Place is often defined as any location that can welcome humans and their social interactions, and to which humans can develop forms of attachment (Cresswell, 2014: Ch. 1). Contemporary geographical thinking on place originated in Tuan (1977) and deeply influenced psychology (Manzo and Devine-Wright, 2021), sociology (Low, 2016), and other disciplines. Complex theories of place(s) and relations between individuals and these concepts therefore emerged (e.g., Scannell and Gifford, 2014). Formal treatments of the properties shaping places were proposed in psychological frameworks (Canter, 1977, 1997). Formalizations of these notions also emerged in Geographic Information Science (GIS) and exploited formal ontologies, first order logic-based specifications of conceptual domains (Hamzei et al., 2020). Place is thus an increasingly refined notion that finds a mentalist counterpart in the general concept of “place,” and in its more specific declinations (e.g., “country,” “city”).

Linguistic disciplines, however, present a different picture. Toponomastics studies place names and their etymological roots (e.g., Sydney, Pitt Street: Tent and Blair, 2011; Perono Cacciafoco and Cavallaro, 2023). Linguistic typology offers cross-linguistic analyses of place names’ grammatical properties (Stolz and Levkovych, 2020). Though both frameworks have made important discoveries on place names, neither framework connects these discoveries to theories of place. Furthermore, the conceptual semantics and generative syntax frameworks use the term “Place” to analyse adpositions and their senses (respectively Jackendoff, 1983; Svenonius, 2010). The consequences of this theoretical mismatch between linguistics and other disciplines create the empirical impasse that we illustrate via (1)–(3).

The sentence in (1) describes an event of a plane landing in the Australian city of Sydney; (2) describes a certain street, as a place within Sydney’s business district. Sydney and Pitt Street are the place names that allow individuals to refer to these places. Instead, (3) describes some boys playing in a nameless park, i.e., a location introduced in discourse without a specific “platial” (i.e., place-like, Tenbrink, 2020) status. Therefore, (1)–(2) are sentences offering semantic information about two closely related and unique places; (3) offers information about some non-specific location identified as a park.

Geographic and psychological frameworks suggest that (1)–(2) convey information about places via place names, mediated via mind-internal place concepts (Kuhn, 2012; Low, 2016). Instead, most toponomastic works suggest that this reference is direct: individuals do not have concepts about places (Coates, 2017). Generative and conceptual frameworks would consider the P(repositional)P(hrase)s in Sydney and in a park, in (1)–(3), as equally referring to places via place concepts (Svenonius, 2010). Linguistic frameworks do not attempt to capture the content that place names carry; some frameworks do not even semantically distinguish place names from other NPs and spatial PPs. Hence, place names and place concepts receive short thrift in linguistics.

This neglected status of place names in linguistics has inter-disciplinary ramifications. Psychological research suggests that individuals can develop rich place concepts, and associate them to place names (Scannell and Gifford, 2010; Davies, 2013). Some individuals may use (1)–(2), and thus the name Sydney, to describe situations and facts about a city they love; others, for a place they are unfamiliar with. Other uses are possible; individuals can express their own conceptual views about places via place names, and continually develop rich mind-internal representations (i.e., concepts) of places. GIS works instead suggest that place names retrieval algorithms need detailed semantic and morpho-syntactic representations to inform place models (Buscaldi, 2011; Zhang and Gelernter, 2014). Linguistic theorising on place names, as we gleaned via (1)–(3), lacks a morpho-syntactic analysis of place names, and mostly explains away their semantic properties. Therefore, the cognitive mechanisms by which place names convey information about places remain mostly unexplored.

The goal of this paper is to ameliorate this situation by proposing a linguistic model of place names integrated with a conceptual model of place. We use Discourse Representation Theory (DRT: Kamp et al., 2011) as a semantic framework to model the content of Sydney, Pitt Street, and other place names. We choose this framework because current DRT architectures use a general format (“Discourse Representation Structures”) to model any mental content, irrespective of the modality (e.g., Pross, 2010; Maier, 2015; Kamp, 2022). Therefore, DRT provides a formal language that can model the multi-modal cognitive content of place and place names into a general, if preliminary theory of “place.” We integrate DRT as a Domain-general Representation Theory with a generative analysis of place names. We then show that the lexical properties and relations connecting place names can be modelled by integrating this multi-modal model (DRT plus syntax) with a frame semantics analysis. We therefore show that our model of “place” can pave the way for a novel, inter-disciplinary approach to platial research (Tenbrink, 2020).

To reach this goal, we first analyse “place” as a psychological concept of a geographical notion (Section 2). We then discuss GIS formal treatments (Section 3). We problematize the uses of the label “place” in Linguistics and outline desiderata for a linguistic analysis (Section 4). We develop a multi-modal, representational model of place concepts and place names. We achieve this result via domain-general DRT’s anaphoric relations between cognitive domains (“anchoring relations”), coupled with a generative view of place names’ grammatical properties. We then address a number of recalcitrant problems stemming from the lack of a theory of place names informed by theories of place. We thus offer an account of the linguistic properties of place names, how they represent place concepts and networks, and how individuals can share place concepts and relations in discourse and communication (Section 5). Section 6 discusses the consequences of this proposal for a theory of place. Section 7 concludes.

The goal of this section is to discuss how psychology has absorbed the geographical notion of place. First, however, we establish our terminology. We use the term place concepts to refer to concepts for cities, streets and other entities that can be considered particular instantiations of the general concept of place. “Sydney,” “Pitt Street,” are possible labels for singular place concepts: concepts for one (quantity-wise) unique place. We standardly conceive concepts as multi-modal, mind-internal representations about entities in the world (Laurence and Margolis, 1999; Margolis and Laurence, 2007; Carey, 2009; Ursini and Samo, 2022). We use place(s) as a label for geographic entities, the mind-external counterparts of Place concepts. We use italics for place names: Sydney, Pitt Street, respectively, are the names for the “Sydney,” “Pitt Street” place concepts. Given this distinction between mind-internal place concepts and mind-external places, we will discuss how reference to either entity type can come about in Sections 4–5, once we introduce the relevant notions.

The emergence of place as a theoretical construct originates in the work of geographers offering philosophical reflections on humans and their environment (Relph, 1976; Seamon, 1979; Tuan, 1979, 1990). These geographers suggested that humans develop place concepts via their interaction with places, thus imbuing these concepts with subjective meaning. Subsequent works sought to explore the psychological and social dimensions that could shape these concepts (Couclelis, 1992; Agnew, 1993; Harve, 1994; Thrift, 1999; Hollenstein and Purves, 2010). Hence, they suggested that the content of place concepts can develop and vary considerably over time, depending on individuals and groups’ emotive and cognitive connections to places. For instance, local dwellers can develop love for Sydney as their home; international rugby fans can love or hate this city when visiting it to experience rugby matches; manyfold possible forms of attachment are certainly possible. Current theories in geography (e.g., Agnew, 2011), philosophy (e.g., Malpas, 2018) and psychology (e.g., Smith, 2017) thus accept place and place concepts as mirror-like notions about locations that are meaningful to humans.

Research on place concepts abounds (cf. Agnew et al., 2015: Ch. 1; Mocnik, 2022); however, proposals tend to diverge on working terms and notions. The notion of “place attachment” has received the most attention in psychology (Smith, 2017; Manzo and Devine-Wright, 2021), sociology (Low and Altman, 1992; Low, 2016) and anthropology (Fitch, 2014; Harun et al., 2015), among others. Place attachment studies investigate how groups and individuals can develop forms of attachment to places by dynamically fostering (place) concepts. Attachment relations can involve manifold emotional, social and cognitive factors interacting in complex manners (e.g., Lewicka, 2008, 2011). These relations can be shaped via direct experience of a place or via indirect information gained via multi-modal sources (e.g., respectively Tang et al., 2021; Merciu et al., 2022). For instance, Sydney locals and tourists can develop attachment via experiences, alone and in groups (e.g., as rugby fans), and via books, videos, social media and any other relevant media.

Though place attachment studies may abound, a dearth of formal models seems attested (cf. Turton Catherine, 2016; Inalhan et al., 2021). One exception is the “tripartite place attachment” model of Scannell and Gifford (2010, 2014, 2017). In this model, place attachment is a triadic relation between “People,” “Places” and “Processes.” In the “People” domain, the model studies whether individuals or groups develop forms of place attachment (cf. also Manzo, 2005). In the “Places” domain, the model studies the facets, i.e., features and dimensions of variation constituting place concepts (Canter, 1977, 1997, 2012). Place facets can represent social and physical properties of places as two distinct types (Uzzell et al., 2002). In the “Processes” domain, the model studies attachment relations between individuals and places via emotive, cognitive and social bonds (Brown et al., 2003). The model suggests that place attachment arises when people (e.g., local inhabitants) develop relations towards places (e.g., love for Sydney) via processes (e.g., living “there”), and via place concepts.

Other notions attaining to place have also been thoroughly investigated, such as place identity, place memory, place formation, and others falling under the “place-making,” “sense of place” rubrics (Seamon, 2018). Place identity studies, for instance, investigate how individuals can shape their identities via the places they interact with (Proshansky et al., 1983; Hay, 1998; Pretty et al., 2003; Banini and Ilovan, 2021). These studies also investigate how the physical and social features of places shape the identities of individuals and the groups they belong to (Bonaiuto et al., 1996; Twigger-Ross and Uzzell, 1996; Stedman, 2002; Clayton, 2003; Knez, 2005). For instance, Sydneysiders can define themselves as city dwellers and beach-goers because Sydney is an Australian city blessed with several beaches. In general, individuals internalise places’ features as mental facets shaping place concepts that they use to develop complex place attachment relations. Place-making is therefore an ever-going process by which individuals imbue place concepts with meaning, thus defining their “sense of place” via the facets constituting these concepts (Seamon, 2012, 2018, 2021).

As our necessarily concise summary suggests, psychological theories of place take a dual stance inherited from Geography (Cresswell, 2014). While “place” describes mind-external entities, place concepts describe internal, multi-faceted and subjective representations that individuals have of places (e.g., Marzano and Siguencia, 2016). Psychology and other cognitive disciplines usually propose model-specific terms and assumptions about place concepts (Seamon, 2018). However, they converge in assuming that place concepts reflect social and physical/geographical aspects of places via facets, and guide the formation of place attachment relations (Scannell and Gifford, 2014). Furthermore, facet theory analyses suggest that place facets can be operationalised via formal languages such as first order logic (e.g., Canter, 1977, 1997, 2012). However, they leave the development of these tools mostly underdeveloped. Psychological models thus offer formally vague models of place concepts. At the same time, they also generally gloss over linguistic matters.

The goal of this section is to analyse how GIS has developed formally explicit treatments of place concepts. Within this field, the development of place models has moved in parallel with the exponential growth of information in two domains. One includes on-line gazetteers and geographical databases (Abdelmoty et al., 2007; Baglioni et al., 2007; Jones et al., 2008; Winter et al., 2009; Allen et al., 2016). The other includes “volunteered geographic information” (Goodchild, 2007; Keßler et al., 2009; Southall et al., 2011; Sui and Goodchild, 2011; Derungs and Purves, 2014; Dunn, 2017). On-line, open-access gazetteers and social media (e.g., OpenStreetMap, FourSquare) exist thanks to the spontaneous (“volunteered”) information that users offer about Places (Williams et al., 2012; Janowicz et al., 2013; Gao et al., 2017; Hu et al., 2019). GIS studies build models to represent these views as expressions of individuals’ place concepts and their inter-subjective variation.

Researchers have hence developed geo-ontologies, structured databases in which platial, individual-based content is explicitly represented. Most works use variants of first order logic (Frank, 2003; Agarwal, 2005a,b; Bishr and Kuhn, 2007; Fonseca, 2008; Fonseca and Câmara, 2009; Couclelis, 2010; Vasardani and Winter, 2016). Furthermore, several proposals use “geo-atoms” as the building blocks of these ontologies (Kuhn, 2005, 2012; Goodchild et al., 2007; Winter and Freksa, 2012; Hu, 2018; Purves et al., 2019). The core properties of geo-atoms can be defined as follows. First, geo-atoms are based on objects, usually defined via crisp boundaries. Second, geo-atoms represent the locations of these objects defined via a spatio-temporal coordinate system. Third, geo-atoms represent the properties and corresponding values associated to objects and locations. The use of geo-atoms allows researchers to represent place concepts via the format in (4)–(5) (cf. Kuhn, 2012).

As (4) shows, a geo-atom A represents an object x, a spatio-temporal location l for the object, a set of properties S individuating the object, and their respective values for the object at its unique location [i.e., S(l(x))]. The location function l(x) represents a “geo-object”: an object occupying a single spatio-temporal location. Consider now (5): this geo-atom offers a compact formalisation of a possible singular place concept for Sydney. For instance, Sydney can have 5 million inhabitants and a certain number of districts. It can also be popular among tourists and be a great or awful city to live in, according to its citizens. These and other properties can represent how individuals may develop place concepts for places, and connect these concepts into complex conceptual networks (Hu, 2018; Purves et al., 2019).

Research on GIS ontologies has also focused on two sub-goals: the formalisation of place and place concepts’ properties, and of relations among places (Ballatore et al., 2014; Ballatore, 2016). For instance, Hamzei et al. (2020) formalises facets as the properties and corresponding values that allow individuals to individuate places in space and time (cf. also Entrikin, 1994, 1996, 1997; Canter, 2012). This work divides place concept facets into two macro-types: geographic and anthropocentric, i.e., affordance-centred, facets (Gibson, 1979; Egenhofer and Mark, 1995; Galton, 2010; Jonietz and Timpf, 2015). Both facet types find their justification in volunteered descriptions of Places. For instance, WeChat users can tag pictures of Sydney’s Pitt Street and describe it as a peaceful central district street, providing information on geographical (i.e., location) and anthropocentric (i.e., peacefulness) facets they associate to this street. Representational systems such as geo-atoms can therefore represent place concepts, and can also represent facets as formal properties shaping these atoms [e.g., extension in (5)].

GIS works have also introduced the notion of “place network,” which requires some motivation for its current discussion. Several works have suggested that the formal modelling of concepts also passes through the study of relations (spatial, conceptual) among Places (Montello, 1993, 2009; Golledge, 2002; Montello et al., 2003; Tenbrink and Kuhn, 2011). Recent works on place information retrieval have shown that geo-tagged sources often offer information about “what” facets individuate places, but also “where” these places can be located (Vasardani et al., 2013; Zhou et al., 2016; Chen et al., 2018; Hamzei et al., 2021). For instance, Instagram users can tag pictures of Pitt Street and Hyde Park while being in Sydney. They therefore provide information about this street and this park being “in” Sydney, and about Sydney being a place that includes other distinct places. The place concepts and models that individuals can share in discourses must represent these manifold relations (Chen et al., 2018; Maren et al., 2020).

These insights are summarised in Purves et al. (2019)’s “ontological commitment” about place concepts. Formalisations of place concepts (e.g., geo-atoms) must represent objects occupying locations and instantiating manifold facets (i.e., places); they can enter relations and participate in events (i.e., define networks). To see why this ontological commitment finds empirical justification, consider (6)–(7) and how they convey information about two places, Sydney and Brisbane.

The sentence in (6) suggests that Sydney and Brisbane stand in a (approximate) “South-of” spatial relation. Instead, (7) suggests that the two Australian cities are connected via a type of social relation: the “rival-of” relation. Sports teams and musical bands have participated in various events that can have motivated the growth of this rivalry over time. Individuals who offer such descriptions of these places thus also offer information about their “Sydney” and “Brisbane” concepts, and how they subjectively connect these concepts. Hence, Purves et al. (2019)’s commitment can be interpreted in this way. Place concept formalisations (e.g., geo-atoms for “Sydney,” “Brisbane”) must represent objects (e.g., the physical cities) and their properties in their locations. These concepts can enter relations (e.g., being “South of” another place), and participate in events (e.g., “being in a rivalry”). Crucially, (6)–(7) also suggest that place names and other categories (e.g., adposition South of) introduce reference to these concepts. However, place names refer to places; spatial adpositions, instead, refer to relations between places and/or locations.

The goal of this section is to assess the status of “place” notions in linguistics; we begin our analysis from Toponomastics. Most works focus on the etymology of place names, and practices of use among communities (Kadmon, 2000; Watts, 2004; Ainiala et al., 2012: Ch. 2). Studies investigating place names’ grammatical properties usually suggest that place names can include generic terms and specific terms. While Street is a generic term, Pitt in Pitt Street is a specific term (Zinkin, 1969; David, 2011; Tent and Blair, 2011; Blair and Tent, 2015, 2021; Tent, 2015; Perono Cacciafoco and Cavallaro, 2023). Some works argue that place names represent individuals’ conceptualisation of their environment (Reszegi, 2012, 2020a,b). Most works however propose that place names only introduce “proper” (i.e., pragmatic) reference conditions to mind-external places (Coates, 2006a,b, 2013, 2017; after Kripke, 1972). Hence, toponomastics studies mostly ignore place concepts and their relations to place names.

Typology has instead uncovered three key morpho-syntactic properties of place names across languages (Cablitz, 2006; Nash, 2015; Nühling et al., 2015; Rybka, 2015). First, place names only including specific terms act as bare NPs (e.g., Sydney). Place names including generic terms are phrasal compounds (e.g., Pitt Street; Schlücker and Ackermann, 2017; Schlücker, 2018, 2020). Second, place names usually refer to unique discourse entities, and thus incorporate definiteness and specificity features (Anderson, 2004, 2007). Hence, they are a highly distinctive NP sub-type across languages (Van Langendonck, 2007, 2017; Köhnlein, 2015; Van Langendonck and Van de Velde, 2016). Third, place names often combine with case markers (e.g., locative case in Finnish) and may license the omission of spatial adpositions (Stolz et al., 2017; Stolz and Warnke, 2018; Haspelmath, 2019; Stolz and Levkovych, 2020). Typological studies thus offer fine-grained grammatical analyses of place names but explain away their semantic properties, and their relation to place concepts as psychological constructs.

While toponomastics and typology are frameworks that focus on functional, social and thus mind-external uses of languages, conceptual semantics and generative syntax take a mind-internal stance. Conceptual semantics attempts to individuate the core conceptual categories underpinning the senses of linguistic units (Jackendoff, 1983, 1990, 1997, 2002, 2010; van der Zee, 2000; van der Zee and Slack, 2003). The framework suggests that adpositions (e.g., in, behind) find their senses in PLACE-functions, which map the senses of NPs, i.e., concepts for objects, to concepts for locations: “Places,” in this framework. For instance, the PP in a park, from (3), includes an adposition, in, and an NP, a park. The sense of the NP is a concept/function individuating a park in discourse, as a landmark object. The sense of in is a concept/function mapping the “park” concept to an internal location, i.e., an “in” place. Other functions (e.g., “front” for in front of) may select specific place concepts defined with respect to landmark objects. In this framework, place concepts model adpositions’ senses. Therefore, these senses are not unique, shaped by individuals’ views nor associated to place names.

Works within generative syntax built theoretical analyses on these insights. These works adopt “Place” as a label for a category projecting from spatial adpositions and case markers (e.g., in, at: Wunderlich, 1991, 1993; Koopman, 2000; Kayne, 2004). This category individuates concepts for regions defined via a landmark object (e.g., “in,” “front” regions). The emergence of the Syntactic Cartography framework introduces a refinement of this analysis (e.g., Cinque and Rizzi, 2010; Garzonio and Rossi, 2020). Cartography suggests that each morpheme in complex adpositions can project a distinct category. If we take in front of as an example, the morpheme front introduces a so-called Ax(ial)Part category, which introduces a “front” concept. The morpheme in projects the Place category, and the “place” concept associated to the “front” concept (Svenonius, 2006, 2008; Franco, 2016). The morpheme of projects a Kase category, and a relation between “front” location concept and landmark concept. We discuss the upshot of this account via (8).

According to this analysis, the PP in front of the table introduces a “Place” concept individuating a front location via a given table. If we take the use of this term as in geography and psychology, individuals may use PPs and sentences such as (8) for locations to which they have affective and social relations. Aside this being a problematic claim, it would entail that in front of the table could be treated as a place name for a generic location unoccupied by objects, i.e., not a “proper” place. Thus, these frameworks create wrong predictions regarding place concepts and their mapping to place names.

Formal, model-theoretic semantics offers a parallel set of theoretical issues when addressing the notion of “place.” For instance, Cresswell (1978) suggests that English adpositions (e.g., over) can denote “places” as possible destinations of moving entities, but does not define this notion. Model-theoretic formalisations of conceptual semantics propose that PLACE-functions denote unique regions in space, or Eigenspaces, via the location function loc’ [i.e., we have loc’(x) = l]; Wunderlich, 1991, 1993; Zwarts and Verkuyl, 1994). Subsequent works propose that adpositions can denote regions or vectors (e.g., respectively, Nam, 1995; Aurnague et al., 1997; Krifka, 1998; Zwarts and Winter, 2000; Aurnague, 2004; Matushansky and Zwarts, 2019). Variants of DRT offer analyses of adpositions and spatial relations (e.g., Asher and Sablayrolles, 1995; Maillat, 2001, 2003; Roßdeutscher, 2013, 2014; Haselbach and Pitteroff, 2015; Haselbach, 2017). Thus, model-theoretic works address neither place Names nor place concepts.

Formal ontology works investigating linguistic issues present a different picture. Guizzardi (2005) suggests that adpositions can express mereo-topological (i.e., spatial “part-of”) relations (cf. also Egenhofer and Mark, 1995; Smith, 1995; Guarino, 1998). Smith and Mark (1998, 2001) suggest that generic terms in place names (e.g., street) may refer to “fiat” locations. Fiat (i.e., arbitrary) locations are such because humans can make complex assumptions on the properties/facets that individuate these locations (e.g., the postulation of boundaries: Smith and Varzi, 2000). These proposals treat place as a fiat type concept (Yue, 2013). They suggest that place names (e.g., Sydney, Pitt Street) encapsulate unique place concepts (e.g., “Sydney” as a unique city in Australia). Thus, these works analyse place names and adpositions as distinct categories realising distinct concept types within a rich ontology of “Space.” However, they do not explicitly state how one can analyse place names and concepts via these distinctions.

We can now summarise the picture emerging from Sections 2–4. Psychological models of place (e.g., Scannell and Gifford, 2010) converge on two assumptions. First, place concepts are shaped via the features that define places, though these features can be subjective and internalised as facets. Second, individuals and groups can internalise relations (e.g., forms of attachment) with these concepts. Crucially, facets theory and GIS models (e.g., Purves et al., 2019) suggest that place concepts and facets can be formalised via first order logic tools: geo-ontologies, representational tools (e.g., geo-atoms), and place networks. Third, linguistic frameworks do not lend themselves to integration in this picture. Some frameworks focus on pragmatic matters, thus explaining away the semantic properties of place names (Coates, 2017; Haspelmath, 2019). Other frameworks erroneously associate place concepts to adpositions (Jackendoff, 1983; Svenonius, 2010). Furthermore, ontology-driven and model-theoretical frameworks (Guizzardi, 2005; Haselbach, 2017) do not explore place names and their relations to place concepts. Therefore, (1)–(3), (6)–(8) and a unified model that makes correct predictions about place concepts, place names and their relations are still outstanding.

The goal of this section is to outline a unified, multi-modal model of place names and place concepts that integrates different frameworks into a single theoretical architecture of cognition. Before we proceed, we clarify which version of DRT we implement and why.

DRT is a dynamic, representational, modular theory of meaning (Kamp et al., 2011: Ch. 1; Nouwen et al., 2016; Geurts et al. 2020). Dynamic, because it offers an incremental model of discourse interpretation that captures the role of extra-linguistic and linguistic context. Representational, because it manipulates Discourse Representation Structures (DRSs), which represent meanings as mental objects that speakers process as discourse unfolds. Modular, because it assumes that the interpretation algorithm takes structures (e.g., syntactic trees) as inputs, and returns DRSs as outputs. These structures can be connected with non-linguistic structures (e.g., concepts) via formally defined mappings, while being evaluated against a discourse context.

DRT comes in dozens of variants: many of these variants offer “layered” representations of linguistic content (e.g., Maier, 2006, 2015, 2017), and psychological states (e.g., Kamp, 2021, 2022). Crucial for our concerns is the fact that DRSs can be used as cognitive Domain-general Representation Structures.

Thus, via DRT, we can simultaneously model linguistic content, i.e., the senses of place names, and non-linguistic place concepts. In so doing, we can also integrate with DRT different theories on place (e.g., facets theories, formal GIS theories of place) that propose first order logic as a formal language. Therefore, we choose this framework as our starting point for two reasons. Theory-wise, DRT allows us to use one meta-language to connect different frameworks for studying place. Data-wise, it allows us to solve recalcitrant problems involving place names and place concepts via a multi-modal implementation of anaphoric relations.

We proceed as follows. We first introduce the basic principles of DRT as a theory of meaning (Section 5.1). We show how DRT permit us to formulate domain-general representations of meaning, and connect these representations via referential (“anchor”) relations (Section 5.2). We then move to the empirical applications, which cover place names’ grammar, semantic and lexical content, and their inter-subjective use in discourse (Section 5.3).

We believe that our proposal offers key solutions to the recalcitrant problems outlined in the introduction. We expose these solutions via five discussion points.

First, we offer a novel morpho-syntactic account of place names. This account solves the problem of “category conflation” emerging in previous generative models (e.g., Svenonius, 2010). We model place names as corresponding to either phrasal compounds or bare NPs (cf. Köhnlein, 2015; Schlücker and Ackermann, 2017). This analysis hinges on identifying two morpho-syntactic categories: Place as a classifier category, Name as a content category. These categories form place names as an NP sub-type, distinct from other NPs (e.g., indefinite a park) and from PPs (e.g., in a park). We then embed our novel proposal in current formal (generative) treatments of proper names and NPs (e.g., Anderson, 2007; Acquaviva, 2019). As a welcome result, we can account increasingly complex places names (e.g., New South Wales, Pitt Street and City of Sydney), but also “bare” place names (e.g., Sydney). This is the case because we offer a generative, recursive and empirically justified morpho-syntactic analysis of place names.

Second, we offer a novel semantic account of place names. This account solves the problem of place names’ content being reduced to their referential properties (e.g., Coates, 2017). We use a multi-modal DRT analysis to achieve this result (cf. Pross, 2010). We assign a semantic interpretation to place names that captures their descriptive content and referential potential, via DRT’s treatment of proper names (Kamp, 2015; Maier, 2017). According to this interpretation, place names have senses that allow individuals to refer to places via the rich descriptive content they associate to these places. We also show that place names as NPs introduce reference to the baptism events in which a place receives its name (Kamp, 2022). We thus formally reconstruct accounts that address how individuals associate place names to rich senses (e.g., Blair and Tent, 2021; Perono Cacciafoco and Cavallaro, 2023). More importantly, we can capture the compositional contribution of the categories forming place names (e.g., Pitt, Street in Pitt Street). We also reconcile descriptive and referential theories of proper names under one multi-modal DRT model (cf. Kamp et al., 2011; Kamp, 2022).

Third, we develop a formal treatment of lexical relations among place names that reflects places’ conceptual relations (cf. Reszegi, 2020a,b). This account solves the problem of offering a full-fledged treatment of place networks as conceptual/semantic relations among place names and the place concepts they reflect (cf. Purves et al., 2019). We achieve this result by extending our approach with a frame semantics analysis of facets and place concepts (e.g., Löbner, 2014, 2021). We show that place names and terms (e.g., alley and street) can enter hierarchical types of relations in conceptual networks. We thus connect standard insights from lexical semantics (e.g., Cruse, 2000; Hanks, 2013) with novel results in GIS (e.g., Purves et al., 2019; Ursini and Samo, 2022). We also reconstruct a view of place names implemented in gazetteers as nodes of complex networks (e.g., Hill, 2006: Ch. 4; Hu, 2018). Overall, we sketch an initial formal model of the lexicon of place names (Reszegi, 2020a,b). Hence, we account why and how place names can refer to places, their facets and their relations.

Fourth, we solve the problem of having a full-fledged and cognitively plausible linguistic theory of place names. We model place concepts as domain-general DRSs: enriched versions of the geo-atoms found in GIS frameworks (e.g., Kuhn, 2012; Hu, 2018). We then offer a formal treatment of place concepts that is consistent with their analysis in geography and psychology (e.g., Cresswell, 2014; Manzo and Devine-Wright, 2021). We achieve this result by modelling facets, the building blocks of place concepts (e.g., Canter, 2012; Hamzei et al., 2020), as conditions forming conceptual DRSs. We then connect facets to a rich ontology underpinning psychological models (e.g., Scannell and Gifford, 2014; Seamon, 2018) and GIS formal models (e.g., Purves et al., 2019; Hamzei et al., 2020). We thus reconcile linguistic analyses of place names with non-linguistic analyses of place as a general concept with manifold declinations (e.g., “city,” “street,” ‘Sydney” and so on).

Fifth, our model achieves a considerable degree of inter-operativity. It connects linguistics, psychology and GIS via the notion of place, thus reaching a theoretical desideratum (e.g., Ballatore, 2016). This is a consequence of DRT integrating apparently opposite (e.g., cognitivist and formalist) perspectives (cf. Hamm et al., 2006; Pross, 2010; Maier, 2017). The model can also achieve a degree of inter-operativity with cognitivist models of communication (e.g., Heller and Brown-Schmidt, 2023; cf. also Maren et al., 2020; Mocnik, 2022). This is the case because our model also uses reference relations and multi-modal representations as the central tools to represent cognitive con tent. Considerations are however offered in the Supplementary file, due to length requirements.

The goal of this paper has been to propose a linguistic model of place names (e.g., Sydney) integrated with a conceptual model of place (“Sydney”). We have shown that such an account is possible once we offer a formal account of place names’ morpho-syntactic and semantic properties. Our account models place names as a sub-type of NPs involving rich descriptive content that can refer to the subjective or inter-subjective place concepts that individuals entertain in discourse. Therefore, our account can solve the recalcitrant problems illustrated via (1)–(3), (6)–(8), and (16). We obtain this result by introducing DRT as a multi-modal theory of mental content and as a variant of first order logic that is compatible with other frameworks. We then build a full-fledged linguistic and psychological model of place names and place concepts on this framework. Overall, we suggest that formal linguistics can offer key inter-disciplinary insights and tools to platial research (cf. again Tenbrink, 2020). This can be the case, however, insofar as linguistic analyses of place names maintain inter-operativity as a clear-cut research goal.

The account also invites the observation that other problems now become tractable (e.g., formal, cross-linguistic analyses of place names, cf. Köhnlein, 2015), though still outstanding. The approach also obtains at least two interesting secondary results. First, we may have the prolegomena of a novel model of adpositions and their senses, fully compatible with our model of place names (cf. also Ursini, 2020; Ursini and Tse, 2021). Second, the model also appears compatible with recent approaches to place names based on construction grammar (e.g., Leino, 2007, 2011) and cognitive linguistics (e.g., Sjöblom, 2011; Reszegi, 2022). Though these frameworks use different assumptions from the generative and DRT assumptions we implemented in this paper, they also propose rich representational analyses of place names. Third, our model may be consistent with multi-modal models of memory (e.g., Tulving, 2002; Cowan, 2005; Bush et al., 2014). In such models, memorising information about events and individuals occurs by tracing their “place” in the world, via dedicated sets of so-called “place cells.”

Our model thus seems potentially extendable to domains and frameworks that may offer further views on place names and places. However, we leave the exploration of these extensions for further research.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

F-AU collected the literature (psychology, linguistics, GIS), developed the DRT and frames parts of the article, and wrote the article. YZ collected the literature (psychology, linguistics), developed the frames parts of the article (visual notation, diagrams, figures), and edited the article. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.