Abstract
Concessivity in English is often associated with structures like Although it was hot, she went running. Although introduces a subordinate clause. There are also several markers of concessivity associated with coordinate structures, among them, all the same and nevertheless. The development of coordinate markers of concessivity has received little attention. Here I compare the histories of three markers that arose from expressions of identity: all the same, just the same, and at the same time. All arose at various times in Late Modern English. Rich rhetorical contexts for the development of concessive uses of all the same ‘exactly the same’ are attested in corpora. Contexts for the later development in the US of just the same are less rich, and analogy to all the same is probable. Both came to be well entrenched in concessive use and are used in clause-final as well as clause-initial and occasionally clause-medial position. By contrast, at the same time functions only marginally as a concessive. It is also used as an elaborator and is not attested in clause-final position. Data are analyzed from the perspective of Diachronic Construction Grammar. Implications for some current debates are discussed, specifically for establishing when a construction has come into being, what kinds of contexts enable this emergence, and what level of abstraction is appropriate for analysis of sets of related constructions. Finally, some modifications of hypotheses about position of connectives relative to the clause are suggested.
1. Introduction
There are several kinds of markers of a concessive relationship between clauses. Some have been shown to originate in expressions of similarity or identity, for example, all the same (see König 1986, 2020; König & Siemund 2000). However, little attention has been paid to the semantic-pragmatic properties and histories of expressions of identity that are used to cue concessive relationships. In this paper, I characterize the meaning, syntactic distribution, and historical development of three expressions with same that are used in Late Modern English as coordinate clause connectives with concessive meaning: all the same, just the same, and more marginally, at the same time. Investigating expressions that share a unit, in this case same, gives insights into the ways in which interlocutors may come over time to use similar concepts to express related connective relationships at different times, enriching the options available in the flow of speech or writing. My approach is cognitive-functional, constructionalist, and diachronic. Therefore, the study is concerned with the semantics and pragmatics, not the logic, of concessive uses of the same-expressions.
There are many kinds of concessive markers. A prototypical marker of concessive relationships is the subordinator although (cf. Mittwoch, Huddleston & Collins 2002:734; also Quirk, Greenbaum, Leech & Svartvik [1985:484], where the variant though is cited). Other subordinator concessives include while, whereas, and if in some of their uses. All can be used to mark the first clause in a “concessive marker p, q” configuration, where p stands for the clause presenting the state of affairs that is conceded and q for the clause presenting a conclusion that might normally not be expected. In Although it was hot, she went running, for example, it was hot is the state of affairs p, and she went running is q. Although marks she went running as normally unexpected, given the situation it was hot. There are also coordinate concessive markers, which include the same-expressions under discussion, and also nevertheless, still, and under some circumstances, yet and and (König 1986; for a considerably longer list, see Iza Erviti 2021:Chapter 5.2). They mark the second clause q in a “p concessive marker q” configuration. The structure of coordinate concessives tends to be informationally looser than that of subordinate concessives, partly because coordinate structure allows longer and more complex stretches of text in p and q than the subordinate structure usually does. Although gives weight to p before the countering comment in q. In coordinate structures, however, more weight is given to q and attention is drawn to the new countering comment. Compare Although it was hot, she went running, where it was hot is in focus and receives prosodic prominence, with It was hot; just the same she went running, where she went running is in focus and receives prosodic prominence.
Examples (1)-(3) illustrate use of the markers all the same, just the same, and at the same time, which are the topic of this paper. In these example, the markers appear as connectives in coordinate structures clause-initially before q to express concession:
(1) “and your overpowering dignity was a sight for gods and men. All the same you were a darling.” (COHA, 1896, Holiday stories for young people)
(2) Certainly this declaration stirred her deeply. Just the same, she took it evenly and asked: “How do you know that?” (COHA, 1994, Gilligan, Voyage of the Golden Hind)
(3) People involved in an organization are expected to embrace their roles (that is, to dedicate themselves fully to their performance while working in and for the organization, to identify themselves completely with the role they are performing at that moment), yet at the same time to distance themselves from them (that is, to remember all along that it is merely a role they are playing. . .). (COHA, 1990, Bauman, Thinking sociologically)
In (1) all the same evokes the assumption that someone behaving with overpowering dignity is likely not to be a darling. Likewise in (2) just the same evokes the assumption that someone who is deeply stirred will react emotionally, not evenly. And in (3) yet at the same time evokes the assumption that people who are expected to embrace their roles in an organization will also be expected to feel close to that organization. In each case the evoked assumption is rejected. As discussed in section 5.2, cooccurrence with a contrastive marker such as but or yet, as in (3), strengthens concessive meaning and may have contributed to its emergence, but is not necessary for concessive interpretation, as seen in (1) and (2).
The phrases all the same, just the same, and especially at the same time are often used with referring (“phoric”) functions, as in (4).
(4) You, Nat, St. Claire, you’re all the same. That’s right. We are all the same. We all have responsibilities. (COHA, 1992, Beverley Hills)
In (4) the first all the same refers back to the three individuals mentioned and all is a quantifier (‘each one’), the second, echoic all the same refers back to we, as well as to the first use.
As is detailed at some length in various chapters in Huddleston and Pullum (2002), same can be used with two related meanings:
‘Identical’ (“Kim and Chris are in the same class”). This use of same is typically anaphoric to a term recoverable from the overt context or from general knowledge (e.g., that a class is a set of students) (Stirling & Huddleston 2002:1545); and
‘Similar.’ This use is comparative (“X is same as Y, X is similar to Y”) and implicates scalar (more or less) equality (Huddleston 2002:1104).
All can also be used with multiple meanings (Pullum & Huddleston 2002:549):
‘Every,’ a quantifier use that is favored when the subject of a copula verb is plural (“we are all the same”) or the nominal is plural (“all the same responsibilities”).
‘Exactly,’ an adverbial “degree modifier” use that is favored with singular subjects, as in “It’s all the same to me.”
In (4) world knowledge suggests that all the same asserts similarity, not identity, of each individual while at the same time allowing the interpretation ‘we are exactly the same in certain respects.’ 1
As Breban (2010a, 2010b), on which I draw, argues, the fixed phrase the same as in the same three houses functions as a complex premodifying determiner. Breban (2010a) shows that, from later Middle English (c1400) on, the same has been used to indicate “an anaphoric relation of identity” (Breban 2010a:205). It is an emphasizing use that serves to “‘enhance’ the act of reference by confirming the correctness of the identification” (Breban 2010b:71). As Breban (2010a:236n) notes, saying that two things are the same “always implies non-identity of reference in order to assert the identity of the referent.” In other words, use of same evokes alternatives and the possibility of not identifying the antecedent correctly. I will suggest in sections 4 and 5.2 that this implicature of non-identity may have contributed to the shift to concessive use. The premodifier uses are grammatical as they serve determiner functions, but they are not concessive. I call them “phoric” uses.
Non-concessive phoric uses of the expressions under study are compositional: the parts are understood to be combined: (i) all or just + the same adverbial phrase meaning ‘exactly the same way’ in the context of a verb of action like do; (ii) all or just + the same anaphoric pronominal predicate phrase meaning ‘exactly the same thing’ in the context of a copula verb like be; and (iii) at + the same + time, a temporal phrase. One can ask “same as what?” as in (4). However, when all the same, just the same, and at the same time are used as unitary connectives, as in (1)-(3), such a question is inappropriate, and there can be no obvious answer to it. The concessive same-expressions function as units and are not compositional, although they are analyzable. That is, the meanings of the separate parts are not identifiable, but the component morphemes are (see Bybee 2010:45). That they are spelled as separate words is a function of their late development in the nineteenth century (contrast nevertheless, which has been used since the fourteenth century; see OED, s.v. nevertheless).
In Fraser’s (2006) terminology, concessive same-expressions are members of a category called “Discourse Markers,” which signal the relationship between discourse segment 1 and discourse segment 2. 2 They are constructions, complex pairings of form and function that are used as a unit or “chunk” (see Goldberg 1995, 2006). Syntactically, they are “conjuncts” (Quirk, Greenbaum, Leech & Svartvik 1985:636). Functionally they are “procedurals” in the sense that they “constrain inferential procedures” (Blakemore 2002:79) and are used to guide interpretations in context and point “hearers to particular – more or less – schematic frames of interpretation for the utterance hosting such expressions” (Hansen 2012:595).
In this paper I outline the history of the three same-concessive expressions from a diachronic constructionalist perspective (see Traugott & Trousdale 2013; Barðdal, Smirnova, Sommerer & Gildea 2015; Hilpert 2021; Traugott 2022a, 2022b). All three concessive uses arose in the Late Modern English period (1700-1920). In her study of the development of “rhetorical adverbial connectors,” Lenker (2010) found that the periods of Early Modern English (1500-1710) and Late Modern English were especially rich in the development of new markers in the set she calls “contrast/concession” (e.g., after all, however). Lenker (2010:98) lists concessive use of at the same time in the Late Modern English 1 period (1710-1780) and of all the same in the Late Modern English 3 period (1850-1920) (just the same is not mentioned). The present study follows up on Lenker’s (2010) findings, drawing attention to the referring origins of the concessive expressions and the discursive contexts in which the concessive meanings arose. The study of these developments contributes to a richer understanding of the history of English in the last three centuries, which have been the center of attention only recently. As discussed in Lewis (2012), most studies of Late Modern English focus mainly on syntactic rather than pragmatic and discursive changes (see e.g., Kytö, Rydén & Smitterberg 2006; Mair 2006; Hundt 2014). This continues to be true.
The history of the clause-introducing use of all the same is outlined in Traugott (2022a:127-135, 2022b:128-135) and of the post-nominal phrasal use of all the same and just the same, as in fear just the same, in Traugott (2021). In these studies, there is no discussion of at the same time, or of distinctions between British and American usage. In the present study, the three expressions are compared, and the possibility of varietal distinctions is explored. Comparing the three constructions provides the opportunity to better understand some issues in diachronic construction grammar, specifically:
(a) How do we recognize when a construction has come into being, and what types of contexts enable this coming into being (Petré 2019)?
(b) What level of abstraction over constructions is plausible (Hilpert 2013:155-203)?
(c) What is the role of position relative to the clause (e.g., Lenker 2010, 2014; Traugott 2022a:203-224, 2022b:172-179)?
The structure of the paper is as follows. Background on the approach to concessives and to diachronic linguistic analysis is sketched in section 2. Section 3 outlines the data and methodology. Section 4 presents diachronic accounts of the development of the three concessives with same. Section 5 addresses the theoretical issues (a)-(c) mentioned above. Section 6 concludes and suggests some areas for further research.
2. Background
Here I briefly provide some background for later analysis and theoretical discussion. In section 2.1, I introduce a cognitive-functional approach to concessives, and in section 2.2 a usage-based approach to constructions and how they come into being.
2.1. A Cognitive-Functional Approach to Concessives
In the cognitive-functional literature, there is a body of research on concessive clauses, both relatively less formalist (e.g., König 1986, 2020; König & Siemund 2000; Hilpert 2013) and more formalist (e.g., Gast 2019). Hilpert (2013) and König and Siemund (2000) mention just the same and all the same among concessive markers, and König (2020:54) contrasts use of all the same as an independent concessive marker with use of at the same time as a contextually dependent concessive. He also mentions the concept of similarity as a source of concessive connectives in French (tout de même ‘all the same,’ from ‘all of same’) and German (gleichwohl ‘nevertheless,’ from ‘identical to be sure’).
In early work on concessives, König focused on incompatibilities between the states of affairs described in p and q. For example, König (1986) addresses similarities and differences between prototypical conditionals (“if p, then q”), concessive conditionals (“even if p, q”), and concessives (“although p, q”). He analyzes both p and q as factual. For example, in Although he studied hard, he failed the exam, both He studied hard and He failed the exam are assumed to be factual. P and q are conceptualized as presuppositions in an “if p, then normally not q” configuration. For example, in if one studies hard normally one will pass, not fail, the exam, failing is considered to be incompatible with studying hard. Q counters this presupposition.
More recently, Gast (2019:157-158), among others, has proposed that a concessive connective signals the degree of potential probability between the content of p and q, rather than potential incompatibility. Commenting on this proposal, König (2020:61) says that if Gast’s analysis “is to have a certain cognitive plausibility we must assume that we are constantly making probability judgements about the co-occurrence of situations.” Finding this assumption implausible, but recognizing the value of a positive rather than a negative analysis, König (2020:61) proposes instead that concessives involve “assumptions about what typically goes together,” in other words, “generalizations about compatibility.” For example, one studies hard and one does not fail asserts a more typical experience than one studies hard and one fails the exam. While such generalizations are often based in experience and world knowledge, they may also be based in the speaker’s point of view in a particular situation.
I propose that, from a discursive perspective, concessives cue that the speaker/writer perceives q to be more or less compatible with p despite possible counter-expectations and invites the addressee/reader to access the same point of view. Example (2) above “Certainly this declaration stirred her deeply. Just the same, she took it evenly” can be said to cue that the writer regards she took it evenly as unexpected given that she was stirred deeply, but nevertheless relatively compatible with this declaration stirred her deeply. This is pragmatically more appropriate than interpreting she took it evenly as incompatible with this declaration stirred her deeply, given that both p and q are factual. Furthermore, the contrast between the content of p and that of q in same-concessives is often not very strong. This appears to have been the case throughout their histories.
The three concessive expressions with same under discussion—all the same, just the same, and at the same time—are structurally different from although-concessives in the following ways:
They are conjuncts (Quirk, Greenbaum, Leech & Svartvik 1985:636) associated with coordinate structure, as are some other concessive markers like nevertheless and however.
The order of p and q cannot be varied. This follows from coordinate structure (contrast the possibility of using both Although he studied hard, he failed the exam and He failed the exam, although he studied hard.)
In the case of although, because it is a subordinator, the marker precedes p; it is not used in clause-final position. In the case of all the same and just the same, which are coordinators, the marker usually follows q and so is used clause-finally, but can be used clause-initially and occasionally clause-medially (see section 4).
All three have homonyms that are literal and are used phorically. They are attested with various degrees of frequency in the data investigated.
Despite the fact that although and same-concessives are different structurally, pragmatically, and discursively, they are functionally similar. In both the subordinate and coordinate types of concessive, there are two components, a background assumption and a relevant opposing assertion. The meaning similarity justifies calling both the subordinate and coordinate types “concessive” (König 2020:59).
Opposing assertions contrast with the background assumption. As Sweetser (1990), Rudolph (1996), and Blakemore (1987), among others, point out, contrast does not exist in the world. Rather, it is established by the interlocutors when comparing states of affairs. Likewise, concessivity does not exist in the world; it is constructed in negotiation of meaning between a speaker/writer/signer (speaker/writer for short) and an addressee/reader. Consciously or not, a speaker/writer chooses a marker that reflects their own point of view on the connectivity and degree of compatibility between discourse segment 1 (p) and discourse segment 2 (q), and that they wish the addressee/reader to infer. In other words, speakers/writers take a stance toward the texts they are creating and invite addressees/readers to interpret this stance in the same way (Du Bois 2007; Traugott 2022a; also Haselow 2013, 2019). Discussing adversative connectors, Lang (2000:249) says but is used to invite the addressee/reader “to look for an appropriate CONTRASTING ASSUMPTION” (caps original). The concept of addressees/readers being invited to search for an appropriate assumption is key to the interpretation of a concessive relationship. As Lang (2000) cited above suggests, speakers/writers invite addressees/readers “to look for an appropriate [background] assumption” regarding norms based in both general knowledge and the particular discourse or situation at hand.
2.2. A Usage-based Account of Constructions and How They Come into Being
Turning now to the constructionalist view of language that I take in what follows, there are several types of construction grammar (see Hoffmann 2022:256-271). The one adopted here is the usage-based model developed in Goldberg (1995, 2003, 2006). Like Goldberg, I assume that knowledge of language is grounded in “instances of a speaker’s producing and understanding language” (Kemmer & Barlow 2000:viii; cf. Bybee 2010). I assume also that knowledge of language is flexible. To make a useful distinction proposed in Hopper (2011:26), in my view language is dynamic and “emerging.” It arises out of relatively stable phenomena, but is not intrinsically ephemeral and “emergent.”
A basic tenet of usage-based construction grammar is that constructions are form-function pairings (see Goldberg 2003, 2006). 3 Form and function are assumed to be linked symbolically. Another basic tenet is that constructions are linked in a network which consists of both particular (“substantive”) constructions like same, nevertheless, and function-based abstract, schematic constructions like the “Concessive.Schema,” a function-based abstract unit that can be posited as a generalization over constructions with concessivity-marking function. Such abstract constructions are linguists’ generalizations.
Most constructionalist research is synchronic. How best to think about the emergence of constructions (“constructionalization”) has been the topic of work such as Traugott and Trousdale (2013) and Petré (2019). I adopt the following characterization of constructionalization, updating definitions in Traugott and Trousdale (2013:22) and Trousdale and Traugott (2021): Constructionalization is the establishment of a new symbolic link between form and function which has been replicated across a network of language users, and which involves an addition to the constructicon.
Key here are “establishment” and “network of language users.” “Establishment” (also known as “conventionalization,” see Croft 2001:19; Schmid 2017:3) means that at least two speakers share knowledge of the new construction. The “constructicon” is a network of constructions that captures “[t]he totality of our knowledge of language” (Goldberg 2003:219). Leading up to and following constructionalization is the accumulation of small constructional shifts that do not involve addition to the constructicon. Earlier these were referred to as “constructional changes” (Traugott & Trousdale 2013:26-27). “Replicated Constructional Shifts” is a more accurate term.
Change is assumed to be not innovation but the development of shared, conventionalized usage (see, e.g., Weinreich, Labov & Herzog [1968] 2017; Milroy & Milroy 1985; Schmid 2017). It is gradual, except in the case of word formation patterns. Evidence for this view of change and of constructionalization is discussed in detail with respect to connectives such as after all, all the same, but, and by the way in Traugott (2022a, 2022b) and will not be repeated here.
In the approach adopted here, discursive contexts are especially important in enabling change. Concessive relationships involve two clauses and, in the early stages of development, use of a negative or semantically negative unit in the immediate local context of either discourse segment 1 or discourse segment 2 often activates concessive inferences. So may contrastive discourse markers like but or yet immediately preceding the second clause q. These are not redundant, but reinforce the concessive use of the connectors.
Example (5) illustrates discourse structuring of two immediately juxtaposed clauses, the first of which is negative in semantic content (I despise flattery), the second of which is introduced by but and is positive in semantic content (I love it). The speaker draws attention to his awareness of the potential incompatibility between the two assertions with all the same appended in final position.
(5) I despise flattery, my dear, but I love it all the same. (CLMET 3_0_3_195, 1885, Linton, Autobiography of Christopher Kirkland)
But textual concessivity is not limited to local connectivity of the kind in (5). Juxtaposition of a clause with “a whole preceding text unit which may be composed of various sentences” (Crevels 2000:319) may evoke less than full compatibility, as will be seen in many examples in section 4.
3. Data and Methodology
Because Lenker (2010) found that new markers of contrast/concession arose in the Early and Late Modern English periods (see section 1), I consulted three electronic corpora covering 1500-1920: (a) Early English Books Online 1473-1700 (EEBO), with about 755 million words from printed books of the period; (b) the Corpus of Late Modern English Texts, 1710-1920, version 3.0 (CLMET 3.0), with about thirty-four million words 4 ; and (c) the Corpus of Historical American English 1820-2019 (COHA), 2021 version, containing about 475 million words. EEBO and CLMET 3.0 exemplify texts written in Britain, whereas COHA exemplifies texts written in the US. CLMET 3.0 and COHA data used for this study were collected in August-October 2022. Spelling variants were not found, as can be expected from Modern English data.
No examples in EEBO could be analyzed as unambiguously concessive. However, concessive uses of the three same-expressions appear in the Modern English corpora CLMET 3.0 and COHA. Given the large number of hits of the three strings, especially at the same time (see Table 1), a subset of examples from these two corpora was hand-collected and entered into Excel files. To achieve at least semi-randomness, all examples from every fifth file in CLMET 3.0 were collected (a total of sixty-eight files). In the case of COHA, data for all the same and just the same are sparse per decade up to the 1870s: fewer than one hundred but over one hundred per decade after that. The first and fifth examples in each decade were selected up to the 1870s, and, after that, the first and tenth examples in each decade. In the case of at the same time, data are ample. There are 699 examples in the 1820s in COHA, and over a 1000 in each subsequent decade. To avoid bias to the same text type, the first and tenth examples of the first and last one hundred hits in each decade were selected to allow for some comparability with data for all the same and just the same. 5 The total data set for the three expressions consists of 1863 examples: 1247 hits in COHA and 616 in CLMET 3.0.
Number of Hits Retrieved in Data Set Compared to the Total Number of Hits in COHA and CLMET 3.0
Table 1 provides the number of hits selected for each expression out of the total hits or files in the corpora. For example, the all the same column gives the following information: 411 hits were retrieved from a total of 4168 hits in COHA. In CLMET 3.0, 165 hits were retrieved from twenty-three of the sixty-eight selected files (forty-five of the sixty-eight files contained zero hits of all the same). The final column lists the total number of words or files in each corpus.
Drawing on my earlier studies of all the same and just the same, examples were considered potential candidates for use as concessive markers provided they serve a linking function and evoke less than full compatibility. Because the data are written, prosody could not be used as evidence. Paraphrase was used to assess plausible phoric determiner or concessive status. The paraphrases used to establish determiner meaning were ‘all the same thing/way,’ ‘just the same thing/way,’ and ‘simultaneously.’ The paraphrases used to identify concessive meaning were ‘nevertheless’ and ‘despite that.’ In some cases at the same time appears to function like an elaborator rather than as a concessive; therefore the paraphrase ‘in addition’ was also used for this expression. In a few cases, category status was undecidable; this was coded with a query (‘?’).
As has often been pointed out in the literature, it can be “difficult – from a formal and semantic point of view – to make a clear-cut distinction” between concessive and other functions of an expression: “only the context may disambiguate the function of the coordinating elements” (Ramat 2018:433). This is especially true of examples with at the same time. Several instances only possibly draw attention to a lack of full compatibility or of incongruity between the content of the conjoined clauses. For example, does (6), which contrasts business preferences for new train designs, one with a dome-car, the other without, exemplify literal temporal use, concessive use, or possibly both?
(6) Fred G. Gurley (1889-), president of the Santa Fe [. . .] thought so much of the dome idea that in 1956, with his “El Capitan,” he introduced the new concept of a “highlevel” train. The new train provided an unobstructed view for all passengers as they rode at dome-car height. [. . .] At the same time, the New York, [. . .] James M. Symes’s Pennsylvania, and the New York Central [. . .] were all excited about several new lightweight lowlevel trains, such as Pullman Standard’s “Train X.” (COHA, 1961 Stover, American railroad)
While the train companies’ different interests were no doubt at least partially contemporaneous, and a temporal ‘at the same period’ interpretation is possible, juxtaposition of Gurley’s and others’ train preferences implicates that the opinions were not fully compatible (a concessive interpretation). One plausible paraphrase of the content of (6) is ‘Although Gurley thought so much of the dome idea that he introduced the new concept of “a highlevel” train, the New York . . . and New York Central were all excited about several new lightweight lowlevel trains.’ In fact, it does not really matter which reading is preferred. Undecidable examples like (6) were coded as “Temp, C?” or, in the case of elaboration, “Temp, E?”.
All hits of the relevant strings (all the same, just the same, at the same time) were collected from CLMET 3.0 and COHA, including some with typos and content that would be considered unacceptable today. Each was coded for one of the following category statuses:
(a) Quantifier in a noun phrase configuration (“all the same symptoms”) (Quant N),
(b) Adverbial phrase (AdvP) use (“let things go on just the same (way)”);
(c) Pronominal predicate phrase use with a copula verb (“
(d) Temporal prepositional phrase (Temp) use (“called at the same time”);
(e) Concessive (C) use;
(f) Elaborative (E) use (in the case of at the same time);
(g) Undecidable (?).
Since degree of contextual dependency is of interest given König’s (2020:54) comment on at the same time cited in section 2.1, use of an immediately prior discourse marker such as but, yet, or still was noted for concessive (or potential concessive) interpretation.
All hits were also coded for position: clause-initial (I), medial (M), and final position (F), because these contribute to discourse structuring and to distributional differences among the expressions. “Initial” is understood as first constituent of the clause after an optional discourse marker, and “final” as last constituent of the clause. Lenker (2014) found that medial use of adverbials like however and therefore greatly increased in frequency in academic prose in the late eighteenth and early nineteenth centuries. Building on Greenbaum (1969) she suggests that medial use of adverbials partitions the clause: an adverbial occurring after the first constituent (the subject or a frame-setting adjunct such as in Rome) focuses and contrasts what precedes, whereas an adverbial occurring after the finite verb focuses the remainder of the clause (Lenker 2014:31). The number of medial examples of the same-concessives is too small (a total of thirty) to lend itself to significant generalizations based on Lenker’s (2014) fine-grained analysis, in which five different medial positions are posited. So only a coarse-grained distinction is made in coding between Med1 (post first constituent) and Med2 (post finite verb). Assessment of the data for medial position in the light of Lenker’s (2014) hypothesis is the topic of section 5.4.
4. A Diachronic Account of the Three Same-Concessives
In this section I give a brief diachronic account of the three same-concessive expressions: all the same (4.1), just the same (4.2), and at the same time (4.3). Section 4.4 provides a summary comparison.
Breban (2010a, 2010b) argues that the fixed phrase the same was used around 1350 as an emphasizer meaning ‘precisely this one referred to,’ which invites the addressee to search for a prior co-referent. Initially it functions as an “identifier.” But it was soon extended to determiner use to indicate “an anaphoric relation of identity” (Breban 2010a:205). This is particularly striking when the co-referents are personal names (7). Such emphasizing uses “‘enhance’ the act of reference by confirming the correctness of the identification,” as in (8). As mentioned in section 1, this phoric use is grammatical as it serves a determiner function, but is not concessive. The purpose of this section is to show how the phoric complex determiner use came in certain contexts to be reinterpreted as a concessive connective.
(7) “As the same Senek seith” ‘as the aforementioned Seneca says.’ (Chaucer Canterbury Tales, Melibee 2640; cited from Breban 2010a:204)
(8) “Rekne [. . .] which is the day of thy month, and let they rewle upon that same day” ‘Calculate [. . .] which is the day of the month, and place your rule on that same day.’ (LEON 1351-1420; cited from Breban 2010b:71)
4.1. The Development of Concessive all the same
There is a total of 1423 hits in the Early Modern English Corpus EEBO of the string all the same, in all positions where a noun phrase can occur. In all examples but one, all the same functions as a complex determiner in a noun phrase. All the same can be followed by adjectives as in (9).
(9) and yet to teach, that all the same little branches [. . .] are always enlarged. (EEBO, 1664, Helmont, Van Helmont’s works).
Corpus evidence from CLMET 3.0 and COHA suggests that concessive all the same has its origins in the degree modified expression alladv the same ‘exactly the same’ used in an adverbial phrase (AdvP) meaning ‘exactly the same way’ (see e.g., example 11). This is consistent with the fact that a large number of discourse connectives in English derive from adverbials, for example, after all, but, further, and instead. However, multiple sources of constructions can often be identified (see, e.g., Van de Velde 2014; Petré 2019), so what Halliday and Hasan (1976:105) call “substitute uses,” or pronominal predicate uses meaning ‘exactly the same thing’ cannot be ruled out as contributing to the rise of all the same in addition to AdvP uses.
Early examples of all the same in CLMET 3.0 are quantifier N, pronominal predicate, or AdvP uses. None suggests a plausible shift to concessive function has occurred that evokes and counters less than full compatibility. The first probable example of concessive use in the data set appears in CLMET 3 (Period 2) in the nineteenth century in the Med1 position after the subject, as shown in (10).
(10) I have no wish to excite your pity, gentlemen, or to gain your silence, by practising upon your feelings. Be silent. I am not the less ruined, not the less disgraced, not the less utterly undone. Be silent; my honour, all the same, in four-and-twenty hours, has gone for ever. I have no motive, then, to deceive you. (CLMET 3_0_2_165, 1826, Disraeli, Vivian Grey; see Traugott 2022a:131)
The context is negative and scalar (not the less ruined. . .disgraced). There is nothing in the text that suggests an anaphoric reference, nothing to ‘be the same thing as’ (pronominal predicate use) or to be done ‘the same way as’ (AdvP). Therefore, it is reasonable to analyze all the same in (10) as a concessive marker. There are, however, no ambiguous or undecidable earlier examples to support this hypothesis.
Letters from the years 1845 and 1846 between Robert Browning and Elizabeth Barrett in CLMET 3.0 provide twelve telling examples, one phoric (11), the eleven others concessive. Example (11) confirms the hypothesis that phoric AdvP ‘in exactly the same way’ used in contexts where ‘not exactly the same way’ is implied may have been the source for concessive uses.
(11) if you let me, love, I will not again, ever again, consider how it came and whence, and when, so curiously, so pryingly, but believe that it was always so, and that it all came at once, all the same; the more unlikelinesses the better, for they set off the better the truth of truths that [. . .] here is the whole wondrous Ba filling my whole heart and soul. (CLMET 3_0_2_185, 1846, Browning, Letters)
In (11) all the same is parallel syntactically with the manner adverbials curiously and pryingly and is best interpreted as ‘all the same way.’ The “unlikelinesses” (that Browning does not love his addressee, Barrett, whom he affectionately calls Ba and is courting) are explicitly presented as not true: they “set off” or distinguish the truth. Although (11) is a discourse about alternatives, all the same is not used concessively as it cannot plausibly be paraphrased by ‘nevertheless,’ or ‘despite that.’
In addition to (11), there are five examples in the file in which all the same is used with concessive meaning in initial position, for example, (12), and four in final position, for example, (13).
(12) My friend the Countess began proceedings [. . .] by asking ‘if I had got as much money as I expected by any works published of late?’ to which I answered, of course, ‘exactly as much, grazioso!’ (All the same, if you were to ask her [. . .] how much the stone-work of the Coliseum would fetch [. . .])[. . .] (CLMET, 3_0_2_185, 1845, Browning, Letters)
(13) you heard the playful words which had meaning all the same. (CLMET 3_0_2_185, 1845, Browning, Letters)
Two examples occur at Med2, for example, (14). Example (14) includes a negative and contrastive but, which discursively support a concessive interpretation ‘although you did not think of doubting me, nevertheless you thought that . . .’
(14) I perfectly understand that you did not think of doubting me, so to speak! But you thought, all the same, that if such a thing happened, I should be capable of doing so and so. (CLMET 3_0_2_185, 1846, Barrett, Letters)
Turning to American English data as represented in COHA, the string all the same appears initially in pronominal predicate use, as in (15).
(15) Therefore whether his life be for a day, or for eternity, it matters not, because, for the present, it is all the same to him. (COHA, 1820, Ballou, Letters)
All the same begins to be attested with concessive function in COHA in the 1830s and 1840s, in other words, at approximately the same period as in Britain, as represented by the CLMET 3.0 data. Example (16), a translation from French, is the earliest example of concessive use in the COHA data set used for this study, 6 and does not appear in the data set. In (16) the narrator comments in clause-final position that the sailors’ insistence on setting out was not always fully compatible with weather conditions.
(16) The coast in those days was not inhabited, and the roads, at that season, were not safe. Whatever may have been the appearance of the weather, the beings who were going to sail away in the Biscayan [a]rea, pressed on the hour of departure all the same. (COHA, 1833, Hugo, By order of the king)
As in CLMET 3.0, there are no ambiguous or undecidable examples in the COHA data preceding the attestation of concessive use. However, an ambiguous example (17) appears in COHA twenty years later than (16).
(17) [Lamenting the death of a stepson] The Lord gin me two gals, and then he sent me as noble a boy as ever was, I don’t care where t’ other comes from. He wasn’t mine, but I loved him all the same. (1854, COHA, 1854, Holmes, Tempest and sunshine)
This can be interpreted as ‘loved him the same way as I loved the girls’ (AdvP), or as concessive ‘loved him although he was not my own son’ (concessive).
In the COHA data, final all the same occasionally appears in phrasal X all the same uses such as in (18).
(18) A girl like that could be trouble, a different kind of honeyvoiced trouble from Harry’s brother, Morris, but trouble all the same. (COHA, 1954, Himes, Third generation)
In patterns like this, all the same has scope over the preceding phrase and is used to affirm an earlier choice of words after calling it into question (Traugott 2021). A formulaic phrasal pattern Thanks all the same also arose in the 1900s used to politely refuse an offer. Since these are phrasal, not clausal expressions, they are not of concern in this study, except to note that they came to be a conventionalized part of speakers’ knowledge of how to use all the same, and to provide a point of comparison with other same-concessives.
Use in clause-final position predominates in the data set until the end of the nineteenth century, when clause-initial use comes to be attested with slightly greater frequency. In addition to the clause-final and clause-initial position uses, there are four clause-medial examples in COHA, two at Med1 (e.g., 19) and three at Med2 (e.g., 20).
(19) But she should be very sorry not to see Brook any more, never to hear him talk to her again, never to look into his eyes – which, all the same, she so unreasonably dreaded. (COHA, 1896, Crawford, Adam Johnstone’s son])
(20) [After a volcanic eruption] the order that begins to establish itself as soon as chaos is confessed took possession of the ruin. But it was all the same a ruin and a calamitous conclusion for the time being. (COHA, 1877, Howells, The minister’s charge)
The CLMET 3.0 and COHA data sets for all the same are summarized in Table 2. The number of ambiguous/undecidable examples is given in parentheses. Note that seventy-six of the total 165 all the same hits in CLMET 3.0 are concessive in function (46 percent) and 239 of the total 411 all the same hits in COHA are concessive in function (58 percent, rounded).
Number of Concessive Hits of all the same in Different Positions in CLMET 3.0 and COHA
4.2. The Development of Concessive just the same
Like adverbial all, just is used in English as a degree adverbial meaning ‘exactly’ (cf. “She did it just the same way”). Just is usually analyzed as a focus marker like only or pure(ly) (e.g., Pullum & Huddleston 2002:587), but the degree modifier analysis is more appropriate for the phoric determiner source of the concessive, especially in light of the similarity with all the same. In EEBO the string just the same is first attested in the 1620s. There is a total of 389 hits in EEBO. Many are used as phoric determiners, as in (21).
(21) [About former and at the time current pricing] if they gaue a groat for that which wee now giue a shilling, they gaue just the same price which wee now giue. (EEBO, 1664, Hakewill, An apologie)
The majority, however, are used as pronominal predicates that are paraphrasable as ‘just the same thing’ (22) or as AdvP that are paraphrasable as ‘just the same way’ (23). In all cases just the same is anaphoric and comparative in reference.
(22) [About pastors and priests] the practise among the one and the other is just the same. (EEBO, 1621, Calderwood, The altar of Damascus)
(23) words no sooner spoke, but vp she flies: where seene, and question’d how she thither came, she opens the whole matter (just the same as was before related). (EEBO, 1635, Heywood, The hierarchie of the blessed angels)
The data set from CLMET 3.0 suggests that just the same was not used in British English as a concessive connector. It is used with the same functions as phoric all the same: quantifier N, predicate phrase, and AdvP. Concesssive just the same appears in only two examples, in one text from 1901 (Kingsley, History of Richard Calmady, CLMET 3_0_3_255), see (24) below, and can therefore not be considered to be conventionalized as a concessive in that data set. The absence of just the same in Lenker’s (2010) list of contrastive/concessive connectives that arose in the Late Modern English period lends support to the conclusion that it is not a widely used expression in British English.
Example (24), one of the two examples mentioned above, by the same author in the same text in CLMET 3.0, is an outlier best interpreted as a concessive because phoric AdvP ‘just the same way’ is excluded for people who always had something wrong. The negative (hadn’t let it interfere) and but support the interpretation of just the same as evoking ‘normally one would expect people to let their disability interfere with doing things, but these people did things nevertheless.’
(24) He told me about a number of people he’d known who had got smashed up somehow, or who’d always had something wrong, you know, and how they’d put a good face on it and hadn’t let it interfere, but had done things just the same. (CLMET 3_0_3_255, 1901, Kingsley, History of Sir Richard Calmady)
As mentioned in section 5.2, all the same is used twenty times in Kingsley’s text. It is possible that the author analogized concessive just the same on concessive all the same.
The data set from COHA shows that prior to the 1870s, just the same was used only with phoric function and in the same functions as phoric all the same. Of these functions, as in the case of all the same, the adverbial use appears to have been most important in the development of the concessive. Just the same begins to be attested with concessive meaning in COHA in the second half of the nineteenth century in several texts. In other words, it is conventionalized in the COHA data set. An early example is (25) from 1870.
(25) [About a prank making the birthday cake disappear, which dismays the party-goers] [. . .] burst out laughing in spite of her woe. “It was so funny to see him spin round and walk on his head! I wish he’d do it all over again; don’t you?” “Yes: but I hate him just the same.” (COHA, 1870, Alcott, Under the lilacs)
From the 1920s on, concessive use comes to be used with increasing frequency and is dominant from the 2000s on. Like all the same, it can be found in phrasal X just the same expressions as in (26) and in Thanks just the same formulae. Lack of evidence for use in identifiable new contextual assemblies, such as are found prior to the development of all the same concessives, suggests that concessive use of just the same may have been analogized to all the same.
(26) proposing that all society should be communal. Not militaristic but communal just the same. (COHA, 1982, Bellow, Dean’s December)
As in the case of all the same, there is a small number of undecidable cases of just the same. These are attested after the first unambiguous case, but within such a short time frame that this is likely to be a happenstance of the data set. For example, (27) is dated 1875, and the earliest concessive example in the data set is dated 1867.
(27) “And do you not know, Mrs. Temple [. . .] that, if your child has been baptized by Father Duffy, that is sufficient? There is no need for our ceremony to-morrow” [. . .]she was intent on having her baby christened at St. Mark’s. “But, Dr. Browne, nobody knows my baby has been baptized. Can not the christening go on just the same?” (COHA, 1875, Lee, Hubert’s wife)
In (27) just the same can be interpreted as an AdvP (‘in just the same way as planned’) or as a concessive ‘nevertheless’ countering ‘that is sufficient.’
Regarding position of concessive uses of just the same, relative to the clause, clause-final position is preferred throughout COHA (see 25 above and 28 below). There are, however, also a number of instances of use in initial position, as can be seen in (29). None was found in Med1 position, but eight appear in Med2 position, as in (30), where, like all the same, they focus what follows (typically a complement clause).
(28) “We know he’s a spy and a traitor, [. . .] But we deal with him just the same – because we have to.” (COHA, 1913, Dana, Within the law)
(29) “I – I – didn’t mean to – to – kill him.” “I know you didn’t. Just the same that is a dangerous river.” [. . .] (COHA, 1905, Stratemeyer, The Rover boys)
(30) the socks were two sizes too large, but he wore them just the same to give her pleasure. (COHA, 1957, Witness for the prosecution)
The CLMET 3.0 and COHA data sets for just the same are summarized in Table 3. As in Table 2, the number of undecidable examples is given in parentheses. Note that, as mentioned above, the two instances of clause-final concessive use of just the same that appear in in CLMET 3.0 in 1901 do not occur in other texts in the data set, so concessive use of just the same is considered not to be conventionalized in the CLMET 3.0 data set. However, in COHA there are 197 instances of concessive use out of 414 hits (48 percent, rounded) in several texts from 1845 onward. This shows that just the same is conventionalized in the COHA data set.
Number of Concessive Hits of just the same in Different Positions in CLMET 3.0 and COHA
4.3. The Rise of Concessive Uses of at the same time
Among the many prepositional phrases with same that were in the repertoires of Early Modern English speakers, one, at the same time, came to be used frequently not only in its lexical, temporal sense, but also, from the nineteenth century on, occasionally with concessive meaning. According to the OED (s.v. time P3, c. iii), at the same time can be “used in introducing a reservation, nevertheless, yet, still.” An additional meaning cited is “as should be borne in mind.” Similarly, Google Dictionary Online lists as a second meaning “nevertheless (used to introduce a fact that should be taken into account).” Both definitions imply that at the same time can be used to expand on an argument with a contrastive elaboration that draws attention to lack of complete compatibility between clauses. As is discussed below, such elaboration, paraphrasable by ‘in addition, furthermore,’ is attested in both CLMET 3.0 and COHA, especially in earlier texts.
The corpora investigated include a large number of hits of at the same time, most of them lexical. Lexically, time can be used to refer to clock time (“what time is it?”), or a period of time (“winter is the time to ski”). At the same time can likewise be used lexically to refer to simultaneous moments in time (“they laughed at the same time”), or to similar, usually overlapping, periods of time (“We can’t fight a war in the Mid-East and at the same time fight a war with your Congress”). It is the latter that by hypothesis underlies concessive use.
Data from the last five decades of EEBO (1650-1700) show an increase in the use of at the same time in raw numbers from 703 in the 1630s to 7251 in the 1690s. Of these, a scattering appears to be used textually, to “introduce a fact that should be taken into account.” (see the quotation from Google Dictionary Online above) They can almost always be understood as expressing overlap in time. However, temporal overlap may be less important in the development of the discourse than drawing the reader’s attention to how to interpret the text and conclusions to be drawn from it. Consider (31), which is part of a series of warnings and advice.
(31) but to return to your scuffle: your imparting that affair to the lord bishop of clogher, and the speaker of the house of commons, seems to be a very good imprimatur; but at the same time, it is incumbent upon you to take care, that there be nothing in it, too mean to deserve such a license. (EEBO, 1699, Dunton, The Dublin scuffle)
Since the addressee already plans to impart the affair, it is incumbent on you pertains to the future, not a simultaneous state of affairs. It appears that the writer uses but to draw attention to the contrast he is making between praise for the plan and recommendation for its execution, and at the same time not to draw attention to the simultaneity of praise and recommendation, but to the importance of the upcoming recommendation. The discursive structure invites not only an elaborative ‘furthermore’ interpretation, but also a weakly concessive one (‘although your telling what happened seems to be a good plan, nevertheless take care how you present your case’). This kind of underspecification is typical of examples of at the same time.
In the last decades of EEBO there are several examples of at the same time that, a complex phrase with subordinating, backgrounding function. There is no evidence in the data that all the same or just the same were used as subordinators in this way at any point in their histories. While most of the hits of at the same time that are temporal, in philosophical and argumentative works conflicting points of view may be presented, implicating concessive reasoning, as in (32).
(32) though an ill man can not by praising confer honour, nor by reproaching fix an ignominy, and so they may seem on equal terms; yet there is more in it; for at the same time that we may imagine what is said by such an author to be false, we conceive the contrary to be true. (EEBO, 1673, Hickeringill, Gregory, Father Grey-beard)
Example (32) is structured as a reasoned discourse with similar syntactic patterning (“though X, yet Y”; “for X, Y”). At the same time that following for can be paraphrased as although.
Five subordinating uses of at the same time that are attested among the 415 hits in the CLMET 3.0 data set, as in (33).
(33) There is, in reality, nothing to prevent our regarding man as specially endowed with an immortal spirit, at the same time that his ordinary mental manifestations are looked upon as simple phenomena resulting from organization. (CLMET 3_0_2_160, 1844, Chambers, Vestiges of the natural history of creation)
Here the subordinate ‘although’-type clause follows the first assertion and pragmatically backgrounds his ordinary mental manifestations are looked upon as (p) to there is nothing to prevent (q). The texts in which subordinate at the same time that appear are argumentative and reasoned.
Most of the remaining 410 instances of at the same time in the CLMET 3.0 data set are used in coordinate contexts that could be interpreted as temporal AdvPs, referring to overlapping periods of time. However, sixty-three can be interpreted as either elaborative or concessive connectives. Most occur in the context of metatextual comments on the text, such as explicit mention of evidence providing (it must be observed that), or text creation (it must be confessed/remembered that). In (34) we find clearly, evidence, it will be observed, and we could only expect, all from the domain of reasoned metatextual discourse.
(34) If there be, it clearly would form a strong evidence in favour of the doctrine [of creation by law], . . . It will at the same time be observed that, the earth being now supplied with both kinds of tenants in great abundance, we only could expect to find the life-originating power at work in some very special and extraordinary circumstances. (CLMET 3_0_2_160, 1844, Chambers, Vestiges of the natural history of creation)
Temporal meaning is weak here. Chambers appears to be using at the same time to draw the reader’s attention to what follows, which is the statement that there are exceptions to evidence in favor of the doctrine; in other words, it draws attention to at least weakly concessive meaning and assesses the truth of the first statement.
Of the sixty-three metatextual uses, five appear to be primarily elaborations, as in (35). Note the use of the coordinator and, which is coherent in an elaborative discourse.
(35) In the various and numerous memoirs, which have been published of the late Mr. Coleridge, I have been surprised at the accuracy in many respects, and at the same time their omission of a very remarkable, and a very honourable anecdote in his history. (CLMET 3_0_2_120, 1847, Cottle, Reminiscence of Samuel Taylor Coleridge)
The first likely example of concessive use in the CLMET 3.0 data set is (36) from 1809. The concessive interpretation is dependent on but, and the following conditional, modal should and negative (if it should not suit you), all of them supporting concessive meaning:
(36) I was tempted by a pretty coloured muslin, and bought 10 yds** of it, on the chance of your liking it;-but at the same time if it shd* not suit you, you must not think yourself at all obliged to take it. (CLMET 3_0_2_128, 1809, Austen, Letters to her sister; asterisks original)
In (36) a temporal reading is dispreferred, as is an elaborative one. Few examples are as unquestionably concessive as (36), as few have as richly concessive contexts. For example, (37), from the same file as the highly specified example (34), is underspecified for concessivity.
(37) But I believe my doctrines to be in the main true; I believe all truth to be valuable, and its dissemination a blessing. At the same time, I hold myself duly sensible of the common liability to error, but am certain that no error in this line has the least chance of being allowed to injure the public mind. Therefore I publish. (CLMET 3_0_2_160, 1844, Chambers, Vestiges of the natural history of creation)
Example (37) is interpretable as (a) temporal (‘I simultaneously believe all truth to be valuable and hold myself duly sensible . . .’), (b) elaborative (‘I believe all truth to be valuable and furthermore hold myself duly sensible . . .’), or (c) concessive (‘I believe all truth to be valuable; nevertheless, I hold myself duly sensible . . .’). Of these, the first appears to be the least likely, but nothing excludes it. At issue here is not ambiguity, the possibility of two or more distinct interpretations, but rather pragmatic inferences that arise out of the textual design and lack of discursive specification.
In COHA, as in CLMET 3.0, a few examples (thirteen) seem to be more plausibly interpreted as elaborative than temporal, such as (38).
(38) [About a statue of the Roman emperor Augustus] It is a noble work, and every minute detail of the ornamentation has a force and meaning that can be explained. At the same time the whole work is full of strength and dignity, which comes from the character of the man himself. (COHA, 1887, Water, A history of art)
A “noble work” could be expected to be full of strength and dignity, so the content of the sentence introduced by at the same time is fully compatible. However, the relative clause which comes from. . . provides new information, and the function of at the same time appears to be to point to that elaboration.
Also, as in CLMET 3.0, likely concessive uses appear in assessing contexts in COHA. In (39) note but, I say, in sober judgment, and the negative stance of it will not do to treat too lightly. At the same time evokes ‘those who believe the Constitutional Union too strong to be shaken will treat dangers to it too lightly’ and counters the evoked belief.
(39) I honor that feeling which believes the Constitutional Union too strong to be shaken. But at the same time I say, in sober judgment, it will not do to treat too lightly the danger which has beset and which still impends over us. (COHA, 1859, Davis, Speeches)
Although twenty hits of at the same time in COHA were coded as concessive, a pragmatically inferable temporal reading can in most cases not be excluded. The corpus examples confirm König’s (2020:54) comment that at the same time is not a “dedicated” concessive, by which he means that concessive interpretation is underspecified and contextually dependent.
The corpus data sets for at the same time are summarized in Table 4. This Table shows how infrequent clear concessive uses of at the same time are in the data set: forty-one examples of concessive use in CLMET 3.0 represents 10 percent (rounded) of the 415 hits in the British English corpus, and twenty examples in COHA represent only 5 percent (rounded) of the 422 hits in the American English corpus. As in Tables 2 and 3, the number of undecidable examples is given in parentheses. Unlike in Tables 2 and 3, an elaborative function is listed in addition to a concessive function.
Number of Concessive and Elaborative Hits of at the same time in Different Positions in CLMET 3.0 and COHA
Most concessive and elaborative examples appear in clause-initial position, but a few appear in both Med1 and Med2 position in CLMET 3.0. Three examples each of concession and elaboration occur at Med2 in the same text from 1844 in CLMET 3.0, Vestiges of the natural history of creation, a treatise by Robert Chambers. Apparently, concessive use at Med2 was a special feature of Chambers’s style as it is not attested in any other text in the CLMET 3.0 data set. There are no examples in either corpus of concessive use of at the same time in clause-final position. Therefore, whereas all/just the same may be used in polite rejections of an offer (thank you all/just the same), at the same time is not used in this formula.
4.4. A Summary Comparison of the Rise of the Same-Concessives
In this section I summarize the main findings about the similarities and differences between the three same-expressions in CLMET 3.0 and COHA. First, although it is possible to interpret some context-dependent uses of at the same time as potentially concessive in the Early Modern English period, the emergence of non-context-dependent, conventionalized and constructionalized concessive uses of the constructions is a Late Modern English development (see Table 5). In both CLMET 3.0 and COHA, probable conventionalized concessive use of at the same time is attested in the early nineteenth century, later than Lenker (2010:180) mentions (1710-1780).
Appearance in the Database of Conventionalized Use of the same-concessives
As earlier Tables have shown, all the same is the most productively used of the three constructions, and concessive use of at the same time is relatively infrequent.
Other findings in sections 4.1-4.3 are:
All three concessive expressions arose out of phoric expressions used to identity a prior referent.
The phoric expressions all the same and just the same function as AdvP and sometimes as a pronominal predicate phrase. By contrast lexical at the same time is an adjunct prepositional phrase and functions only as AdvP.
By hypothesis, the main source of all the same and just the same is an AdvP; the only source of at the same time is AdvP.
Concessive use of all the same is repeated by one individual British author from the 1840s on in the CLMET 3.0 data set. There is no evidence in the data set that just the same was conventionalized as a concessive in Britain.
All in phoric all the same is ambiguous between the quantifier ‘every’ and the degree modifier ‘exactly,’ but just in phoric just the same is unambiguously a degree modifier meaning ‘exactly.’ At the same time does not involve a degree modifier.
At the same time that used as a concessive subordinator is attested in the nineteenth century, but neither all the same nor just the same are used as subordinators.
Just the same is preferred in clause-final position. At the same time is not used there. All three are used clause-initially and occasionally clause-medially.
Concessive use of at the same time is interpretable mainly when but is present or content is clearly contrastive. In other contexts, it may be used as an elaborator of the argument. Many non-temporal uses are discursively underspecified.
5. Discussion
The rise of concessive uses of all the same, just the same, and at the same time supports recent hypotheses that concessivity concerns degree of compatibility rather than incompatibility (see discussion of Gast’s [2019] and König’s [2020] proposals in section 2.1). It also has implications for several theoretical concepts in diachronic construction grammar. Here I discuss briefly how the three expressions provide evidence for: (a) how we recognize a new construction (constructionalization) in a data set (section 5.1), (b) types of enabling contexts for constructionalization (section 5.2), and (c) the concept of a “family” of constructions (section 5.3). Finally, in section 5.4, I consider (d) the role of position relative to the clause, and to what extent the data support Traugott’s (2022a) hypothesis about the role of topicalization of expressions for the rise of discourse structuring markers and Lenker’s (2014) hypothesis concerning medial position.
5.1. Recognizing Constructionalization
As mentioned in section 2.2, how best to think about constructionalization (i.e., the coming into being of a new conventionalized construction) has been debated. The approach adopted here is that constructionalization is the conventionalized establishment of a form-function pairing (Traugott & Trousdale 2013). Conventionalization requires use by at least two speakers/writers (any larger number is arbitrary). The CLMET 3.0 data for just the same is noteworthy in this regard. As stated in section 4.2, there are only two instances of just the same used concessively, in 1901, both in the same text by the same author (Mary St. Kingsley, whose pen-name was Lucas Malet). Based on the data set, it must be concluded that concessive use of just the same is not conventionalized in British English. However, a search for just the same in later texts in CLMET 3.0 that are not in the data set reveals scattered concessive uses. It must therefore be concluded that just the same was in fact conventionalized as a concessive in British English. However, because it is used infrequently, it is only marginally so. This result serves as a reminder of the data-dependency of findings.
Another criterion for constructionalization is that the new form-meaning pairing is attested independently of context. This is easy to show with respect to formal contexts. All three same-expressions cooccur in the formal context of the preceding discourse markers but and yet (e.g., examples 14, 28, and 36). All three also occur without a contrastive marker and can be considered to be established pairings of form and function that have become non-compositional and have been replicated across a network of language users. They are part of speakers’/writers’ knowledge of language. They are polysemous with the phoric constructions of the same form, but have different syntax (conjunct) and semantics-pragmatics (concessive and, in the case of at the same time, elaborative).
Independence from discursive context is less easy to show, since coherent discourse is rhetorically structured (see, e.g., Mann & Thompson 1988, 1992). Argumentative, reasoned discourse is moved forward by concatenations of clauses expressing cause, condition, concession, and contrast, all of which can be marked linguistically, but do not need to be (Couper-Kuhlen & Kortmann 2000). The extent to which a specific marker activates a particular meaning absent larger discursive context helps the analyst assess degree of independent constructionalization. In the constructed example (40), at the same time does not readily activate a prior context of a state of affairs that might seem to contradict I love you. Concessive use of at the same time can be hypothesized to be only marginally constructionalized.
(40) At the same time, I love you.
However, in (41) and (42), all the same and, in American English, just the same, can readily activate a prior context such as a reason why the speaker/writer might not love the addressee. The two connectives can therefore be hypothesized to be fully constructionalized as concessives.
(41) All the same, I love you.
(42) Just the same, I love you.
5.2. The Rise of the Same-Concessives and Factors that Could Have Enabled It
What might have enabled the development of concessive uses of the three same-expressions under discussion? Four potential enabling factors are mentioned here: the inherent meaning of same; ambiguity of meaning; analogization to extant concessives; and the role of individual authors. As shown below, none of these alone was deterministic, but together they seem likely to have contributed to the changes discussed in section 4.
With regard to the inherent meaning of same, recall Breban’s (2010a:236n) comment, cited in section 1, that saying that two things are the same “always implies non-identity of reference in order to assert the identity of the referent.” In other words, use of same evokes alternatives. This evocation of alternatives was by hypothesis reinforced in contexts of negation and contrastive content, often marked by a contrastive discourse marker such as but or yet. However, the inherent meaning of same was probably not determinative, as other multifunctional idiomatic expressions with same such as in the same way and by the same token are not interpreted as concessive, although in certain contexts the latter can be interpreted as contrastive (see Pinson 2023), where evidence from repairs in conversation shows that at the same time sometimes “interferes”).
A generally recognized enabling factor in change is ambiguity between an older and a later meaning. In fact, the enabling power of ambiguity has been a tenet of much historical work, including grammaticalization (see, e.g., Diewald 2002). As mentioned in section 3, whether there is ambiguity in form can be relatively easily determined, but whether there is discursive ambiguity is harder to assess. As corpora represent only a subset of texts, and, in earlier centuries, mainly texts written or read by a small elite group of literate language-users, it is often not possible to ascertain what resources were available prior to the texts in the corpora we have. Suffice it to say that formally ambiguous contexts do not appear in the data set for all the same and just the same prior to the earliest examples that can be analyzed with some certainty as conventionalized units. However, a few do appear later, for example, examples (17) in section 4.1 and (27) in section 4.2. In sum, the data set investigated provides no evidence that concessive uses of either all the same or just the same arose as a result of formal ambiguity or discursive underspecification, although it is not inconceivable that they enabled the changes. By contrast, there is ample evidence for ambiguous and underspecified use of at the same time both before and after the emergence of its use as a concessive (see example 37 in section 4.3). We may conclude that although ambiguity and underspecification may be factors in change, and indeed are likely to be, they may not be necessary for change to occur (at least in written texts). The content and function of discursive contexts may be more important drivers of change.
Other likely factors in the development of the three expressions are analogy and pattern match. There was an already available set of concessive markers that I have called a Concessive.Schema (see section 2.2). This set presumably enabled analogical thinking. Two subordinating members were while (in its concessive function) and although. Concessive use of while originated in a temporal expression the while that ‘the time that’ and may have been a model for the subordinating pattern at the same time that. However, the potential for further development appears not to have been pursued for at the same time that, possibly because while and although were already well entrenched.
Another member of the Concessive.Schema was the contrastive/concessive coordinator, nevertheless. As corpus searches show, nevertheless is attested in EEBO from the 1610s on. Historically it involves a fixed unit (the less), like the same. It could also be used in a Determiner N phrase (see 43) or as a pronominal predicate (see 44). Negative never evokes an extreme end of a scale. An additional coordinating connective that was used with concessive function from about 1700 on, especially in clause-final position, was after all. In short, in the 1800s, the period when concessive uses of all the same and at the same time were developing, there were already coordinate concessive constructions in the repertoire of language-users of the time. These are likely to have allowed for analogical thinking and to have served as partial models for the new uses.
(43) if he preaches treason or sedition, his opinion is not his excuse, because it brings in a crime, and a man is never the less traitor, because he believes it is lawful to commit treason (EEBO, 1648, Taylor, Treatises)
(44) humane judicatures will punish those sins which are most secretly committed, when they happen to be discovered; for the sin is never the less, nor does it less deserve to be punished, for being secret (EEBO, 1692, Sherlock, A practical discourse)
In the nineteenth century, all the same and just the same were added to the Concessive.Schema. As mentioned in section 4.2, the late development of just the same suggests that it was probably analogized to the already conventionalized concessive uses of all the same. Such exemplar-based analogization (Bybee 2010:73-73) is particularly plausible because all, like just, is a degree modifier meaning ‘exactly’ in one of its interpretations.
At the same time can be thought of as a somewhat marginal member that is linked to both the Concessive.Schema and to the “Elaborative.Schema,” which, like the Concessive. Schema, is a member of a larger, more abstract “Connector.Schema” (Traugott 2022a:228). The network of constructions that is hypothesized to be stored in a speaker’s mind is complex and multifaceted (Diessel 2019), so there are many additional links, for example to lexical constructions such as time.
Finally, another factor that may have enabled development in the case of concessive all the same is that some individual authors may have been especially influential, and their discursive patterns may have been adopted by a wide number of other speakers and writers. A striking feature of data from CLMET 3.0 is that some authors repeat concessive use of all the same to the exclusion or near exclusion of the phoric use. For example, Browning and Barrett (CLMET 3_0_2_185, 1845-1846) use all the same with concessive use eleven times in their correspondence and non-concessive use only once; Linton, in her Autobiography of Christopher Kirkland (CLMET 3_0_3_195, 1885) uses it sixteen times, all with concessive meaning; and twenty out of twenty-one uses are concessive in the excerpt from Kingsley’s The History of Richard Calmady (CLMET 3_0_3_255, 1901). On the other hand, although all three writers use at the same time, they do not use it concessively. The disproportionate number of examples of concessive all the same found in individual texts in the CLMET 3.0 corpus supports recent findings that individual usage may impact community knowledge (Petré & Anthonissen 2020).
5.3. The Concept of a “Family” of Constructions
It has been suggested that constructions with similar form and function may be organized in various ways. What level of abstraction is appropriate for analysis has been debated. One conclusion might be that the three same-concessives under discussion are equal members of the higher-order, more schematic Concessive.Schema. If so, they could be assumed to be equally constructionalized, to occur in similar contexts and with similar frequency (see Hilpert 2013:209). An alternative analysis, and the one adopted here, is that they may form a loosely-knit cluster or “family” of closely networked alternatives (Hilpert 2013; Iza Erviti 2021). Another metaphor about closeness of relationship is formation of a “paradigm,” a set of available options for expressing the function at hand, of which one is used more frequently than the other (Ruiz de Mendoza Ibáñez, Oyón & Pérez-Sobrino 2017:4). All the same and just the same appear to be alternative options in a small paradigm in American English of which all the same is the more frequent, “unmarked” member, because they have similar distributions and can therefore be considered to be closely networked. However, at the same time is sufficiently divergent in distribution to be considered an outlier and not fully integrated with the other same-expressions (e.g., it is not used in clause-final position; see section 5.4). It may be concluded that the three same-concessives form a “family.” Whether other coordinate concessive connectives such as nevertheless can be considered to be part of the family remains to be investigated, again on grounds of distribution.
5.4. Same-Concessives and Hypotheses about Position of Connectives
Where a connector occurs relative to a clause has drawn considerable attention recently. Biber, Johansson, Leech, Conrad, and Finegan (1999:890-891) point to genre-based factors in the distribution of connectors other than those that only occur clause-initially in standard speech and writing (e.g., and, but, and so). They found that connectives that are mobile, like however, are favored in clause-initial position in academic prose, where clause-medial use is also relatively common. In conversation, however, clause-final use is relatively frequent and medial use is rare.
Lenker (2010:198) investigates the historical development in the eighteenth century of adverbials like however and proposes that, while use in clause-initial position guides the addressee through the text, “final connectors expressing relations of CAUSE, CONTRAST and CONCESSION [. . .] force a re-processing or even reinterpretation of the preceding assertion.” “Force” might imply that the addressee always understands the function of the clause-final use and processes the associated clause as a concessive. “Invites a re-processing or re-interpretation” seems more appropriate, given the asymmetry of production and perception and given the contextual underspecification of several examples. In her 2014 synchronic study of medial uses of connectives like however, Lenker proposes that in medial position the connectives serve an information-structuring function. At what I call Med1 (after topicalized adverbial or subject) they demarcate topic (what the clause is about, usually familiar and accessible material). At what I call Med2 (after the finite verb), they demarcate the focus, also often contrastively.
Taking a different but related approach to the clausal position of connectives in Traugott (2022a, 2022b), I observe that many connectors in English derive from adverbial adjuncts, like after all. These source adverbial adjuncts are readily topicalized in clause-initial position to frame what follows. EEBO evidences several examples of topicalized lexical at the same time. Most are found in translations from Latin or French, as in (45).
(45) made a great noise in the teple of hector: at the same time the dores of hercules temple at thebes being fast shut with barres, opened sodainly of their owne accord. (EEBO, 1572, Harrison, Of ghosts and spirites)
‘made a great noise in the temple of Hector. At the same time the doors of Hercules’s temple at Thebes, although shut fast with bars, opened suddenly of their own accord.’
This is the kind of context that I hypothesize is a necessary (but not sufficient) step in the shift from adjunct PP to conjunct discourse marker status (Traugott 2022a:235). However, there is no evidence in the data set that adverbial all the same (way) or just the same (way) were used in topicalized position prior to the development of concessive uses of these expressions. Topicalized instances in COHA of in just the same way (46), are not counterexamples as they have PP structure; furthermore, concessive uses of just the same are clause-final. It must be concluded that my hypothesis that topicalization of an adverbial is a necessary step in the development of discourse structuring connectives is too strong. Topicalization may be an enabling factor.
(46) Thus a man is honorable when he is true to his own personal convictions; pious, when he is obedient and true to God, and in just the same way he is loyal when he is devoted and true to his country (COHA, 1863, Bushnell, The doctrine of loyalty)
When it is first attested as a concessive in the data set, all the same is used in both initial and final position, the position typical of its phoric AdvP source. However, as mentioned above, just the same is initially attested only in clause-final position. Its later use in initial position as well supports the hypothesis that concessive just the same was analogized to all the same. The clause-final concessive uses of all the same and just the same support Lenker’s (2010:198) hypothesis that the addressee is invited to re-process the relationship between the preceding assertions. For example, on reading (47), which is adapted from (25) without the concessive marker, the reader might interpret the relationship between Yes (I wish he’d do it all over again) and I hate him to express ambivalence about contrasting viewpoints (see but). Further, whether I hate him applies generally or at the moment is left open.
(47) “It was so funny to see him spin round and walk on his head! I wish he’d do it all over again; don’t you?” “Yes: but I hate him.”
With clause-final just the same, the addressee is explicitly invited to reassess possible interpretations and to conclude that the speaker meant ‘although I wish he’d spin around and walk on his head again, I hate him (for making the cake disappear).’ The stance is contingent on the specific occasion. As discussed in section 4.3, at the same time is favored in initial position and does not occur with concessive meaning in final position. This is one of the factors that suggests it is an outlier in a family of coordinate concessive constructions, not fully integrated into a schema.
Given that the corpora include academic prose (categorized as “Treatise” in CLMET 3.0), there are surprisingly few examples of same-connectives in clause-medial position. There are too few examples of medial use in the data set to draw any solid conclusions whether the same-concessives support Lenker’s (2014) hypothesis about the division of labor between pre- and post-verb function. It should, however, be noted that one of the pre-auxiliary examples of all the same that appears in COHA does not support her criteria (shown in 48).
(48) Construct for me a code B of laws which permit texts to be assessed as “anti,” and I will undertake, for any text which is so assessed, to construct a text which can not be assessed according to code B, but which all the same will be understood as an opposition text. (COHA, 1984, Watzlawick, The invented reality)
In the larger context of the sentence, the but which all the same. . . clause parallels the preceding text which can not . . . clause. In this context all the same focuses will versus can not; in other words, it focuses what follows, not what precedes, as Lenker (2014) would predict since which is clause-initial and functions as subject. This is a one-off in COHA, but if more examples of this type are found, the division of labor Lenker (2014) proposed may need some modification.
6. Conclusion
Concession is a relationship signaling degree of compatibility between the content of two or more contributions in a discourse. Investigation of the current uses and the development of all the same, just the same, and at the same time has thrown the spotlight on the ways in which language users can draw on a phoric construction expressing identity, in this case the same, to mark this complex cognitive capability.
Cross-linguistically, sources of many concessive markers involve simultaneity and similarity, as mentioned in section 2.2. Further research could investigate the cross-linguistic generalizability of contexts enabling concessive meaning identified for the same-concessives, and whether any borrowing has occurred.
As discussed in section 4.2, the data set from CLMET 3.0 provides no evidence that concessive use of just the same was conventionalized in British English. However, there are a few examples in texts other than those in the data set. 7 Whether minimal use of concessive just the same in British English is a happenstance of the corpus or a real difference between American and British English deserves further inquiry using a larger data set.
As the data used for this study are written, prosody could not be investigated. However, prosodic differences in different positions are likely to be found in historical corpora of speech and are worthy of comparison with contemporary speech. Clause-final prosody across varieties of English can be expected to be a rich locus for analysis as clause-final position has been shown to convey different pragmatics than initial and medial uses of some connectives such as anyway (for an early study, see Ferrara 1997).
Footnotes
Acknowledgements
I would like to thank Graeme Trousdale for discussing the family of same-concessives. I am particularly grateful to the editors, Alexandra D’Arcy and Peter Grund, and four anonymous reviewers for their substantial and constructive feedback, which greatly improved this work.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
