Some of our most common technical terms that we all take to be widely understood don’t have a clear meaning – I observed this earlier for the terms “morph”, “bound form” and “affix”, and in this blog post, I will discuss “existential clause” (or “existential sentence”; the difference between these does not matter in the present context). There are some stereotypical cases that everyone would agree are existential clauses, like (1) and (2) but beyond that, the precise extension of the term is not clear. In the spirit of my proposal to provide retro-definitions for widely used terms, I will propose a definition here.
(1) English: There is a cat on the mat.
(2) French: Il y a un chat sur le tapis. [it there has a cat on the mat]
In the earliest typological work that uses the term existential (sentence), Jespersen (1924: 154-156) and Clark (1978) (see also Lyons 1968: §8.4.3), such clauses would clearly count as existential (Lyons uses the example There is a book on the table) – even though they do not talk about “existence” in the literal sense. But it is fairly common in grammatical terminology that our terms are not fully transparent (my favourite example being re-duplication – wouldn’t duplication be much more transparent?), so this doesn’t worry me. In actual practice, the Jespersen-Lyons-Clark definition seems to be the one that most people assume – also Croft (1991) in his influential paper on the negative existential cycle, as well as Veselinova (2013) in the first systematic study of negative existential clauses (which often show peculiarities).
So following the practice of Jespersen, Lyons, Clark, Croft and Veselinova, I would define existential clause as in (3).
(3) existential clause construction: An existential clause is a clause in which an indefinite and discourse-new nominal phrase (the locatum, sometimes called pivot) is said to be in some location that is generally expressed by a locative phrase.
The location can be implicit in the context, so the locative phrase is not obligatory. For example, we could naturally say: A: Why aren’t the children playing on the lawn? B: Well, there’s a big dog… [i.e. on the lawn]. This would still be an existential clause of the same type, even though the location is not overt.
Now there are several kinds of questions that a reader might ask – and as always readers should feel free to add further questions and comments below this post.
Question 1: How can you simply decide on the definition by fiat – isn’t this a research question for the whole discipline? Indeed, many linguists seem to think that the definition of grammatical terms is the result of research, rather than a prerequisite of research. For example, Moro’s (2006: 210) overview article starts out with the observation that there are different English constructions to denote “literal existence” (unicorns exist, there are unicorns, there exist unicorns), and then he says:
“Establishing the defining structure of an existential sentence, thus, has become a major goal of syntactic theory.” (Moro 2006: 210)
If all languages are ultimately made up of the same innate building blocks (or natural kinds), then this approach makes good sense: “Existential construction” could be an innate component of the grammar blueprint (“universal grammar”), and if so, then it is a major research question what these innate building blocks are. And once we have found an innate building block, it may make good sense to attach a traditional label to it.
However, in practice, the search for innate building blocks is a research programme that has not given any definitive results yet, so general linguists need to work with comparative concepts, which are defined by linguists. There is no closed class of comparative concepts, so anyone can propose the concepts and terms they find useful.
Question 2: Isn’t it better to regard these clauses as kinds of locational clauses rather than as “existentials”? Indeed, existential clauses like There is a cat on the mat include a locative phrase and are thus locational, and we could say that the contrast between (1) and (4) is one of information structure. Koch (2012) therefore talks about rhematic vs. thematic location (for the latter, I would use “predicative locative” or predlocative).
(1) rhematic location: There is a cat on the mat.
(4) thematic location: The cat is on the mat.
But this does not mean that we could not also use the Jespersenian term existential in this sense: Having a single term that is widely known is probably better than having to use a compound term each time one wants to refer to existential clauses. Moreover, as Creissels (2019) points out, it is not sufficient to say that (4) is distinguished from (1) by the rhematic (or focused) status of the locatum in (1), because one could have focus on the locatum in (4) (THE CAT is on the mat), and likewise, the locative phrase (rather than the locatum) could be focused in (1) (There is a cat ON THE MAT). For this reason, my definition in (3) specifies that the locatum is indefinite and discourse-new, and it does not say that it needs to be rhematic (though most commonly, it is rhematic/focused).
Question 3: Why should the locatum have to be indefinite? – this does not seem to apply to French and Russian. Indeed, French allows proper names in its well-known Existential construction with il y a, as noted by Creissels (2019: 45).
(5) Il n’y avait pas Jean au cours. [it not there had not Jean at course] ‘Jean was not there at the lecture.’
Creissels notes that Russian also allows names in its Existential construction (Ivana ne bylo ‘Ivan was not there’), and he draws the conclusion from this that the French and Russian construction type is defined by perspectivization, not by definiteness or discourse-newness. This is an interesting suggestion (for which Creissels credits Partee & Borschev 2007), but it can hardly be applied in cross-linguistic research. Creissels (2019) reports on an impressive study of existential constructions (of the inverse-locational type; see below) in hundreds of languages, and it is clear that he did not have information on the contrast between discourse-oldness and perspectivization in all these languages – by contrast, it is not unrealistic to get information on indefiniteness in many languages because grammars written in English, French or Spanish invariably give information on definiteness in translations of examples.
Thus, the range of usage possibilities of the French il-y-a-Construction (and similarly the Russian net/ne-bylo-Construction) is not coextensive with the definition of “existential clause construction” in (3), but since the function described in (3) is certainly a prominent function of the il-y-a-Construction, it is of course quite appropriate to say that the il-y-a-Construction is an existential construction (in addition to being a perspectivization construction, a sentence-focus construction, etc.).
Question 4: Isn’t there an important difference between (permanent) existence and temporary location that should be reflected in the definition? Yes, it is true that some languages use different constructions for permanent location and temporary location, for example German.
(6a) Auf dem Rasen ist eine Katze. [on the lawn is a cat] ’There is a cat on the lawn.’
(6b) Es gibt Löwen in Afrika. [it gives lions in Africa] ’There are lions in Africa.’
While English and French do not distinguish between these two situations (temporary vs. permanent), they are treated very differently in German: It is neither possible to say *Es gibt eine Katze auf dem Rasen for (6a), nor ??In Afrika sind Löwen for (6b). (The latter is not completely impossible, but very hard to contextualize, because it can only mean that there is a temporary presence of lions on a whole continent.)
Koch (2012: §2.2) wants to restrict the term “existence” to permanent location, but this seems to be a German-centred perspective. Just as some languages (like French and Russian) have existential constructions whose usage is broader than (3), others have several different constructions that make more fine-grained distinctions. This is a ubiquitous situation in typology, and it reminds us that we should be careful with our interpretations of the significance of comparative concepts that have well-known labels. The definition in (3) corresponds fairly closely to the English there-is-Construction (though English also allows There’s John under specific circumstances), and this is quite possibly due to the influence of English on Jespersen, Lyons and Clark. So it is good to have additional terms for comparative concepts (like Creissels’ inverse-locational) that correspond to salient distinctions, and maybe we should have the additional terms permexistential and tempexistential to remind us that some languages make more fine-grained distinctions.
(Creissels’ (2019) definition of inverse-locational is similar to my tempexistential, but he additionally requires the condition of “perspectivization” (in terms of Partee & Borschev, as briefly mentioned earlier), and he wants to exclude sentences that do not deviate substantially from the way in which predlocative clauses are expressed – so he wants inverse-locational clauses to be “somehow special”, as also discussed under the next question.)
Question 5: The definition in (3) does not mention the form of the sentence, but doesn’t a construction have to be somehow special to be called existential?
I would answer no, and I would include sentences like Finnish (6), even though it differs from (7) only in word order (the example is from Tham 2013: 304). (But this is not universally accepted, as discussed below.)
(6) Mato-lla on kissa. [mat-ADESS is cat] ‘There is a cat on the mat.’ (Lit. ‘On the mat is a cat.’)
(7) Kissa on mato-lla. [cat is mat-ADESS] ‘The cat is on the mat.’
The definition in (3) is a functional definition that does not mention the form of the construction at all – we can call it a construction-function. Alternatively, comparative concepts for constructions may refer to the form of a construction, which are often called called (construction-)strategies. For example, there is a construction-function “predicative possessive clause”, which may be expressed by a variety of strategies, e.g. “transpossessive clause”, “adnominal possessive clause”, “locative possessive clause” (cf. Stassen 2005).
One could therefore try to restrict the meaning of the term existential clause to constructions that use a specific form to render the function in (3), i.e. to a specific strategy (or set of strategies), as is done by McNally (2016). But if one required that an “existential clause” must be English-like not only in its function (as in (3)), but also in its form (so that Finnish (6) is excluded because it is not English-like enough), how far would one go?
As Creissels (2019: 55-85) explains in great detail, there is a fascinating variety of construction-strategies for existential clauses. Quite a few languages have existential clauses that include a prolocative form like English there (e.g. y in French, ci in Italian), but many others have existential clauses based on a transitive ‘have’ verb, or a special existential item (e.g. Hebrew yeš ‘there is’), or a pronoun-derived expletive (e.g. Swedish det). These possibilities are quite similar to the variety of construction-strategies for predpossessive clauses, and in fact Creissels discusses the interesting coexpression (“alignment”) patterns at some length.
Which of these strategies would count as “existential” if one wanted to restrict the definition not only functionally, but also to specific strategies (as McNally wants to do)? There is no clear answer. McNally (2016: 212) says:
“I use the term [existential] here to refer to sentence types that are “noncanonical’’ in structure, whether due to some aspect of their syntax or the presence of a distinguished lexical item (e.g., Spanish hay), and that are invariably accompanied by what appears to be a special semantics or discourse function related to introducing the presence or existence of some individual(s).”
McNally’s functional definition is oddly vague (“a special semantics or discourse function related to…”), but what I focus on here is the vagueness of her formal requirement: “noncanonical due to some aspect of their syntax” can mean all kinds of things, because there is no uniform definition of what is “canonical”. Is the Finnish sentence in (6) (lit. ‘On the mat is a cat’) with its postverbal subject canonical or noncanonical? Under appropriate circumstances, presentational subjects can be postverbal in Finnish, also outside of an existential context.
In general, comparative concepts for typology must be defined in the same way for all languages (Haspelmath 2018), so a definition that includes “non-canonical in some way” is not a well-defined comparative concept. This is one of the reasons why I favour the definition in (3), which defines existential clause as a construction-function and is completely neutral with respect to the strategy that is used.
Another reason is the fact that Jespersen, Lyons and Clark had the construction-function in mind, as is clear e.g. from the fact that Jespersen (1924: 155) includes Russian sentences like byl mal’čik [was boy] ‘there was a boy’, and Lyons (1968: 390) includes English sentences like Coffee will be here in a moment.
Question 6: Must existence be associated with location? What about “pure existence” where no place is implied? This is perhaps the most obvious question – after all, languages can also say things like ‘God exists’ or ‘There are no unicorns’, where location plays no role and which are therefore not covered by the definition in (3).
So the simple answer is: Yes, languages have clauses expressing “pure existence”, but these are not existential clauses (recall from above that grammatical terminology need not be transparent).
But if they are not existential clauses, then what are they? I would like to propose a new term for them: We can call them hyparctic clauses, based on the Greek word hyparxis (ὕπαρξις), which means ‘existence’ (according to the German Wikipedia article on Existenz, Latin existentia was coined in 360 CE to render Greek hyparxis).
Hyparctic clauses may be coexpressed with (= use the same strategy as) existential clauses, as is the case with English There are no unicorns, or they may be expressed in a special way (the English verb exist ist largely confined to hyparctic clauses, although permexistential clauses like There exist lions in Africa seem possible). But in any event, there does not seem to be a way to subsume existential and hyparctic clauses under a single (more abstract) construction-function – note that hyparctic clauses may have a definite argument, as in God exists, whereas this is excluded for existential clauses.
References
Clark, Eve V. 1978. Locationals: Existential, locative, and possessive constructions. In Greenberg, Joseph H. (ed.), Universals of human language, Vol. 4: Syntax, 85–126. Stanford: Stanford University Press.
Comrie, Bernard & Smith, Norval. 1977. Lingua descriptive studies: Questionnaire. Lingua 42. 1–72.
Creissels, Denis. 2019. Inverse-locational predication in typological perspective. Italian Journal of Linguistics 31(2). 37–106.
Croft, William. 1991. The evolution of negation. Journal of Linguistics. Cambridge University Press 27(1). 1–27. (doi:10.1017/S0022226700012391)
Croft, William. 2016. Comparative concepts and language-specific categories: Theory and practice. Linguistic Typology 20(2). 377–393. (doi:10.1515/lingty-2016-0012)
Croft, William. 2022. Morphosyntax: Constructions of the world’s languages. (http://www.unm.edu/~wcroft/WACpubs.html)
Haspelmath, Martin. 2018. How comparative concepts and descriptive linguistic categories are different. In Van Olmen, Daniël & Mortelmans, Tanja & Brisard, Frank (eds.), Aspects of linguistic variation: Studies in honor of Johan van der Auwera, 83–113. Berlin: De Gruyter Mouton. (https://zenodo.org/record/3519206)
Jespersen, Otto. 1924. The philosophy of grammar. London: Allen & Unwin.
Koch, Peter. 2012. Location, existence, and possession: A constructional-typological exploration. Linguistics. De Gruyter Mouton 50(3). 533–603. (doi:10.1515/ling-2012-0018)
Lyons, John. 1968. Introduction to theoretical linguistics. Cambridge: Cambridge University Press.
McNally, Louise. 2016. Existential sentences crosslinguistically: Variations in form and meaning. Annual Review of Linguistics 2(1). 211–231. (doi:10.1146/annurev-linguistics-011415-040837)
Moro, Andrea. 2006. Existential sentences and expletive there. In Everaert, Martin & van Riemsdijk, Henk (eds.), The Blackwell companion to syntax, volume II, 210–236. Oxford: Blackwell.
Partee, Barbara H. & Borschev, Vladimir. 2007. Existential sentences, BE, and the genitive of negation in Russian. In Comorovski, Ileana & von Heusinger, Klaus (eds.), Existence: Semantics and syntax, 147–190. Dordrecht: Springer.
Stassen, Leon. 2005. Predicative possession. In Haspelmath, Martin & Dryer, Matthew S. & Gil, David & Comrie, Bernard (eds.), The world atlas of language structure, 474–477. Oxford: Oxford University Press. ((http://wals.info/chapter/117))
Veselinova, Ljuba. 2013. Negative existentials: A cross-linguistic study. Rivista di Linguistica 25(1). 107–145.
OpenEdition suggests that you cite this post as follows:
Martin Haspelmath (July 2, 2021). What do we mean by “existential clause”? Diversity Linguistics Comment. Retrieved February 17, 2025 from https://doi.org/10.58079/nswb
How should one handle sentences which are formally “definite”, but semantically “indefinite”? For example, “the lion lives in Africa”, where “the lion” is actually referring to lions in general, rather than a specific, “definite” lion?
I think you’d ignore the fact that this is formally definite in English, since definiteness is intended to be a comparative concept (and saying “a noun phrase is definite if it is preceded by “the”” does not work cross-linguistically), but I just thought to check.
Good question – I don’t know. I have not thought deeply about how to define “definite” as a comparative concept, but I’m aware that it is not straightforward. I had not thought of sentences like “The lion lives in Africa”, and intuitively, they should not count as existential, because “the lion” is a kind of topic here. The same applies to “Lions live in Africa”, which is indefinite (and thus comes even closer to fulfilling the criteria of my definition). So I should perhaps think about adding some condition that excludes such generic sentences.
Here’s a possible definition for “definite” and “indefinite”:
A nominal is said to be “definite” in a particular discourse context if the speaker *believes* that the listener will be able to unambiguously determine what entity the nominal is referring to within that context. A nominal which is not definite is said to be “indefinite”.
Note that under this definition, “definiteness” is dependent both on the context of the utterance, and the speaker/hearer relationship. This is intentional.
When a citizen of England says something like “The Queen visited Russia today”, without any context we don’t know who “the Queen” refers to; after all, many queens exist. However, knowing that the speaker is a citizen of the UK, we can tell that “the Queen” probably refers to the Queen of England, which explains using the definite article. I think that this sort of contextual knowledge is important for understanding how definiteness is used in real life, so it seemed fitting to incorporate it into the actual definition.
Yes, I think that that’s a good proposal. I think there’s a tradition of associating definiteness with familiarity and with unique identifiability. But familiarity is not required for many definites, and unique identifiability is generally sufficient – that’s probably the same as your “unambiguous determination”. In the 1980s, I found Hawkins’s 1978 book compelling, but it wasn’t very cross-linguistic. A recent dissertation is Laura Becker’s “Articles in the world’s languages”: https://books.google.de/books/about/Articles_in_the_World_s_Languages.html?id=n_ppzgEACAAJ&redir_esc=y, and an excellent typological paper is by Dryer (2014).
Thanks for your response!
“Unique identifiability” is indeed the same idea as my “unambiguous determination”, but more clearly worded. 🙂
So the definition could be changed to
“A nominal is said to be “definite” in a particular discourse context if the speaker believes that the listener will be able to uniquely identify what entity the nominal is referring to within that context. A nominal which is not definite is said to be “indefinite”.
In your English examples of existential clauses, the only verb that you use is the rather semantically empty verb “is” (a cat *is* on the mat, there *is* a cat on the mat). Is this significant? Is it possible to use other, semantically richer verbs in existential clauses?
For example, would “a cat sits on the mat” count as an existential clauses? I guess this comes down to what exactly it means when you say
An existential clause is a clause in which an indefinite and discourse-new nominal phrase (the locatum, sometimes called pivot) *is said to be in some location* that is generally expressed by a locative phrase.
Posture verbs are often required in locational clauses (both predlocative and existential), e.g. in German (Das Buch liegt/?ist auf dem Tisch; Auf dem Tisch liegt/?ist ein Buch). So they should definitely be included in the definition. However, the verb should not express much more than location. For example, “There is a technician working in the hangar” is not an existential construction because it primarily expresses a working event, not a location situation.