This paper offers a number of semiological reflections on proper names. It contrasts the Saussurean approach to names with the related socio-onomastic (i.e. Labovian) approach and draws conclusions about their theoretical coherence and empirical viability. It further argues that an ‘informationist’ approach to names, which introduces a conception of the sign compatible with the cognitive sciences, does not advance our understanding of either semiology or onomastics, being fixated on a questionable analogy of the human mind/brain to the computer. Instead, the paper promotes an alternative approach to names based on an integrational semiology as developed by the linguist Roy Harris. The second part of the article revisits a study on colonial New England titles of civility and suggests that sociohistorical onomastics, like socio-onomastics, is founded on a dubious metaphysical assumption concerning the ontology of ‘language’.

Adrian Pablé

In the introduction to the volume Socio-onomastics: The pragmatics of names, the editors Terhi Ainiala & Jan-Ola Östman (2017:1) declare the aim of their book to be the study of ‘names as elements in language’: not only in their function as ‘identificatory or reference devices’, but more importantly in terms of how they accomplish ‘a variety of culturally, socially and interactionally relevant tasks’. The editors (2017:6) define socio-onomastics as the systematic study of ‘the way speakers actually use proper names in their daily activities’, i.e. how names become pragmatically (rather than semantic­ally) meaningful (2017:5). The sociopragmatic approach to proper names, as envisaged by Ainiala & Östman, seeks to integrate the analysis of proper names with analyses of language use in general (2017:6). At the same time, however, they insist that socio-onomastics is ‘really a different perspective on communication’ (2017:16), drawn from onomastic research – an independent disciplinary field, albeit with strong ties to general linguistics.

This paper wishes to introduce such a ‘different perspective on com­munication’, though not exactly in Ainiala & Östman’s sense. It is a perspective that also entails a different conception of the proper name itself. In the first part of the paper I shall invite the reader to engage in a theoretical reflection, which deals with questions in the philosophy of language, in particular the relation between socio-onomastics and semiology. In the second part the focus will be on a sociohistorical onomastics, for which a previous study of mine will serve as the point of departure: the said study reconstructed how the use of early New England titles of civility was structured sociolinguisti­c­ally, as revealed by late seventeenth-century courtroom transcripts (Pablé 2009a). I shall argue that socio-onomastics as an empirical (i.e. sociolinguistic) discipline presupposes a certain language philosophy, which includes a broad conception of ‘linguistic codes’ and a ‘telementational’ understanding of verbal communication (the idea that communication serves the purpose of transferring concepts or ideas, from the speaker’s mind to the hearer’s). Socio­historical onomastics is founded on the same metatheoretical presuppositions as socio-onomastics, though their respective epistemological points of departure are notably different: historical approaches to language cannot rely on the researcher’s linguistic intuitions or on informants’ linguistic experiences. Trust is placed in the researcher as a competent detached analyst and texts are assumed to play the same role as real-life informants: in fact, as Fleisch­man (2000:46) points out, the latter act as ‘native speakers’ of so-called ‘text languages’.

I am not concerned here with typological disputes about what counts as a proper name and what not, but the focus on nominal titles (of address) in the second part of the paper requires touching upon the subject, however briefly. There has been a lack of consensus among onomasticians as to whether forms of address, titles and honorifics are ‘proper names’: given their ‘lexical’ nature and having a general meaning, they are different from proper names identifying individual persons or animals, individualized objects, particular places or events, etc. Their onymic character would seem less disputable when they occur together with a personal name, thereby becoming part of what is commonly referred to as a ‘proper noun phrase’ (Allerton 1996). For example, a chapter on forms of address and titles (Taavitsainen & Jucker 2016) was included in the recent Oxford Handbook of Names and Naming (Hough 2016). As in the case of ethnonyms (Koopman 2016), onomastics struggles with questions concerning disciplinary boundaries, among them the question about the nature of the entity designated by a proper name, i.e. do names identify as their designata entities considered collectively? In translation studies, proper names and titles of address are considered as a common category, namely ‘linguistic realia’ (Rühling 1993), and hence primarily as a ‘language-specific’ phenomenon (Zimmer 1981). For Saussureans, proper names are in fact specific to ‘the language’, while onomastic scholars influenced by structural (i.e. Labovian) sociolinguistics have preferred to assign names to a separate ‘onoma­sticon’, though with strong links to the lexicon. In turn, the pragmatist will remind us that as vocatives, titles of address fulfil the same communicative function as proper names, whether in conjunction with individual names or without them: that is, they address a particular person (or group of people) in a particular situation.

It is important to point out in the context of this contribution that I shall not be concerned with any traditional grammatical or typological distinctions – or with any ontological questions arising out of those very distinctions. In fact, I shall propose that the grammarians’ nominal classification of proper names and common nouns is the result of a particular (ethnocentric) view of language and of the non-linguistic world, and how the two allegedly relate to one another, which the linguist Roy Harris has termed ‘the Myth of Reference’ (Harris 2009a): the belief that words identify entities in the real world in a stable one-to-one relation. Part of Harris’s critique of this myth thus concerns the question of how proper names signify, the answer varying significantly depending on who provides it: structural linguists, philosophers, historians or neuroscientists. However, irrespective of these mutually opposed semantic doctrines, my critical focus here will be on a particular conception of the sign common to all of them, namely a view of the sign as abstract and detachable from the sign-making individuals themselves. The abstract sign is the hallmark of all theoretical orientations labelled ‘segregational’ by Harris (1996). As a counter-perspective to the segregational approach, this paper will introduce Harris’s ‘integrational’ conception of the sign (Harris 1981; 1996; 1998) and present a critical account of its empirical consequences. In fact, I shall argue that from an integrational perspective, there is no ‘socio-onomastics’, just as there is no ‘sociolinguistics’. Integrational linguistics only acknowledges the first-­person perspective and declares the third-person perspective – the foundation of empirical linguistics – to be a theoretical fiction. This radical position follows directly from accepting an integrational conception of the sign as ‘indeterminate in both form and meaning’ (Harris & Hutton 2007).

Socio-onomastics orients towards a Saussurean-inspired sociolinguistics and as such is based on three presuppositions, which Figueroa (1994) defines as being of a ‘metatheoretical’ nature, namely that (i) proper names are abstract linguistic signs, which is why they are shared by the members of a linguistic community, (ii) the use of proper names correlates with socio­logical variables of various kinds, which is why some names are used by some speakers and not by others, and (iii) the members of the community share the same evaluative norms concerning the use of these names. Socio­historical onomastics projects these presuppositions onto the past, based on the assumption that past linguistic usage is subject to the same constraints as present-day usage, as postulated by the ‘uniformitarian principle’ (Labov 1972). I consider sociohistorical onomastics (or historical socio-onomastics) to be a sub-discipline of sociohistorical linguistics (or historical sociolinguistics) – the sociolinguistic study of now extinct varieties of a language (Romaine 1982; Nevalainen & Raumolin-Brunberg 2017) – applied to the special case of names.

Saussure’s posthumously published Course in General Linguistics (1983) was a revolutionary text in many ways, not least because of the author’s radical conception of the linguistic sign as a purely mental unit consisting of two inseparable parts: a signifier (sound pattern) and a signified (mental concept), whose value is determined holistically in relation to the other signs of the language system. Saussure considered proper names to be part of this system (‘la langue’), i.e. he considered them to be linguistic signs, just like all other signs of the language. In fact, if they are not ‘linguistic’, it is not quite clear what kind of signs they would be. In other words, from a Saussurean perspective, proper names must be part of the linguistic system, i.e. they are bipartite signs (signifier–signified) like any other linguistic signs, regardless of how onymic meaning allegedly differs, referentially speaking, from appellative meaning. As part of a bold theoretical move, Saussure in fact rejected the nomenclaturist view of languages. By severing the linguistic sign from the non-linguistic world, Saussure ingeniously managed to explain how different people across different situations ‘knew’ the same linguistic signs and interpreted them identically. Meaning, for the Saussurean, is a matter of how a sign relates to the other signs within the system, rather than of linking a sign to something non-linguistic. From the point of view of Saussurean semiology, form and meaning are thus inseparable, which contrasts with what Harris (1996) has called the ‘surrogational’ perspective, whereby the sign is only the form (what in lay linguistic usage is commonly conceived as ‘the word’) and its meaning is a material thing in the external world or an idea formed in an individual’s mind. Thus Saussure refuted the widely accepted surrogational nature of signs, i.e. the notion that a sign is something that stands for (is a surrogate for) something other than itself. At the same time Saussure did not want to deny that speakers do indeed ask questions like ‘Of what is this a sign?’ and that they do explicitly link words to things. By assigning such instances to the level of parole, i.e. actual linguistic usage, however, Saussure was able to safeguard the notion of ‘a language’ as something purely psychological, collective and self-contained. According to Saussure, the linguistic sign manifests itself in parole as a ‘word’ (through the activities of speaking and listening). What is uttered or heard, however different, must be relatable to the same linguistic abstraction in langue if speaker A and hearer B are to encode and decode the same sign. In turn, the structural sociolinguist William Labov, a semiologist and Saussurean (Figueroa 1994:75), introduced the notion of the variable so as to free Saussure’s abstract linguistic sign from its confinement to langue. The abstraction of interest in Labov’s theory is the variable underlying the different – semantically identical – variants used in spoken interactions with varying degrees of frequency depending on both extralinguistic factors (speakers’ sociological profiles, degree of formality) and intralinguistic factors (e.g. phonotactic environment, syntactic environ­ment). Parole thus became the true protagonist of post-Saussurean sociolinguistics, though the aim of describing the language system remained, however with the focus shifting to the various ‘subsystems’.

Socio-onomastics has traditionally adopted a Labovian macrosocial approach to society (Akselberg 2012), in which language use is taken to reflect the social stratification characteristic of a speech community (Cameron 1990). Onomasticians of a structuralist mindset (e.g. Walther 1971; Nicolaisen 1995) proposed that proper names exist as signs in a separate collective system, named the onomasticon, which retains close ties with another collective system, the lexicon. To them, Saussure’s rigid notion of langue as a fixed code containing all linguistic signs failed to take into account that proper names are different from other words: they are not shared by members of the community in the same way that members share the lexicon. As Walther (1971:54) puts it: ‘As regards name usage, the community is split up into infinitely many groups of name users, which differ in both kind and number with every name’ [Beim Namengebrauch ist die Gemeinschaft [...] in unendlich viele, bei jedem Namen verschiedenartige und verschieden grosse Namenbenutzergruppen aufgesplittert]. Nicolaisen (1995:389) echoes Walther’s proviso, adding ‘competence’ to ‘use’: ‘[A]part from potential dialects and idiolects, the user of such onomastica will display different levels of competence in their onomastic range, their precision of usage, and in the act of naming’. Arguably, this is not only a watered-down version of Saussurean linguistic theorizing, but also a view hardly compatible with structural sociolinguistics, where the communal fixed code turns out to be made up not only of ‘social dialects’ and ‘style lects’, but of an infinite number of ‘mini-codes’ (shared by a minimum of two speakers) and ‘idiolects’ (one-person codes). The notion of ‘personal fixed codes’, as well as the notion of ‘mini-codes’, must be rejected by the Saussurean and the Labovian alike, for in such a scenario no shared psycho­logical reality can be established as regards speaker A and hearer B (two members of the same linguistic community) and thus mutual understanding between them can no longer be (theoretically) guaranteed. Ultimately it would not even be clear where and how to draw the line between A’s idiolect and B’s idiolect: do they belong to the same ‘language’ or to two different ‘languages’? Or does the notion of ‘idiolect’ render the notion of ‘a language’ redundant? As regards socio-onomastics, it could be argued that these questions do not matter, the onomasticon not being part of ‘the language’ in any strict sense. Such a view, however, presupposes that it is always clear which sign belongs to which abstract system (onomasticon vs lexicon). As the aforementioned case of the ethnonym shows, however, there is no such clear position in theoretical onomastics. Or take the example of the proper noun phrase ‘Monsieur Buche’: are there two systems involved or only one, two intrapersonal idiolects (the idiolectal lexicon and the idiolectal onomasticon) or rather one? Explaining how it is possible for A and B to verbally communicate, however, is precisely what a linguistic theory interested in parole should, arguably, accomplish. Scholars like Walther and Nicolaisen disavowed Saussurean idealism, preferring to heed their own linguistic intuition, which suggested that no two people know the same number of proper names or possess the same knowledge of how to use (i.e. apply) those names. Little did it matter whether the ‘multiple fixed-code’ theory they adopted was theoretically coherent at all. For them, there was no contradiction in saying that the members of a linguistic community, while being speakers of the same language, do not share the same onomasticon. However, an analogous argument could be made as far as the lexicon is concerned: as Harris (2008) noted in his critique of Saussure’s synchronic system known to all the members of the linguistic community: is there a real person who could claim to know all the words of the language at any given point in time?

As Saussure obviously realized, linguistic knowledge does not only encompass words ‘in the language’ (structurally speaking): at the level of parole, speakers would also encounter words that are not part ‘of the language’. What kind of knowledge does knowing such words constitute? For the Saussurean, these cases do not involve langue in any way. On hearing speaker A utter ‘moshpit’, hearer B treats it as a word of the language (because uttered by A, a speaker of the same language). However, at quite a different level of reality, no mental signified and signifier are generated in B’s brain on hearing ‘moshpit’. Instead B might try to analyse the word semantically (e.g. ‘mosh’ + ‘pit’), i.e. B treats it like an individually existing, semantically motivated word. In this scenario, the Saussurean would argue, moshpit is not structurally meaningful: it is a word but not a linguistic sign (though it might be one sometime in the future). Nonetheless, moshpit may already be recorded in the Oxford English Dictionary. However, the problem with dictionaries, from a Saussurean standpoint, is that they do not distinguish strictly between langue and parole. They treat word forms as separable from word meanings. A Saussurean linguist would thus disagree with the statement by the variationist sociolinguists Nevalainen & Raumolin-Brunberg (2017:2) that ‘it is only when an innovation has been adopted by more than one speaker that we can talk about change in the linguistic system’. From a variationist (i.e. Labovian) point of view, in fact, moshpit could be a lexical innovation which is synchronically competing with an older, still predominant form. However, both A and B would have to be aware of the new form (and its social evaluation in the community) if uses of the two variants are to constitute ‘social facts’ (in which case they are of interest to the structural sociolinguist).

In conclusion we could thus say the following: the structuralist theorist faces a dilemma that involves deciding between Saussure’s theoretically coherent yet empirically impracticable paradigm (insisting on the primacy of langue and taking it as one’s starting point) and Labov’s empirically viable yet theoretically incoherent paradigm (insisting that there are multiple – semantically invariable – fixed codes governing speech).

Both the Saussurean and the post-Saussurean take on proper names are thus riddled with theoretical muddles. Roy Harris is the linguist who has identi­fied the muddle and described it most clearly (e.g. Harris 1996; 2003). As Harris (2003:70) puts it, ‘the semantics of proper names is a notorious problem area for structuralist accounts of language’, adding that ‘from Saussure onwards, no theorist of structuralism has ever given a clear account of the status of proper names, or explained how we decide which names merit inclusion in a language’s inventory of linguistic signs’. Harris goes on to state:

This passage is in need of further clarification. Harris is referring here to Saussure’s crucial concept of ‘synchrony’: if eighteenth-century English and twentieth-century English are two different ‘languages’, i.e. two different linguistic systems, we have to accommodate the fact that the latter contains linguistic elements that were not part of the former (and vice versa). If we were to assume (wrongly, from a Saussurean theoretical point of view) that proper names known collectively in both eighteenth-century and twentieth-­century English (e.g. Cicero) meant the same (were the same linguistic signs), then any names used in twentieth-century English but not in eighteenth-century English (e.g. Charles Dickens) could not have been part of langue, i.e. the names were used by speakers here and there (at the level of parole), but had not (yet) been accepted at the abstract level of the collectivity. However, this is an implausible stance to take: Charles Dickens, by Saussurean (i.e. purely theoretical) standards, must have been a proper name structurally accepted into twentieth-century English, i.e. a fact of langue (after all, the whole nation had read Dickens). Admitting a new name like Charles Dickens into the language structure is not simply a matter of addition: in fact, said structural expansion had repercussions on other linguistic signs belonging to the same system, i.e. it changed the meaning of other names, e.g. William Shakespeare. From a Saussurean perspective, the main question to ask as regards a proper name is ‘What is the system-internal value of that name?’ – not ‘Who does the name refer to?’ That is why the Saussurean will fundamentally disagree with the surrogationalist, who believes names to be stable identificatory devices: for the latter, in fact, the name WilliamShakespeare identified the very same English playwright in eighteenth-century English as it did in twentieth-century English.

The conundrum that Saussurean linguistic holism generates, if taken seriously, derives from the common (lay) surrogational assumption that a proper name identifies an individual person or an individual place, whose material existence is not a matter of ‘language’ (as a psychological reality). The surrogational thesis of how proper names get their meaning remains plausible until the communication theorist starts wondering how members of the same speech community manage to apply a proper name to the same referent on all occasions. In fact, this is what we have to assume if we take proper names to be identificatory devices for non-linguistic entities that exist prior to and independently of speakers applying a linguistic label. How do speakers of the same linguistic community achieve such communicational stability? One way of dealing with that question is to take a sceptical position on communication, as philosopher John Locke did. Locke still subscribed to a surrogationalist semantic doctrine, one however in which words were defined psychocentrically rather than reocentrically. According to the psychocentric thesis, words are surrogates for individuals’ private ideas, whereas the reocentric thesis treats words as surrogates for things in the real world (Harris 2005). In Locke’s semiotics, knowledge is thus always knowledge of private ideas and not knowledge of things, which cannot guarantee successful interpersonal communication in the absence of the things themselves. The Saussurean definition of ‘a language’ as a communal fixed code dissolves Locke’s argument that words are imperfect instruments for communication (at least on a theoretical level), as ideas are declared to be collectively shared rather than private. Moreover, there is no attempt on the Saussurean linguist’s part to link words with things. The following example illustrates how the structuralist stance on names cannot possibly be reconciled with a reocentric surrogational stance: there are speakers of one language who identify a place called Bellenz and there are speakers of another who identify a place called Bellinzona, and according to the structuralist these are two separate and unrelated facts of language. The observation that Bellenz and Bellinzona actually identify one and the same location is not relevant from a structuralist point of view, as that observation is made from outside a language system and hence does not align with anybody’s language qua psychological reality. The remark might catch the structural sociolinguist’s attention, however, if it turns out that Bellenz and Bellinzona are signs belonging to the same language (or ‘variety of language’, as the sociolinguist prefers). Now we are dealing with ‘two different ways of saying (referring to) the same thing’: a case of socio­linguistic – rather than linguistic – variation. By positing that languages supply both form and meaning – and not only forms – the structuralist has an advantage over the surrogationalist when it comes to explaining how A and B manage to communicate with one another. However, when pressed by the reocentric surrogationalist about the material world and the various things that make up that world, the Saussurean linguist remains silent, for labelling ‘things’ in the external world has, strictly speaking, nothing to do with linguistic knowledge: it is not a ‘collective’ phenomenon in the structuralist sense and hence cannot occupy a place in Saussure’s linguistics of langue. That is also why it does not make sense, from a strictly Saussurean point of view, to speak of science having ‘its own language’ (Pablé 2020). But if there is no ‘language of science’, how can science, in its orthodox (reocentric surrogationalist) conception, exist at all? It seems that theories of the linguistic sign developed within modern linguistics stand outside the reach of the empirical sciences, which have nothing to say about postulated collective linguistic systems. The problem, obviously, is that there are no languageless sciences (Harris 2005).

A noteworthy critique of the proper name conceived as a bilateral sign, and of the Saussurean conception of signhood in general, was presented by Brendler (2005), who boldly proclaimed the ‘death of the bilateral sign’ based on recent developments in cognitive science, as propounded by the linguist Ernst Hansack (2000). Brendler’s conception of signhood is marked by unilaterality: ‘A name is a word’ [Ein Name ist ein Wort] (2005:109). However, Brendler is not only a declared anti-Saussurean, but also an anti-surrogationalist (in the sense that he does not think names ‘stand for’ material objects): proper names, he tells the reader, have an ‘information-­relatedness’ [Informationsbezug], not an ‘object-relatedness’ [Objekts­bezug] (2005:106). As a consequence of his rejectionist stance against the Saussurean sign conception, Brendler does not subscribe to a ‘telementational’ model of communication: language, as Brendler (2005:102) tells the reader, does not ‘transport’ information, but merely transports ‘access indices to information’ [Zugriffindizes auf Informationen] and instructions on how to connect them. Thus communication, according to Brendler (2005:102), is the exchange of those ‘indices’ that allow speaker and hearer to access quantities of information [Informationsmengen]. The data (or contents) are stored in the individuals’ brains and do not leave them. In this model, mutual understanding is not guaranteed, pace Saussure, but depends on whether the storage addresses [Speicheradressen] on which the sender is relying are equally available to the receiver. On the receiver’s end, abstracted images saved in the memory addresses as prototypes (what Brendler calls ‘Bedeut­ungen’) under the corresponding sound pattern [Wortlaut] are projected into his/her conscious awareness in the form of expressions [Begriffe]. The latter are the concrete, contextualized realizations occurring in the speech acts. As Brendler (2005:103) declares, the sole purpose of the sign is ‘the indexing of a quantity of information’ [die Indizierung einer Informations­menge]. He adds to this: ‘A linguistic sign indexes a prototype meaning. However, under no circumstances can we say that the linguistic sign has, carries, transports or contains meaning’ [Ein sprachliches Zeichen indiziert eine Bedeutung im Gehirn. Keinesfalls kann jedoch gesagt werden: Es hat, trägt, transportiert oder enthält eine Bedeutung] (2005:103, italics added), which is why Brendler considers the Saussurean bilateral sign model to be outdated. He also advises against formulating the sign conception derived from contemporary cognitive science bilaterally (as some name scholars might be tempted to do), hence unifying the signifier with the conscious projection. This, however, would imply precisely that one property of the sign is to ‘transfer’ meaning (the telementational view of communication). For Brendler, both proper names and appellatives are thus such ‘access indices’, the difference being that the sign indexes ‘a meaning relative to a class of single elements’ [eine (Einelementklassen-)Bedeutung] and ‘a meaning relative to a class of plural elements’ [eine (Mehrelementklassen-)Bedeutung] (2005:104–105), respectively, but both index a class (since a single element is said to already constitute a class). In this way, Brendler argues, there is no need to distinguish onymic objects from appellative objects: something gets named onymically not because nature constitutes it as a unique ‘thing’, but because we make this distinction linguistically.

It is evident that Brendler, in announcing the death of the proper name as a bilateral sign, was not too worried about the future of socio-onomastics. His proposal (Brendler 2005:100) for a conception of the sign as unilateral still presupposes a communal fixed code in place – otherwise his distinction between ‘onomastics’ as the language-specific study of names, as opposed to ‘onomastic theory’, which explores ‘the essence and nature of the name’ [das Wesen des Namens] in a more general spirit, would not make sense. Thus members of the linguistic community still share the linguistic signs, now conceived as word forms (or signifiers). Meaning (what Saussure called the signified) is no longer part of the abstract sign, but may vary between speakers, and is neurologically much more complex than a mental concept triggered directly in the brain. As Brendler makes clear, the mental representations, or images, that are saved as prototypes in the individual’s brain [Bedeutungen] need not correspond between one speaker and another, which is why variant forms can no longer be presupposed to be ‘alternative ways of saying the same thing’: langue, in Brendler’s proposal, does not guarantee semantic stability within the linguistic community. Brendler rejects the ‘contractualist’ notion of a language already in place and available as the same language across generations of speakers, as a result of joining the language contract as (involuntary) signatories. This rejection, nota bene, is ‘scientifically’ grounded. Questions about semiology, according to Brendler, can now be answered by the brain sciences. About the history of semiology/semiotics, Brendler (2005:99) concludes that our theo­ries have been based on hypotheses and models rather than scientifically derived knowledge, and hence disagrees with Rudi Keller’s statement that everything about the linguistic sign, which Keller views as a language-­philosophical issue, has been said in the past two thousand years and that nothing new is to be discovered (Keller 1995). But what exactly is ‘scientific’ about Brendler’s (or Hansack’s) theory of the sign? Brendler’s alternative to Saussurean semiology – and indirectly to Saussure’s model of communication – is still committed to what Harris (1996) calls ‘segregationism’ (like Saussure’s theory of langue), i.e. the assumption that ‘language’ is something detachable from the rest of human activities. The neuroscientific approach to language makes grand claims that sound all the more impressive through the use of terminology borrowed from the computer sciences: ‘quantities of information’, ‘access indices’, ‘memory addresses’, ‘stored data’, ‘linking templates’ and the like. As regards the purpose of language, Brendler authoritatively declares: ‘The real purpose of language is to program the brain’ [Der eigentliche Zweck von Sprache ist es, das Gehirn zu programmieren] (2005:102). In the end, the reader might wonder how exactly cognitive onomastics is supposed to be any more ‘scientific’ than Saussure’s postulation of the ‘synchronic method’, i.e. taking a perspective that corresponds to a collective psychological reality. Saussure, too, claimed that his linguistics was ‘science’, i.e. it belonged to the social psychological sciences, but he did not claim that there was a scientific method able to validate his synchronic linguistics: ‘The method is itself the instrument by which the linguistic object is created’ (Harris 2005:93). The method consists in adopting a psychological perspective, namely the language users’ internalized knowledge. In turn, with the advent of the neurosciences in the late twentieth century, the conviction grew that ‘language’ can be shown to be ‘there’ in the human brain (at the same time it was declared that the mind, a Cartesian dualist concept, did not exist). When Brendler (2005:102) cites Hansack’s (2000:187) definition of language as ‘an inventory of signs with linking templates, which is employed in brains as a system of notation to manipulate data’ [ein Zeicheninventar mit Verknüpfungsschablonen, die als Notationssystem für die Datenmanipulation in Gehirnen eingesetzt wird], the reader must be awestricken by the rhetoric: surely, disagreeing with this definition would be irrational given that it is presented as a ‘scientific’ one. After all, this is no longer mere language-philosophical speculation presented as reality. Every word in Hansack’s definition is reocentrically defined, pace Brendler’s anti-surrogational stance, i.e. the words ‘stand for’ what they are. They are not supposed to be interpreted metaphorically – in fact, as Brendler (2005:103) argues: ‘Metaphors do not do justice to the requirement of precise phrasing in linguistic theory’ [Metaphern [entsprechen nicht] dem Erfordernis einer präzisen Ausdrucksweise in der Sprachtheorie].

Segregational theories of language require abstract signs, be they unilateral or bilateral. This is the logical consequence of focusing on ‘language’ as something separate and separable from the rest of (time-bound) human activities. Brendler is right when insisting that we have not explored the (linguistic) sign sufficiently, but he fails to see that the ‘sciences’ have nothing to say about the sign. In turn I will argue that the most reliable source of knowledge when it comes to theorizing the sign is our own lay expertise as daily communicators. Signs, according to this view, can only have a personal epistemological source, which in turn (perhaps somewhat surprisingly) allows for a general theory of the sign, but one which does not rely on scientific (i.e. reocentric surrogational) definitions.

The most important theorist of a non-segregational approach to language was the Oxford linguist Roy Harris (Harris 1996; 1998), who developed an ‘integrational linguistics’ in the late 1970s. The integrational linguist treats signs as ‘made’ (created ex nihilo) and rejects the metaphysical assumption that they exist as abstractions. According to Harris, signs are spontaneous constructions of the human mind that serve communicational functions here and now. They do not, however, exist as mentally isolated phenomena, as signs do not serve ‘language’ in any exclusive way: they serve communication, which consists in the continuous integration of cotemporal activities (some linguistic, some non-linguistic). Communication, from an integrational point of view, is the only human reality, and ‘language’ is but one mode that is integrated with other human activities. Thus from an integrational point of view, ‘language’ (or ‘languaging’ for those who prefer this recent sociolinguistic coinage) is a first-order activity, while ‘languages’ (the objects of study of linguistics) are ‘second-order concepts’ (Love 2017) – the result of reflecting upon, and abstracting from, the continuous sign-making processes that characterize our integrated activities. Integrationists acknowledge first-order activities (languaging) as the only communicational reality. By adopting an integrational semiology, which treats signs as concomitant products of communication rather than its prerequisites, the language-philosophical questions that have preoccupied the minds of philosophers regarding the difference between proper name and common noun are relegated to second-order reflections on decontextualized (i.e. abstract) linguistic signs. The question that the integrationist asks about a sign (and thus about its identity and meaning) is: ‘What integrational purpose does it serve?’ The answer to this question will obviously vary with every situation in which signs are created. Drawing on our personal linguistic experience we could think of many different (macrosocially defined) scenarios in which a ‘proper name’ integrates the ongoing activities (between people) differently. For example, Shakespeare could be a code name that integrates the activities ensuing from someone’s wish to enter the premises of a secret society (say, knocking on the door three times followed by the utterance ‘Shakespeare’) and the resolution by another person to allow that person to enter the premises. Any word, e.g. rhubarb, could be substituted for Shakespeare as a code name, for what matters in the first place is its integrational function with respect to the aforementioned activities. The name could thus be either ‘onymic’ or ‘appellative’ in nature. This is not to claim that Shakespeare and rhubarb are intersubstitutable words in any general sense: we would not expect someone to tell us that ‘rhubarb was born in Stratford-upon-Avon in 1564’, even though rhubarb could indeed be a nickname used by some people to refer to the English playwright. The point that the integrationist wishes to make here is that the Saussurean type of fixed code is not helpful in accounting for the use of Shakespeare as a code name.

The socio-onomastician, in turn, would argue that the code name Shakespeare is shared by a small group of people who have certain macro­socially defined interests in common. Following this idea of the Labovian sociolect, these ‘mini-codes’ can be multiplied by the thousands across any language community. The result of such a scenario amounts to the kind of sociolinguistic situation depicted by the name scholars Walther and Nicolaisen: for every name there are indefinite numbers of groups of name-users (and name-knowers), not counting those names that perhaps only single individuals passively know of, or actively use. The Saussurean fixed code is limited to the speakers born into the native speech community – it is not something that outsiders can acquire (however proficient they become in the language), whereas sociolects admit in principle of new speakers (whether voluntarily or involuntarily). New members of the secret society, for example, will be introduced to the code name Shakespeare, or spies may learn about it without the members of the society knowing that others have acquired the code. Fixed-code theorists of different orientations will in fact disagree about how to define a linguistic code, i.e. it will depend on the theoretical stance one wishes to take: the Saussurean linguist, for one, will not be convinced that ‘open fixed codes’ make any sense for language-­theoretical purposes.

By rejecting the notion of an abstract sign altogether, the integrationist will theoretically accommodate the example of Shakespeare as a code name based on the insight that knowing the meaning of a word is to ‘know what to do with it’ (Harris 1998:63): in other words, what matters from an integrational semiological perspective is to be able to explain how a sign integrates one activity (or kind of activity) with another. However, only the sign-makers themselves can provide such an explanation. If the explanation is offered by a disinterested observer or analyst, it is important to keep in mind that the explanation is the analyst’s: he/she will have drawn on personal linguistic experience, remembered or anticipated, in order to say something meaningful about a number of activities construed as being connected for the interactants under observation. The meaningfulness of the analysis consists in (the reader) being able to relate to the analyst’s explanation. It is not, however, an explanation of what the sign meant for the sign-maker. There can be no such insight.

Integrationally speaking, names are contextualized signs like any other signs, i.e. they are made by someone for a communicational purpose. Signs can be recontextualized (repeated, taken up, reconstructed, reanalyzed, etc.) at any time by anybody, but these signs are always new signs for whoever integrates them here and now (Duncker 2019). Integrationists do not distinguish in any strict sense between ‘speaker’ and ‘hearer’ to explain communication processes. Instead they speak of ‘sign-makers’. The unique, integrated sign has an exclusively temporal existence: a sign only exists for as long as it integrates activities of various kinds, without which the activities, taken separately, would remain unintegrated (merely knocking on the door three times will not grant you access to the secret society’s premises). So what exactly is the person intending to be granted access integrating? He/she is integrating the present situation with past experience and future experience, i.e. the here-and-now only makes sense to sign-makers because they have a personal immediate past and future. Once the secret society is dissolved and there are no active members still alive, the code name loses its (potential) integrational function (because presumably there is no one around any more who would assign that particular semiological function to the name Shakespeare), although it may be reactivated by someone in the future who, say, discovers an extant manuscript describing the secret language of that society. In principle, anybody may encounter any name at any time. Whether it is a meaningful sign for that particular person under the given circumstances, i.e. whether the sign-maker knows what to do with it, is not predictable in any sociolinguistic (Labovian) sense: what a sign ‘means’ is not an empirical question.

As I have argued in a previous integrationist contribution to socio-­onomastics (Pablé 2009b), knowing the name of something is not a ‘competence’ that exists in a communicational vacuum: knowledge always depends on the activity/activities involved. What those activities are is not a scientific question: for example, suggesting that the reader of this article must be currently engaged in the activity of ‘reading’ in order to make any sense of it is a lay insight which will likely not be regarded as controversial. Being a lay linguistic concept, the notion of ‘activity’ is an open one, i.e. it will not always be obvious that something is recognized as an activity by everybody concerned, or that activities are ‘labelled’ in the same way. In my own fieldwork, the activity involved a non-local person (myself) seeking to get the locals (people I randomly stopped in the streets) to identify a landmark and provide directions to it. This activity was itself comprised of an indefinite number of activities on both sides. The task my informants were confronted with involved matching a name of little diffusion (which I presented them with) with some object in the locality (which I pretended was unknown to me). The activities that constituted these various encounters were of a very different nature compared to the activity of collecting data as part of a research project in socio-onomastics. Only the latter kind of activity attempts to find systematic patterns by treating knowledge as decontextualized (relating to types of speakers rather than individuals), i.e. by treating the signs as if they could be detached from the ongoing activities in which they occur, whereas the former kind of activity allows the researcher to experience a particular kind of sign-making (identifying a landmark) as embedded in an ongoing activity. The situation created for my fieldwork was a familiar one, involving a non-local (myself) who was supposed to meet a local (imaginary) friend, who had given me an appointment at a particular place but who could presently not be reached on his phone for further directions. Stopping someone in the street to ask for the whereabouts of a street or building is a very common thing to do (and as such hardly surprising). However, the story that my informants were confronted with put them in a rather unusual position: in fact, although they were supposed to be the ones with local knowledge, they also had to take into account the expertise of my (physically non-present) local friend when venturing their guesses where to direct me to. After all, this friend had told me to meet him at a place identified by a name, and thus supposedly a real place. One way to interpret these encounters is to say that the notion of the ‘sociolect’ corresponds to no reality once we cease to postulate that words are somehow ‘shared’ (as abstractions) between specific group members. What transpired from the various encounters was rather that names are made sense of, or contextualized, by the individual sign-makers in relation to the very specific situations they find themselves in. And although I told my informants more or less the same story, every situation was indeed a completely different one. This is not per se a deep insight – it is part of our lay experience – but it is so obvious that linguistic theorists have never attributed much importance to it. I am not claiming that there are no ‘patterns’ one could see emerging on analysing the transcriptions we made of the encounters after they had taken place, but ‘seeing’ them is itself a semio­logical process: it presupposes a certain communicational purpose behind it, as the very activity of doing so is macrosocially, i.e. institutionally, conditioned.

The integrationist worldview is human-centred: human beings exist separa­tely from the external world which they inhabit (a world made up of all sorts of other living beings and non-living entities). In other words, the external world does not exist, humanly speaking, independently of our continuous sign-making activities. Thus scientists reocentrically label the external world, but by doing so they have already ‘humanized’ it, i.e. integrated it by means of their sensory faculties (and with the aid of artificial extensions). The thing we call a flower is not a ‘flower’ to the bee: it is whatever it is to the bee. There are no neutral, contextless linguistic labels because there are no species-neutral interpretations of anything. Modern science is a complex of connected institutionalized human activities integrated by an overarching human view of the natural world, which relies on a reocentric surrogational semantic doctrine of how language relates to the non-linguistic world (Harris 2005). If there are intelligent life forms in the universe, we may be confident that they would not have ‘science’ as we – languaging beings – know it. All human beings share the same ‘communicational infrastructure’ (Harris 1996), and all human beings are constrained in their communication by factors of three kinds: biomechanical, circumstantial and macrosocial (Harris 1998). The macrosocial constraints vary significantly across cultures, which is why different peoples possess very different ‘cosmovisions’ and ways of thinking about what ‘language’ is, and whether it is a uniquely human prerogative (Pennycook & Makoni 2019). At the same time the integrationist argues that no two individuals ever make the same signs (whether within the same culture or across cultures): the reason being that the three aforementioned factors operate interdependently in every moment of our lives, though we often focus on a single factor to explain a communicational constraint. For instance, I cannot tell someone else the name of a certain place if (i) I do not speak the other person’s language (a macrosocial condition), any more than I can show him/her where the place bearing that name is if (ii) the other person is presently not visible to me (a circumstantial condition), or if (iii) he/she is blind (a biomechanical condition). The signs, for the integrationist, are ‘radically indeterminate’ (Harris & Hutton 2007), which is another way of saying that there is no third-person perspective on signs. The sign means something to someone in a given circumstance. However, the sign, for the integrationist, is not the form, whose meaning is indeterminate. ‘Radical indeterminacy’ concerns meaning as much as form. As Orman (2017) argues, integrational linguistics is the only ‘linguistics’ to subscribe to the thesis of indeterminacy of form. In fact, to say that signs are ‘made’ (in the integrationist sense) is to claim that there is no underlying word form that the sign-maker draws on. Any word uttered or any word encountered is so ‘for the first time’ because every situation is unique.

If something is meaningful to me, I do not assume that someone else present interprets this ‘something’ in the same way. In fact, what I do assume is that he/she interprets it in their own way. We may possess the same (already integrated) sensoria, but that does not mean that any two individuals experience anything identically, even if one were to stand where the other is standing (thus assuming the ‘same’ point of view). If I (as a completely detached observer) were to observe Mr Ulvin open the door right after Mr Stevens rang the doorbell, I would still not know what the sign meant for Mr Ulvin. Why? Because I am not Mr Ulvin. Perhaps Mr Ulvin made no such sign as I suppose he did – even if, from an outsider’s perspective, he behaved exactly as one would expect. Perhaps, in fact, Mr Ulvin is deaf and just happened to open the door to step out of the house right after Mr Stevens rang the doorbell. Conversely, if he did not open the door after Mr Stevens rang the bell, this does not mean that Mr Ulvin did not make the expected sign: i.e. construe one activity as an ‘initiative’ requiring an ‘integrated sequel’ (Harris 1996:63). Perhaps he heard the ringing but decided to ignore the fact that someone was standing at his door, for whatever reasons (maybe he suspected it to be Mr Stevens, from whom he had borrowed money). We will not know the ‘facts’ of the matter unless we ask Mr Ulvin himself, but him telling us is a very different integrational activity than him doing – or not doing – what he did.

If signs are not prerequisites for communication, as the integrationist argues, there is no communicational stability as Aristotle conceived it. Aristotle thought that reality is the same for everyone, just as every speaker of Greek shared the words used to refer to the things in that shared reality. According to an Aristotelian (reocentric) semantics, words ‘stand for’ the static (photographic) impressions gained from perceiving the things that make up reality (Harris 2005). A Harrisian semiology cannot reach any such facile conclusions concerning human integrated experience. Instead Harris (2006) argues that our conviction that there are natural (and man-made) units (classes of things), which can be considered as particular things, is the result of species-specific integrated activities on a biomechanical level. It is because we are biomechanically made the way we are that the apple appears to us as a unit (we can pick it from the tree, hold it in our hands and put it into a basket). As Harris further argues, for the mockingbird sitting on the tree, the apple is something quite different. The integrationist treats human life as a continuum of communication processes to which contextualized signs of various kinds contribute by integrating time-bound activities with one another, whereby the material entities that human activities involve only exist (a semiologically bound existence) to serve those activities. The integrationist worldview is thus anti-reocentric (words are not identificatory labels attached to independently existing things), but neither does it endorse any form of linguistic determinism, given its rejection of ‘languages’ as abstract sets of conventionalized signs.

Suppose that a linguistic theorist were to accept the Harrisian position on what constitutes a sign. Being lay-oriented, it is a position one might be willing to endorse even though only few theorists actually do. When it comes to saying something about the (distant) past, however, it would be tempting for the very same theorist to conclude that only by adopting a segregational semiology can one learn anything about the linguistic past, however partial the results may be. In other words, if linguistic signs are not treated as having an underlying abstraction for the corpus-based analysis of writings from the past, what is there to say of sociolinguistic relevance about them? Studies in sociohistorical linguistics (e.g. Nevalainen & Raumolin-Brunberg 2017) allegedly show that we can explain sociolinguistically motivated variation and change by systematically correlating linguistic variables with sociological variables, though Bergs (2012) rightly warns of the danger of committing anachronistic errors in our interpretations of sociological variables.

Sociohistorical linguistics has to rely on written forms in its reconstruction of oral forms. This, it would seem, comes close to a category mistake in Saussurean thinking, since the Course tells us that the linguistic sign (as a psychological phenomenon) concerns spoken linguistic communication only. However, in a rather surprising move, Saussure (1983:15) claimed that ‘writing can fix [linguistic signs] in conventional images’, whereby every sound pattern (signifier) ‘can be represented by one constant visual image’. Evidently, Saussure wanted to safeguard the idea that it is possible to provide accurate descriptions of dead languages (that is, of their linguistic structures) based on written attestations and reconstructions. From a socio­linguistic point of view, the more important question is to what extent the scribes or writers, in reproducing spoken language, have been influenced by their levels of literacy. ​Therefore, when a scribe rendered the spoken discourse of others in written form, how much of what we read is the speakers’ and how much is due to scribal intervention? These are certainly interesting questions for philologists, but philology cannot solve language-­theoretical problems. The theoretical muddle that historical socio­linguistics creates for itself has to do with its conception of ‘a language’, which is broadly Saussurean, at the same time as it studies speech ​variation exclusively through written forms. In fact, the formality spectrum in historical sociolinguistics comprises both writing assessed as closer to written language and that considered closer to spoken language.

As argued before, the Saussurean and Labovian accounts are incompatible, while paradoxically the latter would not make sense without the former. Thus a Labovian is at the same time a Saussurean (Figueroa 1994), whereas a Labovian cannot be a Harrisian (Harris 1998:126–129). A Harrisian, in turn, has no grounds to turn Labovian, given their incompatible views on what constitutes the linguistic past.

In a sociohistorical study on early New England titles of civility (Pablé 2009a), I committed the same mistake as the aforementioned (imagined) theorist. Although I was already an integrationist, I still wanted to hold on to the idea that it must be possible to say something meaningful about language usage in the past based on in-depth historical research. I started from the (Labovian) assumption that I first needed to understand the sociological reality of a speech community in order to study how language use reflected the social stratification of the community. My earlier interest in (post)colonial New England and its dialectal peculiarities prompted me to reconstruct the sociolinguistics and pragmatics of a feature typically associated with Puritan New England: the (now obsolete) titles of respect Goodman and Goodwife (and the contracted form of the latter, Goody). The assumption on my part was that, though these titles had originated in England, their sociolinguistic meaning in the seventeenth- and eighteenth-century Puritan colonies differed from that found in the motherland. For my corpus I relied on metalinguistic and metapragmatic commentaries by contemporary New Englanders and a collection of court records known as the Salem Witchcraft Papers (Boyer & Nissenbaum 1977), featuring transcriptions of trials and hearings held at Salem Village and neighbouring villages in the early 1690s. The focus of my study necessitated taking into account other titles and honorifics with which Goodman and Goodwife contrasted structurally and pragmatically, in particular the (vocative) prefixes Mister (Mr.) and Mistress (Mrs./Mis(t).), as well as the (non-vocative) postfixes yeoman and husbandman. I do not wish to discuss all the details of the article here. Instead I shall briefly outline its most important insights, followed by a critical integrationist reflection in the next section.

From the sociopragmatic point of view I adopted at the time, the question that I sought to illuminate was the following: given that the titles of civility Goodman and Goodwife were described by contemporaries as well as by historians and lexicographers as markers of colonial New England culture, these titles must presumably have been part of the linguistic community code, i.e. they were abstract linguistic signs of langue. However, as pragmatic markers, members of the New England communities must have known how to use them not only in a general sense, but also in concrete situations towards concrete individuals, who were either addressed (in the case of male members) as ‘Goodman so-and-so’ or ‘Mr. so-and-so’, or by other linguistic means (including using no title at all). In other words, the titles served parole-related functions similar to those of proper names. The strictly Saussurean framework thus needed to be complemented by a Labovian one, as knowing the sociolinguistic value of the pragmatic markers was tantamount to being granted an insight into the reasons why in courtroom situations speaker A addresses hearer B with a certain title (rather than another), or refers to someone else (present in the courtroom or not) using a certain title (rather than another). What a description of the Saussurean langue fails to do, in fact, is to account, at the level of parole, for the speaker’s intentions implicit in his/her choice of a title, which the hearer, in turn, is supposed to recognize. In fact, if A and B are speakers of the same language, the Labovian theorist argues, they will share the same norms of use (i.e. they are aware of how the linguistic variants are evaluated in their community), and are therefore able to communicate successfully with one another (i.e. recognize the intentions behind the words). Thus if A addresses B as ‘Goodman B’ (rather than ‘Mister B’), there is a reason why A does so and B (and/or an overhearer C) is supposed to recognize that intention.

On scrutinizing the manuscripts of the Salem Witchcraft Papers featuring occurrences of Goodman or Goodwife, and comparing them with occurrences of Mr. and Mrs., I determined that occupation did not seem to play any role as far as the choice of the male titles was concerned. Instead the relevant sociological variables favouring the title of Mr. seem to have been related to the person’s degree of civic worth and economic prosperity. At the same time it was clear that neither Goodman nor Goodwife was negatively connoted in the New England colonies, i.e. their stigmatization occurred earlier in England than in New England. Still, it transpires from the records that male Salemites who held an office of dignity and/or a high military rank, or had accumulated considerable wealth through their business and/or owned considerable estates, were commonly referred to and addressed as Mr./Mister – and not as Goodman. When it came to landowning farmers, the distinction seems to have been between those identified in the legal records as yeoman (for whom only the title Mister appears in speech-related texts) and those identified as husbandman (for whom the title Goodman preponderantly appears). To cite one example: Mr. John Putnam and Goodman Robert Pease were both weavers by profession, but John Putnam was the constable of Salem and an affluent landowner who employed labourers (i.e. had authority over others). Addressing Putnam as Goodman Putnam, it might be conjectured, would have been socially inappropriate, especially in formal situations. The female titles seem to have depended on the social status of the husband: thus, as the records attest, Mr. Putnam’s wife is referred to as Mrs. (Ann) Putnam, whereas Goodman Pease’s wife is mentioned as Goodwife (Sarah) Pease. On the other hand, it has been suggested that the titles Goodman and Goodwife were associated (positively) with church membership in Puritan New England (i.e. being admitted to the communion table). While this was probably true for most Salemites (in fact many of the ‘goodmen’ and ‘goodwives’ were church members with an impeccable reputation), there are nevertheless examples suggesting that female villagers held in very low esteem might also have been addressed as ‘Goody so-and-so’. A case in point was Sarah Good, who was divorced and had to beg for food together with her two destitute children. Some of the courtroom interactions as rendered by the town clerks, however, suggest that the use of these titles must have been governed by much more complex factors than the sociological variable analysis reveals. For example, it was found that the servant girl Mary Warren referred to her employer, Goodwife Elisa­beth Proctor, as ‘Mistress Proctor’. She is also reported as having addressed John Proctor – a husbandman according to the records – as ‘Master’ and referring to him as ‘my master Proctor’. The servant girl Mercy Lewis on the other hand, who worked in the Putnam household, refers to Mrs. Ann Putnam, one of the richest women in Salem, as ‘Goody Putnam’. The scattered counter-examples found in the Salem Papers must remind any sociopragmatically oriented researcher that (quite naturally) these titles occurred in infinitely many situations in infinitely many constellations of people, so that no collection of records, however extensive, could possibly cope with this kind of variety.

A sociolinguistic study in any orthodox sense can only be done ‘segregationally’, i.e. by taking a third-person perspective on linguistic signs. Strictly speaking, there is no ‘sociolinguistics’ from an integrational point of view. To conclude, as I did in the study, that Goodman and Goodwife encoded information on a male or female villager’s economic and civic standing is already to decontextualize communication as the ongoing process that it is and instead treat the two linguistic signs as being the ‘same’ signs across different situations. This is the metaphysical price that the empirical (i.e. Labovian) linguist must pay when segregating language from communication. Or as Harris (1996:9) puts it, there can be no ‘amicable division of labour’ between segregational and integrational studies of language.

It is reasonable to assume that not many contemporary anglophones are familiar with the appellations of civility Goodman, Goodwife and Goody. As far as I can tell, they are no longer used in modern (American) English, though such a statement obviously does not rely on any empirical ‘truth’: it is on a par with claiming that contemporary anglophones no longer say doth or hath, which does not mean that these forms are never said or written by anybody any more. As far as I am concerned, my linguistic experience now includes acquaintance with the two titles of civility, namely as a result of my research activities. They turn out to be part of my communicational biography whenever I remember or encounter them. Others will certainly know about them as well: for example, any expert on Nathaniel Hawthorne knows that Goodman, Goodwife and Goody were titles conferred on colonial New Englanders, as the title of Hawthorne’s short story, Young Goodman Brown, indicates. Now you, as readers of the present article, have learnt about the titles and how they were (allegedly) used in a late seventeenth-century New England context. The point is that anybody could encounter these titles at any point in time and make sense of them, i.e. try to find out ‘what to do with them’ as circumstantially required for the communicational purposes at hand. The activity of reconstructing how the titles were used sociopragmatically is one such way of making communicational sense of them. It is an activity guided by the conviction that there are ‘linguistic facts’ that the sociohistorical linguist can establish, provided that sufficient relevant data become available. This activity is thus macro­socially conditioned by the assumption that Goodman and Goodwife possess their own (socio)linguistic history. When does the research end? When the researcher has discovered (sometime in the future) as much as can possibly be known about these titles. Knowledge for the sake of knowledge. How might that differ from a contemporary college student of English reading Hawthorne’s Young Goodman Brown, and finding him/herself with an edition that does not gloss Goodman? When does his/her research into the meaning of Goodman end? As soon as he/she has found out enough to satisfy the communicational requirements here and now. It is doubtful whether any such student would look for a more detailed linguistic account of the titles: the definitions given by the dictionary are sufficient for the purpose of reading and understanding the short story. From the point of view of an integrational epistemology, ‘knowing what to do (and how)’ takes priority over ‘knowing that’ (Harris 2009b).

What kind of experience do I lack when it comes to knowing ‘what to do’ with colonial titles of civility? The answer is: everyday lived experience. I know what to do with contemporary honorifics like Sir, Madam, Miss, Mrs, Mr – in the sense that I know from personal experience that I have used them and so have others who have communicated with me. My use of these honorifics and my understanding of how they are used by others are macrosocially conditioned. If I claim to know how Goodman and Goodwife were macrosocially conditioned in Puritan New England, it is clear that my knowledge on that score is of a very different order. Having done extensive research, I could claim to know that Ann Putnam was entitled to be addressed as Mistress Putnam for socio-economic reasons, but I have absolutely no idea who in Salem would call her Mrs. Putnam and who Goody Putnam and how consistently they would do so. Neither do I know how Ann Putnam herself felt about being called ‘Goodwife’ rather than ‘Mistress’. Perhaps she did not mind when it came to some people, but would have minded when it came to others. Perhaps some villagers rarely used titles outside of the courtroom (towards her, or even in general), addressing her simply as ‘Ann’, or referring to her as ‘Ann Putnam’. The segregationist would object that my study at least suggests ways in which the use of competing titles of civility, i.e. Goodman–Mister and Goodwife–Goody–Mistress, may have been macrosocially conditioned in colonial New England. However, the integrationist would disagree: while individuals may be subject to the same biomechanical and circumstantial constraints – neither speaker A, B nor C is able to physically communicate with hearer D in the latter’s absence – individuals are never identically conditioned by macrosocial factors. Even though hearer A, B or C, being speakers of English, cannot understand speaker D, who only speaks French, it is not the case that A, B and C will all hear the ‘same’ or that they will all automatically understand nothing. How they will integrate D’s utterance (construed by them to be ‘French’, if at all) will vary individually depending on their communicational biographies. For the integrationist, there are no single macrosocial factors that one can artificially isolate as an analyst (e.g. socio-economic factors like wealth, offices held, military rank or church membership) in order to ‘explain’ the occurrence of one linguistic variant rather than another. While in our lay analyses of communication we do indeed focus on specific constraints at the expense of others, these explanations lay no claim to being ‘scientific’. The question that arises given the integrationist’s rejection of segregational linguistics is whether historical sociolinguistics can still be viewed as having a legitimate place, and on what basis. For example, it is because historical sociolinguistics isolates macrosocial factors and treats them as timeless abstractions that its studies may become community resources, outside the strictly academic context. Thus a recent historical novel set in early New England (Youmans 2020) refers in its glossary to my study on New England titles of civility in order to legitimate the author’s heavy reliance on ‘Goodman’, ‘Goodwife’ and ‘Goody’. Moreover, sociohistorical studies on lexical variation may contribute to the making of historical dictionaries, or may provide updated information for new editions. In the Dictionary of American English on Historical Principles (Craigie & Hulbert 1938–44:1145), for example, the entry for Goodman reads:

While this gloss does not directly contradict anything that my study revealed, one could propose to modify it somewhat:

Even though the second gloss might arguably be too detailed for lexicographical purposes, the example illustrates how historical sociolinguistics can be employed in the service of lexicography. Both, in fact, deal in linguistic abstractions, which is also the reason why there can be no dictionary founded on integrational principles. From an integrational point of view, the dictionary is a communicational tool which aims to reduce semantic indeterminacy for specific purposes (Harris 1998). However, it is not a compendium containing the semantic ‘truth’ about words.

Integrationists do not deny the past, i.e. they do not subscribe to some radical form of idealism, but they do not think the past is something that can be studied in a communicational vacuum – as something hermetically sealed from present (communicational) circumstances. The past, for the integrationist, is a product of the present. The same is true for the ‘linguistic’ past. Whatever it is I encounter in an extant written document from the (distant) past is not meaningful by itself. The graphic marks by themselves are not signs, but they may become signs for whoever tries to decipher and transcribe the manuscript. Whatever it is that we ‘see’ when we scrutinize and compare various extant records from the past, the regularities and patterns are not ‘there’ already. They have to be ‘made’. This is not to say that the linguistic past is whatever someone says it was. Crucially, however, there is nobody around to personally remember the distant linguist past. A written document can be interpreted in many ways, depending on the signs that one identifies. Pace Labov, there are no independent linguistic facts (Harris 1998). But if several philologists conclude independently that certain graphic marks are the word Goodman, the historical sociolinguist may assume the same unless he/she has reasons to believe otherwise. By identifying the word Goodman, however, we have not identified a word magically ‘teleported’ from the past: we have created a word here and now, to which we assign a semiological value as part of a certain programme of activities. Macrosocially speaking, that programme may be discursively constructed as having as its goal the reconstruction of a past variety ‘as it was spoken’. However, as the integrationist would point out, no sign can exist separately from the activities in and for which it was created. This, it would seem, is the Harrisian ‘uniformitarian principle’.