of the kind generic in sentences such as ?The girl plays with dolls or ?The Italian loves pasta therefore sounds inappropriate; it makes us see girls or Italians as a species-like behaviour (Radden 2009: 313, my emphasis).

Myrdal's assertion above disregards any type of difference among Americans and brings an archetypal subject to the fore: definite singular generic reference conjures up stereotypical judgments (which is still related to the fact that THE + singular noun designates a kind). On the other hand, uttering the lion has a mane is not in the least controversial: it points to an essential property that can easily be ascribed to lions (males) as opposed to other species (tigers, etc). Finally, it should noted that definite singular generic reference is not compatible with hypernyms: such an NP as the friend cannot be interpreted as a generic, whereas the best friend can, as it creates a subcategory of friends.

In order to further establish the unique status of generalizations over humans, one may note that except, maybe, for definite singular generic reference, which is almost never deemed appropriate, acceptability depends on the class referred to. Admittedly, "An American thinks, talks, and acts under the influence of high national and Christian precepts" is more appropriate than its definite singular counterpart (The American) as it only extracts a member that has the characteristics of the class Americans. However, indefinite singular generic reference might sometimes be regarded as unacceptable: a gloss of Radden's example in the above quote with "?A girl plays with dolls" is improper as it tends to take on a modal value: playing with dolls is what is expected of a girl. Therefore, although the sentence is grammatically correct, grammar is not the key element here: the statement expresses a gender stereotype and is inappropriate, which is why indefinite singular generic reference does not provide a solution for the unacceptability of the definite singular. This proves the unique status of generalizations over humans: grammar is of paramount importance but should not be regarded as corresponding to a set of rules that can be applied blindingly to human nouns. Taking context into account is crucial whereas generalizations over non-human nouns are not as sensitive to judgments of social inappropriateness.

2. Generalizing over humans: different forms for different social readings

I have shown in my introduction that generic reference is essential in our making sense of what is around us. This entails seeing generic reference as expressing a view of a world that is based on a set of immutable rules: ducks will always lay eggs, birds will always fly, etc. Generalizations over humans often do not follow this logic, as they are subject to change; nevertheless, in stereotypical statements, they might persist and be regarded as conveying such an unyielding set.

While the grammatical difference between birds fly and a bird flies has no prominently perceptible semantic consequences, this is not the case when human nouns are involved. We may compare: "A well-mannered host will offer a guest something to drink" (Google Books 2011) and its bare plural counterpart: "Well-mannered hosts will offer a guest something to drink". Here, the indefinite singular (a host) triggers a reading based on social norms and rules (Burton-Roberts 1977: 188) rather than on definitional properties as with the bird example (a bird flies, where the indefinite singular establishes a definition of the kind bird). "A well-mannered host will offer a guest something to drink" claims that every occurrence of the class of hosts should behave in such a way and does so through a member extracted from said class: this member is construed as an individual who epitomises all of the characteristics expected of the class (for instance, a welcoming, friendly and hearty individual). Added to this is the use of WILL, which is part of the category of congruent modals (meaning that the link between the subject and the predicate is not problematic: this is why WILL is used in such a sentence as "I'll take it!", in a context where a phone is ringing; the fact that the speaker is going to pick up the phone is perceived as natural). In our sentence about a well-mannered host, WILL more specifically conveys a judgment of high probability: according to the speaker, the odds that a well-mannered host will offer their guests something to drink are extremely high. This assessment of high probability is related to the notion that offering guests a drink is a disposition of a well-mannered host.

In comparison, the bare plural ("Well-mannered hosts will offer a guest something to drink") "merely makes a generalisation about [the group]" (Burton-Roberts ibid.); hosts form a group and the bare plural erases individuality, with no consideration of one specific member of the class whatsoever. In using a typical individual, indefinite singular generic reference tends to create a scenario and the co-speaker might very well build up the image of a typical host in their minds, whereas the bare plural does have normative force but brings about the norm simply by expressing how most hosts behave (while the indefinite singular expresses how a given host should behave).

Even though, as I have shown, definite singular generic reference with human nouns raises the question of acceptability, such a noun as host actually licenses this use: "the well-mannered host should always take time to guarantee his guests' comfort" (Google Books 2007). This form is acceptable because the well-mannered host is a social construction, associated to a set of established values. We would barely imagine a host offended by the well-mannered host, as they could very well decide that they do not belong to this class — the category of well-mannered hosts is one you can step out of. On the contrary, the following sentence most certainly triggers a reflection on potentially offensive stereotypes: "Thus leading to the conclusion that the American is superficial, flakey, or both." (Blog 2012), because it ascribes several flaws to a group regarded as a kind, with no possible reference to its members. This is more exactly an instance of social stereotype: "ceux-ci peuvent inclure des éléments positifs mais ils comportent souvent une dimension négative" (Waroquier and Klein 2006: 20). What is more, in this case, you do not decide whether you belong to such a class (even though you might assert that you do not belong to this type of Americans). All of this is thus a matter of social perceptions and studying definite singulars through these sentences is a way of illustrating how generalizations over humans push in its limits what lies behind the traditional forms of generic reference. Accordingly, there is no unique phenomenon: taking the wide spectrum of human nouns into account (in this case, nouns linked to social classes and nationality nouns) is a way of adding different layers to definite singular generic reference and furthering the debate over acceptability and social perception.

3. Nationality nouns and expansion of the grammar of generic reference

The reason why generic reference is limited to three traditionally acknowledged forms lies in the fact that such a form as the + plural noun does not usually trigger a generic reading: the birds fly triggers a specific reading, and even in the case of human nouns, the students frequently engage in voluntary activities refers to a given, specific set of students. In fact, definite plural generic reference is mostly used with nationality nouns.

Definite plural generic reference is widely used with nationality nouns because of the value of THE: the definite article is an operator of pinpointing, which means that it designates an element within a wider set — if you say the table inside a room you are designating this object and not another one among all of the objects present in said room. This allows speakers to refer to a given people in opposition to others: "The Americans are boasters who love money" (Blog 2012), where the Americans are mentioned in contrast to humbler nationalities. On the other hand, "Americans are boasters who love money" is a simple statement based on a speaker's belief about the nature of the class considered in itself — be that as it may, in both cases, the sentence is the expression of a stereotype.

To conclude, there most certainly is a biased restriction in existing studies on generics, because they do not (or seldom) consider humans. They typically exclude THE + plural noun and restrict examples to cases with distributive predicates, where "[something] is predicated of the singular individuals making up [the] plural individual" (Landman 1989: 559); lions have manes can thus be glossed as such: lion 1 + lion 2 + lion 3, etc. have a mane. Each individual is then integrated within the general group through the bare plural. Generics with human nouns broaden the set of predicates that can be included in generic reference: this is the case of "The Romans invented concrete" (Gardelle 2023), whose generic reading relies on the idea of "collective responsibility" (Gardelle ibid.): only a few Romans invented concrete, but the entire class of Romans is credited with the invention so that we come to know that it was them and not another people. This case constitutes an instance of collective predicate: "the Romans" are seen as "the same plural individual" (Landman ibid.: 559) and not divided into a set of individual Romans. Accordingly, the Romans act as a collective that is representative of its people and the definite article designates the class as a whole: the Romans as a distinct community invented concrete, not the Greeks. Generics facilitate the processing of information; they enable us to learn about facts without needing such statements as only the Romans whose names you can find on that parchment invented concrete.


BORIK, Olga et ESPINAL, M. Teresa. 2012. « On Definite Kinds », Recherches Linguistiques de Vincennes, n°41, pp.123-46, https://doi.org/10.4000/rlv.2104.

BURTON-ROBERTS, Noel. 1977. « Generic Sentences and Analyticity », Studies in Language. International Journal Sponsored by the Foundation “Foundations of Language”, volume 1, n°2, pp.155-96, https://doi.org/10.1075/sl.1.2.02bur.

CIMPIAN, Andrei et MARKMAN, Ellen M. 2011. « The Generic/Nongeneric Distinction Influences How Children Interpret New Information about Social Others », Child Development, volume 82, n°2, pp.471-92, https://doi.org/10.1111/j.1467-8624.2010.01525.x.

DAVIES, Mark. 2008-. « The Corpus of Contemporary American English (COCA) », https://www.english-corpora.org/coca/.

GARDELLE, Laure. 2023. « Lions, Flowers and the Romans: Exception Management with Generic and Other Count Plurals », in Laure Gardelle, Laurence Vincent-Durroux et Hélène Vinckel-Roisin (eds.), Reference: From Convention to Pragmatics. Amsterdam : John Benjamins Publishing Company, pp.71-87.

JUDD, Charles et PARK, Bernadette. 1993. « Definition and Assessment of Accuracy in Social Stereotypes », Psychological Review, volume 100, pp.109-28, https://doi.org/10.1037/0033-295X.100.1.109.

KRIFKA, Manfred, PELLETIER, Francis Jeffry, CARLSON, Gregory N., TER MEULEN, Alice, LINK, Godehard, et CHIERCHIA, Gennaro. 1995. « Genericity: An Introduction », in Gregory N. Carlson et Francis Jeffry Pelletier (eds.), The Generic Book. Chicago : University of Chicago Press, p.124.

LANDMAN, Fred. 1989. « Groups, I », Linguistics and Philosophy, volume 12, n°5, pp.559-605, https://doi.org/10.1007/BF00627774.

LESLIE, Sarah-Jane et LERNER, Adam. 2016. « Generic Generalizations », in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Metaphysics Research Lab, Stanford University, https://plato.stanford.edu/archives/win2016/entries/generics/.

RADDEN, Günter. 2009. « Generic Reference in English: A Metonymic and Conceptual Blending Analysis », in Klaus-Uwe Panther, Linda L. Thornburg et Antonio Barcelona (eds.), Metonymy and Metaphor in Grammar. Amsterdam : John Benjamins Publishing Company, pp.199-228,  https://doi.org/10.1075/hcp.25.13rad.

WAROQUIER, Laurent et KLEIN, Olivier. 2006. « De la difficulté de se débarrasser de ses habitudes mentales : les mécanismes cognitifs impliqués dans la persistance des stéréotypes », in Olivier Klein et Sabine Poll (eds.), Psychologies des stéréotypes et des préjugés. Loverval : Labor, pp.19-55.

Pour citer cette ressource :

Ismaël Zaïdi, "Generic reference: the exceptional status of human nouns", La Clé des Langues [en ligne], Lyon, ENS de LYON/DGESCO (ISSN 2107-7029), septembre 2023. Consulté le 02/12/2023. URL: https://cle.ens-lyon.fr/anglais/langue/linguistique/generic-reference-the-exceptional-status-of-human-nouns