1. Introduction

In formal syntactic theory, it is commonly assumed that weather verbs such as rain and snow do not assign any theta roles and take an expletive subject, it (Chomsky Reference Chomsky1981). This follows from the fact that meteorological events lack distinct participants such as agent and patient (Eriksen et al. Reference Eriksen, Kittilä and Kolemainen2010). Chomsky uses weather verbs to make a finer distinction among expletives. He notes that in examples like the following, the it of weather predicates can control PRO:

This fact, that weather-it can control PRO, distinguishes it on the one hand from a true pleonastic, like it or there in subject position, which cannot control PRO, and on the other hand from a true pronominal argument which can be replaced by a fully referential noun phrase. He concludes that weather-it is a quasi-argument (p. 325).

In this squib we provide evidence to support the idea that weather-it differs from a true pleonastic pronoun (which we will simply refer to as an expletive from now on). The evidence comes from the occurrence of personal pronouns as subjects of weather predicates in weather reports on CBC Radio One in Manitoba. We focus on examples such as the following, which to our knowledge have not been noted in the literature before:Footnote ¹

In these examples a first-person plural pronoun is used where we normally find weather-it, as shown in the parentheses in (2). These examples are in sharp contrast with examples where a personal pronoun replaces a true expletive:

1. (3)

   1. a. It seems that prices are rising.

   2. b. *We seem that prices are rising.

We argue that the substitution of we for it in weather constructions supports the quasi-argument status of weather-it (cf. Gardelle Reference Gardelle2015).Footnote ²

The squib is organized as follows. In section 2 we briefly review the literature on meteorological expressions and describe uses of personal pronouns with weather predicates that have been noted for English. We add to these uses the construction exemplified in (2) above. In section 3 we discuss how this use of personal pronouns is different from other uses that have been identified in the literature. In section 4 we present our analysis, namely, that the bolded pronominal subjects in (2) are making deictic reference to the location for which the weather is being reported. Section 5 concludes the squib.

2. Personal pronouns as subjects of weather predicates

In their cross-linguistic study of meteorological expressions, Eriksen et al. (Reference Eriksen, Kittilä and Kolemainen2010) state that weather phrases are either a predicate type (e.g., it rains), an argument type (e.g., rain falls), or an argument-predicate type (e.g., rain rains). In other words, the type of expression is determined by whether the weather information is given in the predicate, as an argument, or is divided between the two. English weather phrases, according to Eriksen et al., tend to be of the predicate type with an expletive subject. In languages that pattern like English, subject expletives are either in the form of neuter pronouns like it or spatial adverbs like there. Danish, for example, allow there-expletives with adjectival weather predicates:Footnote ³

Larsson (Reference Larsson2014) similarly notes that in the dialects of some Scandinavian languages, locatives like her/här ‘here’ and der/där ‘there’ can be used instead of the more common det ‘it’ as the subject of weather predicates – for example, här regnar ‘here rains’ in Swedish (Hulthén 1944: 43, cited in Larsson Reference Larsson2014: 66). While these locatives usually have an expletive function, Larsson suggests that här, in particular, may have retained its deictic locative property even in subject position in some dialects.

Following their earlier work, Eriksen et al. (Reference Eriksen, Kittilä, Kolemainen, Helasvuo and Huumo2015) focus on the subjects of predicate type meteorological expressions and note that lexical subjects are possible when they are adverbial in nature, referencing place (e.g., name of a city), time (e.g., today or yesterday), or the atmospheric background (e.g., sky or weather). The authors call these subjects “promoted adverbials”, a point to which we will return in section 4.

With respect to pronouns as subjects of meteorological expressions, Eriksen et al. (Reference Eriksen, Kittilä and Kolemainen2010) note that some languages allow “human pronouns” like he or she to appear as expletive subjects, citing Norwegian and Swedish dialects, Icelandic, and Faroese:

While it is clear from the interpretation that the third person pronoun in (5) serves as an expletive, non-expletive uses of third-person pronouns with weather predicates in English are also attested. Gardelle (Reference Gardelle2015) undertakes a corpus study to identify the conditions under which expressions like She's raining and thundering hard are used. Putting aside cases where the pronoun has a textual antecedent such as Mother Nature, Gardelle proposes that the feminine pronoun with no antecedent adds “emotional involvement” (p. 15) such as feelings of enthusiasm or annoyance. She links these uses to others that do not involve weather predicates such as let ’er rip. She also addresses the question of why it is specifically the feminine pronoun that is used in this way, with reference to theories that have been advanced by others (e.g., Svartengren Reference Svartengren1927, Joly Reference Joly1976, Mathiot and Roberts Reference Mathiot and Marjorie1979, and Stenroos Reference Stenroos2008), though she does not ultimately endorse any of them.

Our brief survey of the literature on weather phrases reveals that first person plural pronouns as subjects of weather predicates have not been discussed. These subjects are not expletive, are not obviously adverbial, and do not have the same flavour of emotional involvement that the third person feminine pronoun contributes. In the next section we briefly survey the literature on non-canonical uses of the first-person plural pronoun in order to show that its use with weather predicates has not been noted in those discussions either.

3. Non-prototypical uses of the first person plural pronoun

The prototypical use of the first person plural pronoun is to refer to the speaker and one or more contextually given associates (see, for example, Ackema and Neelman Reference Ackema and Neeleman2018). However, like all other pronouns, first person plural pronouns may have non-prototypical uses as well (Helmbrecht Reference Helmbrecht2015). In this section we review the non-prototypical uses of we identified by Helmbrecht in order to show that its use in weather reports is of a different nature. We put aside the distinctions between dual vs. plural, and inclusive vs. exclusive first-person pronouns because they are not morphologically marked in English.

Helmbrecht (Reference Helmbrecht2015) considers non-prototypical uses of first, second, and third person pronouns, with most of his examples drawn from German. Here we present his description of first-person plural pronouns, illustrated with either adapted English translations of his invented examples (6a–e) or examples we have invented ourselves (6f). Helmbrecht (Reference Helmbrecht2015: 182–3) notes that a first-person plural pronoun can be used as a third person plural (6a) when a sports fan refers to their team. It can be used as a second-person singular (6b) when the pronoun refers to the addressee (the so-called nursery-we) or as a second person plural (6c) when the pronoun refers to a group of addressees. It can be used as a first-person singular (6d) by someone with very high-ranking such as royal status and/or social or political superiority (the royal we). It can also be used as a first-person singular (6e) when single authors refer to themselves in the plural in academic writing to achieve more authority or formality. While Helmbrecht does not give an English example of the first-person plural pronoun being used as an impersonal or generic pronoun, examples like (6f) arguably show this.

1. (6)

   1. a. We lost again last night. (1pl → 3pl, uttered by the supporter of a sports team)

   2. b. How do we feel today? (1pl → 2sg, doctor to patient)

   3. c. In the last class we learned about c-command. (1pl → 2pl, instructor to students)

   4. d. Yesterday, we gave the order that … (1pl → 1sg, pluralis majestatis)

   5. e. In the previous chapter we argued that … (1pl → 1sg, ‘editorial-we’)

   6. f. We often laugh when we are nervous. (1pl → impersonal/generic)

Helmbrecht convincingly shows that all pronouns can undergo person and/or number shifts for various pragmatic effects. However, we note that in none of the examples in (6) can we be replaced by expletive it because they all receive theta roles from the predicates with which they occur. That is, the pronouns in (6) are fully referential and can be replaced by full noun phrases. In contrast, the we in statements like we are five degrees does not instantiate a person/number shift, does not stand in for another referential pronoun, and cannot be replaced by a full noun phrase. Rather, as we explain in the next section, the we in weather reports refers to the location in which the speaker and their contextual given associates are situated.

4. Analysis

Reports like we are five degrees and we are raining have been heard by the first author on the radio in Winnipeg for at least fifteen years. A small corpus of weather reports gathered from CBC radio in fall 2019 confirmed that the construction is used by more than one CBC announcer in Winnipeg but is not used in other major cities.Footnote ⁴ We are not making any sociolinguistic claims based on its occurrence on one radio station in one Canadian city. We believe that the fact that it is possible to use a first-person plural pronoun as the subject of a weather predicate anywhere is worth investigation.

We are five degrees is one of a number of variations on a theme that are used to avoid repetition while reporting the weather on the radio.Footnote ⁵ As background, meteorological reports in Canada contain information such as the high and low for the day in degrees Celsius, weather conditions (e.g., sunny, cloudy, foggy, etc.), the type, likelihood, and amount of precipitation, the direction and velocity of the wind, plus the humidex at the height of summer and wind chill at the height of winter. Reporting this information on the radio may simply involve some descriptive noun phrases (7a), two noun phrases juxtaposed as a list (7b), or in a topic-comment structure (7c):

1. (7)

   1. a. afternoon high today of 14 degrees

   2. b. showers, risk of a thunderstorm

   3. c. tonight, a low of 14 degrees

Expletive subjects as well as subjects like the high/low are also common:

1. (8)

   1. a. It's 25 in Winnipeg.

   2. b. It's lightly raining.

   3. c. The high is/will be 14 degrees.

The examples in (8) all involve a copula verb be; however, there are a few other verbs that frequently occur in weather reports and that can take first person plural subjects:

1. (9)

   1. a. We will get up close to 30 degrees.

   2. b. We are at 24 degrees.

   3. c. We should wind up around 21 degrees.

   4. d. We are headed to a high of minus 10 degrees.

We propose that these are metaphorical constructions that involve viewing temperatures as locations and “we” as the travellers to those locations.Footnote ⁶ Unlike true weather constructions in which there is no theta role for the subject, we argue that the first person plural subjects in constructions like (9) get theta roles from the predicates (e.g., be at, get close to, wind up around, head to). For this reason, expletive it is not a possible subject in these sentences.

Returning to the constructions that are the focus of this squib, we claim that sentences like we are five degrees are not metaphorical constructions, based on the fact that we can alternate with expletive it as the subject (cf. we are/it is five degrees). Our proposal is that we is the realization of one of the deictic coordinates of the utterance, i.e., the key aspects of the context of utterance that can be encoded by deictic elements. Among the deictic coordinates that are available to interlocutors are the speaker and addressee, typically encoded by first and second person pronouns, respectively; the space occupied by or around the speaker, typically encoded by spatial adverbs like here; and the time of utterance, typically encoded by temporal adverbs like now. These are referred to as person, spatial, and temporal deixis respectively (see, for example, Fillmore Reference Fillmore1997; Levinson Reference Levinson1983, Reference Levinson, Horn and Ward2004; and Saeed Reference Saeed2016).

The location of the person giving a weather report is particularly important, though often unexpressed. This is because statements such as it's sunny, it's raining, it is minus three degrees are not understood to be true of everywhere in the world at all points in time. They are understood to be true of the speaker's location at the moment of utterance (here and now). The speaker's location is understood, in turn, to be the maximal geographical area across which the weather is uniform such as a city or a town. Note that on the radio the location for which the report is being given may, but need not be, expressed overtly as a prepositional or adverbial phrase (e.g., It's sunny and minus five degrees in Winnipeg). Overt expression of location is rarer in conversation presumably because it is evident to all speech participants where they are. In other words, if all participants in a conversation are in Winnipeg, it is unnecessary to add in Winnipeg or here to statements like It's cold.

To summarize so far, we are claiming that a location is a necessary part of meteorological statements and that it can be expressed via an adverbial phrase. In expressions like we are five degrees we see that the realization of this location is innovative in two ways: (a) that the location is expressed as the subject and (b) that the location is expressed via a personal pronoun. With respect to the first point, we have seen that there are languages that allow locative adverbs such as here (Erikson et al. Reference Eriksen, Kittilä and Kolemainen2010, Larsson Reference Larsson2014) or other types of “promoted adverbials” that index location (Erikson et al. Reference Eriksen, Kittilä, Kolemainen, Helasvuo and Huumo2015) to serve as the subjects of meteorological predicates. With respect to the second point, we propose that the first-person plural pronoun is a possible realization of the information encoded in the locative adverb here, because this adverb in turn necessarily includes the location of the speech participants.Footnote ⁷ In other words, here in a given interaction between speech participants, means where we are. In proposing that the implicit person features involved in spatial deixis can be spelled out as a personal pronoun rather than as a DP (e.g., the place we are) or an adverb (e.g., here), we are extending the non-prototypical uses of pronouns identified by Helmbrecht to encompass not only person and number shifts (see discussion of the examples in (6) above) but also a shift from location to person – a type of personification of location.

The personification of location can be extended to the use of second-person pronouns when the location (“where you are”) is clearly identified by name and is distinct from the location from which the report is being broadcast. Thus, alongside reports like We are minus fifteen we get the following:

1. (10)

   1. a. Brandon, you are minus ten. [cf. It's minus ten in Brandon.]

   2. b. Thompson, you are cloudy and a balmy minus five degrees. [cf. It's cloudy and a balmy minus five degrees in Thompson]

Note that Brandon and Thompson are cities in Manitoba but are being addressed vocatively in the examples in (10).Footnote ⁸ These examples bear a striking resemblance to those noted by Piepers et al. (Reference Piepers, Maria van de Groep and de Hoop2021) in their Twitter corpus. Piepers et al. gather examples like Amsterdam, you're raining, which appears in the title of their paper, noting that they involve a vocative and a spatio-temporal addressee personified by a second person pronoun. They argue that the user's goal is to tell an imagined audience about an experience they have had at a certain place and time. However, there are at least two differences between the two types of constructions. First, their corpus includes not only meteorological observations but also positive and negative evaluations (Valentine's Day, you are a bitch). Moreover, what they identify as the illocution of these utterances – evaluating a first-hand experience (p. 107) – is absent in the weather reports we have been discussing. Second, first person plural pronouns are unattested in their corpus while they are prevalent in our examples. While the constructions differ, taken together they show that the personification of spatio-temporal coordinates may not be so rare.

As a final note, while this squib has focused on the use of we for location in weather reports, we have heard at least one instance of we for time as well:

1. (11) [Pilot on Air Canada flight YYC to YWG (AC288, November 17, 2019)]

   Hope to have you on the ground early at 40 minutes past the hour. Right now, we are 17 minutes past.

Again, the use of we as subject in the utterance above is shorthand for it is 17 minutes past where we are. In the absence of any explicit spatial or temporal elements in the utterance, the deictic coordinates are taken to be where the speech participants are and can be referenced through a personal pronoun.