“Good Savage” vs. “Bad Savage”. Discourse and Counter-Discourse on Primitive Language as a Reflex of English Colonialism

In the ideological construction of colonialism and, more widely, of any hierarchy of human communities, a crucial role is played by discourse on language. English nationalism and imperialism, in particular, developed extensive argumentations on language as an interpretation of the encounter with the other, on the basis of internal cultural developments that assigned to language the role of social discriminator. The paper investigates a strand of such argumentations during the period from the seventeenth to the nineteenth century: the concept of “primitive” languages, described in a positive or, more often, in a negative light. The former arguments employ tones related to the idea of the “good savage” and stand in connection with narratives on the “language of Adam” and of the “Welsh Indians”, the latter uses a rhetoric extolling “progress” and “civilization” against the “immaturity” and “backwardness” of primitive languages, a perspective that was later to influence Darwinism.


Introduction
In the wake of previous research (Mazzon 2015(Mazzon , 2016(Mazzon , 2018, this paper looks at metalinguistic discourse as cultural conflict from a constructionist perspective, and focusses on the role that discussions on language had in eighteenth-and nineteeth-century Britain within more general scientific, anthropological and philosophical discourse. These discussions reflected the possibly universal phenomenon of "language anxiety " (Bourdieu 1991;Machan 2009) 1 as expressing political/social/economic insecurities, and brought intellectuals in many countries, especially those that developed their colonial expansion in those centuries and therefore had a more direct encounter with distant populations and cultures, to make of language an ideological and philosophical arena. This refers especially to the battle between universalism (supporting innatism) and adaptationism (highlighting the value of progress and evolution).
These discussions can be studied within the framework of Critical Discourse Analysis, looking at the ways in which ideologies are revealed by seemingly mere "technical" discourse (Fairclough et al. 2011). This approach highlights the ways in which choices of linguistic items and other communicative signals are employed in order to appeal to widespread mind-sets and to contribute to construct them, with the aim to persuade and to convey and reinforce ideological thoughts even without actually stating them. Section 2 of this paper briefly recaps traditional highlights in meta-linguistic discourse as an expression of social conflict, and the ideologies that such discourse hides. Section 3 looks more closely at the discussion on "primitivism" in Western thought, and especially in Britain, and at its connections with colonialism and with language debates. The latter topic is the focus of Sect. 4, which investigates further the discussion on "primitive languages" in the centuries examined and also presents some well-known examples related to the characterisation of Celtic and Native American languages.

3 2 Language Myths and Language Wars
Discourse and counter-discourse on language are deeply connected with the development of a national(istic) narrative as well as with the establishment of a social hierarchy. The construction of a standardized form of the language, propagated by an official educational system and serving as a social discriminator, is inevitably accompanied by forms of purism and by a codification of the "best" forms of a language. This leads to the production of a rhetorical repertoire aimed at extolling the standard language and at depreciating all other varieties of that language, as well as other languages (typically those of neighbouring communities, of minorities, and of hostile nations). As works like those by Crowley (2000Crowley ( , 2005 and Blank (1996) demonstrate, these arguments had already developed in pre-Renaissance times within various forms of public discourse in the British Isles.
As standardisation proceeded and the model of nationstates became predominant in Europe, however, there was a development in this rhetoric that gave rise to actual language myths (Watts 2011), i.e. stories about the history, the qualities, and the forms of a language that each community develops. Such stories tend to spread by way of the media, the educational system, and other means of communication that are emanated "from above", and are therefore expressions of the elite. Examples of these myths are: the myth of the glorious past and of the unbroken tradition, which emphasize the continuity in the linear history of the language and tend to construct its "pedigree", or the myth of authenticity, i.e. the search for true/authentic vs. false versions of a language, leading to labelling, reconstruction of ascendancy, and normalization of variation. Also crucial is the myth of purity, responsible for the long-standing horror of mixture, "contamination", or miscegenation leading to "corruption" (all concepts visible in discussions of language mixtures, from borrowings to pidgins).
These myths tend to be very pervasive, and, at a time when language studies started to acquire a level of prestige they had not previously enjoyed in the eyes of other sciences (Jones 1932, p. 317), they percolated from intellectual discourse into education, journalism and common lore, which has made them still quite persistent today. Their origin, however, seems to lie further back in the past, as episodes in the Bible such as the Tower of Babel and the shibboleth conflict demonstrate (Power 2015). Counter-discourse also tends to develop, as a form of opposition and resistance to the dominant rhetoric, but substantially using the same myths and metaphors. These stories, as well as the "wars" between different languages and different varieties to acquire predominance and prestige, are ubiquitous and widespread, and they are instrumental to the affirmation of political, religious and/ or socioeconomic agendas. One strand of such ideological metalinguistic discourse, and some of its sub-strands, are analysed in the sections that follow.

Primitivism in the British Isles
Spurred on by the rise of Enlightenment on the one side, and on the other by the increasing need to theorise a hierarchy of populations that could justify colonialism and slavery, new forms of discourse developed in the eighteenth century that employed a rhetoric of progress, based on the interconnected oppositions/dichotomies of Barbaric vs. Civilized and Passions vs. Rationality/Logic. The views of rationalism, which saw reasoning, concepts and ideas as the main achievements of humankind, could only be instrumental to nationalism and colonialism by stating that not all humans were equally capable of such mental operations, a view that was to become deeply entrenched in Western thought and that still finds echoes in racism today. A counter-discourse extolling the first items in the above-mentioned dichotomies also quickly developed, eventually feeding Romantic views.
In the anxiety of classification that pervaded most European intellectual debates at the time, stages of civilization also became the object of taxonomies-in the new hierarchies of human beings, levels of "development" of cultures play quite a relevant role, as remarked by Foucault (1966, p. 305 ff.). Savages 2 tend to occupy the lowest level in such classifications, while barbarians, although perceived as inferior to civilized people, are one step higher than savages. The term civilized came into use only at the end of the sixteenth century; until then, there was an equation between savagery and barbarism as forms of primitivism, while barbarians came later to be seen as already having some degree of development. The tripartite distinction/hierarchy arose because the vicissitudes of colonization, the changed attitude of the enlightened bourgeoisie to the peoples of the New World, induce a recognition of the presence even among the "primitives" of autochtonous organizational forms and a process of acculturation: the Americans are no longer the muta animalia described by the intellectuals in the train of the conquistadores (Formigari 1974, p. 278).
The savage was thus pushed to the margins of humankind, often seen as hardly human at all, which triggered the discourse about the discontinuity (or not) between these and primates (see the next section).
Towards the end of the nineteenth century, detailed categorisations of stages of humankind appeared including all three categories: for instance, Morgan (1877) introduced a further distinction into Older (Lower), Middle and Upper Savagery, Older (Lower), Middle and Upper Barbarism, and Civilization. The criteria for distinguishing these stages are interesting for this paper: the savage is "man in the state of nature" (hunter), using a form of pre-language (monosyllables); the barbarian falls within the category of history (shepherd) and employs "primitive languages", while we can only speak of a civilized man after the introduction of agriculture and of the alphabet. The discriminatory character of such classifications still reverberates in contemporary popular conceptions of communities whose languages have no written tradition as "inferior".
The discourse on primitivism and its counter-discourse are present from early on. For instance, many thinkers in the Renaissance hypothesised a middle way between humans and apes for the savages, through miscegenation or through a pre-Adamitic race of brutes: certainly the despised Hottentots (see Sect. 4, footnote 8) could not have descended from Adam (Jahoda 1999, p. 25). The exploration of Virginia originated a different narrative coming from the Puritans, of a good savage not corrupted by civilization (Barnett 1975, pp. 9-10), i.e. the counter-discourse concerning "American Indians" (misrepresentation starting of course with the name; Barnett 1975, p. 4). Their negative characterization is much older, and derives from the time of Columbus: these others came to be depicted as the prototype of savagery (Jahoda 1999, pp. 15-16, 22-23), especially with reference to their "filthy habits", and through employing dehumanizing words, often not only figurative but also literal, with epithets ranging from "animals" to "devils" (Barnett 1975, p. 5).
These discourses employed several rhetorical means, and were not confined to treatises and intellectual pamphlets, but were also conveyed in literary works, thus enabling these ideas to penetrate society at large. The anti-primitivist view is represented by lines such as the following (all italics used in quotations have been added for emphasis): When in a barbarous age, with blood defiled. The human savage roamed the gloomy wild, When sullen ignorance her flag displayed. And raping and revenge her voice obeyed (Falconer 1692, III, pp. 1-4).
Primivitism, i.e. the idea that man is inherently good before being "corrupted" by civilization, represented by as diverse figures as Locke, Swift, Captain Cook, Shaftesbury and Benjamin Franklin, received an early literary treatment in poems, too: I am as free as Nature first made man. Ere the base laws of servitude began. When wild in woods the noble savage ran (Dryden 1669, I.1).
An aspect specific to English debates is the polarisation in terms of barbarism vs. civilisation when it comes to the opposition Celtic/British vs. Saxon; this debate is summarised, for instance, in Kidd (1999, pp. 61-69; see also Kidd (1994), Leavitt (2010, p. 188 ff.), Morgan (2012), and was the ideal arena to convey political messages under the cloak of "science".
As mentioned, a very important role within the whole debate on primitivism is played by observations on language, to which we now turn.

Reflexes on Meta-Linguistic Discourse:
The Linguistic "Savage" The inclusion of language among the most important elements in placing a population in the hierarchy of civilization gradually became more central with the development of notions such as the "genius" of a language, i.e. the idea that a language is deeply connected with the "disposition" of a "race" or nation that employs it (Stock 2011). The Enlightenment (particularly Condillac) introduced a naturalistic view, according to which language originates from natural cries and gestures, and initially only includes monosyllables. This view was part of a primitivism that highlighted the positive aspect of "spontaneity" and "naturalness", but intellectual writing on the origin of language often did not embrace this version, partly because these theories established a strong connection between men and animals, and partly because they created too many problems of compatibility with a religious worldview (Formigari 1974;Leonard 1929, p. 19 ff.). These intellectuals opposed the traditional view, which saw language as God's gift to Adam, and therefore as already complete and perfect from the start, like all things created. The belief in language as a gift from God could not, initially, allow for "primitive" stages-the explanation and solution to the related incompatibility between the perfection of Godgiven language and the "deficits" of a primitive language were often found in the trauma of Babel, which created a discontinuity between the perfect "original" language and the new, imperfect and variegated languages that emerged as a consequence of divine punishment. This narrative follows the same lines as those given after the early discoveries of fossils, which were explained through recourse to biblical narratives like the Flood. One emblematic case is Rowland Jones (1722Jones ( -1774, who wrote extensively about his ideas on the origin of language/s (Jones 1764(Jones , 1768(Jones , 1771. He developed theories aimed at demonstrating the antiquity and "pedigree" of Celtic (which for him meant Welsh, but for other scholars Irish), in which he repeatedly makes reference to the excellence of "natural" language as it develops in its primitive stages, without the "unnecessary complications" and refinements later introduced by grammarians. Jones writes: It is yet the general opinion that human speech derives its origin solely from the arbitrary composition or invention of man, without any connexion with nature or the intervention of Providence. However true such bold and presumptuous doctrines may be with respect: to some of the corrupt compounded parts, which chiefly occasioned the great variety and confusion of languages, yet articulate sounds, the materials of speech, clearly appear to have been the gift of Providence, and always the same in all countries (1768, p. 3).
Clearly, for Jones (1771, p. 23) there is a sort of universal language that reflects the connection of sounds, vocabulary and parts of speech with natural elements, 3 and the later developments of the individual languages are deterioration rather than progress. In contrast, for other thinkers, "a nation must have been far removed from a state of barbarity, before they could have so much as thought of this invention [of language]" (Monboddo 1773-1792, II, p. 242). 4 Also influential was Rousseau's view, summed up by Eco (1993, p. 117) within his discussion of theories on the original universal language: according to Rousseau, primitive languages mostly employed metaphor; therefore, in a primitive language, words did not, and could not, express the essence of the objects they named-such a primitive language was inevitably less articulated. Thus, primitive languages are better suited to express passions, and should have derived from cries related to emotions, a derivation that connects these language forms with poetry and song-as such, they are highly suited to convey the eloquence of the tribal chief, as conceived by primitivists (Monboddo 1773(Monboddo -1792VI, pp. iv-vi;see also Sect. 4.2. below), while civilized languages have more structure and are better suited to the expression of rationality, ideas and concepts.
The connection between man and (other) animals became a hotly discussed topic a century before Darwin, precisely within debates on language. Several scholars wrote about the origin of language as derived from animal cries, especially in relation to the behavior of primates. A particularly significant figure in these debates was Lord Monboddo (1714Monboddo ( -1799, who devoted several pages to wondering why primates do not speak, since their physical apparatus would allow it and since they look like more primitive versions of man (Monboddo I,p. 187 ff.,270 ff.).
A few decades earlier, Linnaeus had been the first authoritative voice to claim a possible continuity between primates and humans, going against the dominant view of his time (Blanke 2014, pp. 35-36). This claim tied in with the views of thinkers like Herder, for whom "the birth of the word is to be sought in the terra incognita that divides man from his animal progenitors" (Formigari 1974, p. 282). Language is, for some thinkers, a decisive, defining feature of humankind, while for thinkers like Monboddo there could be a higher level of continuity with anthropoid apes (Monboddo I, p. 475). The primitive language would however have been very different from "civilized languages", whose development requires the rise of a "political state" and the acquisition of artificial habits (Di Martino 1984, p. 55 ff.). 5 If some were outraged at the suggestion that "primitive" men and languages may be closer to animals and their cries, others employed precisely those arguments to demonstrate evolutionary theories that would later become established, thus positing that language evolved and, consequently, that it must be subject to progress, like other aspects of human life.
The human hierarchies mentioned in the previous section were transferred to language with analogous arguments (Lauzon 2008, p. 259). The encounters with colonized populations, and the study of their languages, as well as the opposition between several different European nations fighting over predominance, gave rise to hierarchies of languages in terms of the degree of "development" of their speakers. This frame of thought became dominant, with very few exceptions: in the seventeenth century, for instance, Wilkins was one of the first to claim that, since reason is the same in all men, language differences are only accidental (Di Martino 1984, pp. 42-44). In the golden age of English normative grammars, the most common terms to refer to despicable forms of language are barbarian/barbarism, but the term primitive is also employed increasingly frequently over the two centuries, often in connection with the analogy between "savages" and children. 6 Condillac and Rousseau both speak about the infant's "instinctive" cries to express feelings, and see in them a possible origin for language (Reed 1965), thus creating a parallelism between ontogenesis and phylogenesis that will then continue to be pursued by several linguists even today (Yang 2013).
The ideas related to a "primitive" language that circulated in the centuries under discussion constructed myths that are related not only to nationalism, but also to the foundation of a colonial racism. The imposition of a standard language directly involves a narrative of its excellence, in opposition to other languages but also to non-standard varieties, a narrative related to the equation "one nation-one language". 7 For English, this construction starts with the unification of the kingdom and with the beginning of overseas expansion, with a thorough exploitation of the myths mentioned in the previous sections to construct a rhetoric of standardization (Mazzon 2016, pp. 11-19), in contrast with forms of speech that are not standardized, or, worse, do not have a writing system, and are thus not considered equally "developed". Myths once again percolated into all levels of society by their being formulated in literary texts, which facilitated their spread: So void of Sense the Hotentot is found, Whose Speech is scarce articulated Sound, That 'tis disputed, if his doubtful Soul. Augment the Human or the Brutal Roll. (Blackmore 1711, ll, pp. 87-90).
[Saying that the English verb system is "simple" is tantamount to saying] that the English Language is nothing superior to that of the Hottentots; and that the wisest and most respectable Body of People upon the Face of the Globe, own a language which is incapable of ascertaining their ideas, or of exhibiting the Soul, and its various Affections. (Buchanan 1762, p. 105). 8 Intellectuals like Swift tried to reverse this discrimination by inventing populations that are more advanced in civilization than humans-in his proposal to fix and ascertain the language, Swift always expressed ideas in terms of barbarous vs. civilized (Pinkerton 1787, p. 101;Mazzon 2016, p. 16), thus producing a reversed rhetoric of the "civilized" language and stressing the counter-ideology of change as decay, not progress.
My only concern is, that I shall hardly be able to do Justice to my master's Arguments and Expression, which must needs suffer from my Want of Capacity, as well as by a Translation into our barbarous English. (Swift 1726, IV, p. 5).

The Reconstruction of a Celtic "Pedigree" and the Language of Adam
As illustrated elsewhere (Gray and Mazzon 2011; Mazzon 2018), the Celtic areas were the first arena for the development of metalinguistic rhetoric, in accordance with the fact that they were the first in which a colonial ideological dynamics evolved. For instance, Ireland became the target of invective and of prejudice already in the Renaissance, as exhaustively illustrated by Crowley (2000Crowley ( , 2005; opposition to this discourse, in turn, created a counter-discourse of resistance and an alternative mythology of the Irish people and language. A crucial point in these arguments is the previously-mentioned idea that a lower level of civilization corresponds to an "inferior" level of linguistic expression. One early example of this construction of Irish barbarism is offered below: [Now that] the vulgar English tongue be universally planted, the Irish language, which is one of the chief and principal causes of the continuance of barbarity and incivility amongst the inhabitants of the Isles and Highlands, may be abolished and removed (Irish Act for the English Order, Habite, and Language, 1537;quoted in Machan 2009, p. 165).
This type of claim continued to be expressed for several centuries: 6 These terms were applied, over time, to illustrious voices of the past -both Chaucer, in the Renaissance, and Shakespeare, in the eighteenth century, were labelled in this way. Thus, the idea of "progress" of the language has also been applied, even before the time of Enlightenment, to the language of one's own nation, at least until the Pre-Romantic movement tipped the scales in favour of primitivism, and created a new appreciation for previous writers in spite of their rough/unpolished/incorrect English. See a collection of such statements in Moore (1910). 7 See, for instance, the evaluative and comparative statements about modern (European) languages by Monboddo in his vol. IV. But the Celtic, I will venture to say, is of all savage languages the most confused, as the Celts are of all savages the most deficient in understanding (Pinkerton 1787, p. 126).
The counter-discourse focused on Celtic populations as the first inhabitants of the British Isles (Malcolm 1738), and on their languages as authentic and pure, thus employing the same myths as English metalinguistic discourse.
What, shall a language [Irish] confessedly derived from one of the first tongues which subsisted among polished nations, be abolished, merely to make room for another [English] compounded of all the barbarous dialects which imperfectly communicated the thoughts of savages to each other? No, my Lord, the natives of Ireland must ever be instructed in their vernacular tongue. (Daniel Thomas, Observations on the Pam-phlets…, 1787, quoted in Crowley 2000. … nor does the Celtic or any incorrupt language need many grammatical rules: but the chief business of grammar, is to aid and supply defects and imperfections in languages (Jones 1764, p. 16).
The first example represents a total inversion of the mainstream discourse (it is English that is considered the "barbarous" language here), while the second is interesting since it concerns the role of grammar, reversing the common view of syntax as the key element for the expression of rational ideas. Mostly, however, the counter-discourse employed similarities in vocabulary, especially through a pre-philological derivation of etymologies, to prove the antiquity of the Celtic language. At a time in which the existence of a genealogical relation between Indo-European languages had not yet been discovered, the similarities between Celtic roots and some Latin and Germanic words offered good ground for speculation on antediluvian and pre-historic relations, as well as for claims of ancient glory. 9 This anxiety to project one's language back into the past spilled over onto debates on the original language or the language of Adam, triggering competition in nationalistic discourse. The most common views were that this original language could be either Hebrew or a language that was lost at Babel (Parry 2015, p. 17). Eco (1993, pp. 103-113) and Bonfiglio (2010) review some of the hypotheses that were put forth responding to nationalistic interests, investigating the roles of figures such as Leibniz, Rousseau and Herder.
Celtic quickly became a contestant in the British version of the debate 10 : There have been many nations, who have put their claim for the honour of the first language; and though the Hebrew, Chaldee, Syriac, Armenian, Chinese, Greek, Swedish, Coptic, Teutonic and Celtic have had their advocates, the Celtic seems to me to support the claim with the best proof. (Jones 1764, p. 29).
…what tends mostly to prove that the Celtic received no alteration at Babel, is its perfectly retaining the first frame and construction, and its defining all ancient names of persons and places, before the confusion, in so natural, clear and rational a manner. (Jones 1764, p. 34).
In the related discussion, British is argued to be related to Phrygian, considered a "parent" population/language of Greek, and where both Cumbri and Saxons came from. Ultimately, this population descended from Noah's son Japhet, who was allegedly a Druid. The original inhabitants of the British Isles were often argued to be related to the Scythians and to the tales of Homer and Virgil, e.g. identifying Gauls with Scots.
Petigrées and genealogies also the Welsh Britons have plentie in their owne toong, insomuch that manie of them can readily derive the same, either from Brute or some of his band, even vnto AEneas and other of the Trojans, and soo forth vnto Noah without anie maner of stop. (William Harrison, An historicall description of the land of Britaine, 1587, quoted in Bolton 1966, p. 16).
Huge efforts were made to reconstruct etymologies in order to justify these claims of connecting the "antiquities" of different populations, to extol the Old British tongue (Malcolm 1738, pp. 13-17), which has given names to all important ("greatest") things. Rowland Jones employed similar etymological arguments to demonstrate that the Cumbri-Galli-Celts were the forefathers of many other peoples; he related even the word Babel to British etymology, and since this language underwent no change after Babel, the "pedigree" is preserved (R. Jones 1764, pp. 23-29).

Trans-Atlantic Metalinguistic Rhetoric and the "Welsh Indians"
These arguments were transported, more or less in the same decades, across the Atlantic, where colonization was offering a whole new set of "Others" that British explorers, colonizers and later settlers confronted in more or (frequently) less peaceful terms. At the same time, North America was becoming an increasingly multi-cultural area, as settlements, conquest and immigration proceeded. 11 Unsurprisingly, also given the high numbers of immigrants from the Celtic areas, this reinforced the idea of the Welsh and Highlanders as a "purer" population but also as "savages", on the same line as discourse on the "Wild Irish", which had been in place for centuries (Langford 2000, p. 20). This idea was supported by arguments such as that "the simplest shape of Language may be found in the remotest Places from the Center of the Dispersion of Mankind, and perhaps not in the Center itself" (Malcolm 1738, p. 36). It is interesting to notice how the languages of native Americans are characterized in accordance with the views of primitivism mentioned in the previous section: An American chief, at this day, harangues at the head of his tribe, in a more bold metaphorical style, than a modern European would adventure to use in an Epic poem (Blair 1763, p. 2).
Their [Delaware] language is lofty, yet narrow, but like the Hebrew; in signification full… imperfect in their Tenses, wanting in their Moods, Participles, Adverbs… And I must say that I know not a Language spoken in Europe that hath words of more sweetness or greatness, in Accent and Emphasis, than theirs (Penn 1683).
Notice that, once again, grammar and syntax are the marks of civilized, developed languages, but at least some positive elements are recognized to primitive languages in terms of poetic and expressive qualities. These arguments were voiced, for instance, by Hugh Blair and other Scottish thinkers, for whom the primitive languages of populations such as those of North America originated in "instincts and imagination" and then developed along with civilization, which entailed a gain in "correctness" but a loss in the expression of the "native effusions of the heart" (quotations in Manning 2000, p. 271).
According to Palmer (2001), this relatively benign portrayal of Amerindian languages as compared to the Celtic ones is due to the high level of "exoticism", which made them less threatening to the British, as they appeared more "distant" (Mazzon 2018, p. 87). There was also an attempt to tie the origin of the "pure" Native American languages to ancient, prestigious languages like Hebrew, and to extol their rhetorical qualities (Dossena 2015, p. 11 ff.). In any case, as shown above, such discourse applied also to the Celtic areas and generally to the counter-discourse of "primitivism", of which the Ossian controvery is a prime example (Mazzon 2018, pp. 93-95), along with the curious story of the "Welsh Indians".
The first studies on "exotic" languages began in the mid-1600 s (Leavitt 2010, p. 39); the discovery of their complexity undermined the arguments about "primitive" languages being simpler. An interesting strand of this discussion is the attempt to tie "primitive" societies to early European populations, aiming at demonstrating a common origin. For instance, around the end of the eighteenth century there was a revival of the legend of Madoc, the Welsh prince who in the twelfth century allegedly landed in America, as also reported in less precise terms by Monboddo (I,p. 589). This legend emerged in medieval chronicles, and was picked up now and again to validate England's claims to the New World against Columbus' primacy for Spain, 12 but was then submerged in the Renaissance (during Elizabeth's reign, at the acme of conflict with Spain, it would have looked too "suspicious"). It was then re-used later to substantiate counter-discourse on the Celts, and thus the legend became a quite regular feature in such discourse, in different versions, with the common aim of proving the primacy of Celtic people in exploring and dominating overseas territories (Willliams 1949).
Numerous first-hand accounts seemed to lend credit to the story. In 1740, the Gentleman's Magazine reprinted (from an original dated ca. 1685) a letter about an alleged witness, Morgan Jones, who claimed that he had met a tribe of Indians "fairer" in colour, 13 who spoke a Celtic ("British") tongue, and had a "sacred book" which may have been a Bible in Welsh. In 1699, another witness wrote that being acquainted with the "Primitive Irish" made it easy to understand the "Indians", because of the "affinity", especially in pronunciation (Winship 1903, pp. 10-1).
These accounts generated ample discussions in intellectual and scientific circles, to the extent that new expeditions were organized with generous funding to prove or disprove the witnesses. Scientific interest provided the ideal coating for what was in reality an essentially political battle.
The defenders of the Welsh Indians uncritically embraced philology and antiquarianism, were religiously and politically radical, and believed that by reviving and spreading the use of Celtic languages a pre-lapsarian golden age might be recovered (Lauzon 2008, p. 266).
As mentioned, language was again central to the debate, with supporters of the "Welsh Indians" trying to prove linguistic relatedness. For instance, John Williams wrote a quite lengthy treatise (1791) to discuss the evidence for the connection with Madoc's story and the primacy of Celtic arrivals in America; according to him, there is "no doubt" about this primacy, even the fact that the Irish were "primitive" is no problem because the Saxons were also "primitive", and yet they ventured into many new lands. Claims to the contrary are "fables, inventions, idle monkish tales", 14 since "that the Language was Welsh it cannot be denied; for one Lewis a Welshman conversed with Indians in their own Language" (Williams 1791, p. 22). 15 The words in common use on different parts of the Continent, which are very near, or undeniably Welsh, in both sound and sense, could not happen by chance, and they could not be derived from any Europeans but from the Ancient Britons (Williams 1791, p. 1).
Tables and long lists of etymologies found place in countless treatises. The story of the "Welsh Indians" was taken up again and again in the nineteenth century, generating toponyms (Madoc is also the name of a town in Ontario) and other forms of naming, as well as literary works, especially poems, by well-known writers such as Southey and Pushkin. Even today, this story inspires works of fiction and non-fiction, as the brief list of titles below exemplifies-this myth has not lost its appeal yet.

Conclusion
This paper has briefly examined some strands of discourse on language in the British Isles in the eighteenth and nineteenth centuries, showing how crucial language can be in conveying ideologies. The construction of language-related myths, although especially developed in the centuries under examination, can be hypothesised to be a universal: mythconstruction is not just a feature of pre-scientific language discourse, and that it is very persistent in popular discourse. Thus, the bold pronouncements by Monboddo, Jones, and Williams can appear dated and unfounded to the modern scientific community, but their influence can still be seen in (explicit or implicit, conscious or unconscious) racist and discriminatory discourse.
A main reason for the effectiveness of this persistence is the employment of rhetorical patterns and metaphorical discourse, as for all persuasive speech. Critical Discourse Analysis and the investigation of language devices can help uncover and defuse such myths, as I hope to have shown in these pages.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.