A corpus-based study of grammatical post-metaphorical expressions

2.1 Principles of Context first and AS IF

Principles of Context first and AS IF are proposed by Yang (2019) chiefly for the purpose of identifying interpersonal metaphor of mood and modality, although principle of Context first could be applied to the other kinds of GM. According to Yang (2019), the specific context, in which the speech event occurs, shall play an important role in determining the occurrence of interpersonal metaphor. Although the Context first principle has not been elicited as a means of identifying GM, this principle de facto has been indicated by Halliday (1994) and some other systemicists (Lassen 2003a, 2003b; White 2008).^[4] Yang (2019: 193) quoted the two clauses in (6), which were employed by Halliday (1994: 366) to exemplify the necessity of considering context while getting a full understanding of them, and expounded that either clause can be regarded as the metaphorical expression of the other. In other words, the imperative (6a), in the context of demanding goods-&-services, fulfills the speech function of command in a natural way, and hence (6b) is an instance of metaphor of mood. However, the declarative clause in (6b), while in the context of giving information, is realized congruently, and (6a), therefore, is metaphorical.

The principle of Context first can also be employed to identify cases of interpersonal metaphor of modality. Considering the two instances in (7), I think in (7a) is an instance of metaphor of probability,^[5] but it, in the situational context of a doctor giving advice to an advice-seeker, can be understood as an interpersonal metaphor of obligation, because, in so doing, it is easier for the advice-seeker to accept the doctor’s advice. Similarly, this explanation also applies to (7b). From the vantage point of linguistic context, the parenthetical construction I think and I hope in (7a) and (7b) respectively attenuate the obligatory force of the corresponding proposition of it is a great time to try the pill in (7a) and you consider posting a follow up in (7b).^[6] Specifically, the modal expressions of probability I think in (7a) and I hope in (7b), together with the linguistic context, realize the modality of obligation denoted by it is obligatory for you to take the pill and it is obligatory for you to consider posting a follow up respectively, and thus metaphor of modality occurs.

Besides the principle of Context first, the principle of AS IF can also be employed to identify interpersonal metaphor. According to Yang, the principle of AS IF can be defined as the determination of interpersonal metaphor “by an AS IF relation between the default congruent utterance and the metaphorical counterpart” (Yang 2019: 194). With respect to speech functions, giving or demanding goods-&-services are congruently realized by an offer in the form of a declarative or a command in the form of an imperative respectively, while giving or demanding information are congruently fulfilled with the forms of a declarative functioning as a statement or an interrogative functioning as a question respectively. The speech functions can find their congruent expressions exemplified in (8), which is a conversation between a detective and a little girl who are talking about the bad guy Max.

The interrogatives Is something the matter here and Are you a policeman perform the function of questions in demanding information, while the declaratives I’m the police officer of the store and He’s a bad man function as statements in (8). The imperative make him go away is an instance of congruently realizing the function of a command.

In social communications, these congruent expressions can be expressed alternatively. To be more specific, a command in the form of an imperative can be employed to function AS IF it was used as a question, a statement AS IF it was a question, or a question AS IF it was a command. For instance, the statement in (9a) can be understood AS IF it was a question do you have any trouble or do you need any help. Similarly, the question can you take him away in (9b) is expressed AS IF it served as a command denoting take him away. Thus, interpersonal metaphors of mood occur.

The principle of AS IF is also applicable to interpersonal metaphors of modality (Yang 2019). Metaphor of probability is defined as the phenomenon that the clause in (10b), which is used in a context of explicit subjective projection, is employed AS IF it was an implicit subjective clause in (10a), and an explicit objective projection clause in (10d) is adopted AS IF it was an implicit objective clause as in (10c). Metaphor of usuality refers to an explicit objective projection clause as in (11b) is used AS IF it was a usuality realized in the form of an implicit objective clause as in (11a). Obligatory metaphor is articulated in a context that an explicit subjective clause as in (12b) is used AS IF it was an obligatory clause which is realized by an implicit subjective clause as in (12a), and that an explicit objective projection exemplified in (12d) is used AS IF it was an implicit objective clause in (12c). Metaphor of inclination^[7] can find its expression in (13b). An explicit objective clause is used AS IF it was an inclination realized by an implicit objective clause as in (13a).

It is evident from these instances that the principles of Context first and AS IF are applicable to the analysis of interpersonal metaphor, but there is little evidence showing its applicability to the analysis of ideational metaphor. The principle of double functionality, therefore, is employed to identify instances of ideational metaphor.

2.2 Principle of double functionality

GM is one of the items de facto on the lexical-grammatical stratum realizing double functions (He and Wen 2017; He et al. 2015), and hence double functionality can be employed as a principle to identify GM (Fan 2007). In other words, GM occurs on the condition that the acquisition of an extra meaning happens in the ideational domain. Examples (14) and (15) will be adopted to expound this type of phenomenon.

(14b) is the metaphorical expression of (14a) while (14c) is not, for the reason that roof in (14b) functions not only as a participant, but also as a circumstantial element, and thus a double functionality; whereas there is no element in (14c) realizing the double functionality, and hence it is not a metaphorical expression. Comparing with (14), which is an instance of double functionality occurring in different elements in the clause rank, (15) can be exemplified as an ideational metaphor rank-shifting from a clause complex to a simple clause. That is to say, the sequence realized by clause complex in (15a) is remapped by a clause in (15b). In other words, the nominal groups, functioning as elements, the exasperation of the patrons and the gambling of all their lives also function as two figures. The lexical verb preceded functions as a material process and a relator as well. Accordingly, all three elements in (15b) are of double functionality.

2.3 GM cline

The concept of being a GM is relative. Compared with a more congruent expression, it is regarded as a metaphorical one, while compared with a more metaphorical one, the same expression may be treated as a congruent one. Therefore, the development, from congruent expression to metaphorical one and then to a new congruent one, constructs a cline (He 2019; Yang 2018, 2019). See example (16).

(16b) is more metaphorical than (16a) in that the two clauses in (16a) is down rank-shifted into two nominal groups in (16b) and the relator because is class-shifted into a verb realizing the material process; whereas (16c) is more metaphorical if compared with (16b) underlying the fact that the clause in (16b) is further realized by a nominal group. Hence the sequence from the most congruent end to the most metaphorical end forms a cline.

It seems that the GM cline, to some extent, can be employed to account for the pathway that the GPM undergoes. Actually, only part of the whole pathway has been enunciated. That is to say, the rank-shifting from clause complex to clause or from clause to group is made clear. However, GM cline precludes the analysis of componential elements in ideational metaphor and interpersonal metaphor. This research therefore aims at investigating GPM by the analysis of componential elements in GM.

3.1 Corpus

In order to explore the pathway from metaphorical expressions to post-metaphorical ones, we need consider the evolvement of GM to GPM along the time. It is the Corpus of Historical American English or COHA that serves this purpose very well in that COHA contains all together 400 million words, spanning almost two hundred years from 1910 to 2009. Furthermore, it incorporates a wide range of genres such as Fiction, Magazines, Newspaper, and Non-Fiction Books. Genres and sub-genres are balanced in each decade. For instance, in each decade, non-fiction (e.g. history, religion, and technology) accounts for approximately 15% of the total number of all genres, and fiction, including its sub-genres like prose, poetry, and drama, takes amount of 48–55% of the total (Davies 2012). These balanced genres and sub-genres will enable the findings of the investigation to be closer to the changes of actual reality or ‘real world’.

3.2 Data collection

In this sub-section, materials employed for investigation are identified, and procedures that are conducted to elicit the research data are explained. As for materials of identifying GPM features, those of examining the post-ideational metaphor are firstly collected, and those for post-interpersonal metaphor ensued.

4.1 From quality to thing

GM of quality to thing refers to the phenomenon that the grammatical category of adjectives (mostly but not always) functioning as quality are transcategorized into nouns functioning as thing. In this study, together with the attested examples by Halliday and Matthiessen (1999: 246–247), the top five occurrences of words suffixed by ‘-ity’ and the other top five suffixed by ‘-ness’ are considered diachronically. The findings demonstrate that, except the GM feature, the grammatical features of countability and properness are also captured with evolvement of the language, which is shown in the first half of Table 2. According to the table, all attested items, transcategorized from quality to thing, are endowed with the grammatical feature of countability, while only speed, security, and capability among these items capture the feature of nominal properness. That is to say, the ensuing dominant grammatical feature after the GM one is countability, and the nominal properness might be a tendency following countability. These phenomena are exemplified by examples (21a)–(21c) ((21b) is reproduced from example (17b)) and further explicated by the principle of double functionality for the purpose of examining whether GM feature still retains after features of countability and properness are obtained.

According to the principle of double functionality, happiness in (21a) should realize both quality and thing but it in this clause in question has only obtained the grammatical feature of countability (here singularity) while its property of being ‘quality’ is blurred and denies any further effortless unpacking. Pertaining to the case of priorities in (21b), its grammatical feature of countability in dominance is foregrounded, and at the same time, the feature of being ‘quality’, although backgrounded, could also be regained, hence the nominal group the pinch of priorities can be unpacked to the prior pinch. With regard to Securities in (21c), the possible grammatical feature ascribed to it is nominal properness, and there seems to be a distant contiguity to the feature of ‘quality’. From the above statements, the idea that grammatical features of countability and properness are GPM features is further confirmed. However, as to when GM of nominalization begins to capture features of countability and nominal properness and when these features are significantly foregrounded, it is necessary, pertaining to these later occurring features, to resort to a further quantitative scrutinization to the data of COHA, which will be delineated in the following paragraphs.

In order to get a clear outlook on the occurrence of GPM features, further examination of the frequency of occurrence is conducted. We firstly carried out an automatic search and then manually dealt with these raw frequencies. Taking security as an example, the specific process of obtaining frequency of GM and GPM features during 1810s are progressed as follows:

• Step 1: with search queries ‘[security].[nn*]’ and ‘[security].[np*]’, occurrences of singularity (N = 153), plurality (N = 9), and nominal properness (N = 0) are obtained. However, total occurrences of singularity are not solely cases of GM, but also include those of countability denoting GPM features, and hence a further step is adopted to elicit the actual occurrences of GM.

• Step 2: retrieve with the query ‘a security.[nn]’ for the occurrences that belong to the cases of countability (N = 4).

• Step 3: occurrences of GM equal to occurrences of singularity minus occurrences obtained in Step 2, i.e. 153 – 4 = 149; occurrences of countability are occurrences of plurality added by the figure in Step 2, i.e. 9 + 4 = 13.

• Step 4: when comparison is needed, the raw frequency is normalized to per million words.

In the same vein, the same method is implemented to all items that denote thing from quality and to all segments of different time periods. Raw frequencies are shown in Table 4.

With respect to nominal properness, although items such as speed, security, and capability have captured this grammatical feature in each of the 20 decadal phases, yet these occurrences are simply sporadic and generally obtained their status in the twentieth century, specifically, the second half of the century (see Figure 1). Pertaining to the GPM feature of countability, attested items such as security, capability, similarity, happiness, and willingness possess this feature with a large number of occurrences throughout the two centuries, and surprisingly with a gradual increase (see Figure 2). As far as the rest of these items are concerned, their occurrences are either sporadic or at the very inceptive phase. Therefore, it is still inceptive or even should be taken as tentative to claim that GPM is characteristic of nominal properness; nevertheless, countability can be regarded as a typical feature of GPM if compared with their corresponding GM counterpart.

Figure 1:

GPM feature of nominal properness of speed, security, and capability in COHA (normalized).

Figure 2:

GPM feature of countability of security, capability, similarity, happiness, and willingness in COHA (normalized).

4.2 From process to thing

The other source of nominalization considered in this study is the items of thing that are realized by derived nominals from verbs. This phenomenon is evidenced by the two most typical suffixes ‘-ment’ and ‘-tion’ together with items attested by Halliday and Matthiessen (1999: 246–247). Similar to the conduction of the previous two suffixes ‘-ity’ and ‘-ness’, retrieving procedures are employed and five items that topped the list of each suffix are further scrutinized. The general features of countability and nominal properness are listed in the second half of Table 2 from transformation to improvement, which demonstrates that all the attested items characterize grammatical feature of countability, and items such as prospect, attempt, desire, education possession, government, and management are characteristic of nominal properness while the others are not. Witness this phenomenon in examples (22), which are expounded in the light of the principle of double functionality.

According to the principle of double functionality, the grammatical feature of nominalization, pertaining to constructions in (22a), can be captured, while its original feature of process realized by the grammatical category of a verb is blurred. That is to say, the clause with construct realizing process could not be recovered from the nominal group headed by constructions. The case in (22b) is to some extent different. constructions in this clause not only characterizes the function of being a nominal group with the grammatical feature of countability, but also has the potential of being used as a process recoverable. Hence this clause could be unpacked to its congruent form as what Medunetskii and the Stenbergs constructed were first shown in January 1921 at “The Constructivists” exhibition. What should be made clear is that the function of being a thing is foregrounded while that of being a process is backgrounded. In (22c), the nominalized Government possesses the grammatical feature of nominal properness solely, and its feature of GM countability could be by no means recoverable. In the light of the data listed in Table 2, it is still not clear with respect to the specific time period and popularity of the two grammatical features that occurred in COHA. The following paragraphs, therefore, are devoted to this respect.

Raw frequency of occurrences concerned with items mentioned above is outlined in Table 5. With respect to the grammatical feature of nominal properness, it can be shown from Table 5, quite surprisingly, that only prospect among all the items is gradually obtaining this GPM feature (101 occurrences) while the others attested are sporadic and inceptive (20 occurrences or even less). Therefore, it could be said that items from process to thing have been undergoing the inceptive phase. However, GPM feature of countability is considerably full-fledged. Except for the three items education, management, and potential, which lag substantially behind their counterparts of GM (see Figure 3), the others are developing in a considerable number of occurrences. Although the three items are falling behind during the past two centuries, yet the general trend is gradually catching up with their counterparts, especially the beginning of the twentieth century has witnessed the rapid growth of this trend (see Figure 4). A subsequent Pearson’s correlation test in RStudio shows that the GM feature and GPM feature of countability in education are not significantly correlated (p = 0.6991 > 0.05), while the two features in both management and potential are significantly positively correlated (p = 0.000 < 0.01; p = 0.000 < 0.01) respectively (see Table 6). This significant correlation indicates that the rapid growth of GPM feature of countability by and large is gradually foregrounded during the past two centuries.

Figure 3:

GM feature for education, management, and potential (normalized).

Figure 4:

GPM feature of countability for education, management, and potential (normalized).

In summary, concerning grammatical features for postmetaphorization of nominalization, the GPM feature of nominal properness is not developed fully or even in its inceptive phase; on the contrary, that of countability has been obviously increasingly captured during the last two hundred years and gradually preferred by language users.

5.1 It Construction

Based on occurrences of It Construction attested in Table 3, we further subcategorized the construction into modalized and non-modalized ones (see Table 7). The latter pattern is what Halliday (1985, 1994 defined as explicit objective GM of modality, and the former refers to the predicate verb in It Construction being qualified by modal auxiliaries. Once It Construction is qualified by modal auxiliaries, it is no longer characteristic of objectivity^[12] (at least is gradually distant from objectivity if subjective-objective cline is considered). The underlying reason is that modal auxiliaries have experienced the process of grammaticalization whose typical mechanism is subjectification (Hopper and Traugott 1993, 2003; Krug 2000); furthermore, the modalized construction is speaker-related and the speaker’s attitude towards the proposition is foregrounded, while the explicit objective of modality in Halliday’s sense is backgrounded and could not be captured easily. This perspective is facilitated by employing the AS IF principle exemplified in (23).

According to AS IF principle, the explicit objective It construction It is possible in (23a) is employed AS IF it was an implicit objective clause possibly the entire unissued authorization may not be devoted to stock dividend purposes at this time, although there is. Therefore, it is a GM of modality in Halliday’s sense. Different from (23a), It construction it would be possible in (23b) does not necessarily display AS IF it was an implicit objective clause. Specifically, it is not a clause denoting explicit objective orientation, but a clause of explicit subjective de facto. That is to say, It construction denotes the GM of modality in the orientation of explicit objective during its non-modal phase, while the construction is being modalized by modal auxiliaries or in the modal phase, it is in explicit subjective orientation, which is also denoted by what will be elucidated in Section 5.2.

With respect to the items attested concerning It Construction, Table 7 profiles their non-modalized and modalized phases and occurrences of each decadal phase. Table 7 tells us that the five modal adjectives in It Construction are all undergoing the phase of GM of modality with the explicit objective orientation while not modalized. What they differ is that scant occurrences for obligatory (N = 11) and usual (N = 147) seem to be still in their inceptive stage. On the contrary, the other three modal adjectives possible (N = 4,069), necessary (N = 3,991), and probable (N = 1,697)^[13] are considerably developed and seem to be full-fledged, especially preferred by language users during 1920s and 1930s, as Figure 5 demonstrates visually. Surprisingly, among the five items, only the most frequent ones, comparatively, possible and necessary have identified It Construction which is qualified by modal auxiliaries, while this GPM feature is unanimously absent from the other less frequent modal adjectives. Figure 6 diagrams normalized occurrences of modalized It Construction and shows that the growth is a gradual and to some extent levelled process, which is consonant with Hopper and Traugott’s (1993, 2003 view of language change. Compared Figure 5 with Figure 6, we found that the two modal adjectives possible and necessary quickly acquired their GM feature at the very beginning of nineteenth century, while the obtaining of their GPM feature keeps plateauing except for the rapid decrease of necessary during the first two decades in the twentieth century. Another interesting phenomenon is that the GPM feature of obligatory also attested once during the 1890s (see example (24)), although its GM feature is still in the inceptive phase. This case is not considered because of its accidental occurrence.

Figure 5:

Occurrences of attested modal adjectives in It Construction in COHA (normalized).

Figure 6:

Occurrences of modal adjectives possible and necessary in COHA (normalized).

In order to demonstrate the trend between GM feature of explicit objective orientation or non-modalized modal adjectives (i.e. it is possible) and GPM feature of explicit subjective orientation or modalized modal adjectives (i.e. it would be possible) in possible and necessary, the Pearson’s correlation test in RStudio is employed. The results are shown in Table 8, indicating that the two features in both modal adjectives are significantly correlated (p = 0.010 for possible, p = 0.009 for necessary). This correlation articulates that both features of GM and GPM are developing in a similar trend.

Although occurrences for possible and necessary are comparatively scant (see Table 7), it could be concluded from above discussion that the GM feature (non-modalized) of modal adjectives in It Construction precedes the GPM feature (modalized It Construction) and the two are developing towards the same direction. It Construction, however, is just one type of GM of modality, and what the other type, i.e. I think type, demonstrates is the motif being articulated in the following section.

5.2 I think type

The five mental verbs (i.e. think, assume, believe, guess, and suppose) have undergone a process from a simple clause to a modal element. They, in clausal rank, realize mental processes together with personal subjects, but the subjects are generally the first singular personal pronoun I if the projecting mental clauses are regarded as modal elements (Fetzer 2008), which is employed mainly to refer to the GM of modality in explicit subjective orientation in Halliday’s sense. By using SQs10–12, occurrences of initial, medial, and final positions of the five mental verbs with the first singular personal pronoun are detected. Table 9 concludes the raw frequency of these items of I think type attested.

As Table 9 shows, occurrences of initial position of I think type items dominate its counterparts of medial and final positions, which is consonant with Fetzer (2008) who identified that occurrences of initial position of I think outnumber those of other positions. With respect to grammatical features, if those items in the initial position are complemented by a clause with the complementizer that, they still denote ideational meanings of being mental processes; if not, they are regarded, by Halliday (1985, 1994, as explicit subjective GM of probability subsumed into GM of modality. When positioned medially and finally, I think type items are gradually losing their identities of being a modal element in the interpersonal domain, experiencing a process of grammatical change as discourse markers. As a discourse marker, it rarely collocates with TAM (tense, aspect, and modality) (Fetzer 2008). Consistently, these TAM features are hardly attested in the medial and final uses of I think type. Such cases positioned medially and finally are illustrated in example (25), in which I think in both clauses serves as discourse markers expressing pragmatic function of attenuating the propositions, instead of fully (it might be partially) denoting explicit subjective GM of probability. According to the principle of AS IF, I think in medial and final positions cannot be expounded AS IF they were denoting implicit subjective, because they themselves hardly denote explicit subjective expressing GM of probability.

Pertaining to I think Type, the grammatical feature of being a discourse marker is regarded as a GPM feature after being used as GM in this study. In order to compare their distribution and paths of development of this GPM feature (being positioned medially and finally in question), a further analysis among those items are progressed. We will illustrate this viewpoint by employing the most typical item I think, and its distribution of occurrences of initial, medial and final position is presented in Figure 7. It shows that the GM use of I think in initial position is preferred by language users gradually and highly excels the other two positions. With respect to its GPM uses, medial position is more preferable than the final position, but both are to some extent increasing gradually. A further correlation test is employed to examine how the paths of development about GM and GPM features are associated. The results show that initial and final positions, although there is no significant correlation between initial and medial positions (r = 0.249 p = 0.208 > 0.05), are extremely positively correlated (r = 0.902 p = 0.000 < 0.01), indicating that the gradually preferred use of GM/initial position is concomitant with the gradual increase of final GPM during the past two centuries in COHA (cf. Table 10).

Figure 7:

Normalized distribution of occurrences of initial, medial, and final I think in COHA (equal totality).

A thorough examination of the other four items of I think type generally corroborates with the trend which is identified from the survey of I think except for that of I guess. Besides, initial and medial positions are also significantly correlated. Consequently, it is safe to conclude that the GM use of I think type in the initial position leads the path of development and the GPM use follows suit.