Metaphors of migration over time

Abstract

This paper aims to cast light on contemporary migration rhetoric by integrating historical discourse analysis. I focus on continuity and change in conventionalised metaphorical framings of emigration and immigration in the UK-based Times newspaper from 1800 to 2018. The findings show that some metaphors persist throughout the 200-year time period (liquid, object), some are more recent in conventionalised form (animals, invader, weight) while others dropped out of conventionalised use before returning (commodity, guest). Furthermore, we see that the spread of metaphor use goes beyond correlation with migrant naming choices with both emigrants and immigrants occupying similar metaphorical frames historically. However, the analysis also shows that continuity in metaphor use cannot be assumed to correspond to stasis in framing and evaluation as the liquid metaphor is shown to have been more favourable in the past. A dominant frame throughout the period is migrants as an economic resource and the evaluation is determined by the speaker’s perception of control of this resource.

Introduction

In this paper, I take a historical perspective to investigating the conceptual metaphors which have been used to frame migration. While there has been extensive analysis of migration discourses in a contemporary time frame, we have relatively little understanding of how longstanding these discourse frames are, how they have developed over time, or what alternative framings might have been lost through time. This is where the kind of long-distance diachronic study made possible by combining corpus linguistics and (critical) discourse analysis can make a vital contribution to understanding contemporary language use. Drawing on a corpus covering a 200-year period, it is possible to track when metaphors become conventionalised and how they have developed. These findings can provide a background to contemporary research and enable insight regarding the extent to which our metaphoric framings are inherited and identification of what is absent from current public discourses of migration. I want to argue that taking a long-distance diachronic approach is not a niche interest but a way to understand the underpinnings of contemporary discourse (see, e.g. Musolff, 2010, 2014 on the state-body metaphor). It can provide a means of showing how present discourses “rhyme” with those of the past (Baker et al., 2013) and whether a ‘representation is new, or simply a modern version of a representation that has deep historical roots’ (2013: 230).

Furthermore, a major affordance of a long-distance view for discourse analysis is the ability to denaturalise the discourse in question and to approach it as a construct. As discourse analysts, when we examine contemporary discourses we are inevitably part of our own object of study and our readers are likely to be in the same position. Thus, it can be difficult to identify and communicate the ways in which discourse represent choices and that, as Fowler (1991: 4) so precisely put it: ‘there are always different ways of saying the same thing, and they are not accidental alternatives. Differences in expression carry ideological distinction (and thus differences in representation)’. When we incorporate historical analyses into a diachronic study, or when we shift entirely to a historical period, we are no longer fully proficient speakers of our texts and this distance can bring insights. We can uncover the ways in which our discourses have been shaped – as when Wilkinson (2019) traces the development of erasure in representation of bisexuality in the Times. We can empirically test claims about modern-day discourse, as in Taylor (2020) who reveals the nostalgic nature of government rhetoric about the UK’s history of welcoming migrants.

Analysing metaphor gives us a practical and theoretical means for operationalising these long-distance goals. In practical terms, metaphor analysis provides a response to the challenge of comparing texts published over 200 years apart which will vary considerably due to the nature of language change. In this context, the identification of comparable units becomes especially salient and we need to abstract out from a purely lexical level (see also Taylor and Del Fante, 2020 on this from a cross-linguistic perspective). For instance, by operating at a macro level, we can trace continuity in a commodity metaphor whether it is lexicalised as conveyance or transport. In theoretical terms, metaphor is an important way into understanding discourse because it acts as a way of understanding the world; thus, the use of metaphor always offers up an interpretation. When we liken one thing to another we do so on a partial basis; or, as Semino et al. (2018: 29) put it, metaphor use ‘highlights some aspects (the similarities that can be established between the two) and backgrounds others (things that are different or irrelevant for the comparison). This helps communicate the [. . .] evaluation, and can facilitate some inferences while making others unlikely’. Charteris-Black (2011: 32) goes further to assert that ‘in political rhetoric the primary purpose of metaphors is to frame how we view or understand political issues by eliminating alternative points of view’ (2011: 32, my italics). An additional component of a metaphor’s persuasive power lies in its ability to evoke emotions. For instance, Chilton (2004: 17) proposes that in metaphor use ‘[c]ertain emotions that can be regarded as in some way basic are evidently stimulated – most obviously fear, anger, sense of security, protectiveness, loyalty’ (Chilton, 2004: 17). The relevance of metaphor to emotional response is also indicated by Abid et al. (2017) who find a greater use of metaphor for describing Syrian refugees in host countries than non-host countries. Furthermore, experimental research has provided evidence for a legitimating impact of metaphor as in Hart (2018) who found that participants presented with fire metaphors in news reports were more likely to consider police use of water cannon as legitimate compared to participants presented with a non-metaphorical text. So, we can see that metaphor helps frame the world and that identification of those frames can give the discourse analyst insight into the viewer’s perspective. Furthermore, when we discuss metaphor in terms of conceptual metaphor we are working with a phenomenon which is cumulative in nature, like discourse. The metaphor (e.g. migrants are guests) emerges when this concept is lexicalised multiple times (e.g. host country, reception centre). What this means is that the analysis can also encompass ‘non-obvious meanings’ (Fairclough, 2001; Partington, 2017); those utterances which taken individually might be unmarked, become salient when they are seen as a group, all fulfilling the same underpinning idea. Or, as Sinclair (1991) put it ‘the language looks rather different when you look at a lot of it at once’ (p. 100).

Metaphors of migration

The large number of metaphors in which migrants are the target discussed in previous research nearly always refer to immigration rather than emigration and so they also often interrelate with conceptual framings of the country which the ‘outgroup’ are living in or moving to. For instance, nation as body, nation as container or nation as family home. The conceptual metaphors discussed in previous research include those which include human as a semantic feature such as invaders, criminals, guests and those which do not, such as water, animals, objects, burden, weeds, pollutants, commodity, parasites, disease, liquid. Montagut and Moragas-Fernandez (2020) establish this humanising/dehumanising distinction as the primary division in metaphors of migration. They see dehumanising metaphors as being intrinsically problematic while humanization metaphors which ‘endow the migrant with political action’ (Montagut and Moragas-Fernandez, 2020: 79) may position refugees either as subjects (positively realised by journey metaphors) or as a source of conflict (negatively realised as conflict metaphors). Similarly, Soto-Almela and Alcaraz-Mármol (2019) emphasise the role of metaphor in dehumanising refugees, and this is a recurring feature in analysis of migration metaphors. Arcimaviciene and Baglama (2018) propose that the metaphors they identified in US and European migration discourses may be grouped into two myths or narratives: the myth of dehumanization and the myth of moral authority. If we consider the list outlined above, we can see that most would be subsumed into dehumanisation while invaders and criminals are clearly encompassed by moral superiority, leaving just guest as potentially ambiguous unless the metaphor is actually unwanted guest.

As Arcimaviciene and Baglama (2018) also point out, these metaphors may function to enhance or suppress emotions, such as fear and empathy respectively. Indeed, a key point which is emphasised in this substantial body of work is that metaphor is an integral part of the discourse. Baider and Kopytowska (2017: 225) conclude that ‘metaphors can perform important social functions: they can be used to dehumanize the Other (in this case refugees and migrants), legitimize and delegitimize verbal and physical actions (e.g. verbal and physical violence against refugees, including hate speech and hate crime), as well as emotionalize (evoke both fear and anger) and desensitize the audience’.

As will have become clear, the vast majority of these studies find metaphor is used to negatively frame migrants. In part, this is likely to be that the focus of most research on migrants has actually been more restrictively concerned with immigrants and is often driven by the need to respond to societal challenges, such as discrimination or xenophobia. An important exception to the negative framing is Catalano’s (2016) work on self-representation by migrants which showed major discrepancies between migrant and media representations with migrants most likely to frame the process of immigration as either a journey or cultivation. Within work on public discourses, we find an exception with reference to the much-discussed water metaphors in KhosraviNik (2009) who notes the presence of common metaphors of large quantities like ‘floods of’ within topoi of ‘humanisation and individualisation’ (p. 21). This finding is supported by Nguyen and McCallum (2016: 168), who find water metaphors in news articles they considered neutral in evaluation and those which advocated for better treatment of asylum seekers, and Salahshour (2016) who finds water metaphors in favourable economic framings (e.g. a ‘flow’ which has ‘dried up’). In these instances, we have an overarching frame of migrants as a natural resource which can be used or exploited. Another recent exception to the negative evaluations is Taylor (2020) who finds that contemporary parliamentary debate frames the Windrush Generation (typically used to describe people who moved from Caribbean countries to the UK, as British citizens, in the period from the late 1940s to the 1970s) as builders perhaps reflecting a wider nation as family home frame.

However, when making these classifications about positive or negative representation, we should be very clear about the difference between what is positive for the speaker and what is positive for the person being described. The intent of the speaker using ‘floods of’ to evoke news values (Bednarek and Caple, 2017) of urgency and superlativeness, does not mean that the problems for the individual subsumed in that metaphor are eliminated. They are still de-individualised, their agency is minimised, and for many, it simply does not reflect their experience of movement to a destination country. The same kind of argument can be made about metaphors which favourably construct migrants as an economic resource and may mask exploitation and erase other kinds of positive contribution made to society. An additional point here is that the speaker’s evaluative intent may not correspond to the reception of the metaphor either because different readers of media texts will have been primed by their previous textual exposures (as, for instance, in ‘dog-whistle’ journalism and politics).

We also need to exert caution in assuming that each speaker using a metaphor has the same intent, or indeed any conscious intent. The point about metaphor, like discourse, is that it is cumulative and pervasive. For a lexical item to have a particular connotation or discourse prosody, it is not the case that every utterance must conform to that evaluative positioning. Nor will all speakers be using language with the same degree of awareness and deliberateness at all times. Drawing on a discourse which surrounds us, especially if naturalised in the society which we inhabit, cannot be simplistically equated with a conscious averral of the ideology which the discourse reflects and construes. If that were the case, we could never, as linguists and discourse analysts, achieve impact in society by raising awareness of the discourse histories and connotations of language we use. This tension around intention may help account for resistance to critiques of language – drawing on their personal experience, individuals may recall times when they have used an expression without conscious intent, or may recall a single counter-example, which means that they reject the thesis about implications at a discoursal level – and this is where empirical work may be able to influence receptive audiences. These challenging concepts of intent and deliberateness have, of course, received ample attention within metaphor studies as well as discourse studies more generally. Notably, Steen (2008) distinguishes deliberate and non-deliberate metaphors from the perspective of the addressee so that deliberate use implies a change of ‘the addressee’s perspective on the referent or topic that is the target of the metaphor, by making the addressee look at it from a different conceptual domain or space’ while non-deliberate metaphors ‘do not draw addressees’ conscious attention to other conceptual domains’ (Steen, 2008: 226). Much of what is covered in this paper is non-deliberate in this sense, because it examines conventionalised expressions. While novel and deliberate use is striking, it is also less insidious because it is more likely to be noted than the banal, conventionalised metaphor.

Methodology

The methodological framework used here combines corpus linguistics and (critical) discourse studies (e.g. Baker, 2006; Mautner, 2016; Partington et al., 2013) and follows McEnery and Baker (2017a, 2017b) in the application of this combination to historical corpora.

The main corpus used in this study was the Times Online. This corpus was created at Lancaster University, using the OCR (optical character recognition) files made available by the British Library. The corpus covers the period 1785–2011 and the current size is c. 10.5 billion words. It was analysed through Lancaster’s CQPWeb interface (Hardie, 2012). The scanned articles are also available to view as images through the Times Digital Archive to enable reading within a wider context. It was supplemented by a search-term corpus covering the period 2012–2018 which was compiled from the Times archive on the Nexis database (c.19 million words). The former was analysed using CQPWeb (Hardie, 2012) and the latter using AntConc (Anthony, 2019).

In any metaphor study, the principle challenges are identification and retrieval of metaphor. The Pragglejaz Group (2007) set out a systematic and replicable process of Metaphor Identification Procedure in which an expression may be classified as a metaphor when ‘(a) its “contextual meaning” contrasts with a “basic meaning” that is more physical and concrete (although not necessarily more frequent), and (b) where the contextual meaning can be understood via comparison with the basic meaning’ (Semino et al., 2018: 5). Regarding retrieval, as I was interested in everyday public discourses of migration, I chose to look for metaphor candidates in collocates. Collocates are understood here as items which have a strong connection with a node (for instance immigrant and wave) where strength is a statistically salient value. By identifying lexical items which occur as collocates, we can assume that there is a degree of conventionalisation indicating that the metaphor was indeed part of the discourse. The process followed was the following:

• (a) Collocates were calculated for node items (aliens, asylum seekers, boat people, colonists, emigres, emigrants, evacuees, exiles, expats, expatriates, immigrants, migrants, refugees, settlers) in each decade of the corpus using the statistical measure loglikelihood and a five-word left/right span. The choice of node items was intentionally broad to ensure that I was not replicating the highly problematic socio-political trend of only considering immigration to be salient to discussion of migration. They are all terms that foreground the migrant status of the people (rather than national, ethnic or religious labels, for instance) and this is to increase confidence that all metaphors identified are metaphors of migration.

• (b) The collocates were searched for metaphor candidates. This was a triangulated and iterative process which involved both scanning the collocate lists and searching them for possible candidates identified by compiling lists from previous research, using contemporary and historical thesauruses, and reading sample concordance lines.

• (c) Sample concordance lines for each metaphor candidate for each time period were then read and checked for metaphor use and coded according to the lexicalisation of the metaphor and evaluation of speaker.

• (d) The final stage addressed the problem of interpretation raised by the analysis of culturally distant texts. Once a metaphor had been identified, the collocate (e.g. wave) was moved to node position and I searched, per decade, for its collocates. In this way, the corpus provides a means of re-contextualising the lexical items by showing the company it kept (Firth, 1957) at different points in time.

Findings

Overview of patterns

In determining whether a metaphor was conventionalised, I determined that there should be at least two lexicalisations (e.g. wave and flood) per metaphor category per decade because different lexicalisations could signal a degree of metaphorical ‘animacy’ and exclude a metaphorical expression which had reached the end of its ‘life cycle’ (Croft and Cruse, 2004) and was now bleached or fossilised. The criteria of two different lexicalisations is perhaps low but this is offset by the fact that each lexicalisation must be sufficiently conventionalised to appear as a collocate. Figure 1 visually displays for which decades the metaphors were conventionalised according to this operationalisation. From this, we can see which metaphors have been consistent (water, object), which have apparently faded and returned (animal, commodity, guests), and which are relatively recent in conventionalised form (enemy, weight).

Figure 1.

Distribution of conventionalised metaphors.

Figure 2 illustrates the number of lexicalisations for each decade which can tell us about the ‘liveness’ or ‘productivity’ of the metaphor (not relative frequency).

Figure 2.

Variation in lexicalisation of metaphors over time.

The findings displayed in Figures 1 and 2 are discussed in the following sub-sections.

Liquid/water

The migrants are liquid metaphor (variously referred to as migrants are dangerous water, migrants are liquid or as a subset of migrants are a natural disaster) has been identified in migration discourses across different national contexts, including Australia (Pickering, 2001) Austria (El Refaie, 2001), France (Van der Valk, 2003), Germany (Petersson and Kainz, 2017), Malaysia (Don and Lee, 2014), Poland (Baider and Kopytowska, 2017), Spain (Montagut and Moragas-Fernandez, 2020; Rubio-Carbonero and Zapata-Barrero, 2017), New Zealand (Salahshour, 2016), Sweden (Petersson and Kainz, 2017), the UK (Charteris-Black, 2006; Gabrielatos and Baker, 2008; KhosraviNik, 2009; Parker, 2015) and the USA (Strom and Alcock, 2017). In the UK context, liquid metaphors have been discussed in work on right-wing election manifestos Thus, their status in contemporary migration discourse is well-established in diverse contexts. In terms of effects, Petersson and Kainz (2017: 54) discuss their strategic use in contributing to Othering, noting they are ‘likely to divert readers’ attention from the perilous journeys undertaken by refugees on their way to Europe [and] while focusing on migratory movements as streams and flows directed northwards, toward Europe, these naturalized metaphorical framings obscure the fact that most people either internally or externally displaced remain outside of Europe in regions adjacent to the crisis areas’.

In the Times data investigated here, the liquid metaphors were found in conventionalised form from 1850 onwards, they collocated with settlers, emigrants, migrants, immigrants, refugees and asylum seekers (but not colonists) and the number of lexicalisations increased over time, as seen in Figure 2. Thus, although long-standing, we have evidence that the metaphor continues to be creative and productive.

However, we cannot assume that continuity in the presence of the metaphor indicates continuity in framing or evaluation and nor can we rely on contemporary associations between metaphor and persuasive function to interpret historical data. In order to explore the framings that would be triggered by the metaphor, I took six of the most frequent lexicalisations and used the corpus to re-contextualise their usage by examining what collocated with each term in each decade. The findings are summarised in Table 1 (see Taylor, forthcoming for more detail).

Table 1.

Semantic associations for lexicalisations of liquid metaphor.

Used across period				Used in later periods
Influx	Pouring	Flow	Stream	Wave	Surge
People who move	Liquid	Liquid	Liquid (earlier period)	Sea	liquid [later period]
People	Money	Money	Money (later period)	crime (later period)	money (later period)
Money	Refugees (later period)
	Troops (earlier period)

Table 1 indicates that the words used to describe people in terms of {liquid}, were also often associated with the semantic field of {money} suggesting a relationship between the two framings. An exception here is wave which is a newer lexicalisation of the migrants are liquid metaphor and is primed for a negative semantic association with {crime}.

If we consider the evaluations which accompany these frames, we can anticipate that those clustering around wave are more likely to be negative from the speaker’s perspective than those which might also co-occur with money which may be evaluated more variedly, as illustrated in examples (1) and (2) from the Times 1850s. In (1), the speaker uses the liquid metaphor to highlight lack of control (see Duguid, 2011 for more on control and evaluation) while in (2) migrants are a resource to be managed for the benefit of others.

(1) If the victorious party should follow up vigorously its successes and press the vanquished, these islands may be inundated with refugees to an alarming extent (Times 1852).

(2) It was good for the West Indies and the Anti-Slavery party if immigrants could be induced to pour into those colonies (Times 1859).

The evaluation expressed by the speaker was coded in three decades in which there was a peak of discussion of migration. As Figure 3 shows, the evaluation changed over the three time periods with the earliest time period showing the greatest number of favourable evaluations and the most recent time period showing the lowest number of favourable evaluations.

Figure 3.

Summary of evaluation in concordance lines of migrants are water.

What this shows is that we cannot retrospectively assign interpretation to metaphors based on current framings because migrants are water has a historic root of evaluating people who move favourably. However, as noted above, this should not be mistaken for a favourable representation from the perspective of the people being described.

Object/commodity

The dominant metaphor in terms of continuity is migrants are objects. This is discussed here together with migrants are a commodity because the two are very closely associated (commodity necessarily invokes object). The migrants are objects metaphor has been noted in contemporary migration discourse with reference to Spain (Montagut and Moragas-Fernandez, 2020) and UK & Bosnia-Herzegovinia (Mujagić and Berberović, 2019). Metaphors of commodity (Petersson and Kainz, 2017; Santa Ana, 1999) and trade (El Refaie, 2001) have been mentioned in previous research but have received relatively little attention to date. However, Arcimaviciene and Baglama’s (2018) analysis of 57 media articles from European and USA online sources (including newspapers) found that object and commodity metaphors accounted for approximately a quarter of the metaphors in their data. They claim that ‘[t]heir use heavily contributes to the creation of social reality based on the mythical narrative that migration is not related to people, their lives and fate, but is rather a process based on the exchange of commodified relations between countries or governments’ (Arcimaviciene and Baglama, 2018: 5–6). Moreover, we might also want to consider the relationship between metaphors of migrants as commodity and trade and enslavement. The slave trade involves a metaphorical conceptualisation of people as economic commodities which results in a completely dehumanised treatment. Lastly, if we consider that the lexicalisations of liquid metaphors were also used to describe financial movement (as discussed in Section 4.2), then we can trace a logical coherence between the two dominant metaphors.

The object metaphors are found across the time period occurring with reference to settlers, emigrants, immigrants and refugees in multiple decades. The metaphors are quite consistent in the number and type of lexicalisations (see Figure 2) which mainly refer to movement of objects (e.g. convey, carry, removal) but also grouping of objects (e.g. batch). It is likely that most of these lexicalisations feed into other metaphors, such as commodity.

The explicit commodity metaphors realised, for instance, by cargo, demand, supply, trade show a distinct rise and fall: as Figure 1 shows, they do not appear in the collocates in the period 1940–1980. In the earlier time period (1830–1930) they collocate most consistently with emigrants, but also settlers and colonists while in the later period (1990–2018) they occur with migrants and immigrants. Thus, we see a shift to focus on people moving to the UK. The two time periods also differ in evaluation. In the earlier period, the evaluation of migrants in the commodity frame was either favourable or neutral, as shown in (3), while in the later period, the evaluation of the trade was negative. The key difference here being who controlled the trade; in the earlier period it is the speaker’s ‘us’ group, in the later period it is others (people-traffickers).

(3) For agricultural labour, however, and for single female immigrants the demand, they say, is practically inexhaustible (Times 1850).

(4) The warning about Albanian criminals came as the service identified a growing trade in illegal immigrants (Times 1999).

In these framings, there is a dehumanisation of the people involved and the resonance with enslavement is evident, as in (5).

(5) with the avowed purpose of loading, on account of a notorious slave-trading firm, a cargo of Negro emigrants (Times 1858).

Indeed, we can hypothesise that the language used to describe forced movement of enslaved people is part of early discourses of migration. If we examine the collocates of slaves in the first decade of the nineteenth century, the top ten lexical collocates are: Negro, importation, number, Colonies, treatment, Indian, board, cargo, Africa, Cattle. Dehumanisation is evident in both the commodity frame and association with animals, and the reference to colonies illustrates the centrality of slavery to empire building. However, as seen in (3) and (6), it should be clear that these commodity frames are not only used to represent those who are ‘other’ in terms of perceived race but class and power are determining factors in the extent to which people who move are presented as commodities for an economic trade. The young women referenced in example (3) and children in (6) are far from the emigrant of popular contemporary imagination.

(6) The Earl of DONOUGHMORE pointed out that a valuable supply of emigrants might be selected from the Irish unions, where there were numbers of young lads, orphans, for whom the guardians had taken care to promote certain knowledge of agricultural occupations; over such emigrants the colonists might have control (Times 1852).

Animals

The animal metaphor has been extensively discussed in previous research (e.g. Baider and Kopytowska, 2017; Mujagić and Berberović, 2019; Santa Ana, 1999) including specific focus on parasites (Musolff, 2012) and vermin (Marshall and Shapiro, 2018). Marshall and Shapiro’s (2018) experimental work showed that people associated conventionalised metaphors of migration (e.g. ‘They have invaded our house, coming in floods that we’ve been unable to stop’) with vermin and migrants. In a second experiment, they manipulated texts so participants were randomly assigned one of two texts where one included vermin metaphors. For instance, ‘At the U.S.-Mexico border, Border Patrol agents work tirelessly to stop [large groups/swarms] of illegal immigrants that [hurry/scurry] across the border’ (Marshall and Shapiro, 2018: 781). They found that metaphor use ‘can yield disgust reactions, with more disgust sensitivity being felt the more participants identified as American’ (Marshall and Shapiro, 2018: 783). The findings confirm the emotional role of these metaphors, and the work of Andrighetto et al. (2016) on the association between animal metaphors and social exclusion.

In the Times data analysed here, the animal metaphors were found in conventionalised form in the 1920 (swarms, flocking), 1930 (flocked, flocking) and 2010 (flocked, swarm) subcorpora. Thus, it was a minor pattern in this data which is quite limited in lexicalisation. However, that is not to underestimate the emotional impact for both readers and those described in these terms or the potential for these uses to resonate with readers who are exposed to these metaphors in more varied forms elsewhere. The evaluation which accompanies these metaphors is largely negative, as shown in (7) where the last sentence underlines the consequences. It also accompanied a sympathetic stance, as in (8) in which the speaker underlines their plight (sick and shoeless) and critiques the treatment of migrants as animals – ironically, while framing them as such (herd).

(7) THE BRITISH AT SHANGHAI The Municipal Council here is virtually a government, responsible for peace and order in an area which is solely under foreign control, but which includes 1,000,000 Chinese, an exceptionally large criminal element, and which swarms with political refugees. This necessitates the maintenance of a highly-organized police force (Times 1920).

(8) After riding in one of the lorries with the refugees – many of them sick and shoeless after being herded from the mountainside border camp at Isikveren – Mr Howell said: “Everywhere that you look these people are being treated like animals (Times 1991).

Strikingly, the animals collocates only occurred with reference to refugees, often the most sympathetic term for migrants, with of the 2010 data where they were also found with migrants.

Enemy

The invader or enemy metaphor has previously been discussed in relation to right-wing migration discourse in a range of contemporary contexts (Baider and Kopytowska, 2017; Böke, 1997; Burke, 2002; Catalano and Fielder, 2018; Chichon, 2020; Parker, 2015; Petersson and Kainz, 2017; Santa Ana, 1999; Van der Valk, 2003). This metaphor does not dehumanise the migrants and explicitly presents them as a threat. As Figure 1 showed, it is a recent arrival as a conventionalised metaphor (two occurrences in each of the three most recent decades), the lexicalisation is restricted (hordes, invasion, invaders) and it only collocated with refugees and immigrants. There is little ambiguity in the evaluation of this metaphor.

Hordes also occurs as a collocate in 1920 and 1940. This lexical item is so highly conventionalised it is somewhat debatable as a metaphor candidate and the contemporary usage is often light-hearted (e.g. the ten most significant collocates of 2010s include tourists, passengers, fans). However, if we look at the collocates at the time this becomes associated with migrant-names, the subjects are far more threatening (e.g. in the 1910s: German, barbarians, Kaiser, invading, Germanic, savages and in the 1920s: starving, savage, troops, soldiers, armed, German) so there is a clear priming for the invader metaphor and association with recent real war. Thus, we cannot transfer the contemporary bleaching of the term to the past and should recognise that there is a more threatening discourse history.

Guests

The guest metaphor, like the enemy metaphor, does not dehumanise the people being described. However, it does not easily fit into the classification of migration metaphors as dehumanisation or moral superiority myths (Arcimaviciene and Baglama (2018) as there is a less implicit negative association. It has not been frequently discussed in previous research, perhaps because it seen as less problematic than those mentioned above, although it is a familiar metaphor that has been conventionalised as guest-worker (Gastarbeiter in the German post-war context). It may be seen as building on the nation as family (or perhaps more accurately family home) metaphor (Burke, 2002; Charteris-Black, 2019) echoing colonial discourses of father land and mother country in which migrant as guests are positioned as less than family but still entailing some expectation of support (Laarman, 2013).

Collocates pointing to this metaphor first occur in 1860 and 1870 (invited) then 1890 and 1910 (welcome). There are two collocates (a potential sign of conventionalisation) from 1920 (welcome, invited), 1930 (guest, reception, welcome, invited), 1950 (welcome, invited) and then a break to 2000 (welcome, invited) and renewed activity with richer lexicalisation in the 2010s (host, hosting, overstay, reception, welcome, welcomed, welcoming, invited). In the earlier period it collocates with settler, while in the later period it collocates with migrants and asylum seekers. The data from the earlier periods shows that for those doing the welcoming, the migrants are favourably evaluated. However, as seen in (9) it may not be so for the people themselves, a familiar pattern by now, and the welcome is often explicitly tied to economic worth, as in example (10) reflecting a theme activated in the water and commodity metaphors.

(9) They kept from him information which they bad, which showed that the golden prospects were nothing, but a sham and a fraud, and that settlers were being invited out to a place where the prospects were nothing or worse than nothing (1930).

(10) David Blunkett is adamant that we need working migrants and must welcome them to boost the economy (2004).

In terms of more recent patterns, example (11) illustrates a broadening of the metaphor to migrants are unwelcome guests also signalled through the presence of the collocate overstay.

(11) [the Prime Minister] also refused further demands to streamline the British visa system unless he agreed that India should do more to take back more migrants who overstay their welcome in the UK (Times 2016).

An additional pattern in the contemporary data is the use of the guest metaphor to refer to past migration movements rather than the present. This is illustrated in the concordance lines in Table 2. Here we see a rhetorical move of asserting a persona of a generous host – which is frequently employed to legitimate a current inhospitable stance (see also Taylor, 2020).

Table 2.

Concordance lines of welcom* and history/tradition in discussion of migration.

London has such a long and rich history ofwelcomingwelcoming immigrants.

Times that Britain had a proud history of having welcomed Jews fleeing the Nazis and must not

“We’ve got a long history of welcoming people from disadvantaged background

that she did not. “The UK has a proud history of welcoming people who flee persecution

Mrs May said that the UK had a proud history of welcoming refugees and economic migrants

This country has a proud tradition of welcoming refugees. But in return, they have to

remind viewers that Britain has a long history of welcoming those fleeing violence abroad,

a promise to honour the proud British tradition of welcoming those most in need.

Note: Welcom* indicates the search included welcome/welcomed/welcomes/welcoming.

Weight

The weight or burden metaphor (e.g. discussed in Baider and Kopytowska, 2017; Catalano and Fielder 2018; Petersson and Kainz, 2017) occurs in conventionalised form in just two decades (1980, 2010) with two lexicalisations (strain, burden) and is used only with reference to refugees (1980, 2010) and migrants (2010). It occurred with a negative evaluation in which migrants were framed in mainly economic terms.

Other recurring frames

There were two recurring semantic patterns which were not captured or accounted for through this level of metaphor analysis. The first is a semantic preference relating to lack of volition in migrants (also discussed in Taylor, 2014) which was evident in collocates of migrant-names from the mid-nineteenth century onwards, as illustrated in (12) and (13).

(12) Cherbourg is thought to be a magnet not only for illegal immigrants but also criminal gangs trafficking people into Britain who will be forced into prostitution and slave labour (Times 2007).

(13) The authorities identified 12 nationalities among the dead, several of them legal or illegal immigrants from Latin America drawn by the booming economy (Times 2005).

This pattern was lexicalised through items such as attract, drawn, magnet, and pull which point to the use of force dynamic image-schema and, more specifically to attraction (Johnson, 1987). Image schema are a cognitive process, like conceptual metaphor, which frame phenomena in terms of our experience of the world and, more specifically, ‘force-dynamic schemas arise from the experience of pressure and motion’ (Hart, 2011: 273). In the attraction schema seen here, migrants are represented as passivated agonists but the antagonist, exerting force on the migrants, often remains under-specified and partial (as with metaphor, there is a foregrounding and backgrounding action at play). For instance, (13) reports on migrants from Latin America to Spain and reports the draw as the booming economy rather than a shared language deriving from the colonial past. There is also a broader partiality in that the attraction-force schema focusses attention entirely on the country of arrival and erases factors forcing migration. As Hart (2011: 279) notes ‘[r]epresentations which distinguish push factors tend to be more sympathetic toward the plight of immigrants and asylum seekers. Anti-immigration discourse more typically focuses on pull factors such as work or welfare which supposedly act as ‘magnetic’ forces on immigrants/asylum seekers’.

The second pattern is a strong semantic preference for quantification (e.g. number, proportion, amount) which has been noted as a feature of migration discourse previously (e.g. Baker, 2006; Van Dijk, 1988). This pattern, which was consistently conventionalised from 1830 onwards, almost certainly plays into the migrants as objects metaphor (quota, bulk, mass) but the boundaries are fuzzy and the pattern needs further attention.

Conclusions

The long-distance view of the metaphors showed that migration metaphors are historically rooted and conventionalised. Some of these are found to be very long-standing (liquid, object), others were more recent (animals, invader, weight) and some apparently dropped out of conventionalised use before returning in modified form (migrants as commodities for others, unwelcome guests). Patterns which have received less attention in previous work were also shown to be frequent, particularly commodity, object, and to a lesser extent guest and weight. Those which have been discussed in the literature, but which did not appear in this investigation include metaphors of migrants as parasites (Musolff, 2012), disease (Demata, 2017), weeds (Santa Ana, 1999) and pollutants (Catalano and Fielder, 2018). This does not mean that these metaphors do not occur at all, but that the relationship is not sufficiently stable for them to appear as collocates.

The evidence from this investigation provides support for a theory of contemporary migration metaphor based on inherited frames. There is continuity visible in the metaphor use (especially liquid and object) and there were no conventionalised metaphors in the earlier period which did not appear in the contemporary period suggesting no metaphoric frames have been lost. The underpinning discourse frame in the dominant metaphors (enemy is the exception) positions migrants as lacking agency, and in most cases the metaphors are non-human actors. The metaphors also showed how the key to evaluation across time periods was the speaker’s perception of control (so, water flow is positively evaluated if managed by an us-group). Furthermore, we see that metaphors of migration are shared across naming choices which today occupy very different discourse spaces (e.g. emigrants, settlers, immigrants). However, there are also challenges to an overly simplistic view of inherited frames. For instance, some metaphors are more recent in conventionalised forms (invader and burden) and, more importantly, the analysis showed that same metaphor may pertain to different discourse frames (e.g. liquid and animals).

The implications of this are varied. In terms of metaphor study, the patterns observed here did not fit the ‘life-cycle’ models of metaphor (e.g. Croft and Cruse, 2004) in which we might expect a progressive narrowing of lexicalisation and fading of metaphoricity (see also Musolff, 2010 for further challenge to this). In terms of migration discourse, the study shows how historical data can help us understand contemporary metaphor use. What we can reveal is that patterns of language used in the current day are not a basic ‘common-sense’ response to specific features of the group being described. Further research may also cast light on the extent to which the shifting evaluations and discourse frames associated with the same metaphor (e.g. liquid) may facilitate ‘deniability’ of negative stance in the contemporary period. The historical data may also hold subversive potential in showing us that ‘we’ (outgoing groups) were evaluated using the same metaphors as ‘they’ are now (e.g. commodity used by a UK newspaper to describe UK emigrants in the nineteenth century) and this work may be employed to raise awareness of language use. However, if we want to use this work for impact, we are also limited because the analysis of past discourses has not offered any conventionalised positive alternative metaphors outside the economic frame.