The English term ‘blending' refers to a specific word formation process by which two (or, in very few cases, more than two) words are merged together to form a new word, a blend. Blends can thus be defined as “two-constituent compounds in which at least one constituent has lost some phonological material” (Arndt-Lappe and Plag, 2013, p. 5). Although in cases such as covidiota both constituents covid and idiota are contained in the blend, the overlapping phonological material (the syllable id in this case) is not repeated in both constituents (covididiota), but shared between them (covidiota). This differentiates blends from compounds, which also constitute a combination of two (or more) lexical units. However, the constituents usually maintain their form and are largely unchanged in the process of compounding. The amount of material that gets lost in blending can differ and depends on whether the constituent words share a sound or sequence of sounds. If they do, the merging usually takes place within that sequence (Ulašin, 2016, p. 164). As a consequence, when morphologically analyzing a blend, it becomes obvious that the constituents do not usually coincide with morpheme boundaries. In blending, matters of prosody, syllabic structure and accent, are of a higher priority than morphological accuracy.

Blending is frequently considered a marginal word formation type not only because of its disregard for morpheme boundaries, but because it is not a very frequent phenomenon (Pharies, 1987, p. 271). To be more precise, a relatively small number of blends only ever become widely known within a speech community, meaning that most words that are formed by means of this particular process never become lexicalized. In spite of its alleged marginality, blending is a relevant mechanism of lexical innovation (Ulašin, 2016, p. 181), mainly because of another very distinctive quality: Blending is a very resourceful mechanism for forming particularly creative neologisms. Apart from meeting the need for linguistic economy (since it enables speakers to express a complex meaning in a condensed form), blending is generally perceived as conveying a certain nuance of playfulness, and is therefore very often and very characteristically linked to a humorous intention (Pharies, 1987, p. 281; Lang, 1990, p. 199; Varela Ortega, 2005, p. 95; Ulašin, 2016, p. 179).

Several authors, albeit without referring to any quantitative, comparative studies of the matter, point out that in English, blending has been much more common than in other languages (e.g., Lang, 1990, p. 199–200; Ulašin, 2016, p. 181) and that an increasing number of blends in several other languages can be traced back to the influence of English on these languages (Roig-Marín, 2017, p. 52). Balteiro (2018, p. 2) points out that “due to English influence blending is evolving from a minor device in Spanish into an important one, not only to cover lexical needs but also for creative and stylistic reasons (to produce humor, irony, punning, etc.)”. When it comes to describing the introduction of covidiota into Spanish, this aspect is relevant because it helps to explain the conspicuousness of the neologism with regard to the Spanish language system: covidiota is a blend, i.e., created by means of a word formation process which stands out in its own right, for the reasons described above. Moreover, covidiota is a blend which was first coined in English and only then introduced into Spanish.

Language on the internet, and thus also language in social media-communication, is in many regards subject to a strong influence from English (e.g., Balteiro, 2012, p. 9; Balteiro, 2018, p. 2), even though the internet has definitely stopped being an all-English medium and has become a multilingual one (Crystal, 2011, p. 78). Following Kachru's categorization, the Spanish speaking Twitter users whose language usage is studied in the present case study can be said to belong to the “Expanding circle,” which is defined by people who learn and use English as a foreign language (Kachru, 1994, p. 136). It is, however, also perfectly possible for at least a part of them to belong to the “Outer Circle,” or even to the “Inner Circle” (Kachru, 1994, p. 136), depending on whether they use English as a second or as their first language. The tweet corpus compiled for the present study comprises tweets from all over the Spanish speaking world, and it is likely that it includes several texts composed by bilingual speakers of English and Spanish, for example in the USA.

Apart from bilingualism as an influencing factor originating from the speakers (Twitter users), the language used in digital communication can generally be said to possess certain characteristics in its own right, which, again, include a strong influence of English. Balteiro depicts this kind of language as basically a “new and rapidly-changing and developing language as a result of a rapid evolution or adaptation to Internet requirements of almost all world languages, for whom English is a trendsetter” (2012, p. 9). She adapts David Crystal's term Netspeak (Crystal, 2004, p. 78; Crystal, 2006, p. 19–25) and thus refers to the variety of Spanish used in digital communication as Spanish Netspeak (Balteiro, 2012, p. 10).

It can be assumed that both aspects, i.e., the bi- or multilingualism of Twitter users and the role of English as a trendsetter in language use in digital communication, contribute to developments such as the creation and rapid spreading of neologisms all over the Spanish speaking Twitter community, as in covidiota. The prominent role of English in this context does not, however, keep English-induced items from being perceived as a conspicuous, unfamiliar, potentially marked, and, in any case, generally new item. Speakers typically have several strategies at their disposal for dealing with such linguistic items. The following section provides a short overview and possible classification of such strategies.

In the initial phase of the introduction of a borrowing into a receiving language, the new item is generally perceived as foreign and unfamiliar by speakers of that language. The fact that it does not (yet) belong to the vocabulary inventory of the receiving language often becomes manifest in the form of certain strategies which speakers apply to deal with the newness and foreignness of the newly borrowed item.

Following the works of Rey-Debove (1978, 1998) and Authier-Revuz (1995), Pflanz (2014) establishes a typology of “marqueurs d'altérité” (‘alterity markers'; Winter-Froemel, 2021). These alterity markers are used in the context of different linguistic strategies applied by speakers to mark borrowings which are still considered to be new and unfamiliar, and therefore not taken to be self-explanatory (Pflanz, 2014, p. 164–165). Pflanz introduces four types of alterity markers. The first type (X montré, Pflanz, 2014, p. 164) comprises visual markers which make the item stand out in a text. For example, this can be achieved by the use of quotation marks. The second type (X commenté, Pflanz, 2014, p. 165–168) refers to metalinguistic comments which are made by the speaker and can either be of an evaluative nature (commentaire axiologique) or rather of an explanatory kind (commentaire didactique). The third type of alterity-marking strategy consists in paraphrasing, or translating, the new item to clarify its meaning (X traduit, Pflanz, 2014, p. 168–173). The fourth type, on the other hand, comprises the explanation of the neologism (X expliqué, Pflanz, 2014, p. 173–175). The fourth type differs from the third mainly in the predominant function associated with it, which is the referential one, while for X traduit the metalinguistic function is predominant (Pflanz, 2014, p. 173).

These categories or types, established by Pflanz, are taken into account at different points throughout the present study and are reflected in several of the research questions outlined in the following section. Since these features are explicitly established as markers of alterity, i.e., otherness, one very interesting question is whether the time span of the observation periods in the case study is long enough for their manifestations to change, as speakers become more familiar with the term at hand.

The first relevant aspect to be outlined is concerned with form rather than with the content of the tweets: Speakers (and, of course, writers) often feel the urge to highlight items in their texts which are considered unusual as they are not part of the general vocabulary of their language because they are of foreign origin, for example. In the graphic medium (which includes, among others, the communication on social media platforms), such strategies comprise capitalization or the use of quotation marks, among others. The highlighting of foreign items corresponds to the first type of alterity markers as established by Pflanz (2014, p. 164–165): visual markers of alterity. The working hypothesis for the case study is that the use of these markers becomes less frequent over time, as the term becomes more widely known and continually loses part of its alienness.

At the same time, graphical highlighting is not only relevant when it comes to flagging recent borrowings, but it is also considered to be one of many different possible humor- or irony-markers available to speakers/writers who (consciously or unconsciously) wish to underline an ironic or humoristic intention of their utterances (“typographic markers,” Burgers and van Mulken, 2017, p. 388). As has been outlined above, blends are typically associated with playful language use and humoristic intention. Covidiota can therefore be considered at least a potentially humorous lexical item. So, if graphical highlighting should prove to be a prominent feature during all three observation periods, this feature would not (only) be used as a marker of newness but as a means to emphasize the speaker's humoristic intention.

A particular kind of graphical highlighting has to be discussed in more detail, since it is a highly characteristic feature of language in social media: the use of hashtags. Apart from highlighting in itself, the appearance of a term in the form of a hashtag often goes hand in hand with that term not being integrated into the syntactic structure of the rest of the tweet. Therefore, the appearance of covidiota or covidiot as a hashtag constitutes one of the categories of analysis.

When it comes to examining the way in which English, or, more precisely, the presence of an English-induced neologism such as covidiot(a) influences Spanish twitter discourse, it seems appropriate to look for other English (and English-induced) elements (words, phrases, indications of code-switching) in these texts. The underlying research question connected with this category of analysis is whether the presence of Engl. covidiot or Span. covidiota might trigger the use of more English-induced linguistic items.

After establishing parameters that reflect the newness of the term covidiot(a) to the Spanish language, I will now propose a number of features that can be considered indicators of integration. These can be found on different levels of language and can be said to reflect a growing tendency of the term to behave as a Spanish word.

Although the neologism Span. covidiota is of English origin, its formation is completely transparent to Spanish speakers and it does not appear absolutely clear whether it is more appropriate to say that Spanish speakers (i.e., Spanish speaking Twitter users) borrow the entire new-built term covidot from English and then adopt it, or that they rather form the Spanish equivalent of the term, following the English example. The latter would be just as plausible as the first possibility, since both constituents COVID and idiot(a) can be considered internationalisms, albeit of very different kinds. While COVID is a very recent creation and was coined in English, Span. idiota (as well as Engl. idiot) is of ancient Greek origin and constitutes a long-established element of the Spanish lexicon.

Both the English and the Spanish versions of the neologism can be described as cruces subordinativos ‘subordinating blends,' which means that one of the two components constitutes the nucleus of the word (Ulašin, 2016, p. 165). In this case, it is the second constituent: a covidiot(a) is a particular kind of idiot(a).

The nucleus is also decisive when it comes to determining the grammatical properties of the blend, for example its belonging to a certain word-class. The English term idiot, according to the OED, can be used as a noun or an adjective (OED online, 2021, s.v. idiot), although the adjective idiotic appears to be the more frequently used alternative. The definitions of Engl. covidiot which are received, adopted and disseminated by Spanish speaking Twitter users during the initial phase of its existence, characterize it as a noun. As for the Spanish term idiota, the Spanish language academy, (RAE), in its Diccionario de la Lengua Española (DLE), classifies it as an adjective, but points out the fact that it is also used as a noun, by adding the annotation “U[sado] t[ambién] c[omo] s[ustantivo]” (DLE, 2014: s.v. idiota). Other dictionaries register idiota as both an adjective and a noun (e.g., Moliner, 2008: s.v. idiota).

The actual behavior of Span. covidiota with regard to word-class will therefore be one of the aspects under investigation, the actual question being: If the English etymon covidiot has been coined as a noun, and Span. idiota can be an adjective as well as a noun, how does this influence the use of covidiota over time? Is it introduced into Spanish as a noun and continually used (mostly) as such, or is there an increase in usage as an adjective?

When dealing with a foreign, unknown lexical item, it can be expected that users feel a need to define it and comment on the new term before they start actually using it. This point relates to the second, third and fourth types of alterity markers as established by Pflanz (2014) and, with regard to the study at hand, leads to the following crucial questions: Is the term covidiota talked about, cited and/or explained, or is it actually used? Does the main function seem to be referential or rather metalinguistic? How does the use of covidiota change over time, with regard to these aspects? The differentiation between metalinguistic commenting, defining, negotiation of semantic content and referencing is thus considered a crucial parameter for the analysis of the tweets. With regard to the above-mentioned typology, a possible hypothesis could be that commenting (second type), translating (third type) and explaining (fourth type) become less frequent functions over time, since all three are considered as alterity markers and thus descend in frequency as the item in question becomes better-known.

Another indicator which is considered highly revealing in terms of integration of a new item into its receiving language is the formation of new phrases, compounds, etc. which include the neologism in question. Therefore, one important point of investigation in the present case study consists in the search for recurring expressions in Spanish tweets which contain the term covidiota, the hypothesis being that such phrases should become more frequent as the term becomes more widely known and more frequently used.

The question of what part of speech covidiota actually constitutes in a specific case of language usage cannot always be answered in a definite way. There are cases which do not allow for a clear allocation to either one of the possible word classes (i.e., noun or adjective) at all. Other instances provide certain indications which can support the analysis. In the following, some examples shall be outlined in more detail. The Nueva gramática de la lengua Española (NGLE), published by the Real Academia Española, is used as a reference in cases of doubt and to provide further information.

Covidiota is very frequently used together with the verb ser ‘to be,' which can generally be followed by an adjective as well as by a noun. Examples (21) and (22) illustrate a significant difference between two different possible constructions with ser: in (21), covidiota is accompanied by the indefinite article un, and thus clearly used as a noun. In (22), however, there is no determiner, nor are there any other indicators which would encourage an interpretation of covidiota as a noun.

Following the descriptions in the NGLE, in the present study, in uses like (22), i.e., constructions of the type ser + covidiota (without insertion of determiners), covidiota is taken to be an adjective. The same applies for the construction por + covidiota, as for example in (23), since these phrases can be interpreted as being essentially a shortened form of por ser covidiota (cf. NGLE, 2009, p. 3466), ‘for being covidiota [‘covidiotic']'.

The identification of word-class in constructions of the type verb + de + covidiota (andar/ir/estar/seguir de covidiota, see Section Recurring Constructions, Patterns of Combined Occurrence) brings about certain difficulties. Here, covidiota is preceded by neither a determiner nor any other element which would allow for its identification as a noun at first glance. Neither is there any clear evidence supporting an interpretation of covidiota as an adjective. Consultation of the NGLE, however, suggests an interpretation of covidiota, in the instances indicated above, as complementos predicativos nominales.

The complementos de régimen, mentioned in the first part of the extract, are those whose preposition is determined by the verb, as would be the case for de in quiero salir de aquí ‘I want to get out of here,' for example. This does not apply to the examples at hand, as in andas de covidiota, sigue de covidiota, etc. Moreover, the interpretation of the syntactic functions within the examples under investigation as copula constructions with predicative complements is highly plausible, even mandatory in the case of estar de covidiota (estar being one of the two central, prototypical copula verbs in Spanish, next to ser). This fact alone does not imply that the item following the preposition de necessarily has to be a noun, since there are also constructions of the type verb + de + adjective, as in acusar de incompetente (NGLE, 2009: 2891) ‘to accuse of being incompetent', also containing a predicative complement, albeit one that relates to the complemento directo. However, as is pointed out in § 38.10c, adjectival elements frequently appear in this kind of construction together with verbs that express judgement (NGLE, 2009: 2890–2891), in this case acusar. The verb semantics and the overall meaning of the phrases from the covidiota-corpus (andar de covidiota, ir de covidiota, etc.) are of a different nature, and actually match the description in the extract cited above very closely: designan alguna actividad u ocupación en la que se ingresa […], se permanece […] o se termina […] – the activity or occupation in our case being that of ‘acting as a covidiot,' ‘going somewhere as/being a covidiot,' ‘continuing to be a covidiot,' etc. Some of the examples from the extract also closely resemble some of the constructions in the corpus:

Thus, with reference to the RAE's assessment as quoted in the extract above, in expressions of the type andar/ir/estar/seguir + de + covidiota, for the part of speech-analysis at hand, covidiota is considered to be a noun. This decision suggests an interpretation following the RAE's description of the elements in question as complementos predicativos nominales, nominal predicative complements.

As a further example, I would briefly like to discuss the elements muy and mucho as accompanying items of covidiota. At first glance, both items seem to be rather unproblematic in terms of part of speech analysis and therefore can be of much help when it comes to determining the word-class of the element to be modified by them: muy, as an adverb, modifies the adjective covidiota, whereas mucho relates to covidiota as a noun (or rather, as covidiota has to be considered a countable noun, the correct form in standard Spanish should be the plural, as in muchos or muchas covidiotas). This holds true for muy in most cases, as is illustrated by the example (26).

Example (27), however, is much more ambiguous. On the one hand, the presence of the adverb muy would usually suggest a classification of covidiota as an adjective. On the other hand, muy is preceded by the definite article la, which provides a strong clue for the interpretation of covidiota as a noun. In the NGLE, expressions similar to the one in question are mentioned in connection with a certain type of emphatic apposition, as in la muy loca de tu prima […] (NGLE, 2009, p. 887). For that the authors refer to processes of neutralization of noun and adjective (NGLE, 2009, p. 887). Although they generally opt for the analysis as a nominalized adjective (cf. also § 13.7c, NGLE, 2009, p. 944), they explicitly point out the possibility of a reverse interpretation, i.e., analyzing the element in question as an adjective, or rather, an adjectival use of a noun, “ya que los nombres que se admiten en esta estructura suelen aceptar el adverbio muy, característico de los adjetivos graduables” (NGLE, 2009, p. 887). For the purposes of the present study, and following the NGLE, covidiota in expressions of this type (la muy covidiota) was analyzed as a noun, bearing in mind the fact that this is not the only possible way of interpretation.

The tweet-corpus does not contain the forms muchos or muchas in the plural. Instead, there are 10 instances of mucho in the singular, followed by covidiota (also in the singular), as illustrated in (28).

The use of the singular instead of the plural form does not alter the fact that mucho as a cuantificador helps identifying covidiota as a noun in these cases. The meaning, on a denotative level, is the same as that of muchos covidiotas. For the connotative level, however, the authors of the NGLE point out that, firstly, these constructions are more common in colloquial than in formal conversations, and secondly, that they often tend to convey a derogatory note (NGLE, 2009, p. 1479).

As these examples show, the analysis of Span. covidiota as a noun or as an adjective involves some particular challenges and cases of doubt. Bearing in mind the fact that uses are not always unambiguous and that decisions were made as outlined above, the numbers obtained in the present study and presented in Table 5 suggest an increase in the proportions of the use of covidiota as an adjective over time.

When it comes to interpreting these results, the question is whether this growing tendency of covidiota to function as an adjective can be considered as a manifestation of increased integration into the Spanish language. This would be the case if the proportions of Span. covidiota as noun/as adjective could be shown to become more similar to those of Span. idiota as noun/as adjective, over time. Online corpora of Spanish were consulted in order to provide the necessary data to make this comparison.

The research in several online corpora of Spanish reveals a varied picture with regard to the use of Span. idiota as a noun or as an adjective. Table 8 shows the results for three corpora of the Corpus del español, Historical/Genre (Davies, 2002), Web/Dialects (Davies, 2016) and NOW (News on the web, Davies, 2012), as well as for the RAE's Corpus del Español del Siglo XXI (Real Academia Española, 2021) and Spanish Web 2018 (esTenTen18) in Sketch Engine (2021)⁴.

As can be seen in the table, the proportions of idiota tagged as an adjective in the corpora range from 11.43% (NOW) to 51.02% (CORPES XXI), those for idiot as a noun from 48.96% (CORPES XXI) to 88.57% (NOW). The reasons for these considerable differences cannot be examined in detail in this study. However, it seems reasonable to assume that the classification is based on different criteria in the different corpora. In order to check, for instance, how ser + idiota is handled in the corpora, a search for the exact phrase es idiota was conducted, in order to test the results for this one specific form of ser idiota. In the Genre/Historical-corpus (Davies, 2002), there are nine cases of es idiota when looking for idiota as a noun, and none at all for idiota as an adjective. The Web/Dialects-corpus shows 520 hits for es idiota with idiota tagged as a noun, and 20 for idiota as an adjective, and the NOW-corpus contains 487 cases of es idiota with idiota as noun, but none with the adjective. In the CORPES XXI, on the other hand, all 17 instances containing the phrase es idiota are delivered when searching for idiota as an adjective, while non is returned when searching for the noun. Thus, there might be a general difference between the Corpus del español and the CORPES XXI regarding the handling of ser idiota in terms of word-class, which could constitute one of the reasons for the strongly varying numbers represented in Table 8.

A short examination of some selected samples reveals that sometimes, cases which appear to show the same structure appear in both groups, as a noun and an adjective. For example, the expression el/la muy idiota (similar to the above discussed la muy covidiota), is shown to appear 26 times in the NOW-corpus. In 24 of these cases, idiota is tagged as an adjective. In two instances, however, idiota is tagged as a noun. It does not become clear from the examples what causes these differences with regard to word-class. There are other expressions which show variation in this respect within several corpora (e.g., por idiota and qué idiota tend to appear in both groups), which cannot be examined in detail here. On the whole, this seems to concern rather a small number of examples, and might just be due to errors in the Pos-tagging. It is therefore rather unlikely to be the reason for the big differences in the numbers shown in Table 8.

In addition to the corpus research for Spanish idiota, a quick consultation of some corpora of English was carried out in order to gain some information about the use of Engl. idiot with regard to word-class. The search in the COCA (Davies, 2008), GloWbE (Davies, 2013) and BNC (Davies, 2004)⁵ all lead to the same result in that none of them delivered any findings of idiot being used as an adjective. Of course, this does not necessarily mean that it actually is not used as an adjective, as other corpora might show evidence for its adjectival use. However, this use can probably be said to be less common than in Spanish, since in three different corpora, no evidence can be found.

The question underlying the parts-of-speech analysis was whether the introduction of covidiota into Spanish brings about an increase of usage of the term as an adjective. The data provides evidence for that. Moreover, the examination of corpus data was carried out in order to find out whether the usage of covidiota becomes more similar to that of idiota over time. A comparison of the numbers gained from the covidiota-data (Table 5) with the numbers in Table 8 shows that the result for covidiota as an adjective in T1 (7.25%) is lower even than that for idiota in the NOW-corpus (11.43%). Apart from being the earliest of the investigation periods, and therefore probably most strongly influenced by English, T1 is the least reliable with regard to information about word-class, since about half of all items contained in it could not be assigned to any word-class at all. The results for T2, i.e., use as an adjective in 33.59%, resemble the numbers from the Genre/Historical-corpus and the Spanish Web 2018 (esTenTen18). T3 comes closest to the result for idiota in the CORPES XXI. This latter finding is certainly influenced by the fact that, in the analysis of the covidiota-data, reference was made to the definitions and classifications published by the RAE (in the NGLE). The CORPES XXI being one of the corpora provided by the RAE, it can be assumed that Pos-tagging is based on the RAE's criteria as well, while the other corpora were created to match different criteria. Also, although the numbers for T2 and T3 with regard to the adjective come close to the numbers of some of the corpora, the numbers for the usage as a noun do less so. In general, comparability has to be considered somewhat limited by the fact that the covidiota-data contains not only the noun- and adjective-group, but also the third category of indeterminable cases which remain without classification.

To sum up, as exemplified above, the part of speech analysis for covidiota is, in some respects, controversial and, in some cases, it is not at all possible to assign a token to one of the two word-classes (see section Parts of Speech: covidiota as an Adjective and/or a Noun). Nonetheless, an increase over time in the use of covidiota as an adjective can be attested (even though the exact proportion might not be entirely uncontested, given the challenges of the analysis mentioned above). While the etymon was coined as a noun in English, it is increasingly used as an adjective in Spanish, which therefore can be considered one aspect of its integration into the target language.

As Table 7 shows, more than half of the tweets in T1 contain a definition of the term covidiot or covidiota. The most commonly found definition in English is shown in Figure 1.

Since this definition is usually attached to the tweet as an image file, the tweet itself offers space for either an accompanying comment, often of a metalinguistic kind, as in (29), or a comment in which the author uses the term for referencing, as in (30) or (31), either to express a statement, demand, etc. of their own, or merely to offer a sample sentence for how to use the term in Spanish.

In other instances, the tweet text only comprises the word covidiota, and is then followed by the English definition depicted above, thus offering the Spanish version of the term together with its proposed meaning and sample sentences in English. There are, however, also cases in which the tweet itself presents a Spanish definition of the term covidiota, which can either be a very close translation, as in (32) (for the first part of the definition), or, otherwise, show considerable deviations from the original, e.g., by adding new aspects and possible meanings, as in (33).

In total, 42 out of the 1,491 tweets (42 out of the 136 in T1, to be more precise) analyzed for this study were accompanied by this particular English definition of the original term Engl. covidiot. Another two comprised either a slight variant of it, which is, however, very close to the first one in terms of content, or only include the first part of the first definition, leaving out the second half.

Spanish versions of the definition, also circulating as image files, arise soon after the English one, and are also passed on through the social media networks. They are, however, less uniform and less frequent. There are only seven tweets within the corpus which contain such a file. The most frequent one is shown in Figure 2. It appears only three times throughout the corpus, but has an almost identical variant (which appears twice). The remaining two have to be considered independent versions. There is only one definition among them which depicts the Spanish word covidiota as either a noun or an adjective, while all the others classify it as a noun.

In most cases however, Spanish definitions appear within the actual text of the tweet, i.e., typed in by the individual Twitter user. As the Spanish definitions of the term during the first days of its existence constitute individually composed texts, as opposed to a “standard” definition which is spread in parallel for the English term, they exhibit a certain range of variation. Most of them agree on one basic meaning of the term covidiota, which corresponds to the first half of the definitions shown above, i.e., someone who ignores the appeal to keep physical distance from others, avoid social gatherings, etc., and thus is considered to further enable the spreading of COVID-19 [see for example (34), but also (35)]. Some of these definitions also take into account the second part of the definition, which describes a person who resorts to panic purchases and hoards large amounts of certain goods at home [e.g., example (36)]. A peculiarity of example (36) is the classification of the first sense of the word as an adjective, whereas the second is registered as a noun.

As has already been pointed out, in a number of uses, more than one of the categories indicated in Table 7 are present in the tweet, as in (37). Here, the speaker first makes a comment about the term covidiota (regarding its origin and form of appearance), which is a metalinguistic comment, and then gives a short definition of the term. Thus, this tweet qualifies for “containing a metalinguistic comment” as well as for “containing a definition.”

(38) is also an example in which metalinguistic use is combined with a definition. This example is particularly noteworthy because the author mentions the Spanish language academy (RAE) and calls upon their lexicographic authority to ask for verification of the term's existence. References to the RAE can be made out in five other tweets, too [see also (36) above].

While tweets as (39) are categorized as pure definitions, there are other instances which appear to contain a kind of definition, or explanation, too. At the same time, they also represent an actual use of the term, an application of the neologism in context, with the function of referencing. Tweets (40) and (41) represent manifestations of this particular type.

Both (40) and (41) are individual Spanish texts which contain the term covidiota and thus qualify for the category “referencing.” On the other hand (40) very explicitly defines the author's interpretation of the term, whereas (41) contains a more implicit clarifying element: the author implies that someone who organizes a Halloween-booze-up is a covidiota. This type of tweet, thus, differs from the actual (“pure”) definitions in that defining the term is not all, or not even the primordial function of the utterance, but only constitutes a part of it. As it seems, in (40) the author also intends to reassure the addressee, saying that he/she is not a covidiota. In (41), the Twitter user expresses a wish to be invited to a Halloween party. In both examples, the definition of covidiota is only part of the content conveyed by the texts. Tweets of this kind are therefore grouped together to form the type “referencing and negotiating semantic content.” Coming back to the types of alterity markers defined by Pflanz, this function can probably be said to correspond to the fourth type: explaining the neologism (X expliqué, Pflanz, 2014, 173–175).

Examples (42) and (43) illustrate the difference between this type and the one depicted as “referencing only.” Here, the author (tacitly) assumes a certain meaning of the term covidiota and uses the word accordingly, without further defining, explaining or discussing it.

The distinction between both categories is made because it provides a certain amount of insight regarding the proportion of tweets which actively and explicitly contribute to the collective negotiation of the meaning of this particular neologism among the Spanish-speaking Twitter community. It is, however, not always an easy distinction to make because it is not always clear where to draw the line between both types. The crucial question is whether a meaning-defining notion can be detected in the tweet or not. If a tweet constitutes a reply to another tweet, for example, it is hard to say whether a statement such as (44) should be assigned to the first or the latter category since we do not know what is meant by eso ‘that.' But it clearly refers to an action or event known to the addressee, and it does clearly represent a positioning of the tweet's author concerning the meaning of the term covidiota. Consequently, it has been assigned to the type “referencing and negotiating semantic content,” in line with statements as in (40).

In (45), it is not absolutely clear whether being in CITY is what would qualify as covidiota, or whether this would only apply to being in CITY and behaving in a certain way. For the latter, (45) could not be said to contain a defining element. Here, the latter option has been chosen.