Lexical and semantic variation in contemporary spoken Portuguese in urban Funchal and rural areas of Madeira Island

This paper aims to provide a dialectal and sociolinguistic approach and description of the use of the lexical and semantic regionalisms in contemporary spoken Portuguese in urban Funchal (the capital of Madeira) and in four rural areas of the island. The study of the lexicon and its semantic variation is based on recent samples of a semasiological questionnaire, applied to stratified informants by sex/gender, age group and level of education. The correlation of the results with the extralinguistic factors, geographical origin of the respondents on the island and sociocultural variables from the qualitative analysis and the quantitative point of view confirm that speakers from urban areas are less conservative than those from rural areas and showed flexibility and variability as well as stability of lexical meanings, two opposite and complementary lexicon characteristics. The traditional lexicon of Madeira studied highlights the coexistence of dialectal forms, some of them common to the Azores and Brazil, with European Portuguese standard variants. Therefore, the qualitative and the quantitative expression of the knowledge of Madeiran regionalisms and their lexical and semantic variations expose their vitality and support the heterogeneity of the European Portuguese language and the identity of the Portuguese spoken on Madeira Island.


Introduction
The main purpose of this paper is to study the lexical and semantic variation of some Madeiran regionalisms in contemporary spoken Portuguese, verifying the knowledge and consequently the linguistic vitality of 39 words and one expression in urban Funchal (capital of Madeira) and in four rural areas of the island. It is also important to know the vitality of these lexical units that belong to the Madeiran heritage lexicon or Madeiran linguistic heritage, according to Rebelo (2014): the linguistic uses of some words in the variety of Portuguese spoken in Madeira, especially regionalisms that correspond to old words that the younger generations do not use or recognize as part of their grandparents' and great-grandparents' way of speaking.
The phenomenon of knowledge and use of regionalisms in spoken Portuguese on the island of Madeira, such as non-standard vocabulary, is directly related to extralinguistic or geographical and sociocultural variables. The social stratification of the sample, in addition to the geographical factor (origin of the speakers on the island), explains some differences between the use of these words in the Funchal area and in the four rural areas of Madeira Island. Thereby, the external variables considered in this study are geographical varieties of spoken Portuguese in Madeira Island (urban vs. rural) and sociocultural (sex/gender, age group and level of education) variation in the knowledge of Madeiran regionalisms. The internal variables studied are lexical forms and semantic variation of the words, describing their variability and flexibility, as well as stability.
The theoretical framing of this study intends to show the articulation between geographical and social variation, more precisely the correlations between lexical dialectology and sociolinguistics. Dialectology is the discipline that describes and identifies dialectal areas inside a national language, as well as the factors that cause their formations, and lexical dialectology studies the vocabulary or lexical variation of the dialectal areas: the use of different words in different communities to designate the same concepts, but also the same words with different meanings. Segura (2013) writes about the lexical varieties of European Portuguese, explaining that dialect, traditionally, is a geographical variety of a language, and that dialects are varieties of simultaneous use inside the same language with their regularities and their systems, with their own standards, even if they are not usually fixed in any normative document. The author states that dialects present the language at different stages of its evolution, some retaining archaic characteristics that the standard language did not retain, others presenting innovative aspects that the standard language did not adopt. Concerning the lexical aspects of Portuguese dialects, Segura (2013) shows that the variation of the lexicon allows different groups of dialects to be distinguished, depending on a large scale of extralinguistic factors, namely historical, cultural and social, although it is possible to find some regularities in their geographical distribution. The author adds that different denominations can be found in the lexicon for the same object or notion, but also different forms that a word can present in the territory.
As stated by Segura (2013), the term variety has tended to replace dialect, after the development of sociolinguistic studies that establish correlations between social variables and linguistic phenomena. The heterogeneity of the language is a part of linguistic economy of the community because it is necessary to satisfy the linguistic demands of daily life. That is why Labov (1982) states that the stratification in language use in society is not chaotic but obeys certain regularities. Traditional dialectology with its classical requirements (a main informant over 50 years old, with little schooling because of their usefulness for retrieving linguistic data not influenced by urban life or centres of prestige) became multidimensional, aiming to capture social variation together with geographical variation in rural and urban areas. It is thus more complete and explains better the correlations between the different external factors of linguistic variation. This means that dialectology and sociolinguistics have become complementary, having some common methodological parameters, namely the collection of samples of spontaneous speech of respondents inserted in their linguistic communities, and standard variants are considered as reference units in the comparison with geographical and social non-standard variants.
In this study, we consider the multidimensional dialectology focused on linguistic variation in the territory of Madeira Island, namely the diffusion of lexical and semantic regionalisms in the speech of informants conditioned by geographical factors (rural and urban origin of the speakers), from both genders, different age groups and schooling, and reflecting social stratification. The research is based on the social dimension of language use, built on earlier geographical dialect methods of language variation, regional and social corpus-based studies of language. Speakers tend to select linguistic variants for use, consciously and/or unconsciously, revealing a usage based on spoken language. Several factors contribute to the probability of one variant being selected and individuals might not be aware of the pattern of their choices, related to diatopic variation (geographical factors), diastratic variation (socially stratified factors) and diaphasic variation (contextual factors). By analysing usage data, we can observe linguistic behaviour and the geographical variants, we sometimes need to add historical and ethnographic information about the Madeiran traditional lexicon under study. As Cunha and Cintra (1999) claim, among the large varieties of a language, the standard is the most prestigious because it is the model and a force against variation. The denotational or referential synonyms (words from the same lexical class with similar meanings) reveal the existence of competition between different lectal varieties of a language with the use of alternative words that describe the same concept/referent. Prestige associated with certain forms in co-occurrence with others can lead to elimination of those considered less prestigious. However, this did not happen with the word semilha, which, rather than disappearing, gained its own Madeiran identity. Verdelho (1982) affirms that lexical regionalisms have a local or regional quality, with many subtle gradations, noting that some of these lexical items may be different from the common language in form and in sense, designating specific and characteristic circumstances of one region, while others can be similar to standard language but used in a particular region with a special meaning, mainly to express situations, facts or regional objects from that geographical area. More isolated regions preserve the local way of speaking better because rural areas normally have difficulty communicating with other localities. Teixeira (2006), discussing variation in lexical semantics, says that standard words regionally correspond to diversified forms. He claims that there is an opposition between the standard lexicon and dialectal words with their popular forms, even if they have the same semantic and referential value, and that the standard words have a superior sociolinguistic status, i.e. more linguistic and social prestige. In this sense, there are standard variants and regional variants (called regionalisms), forms that do not belong to the standard language. Thus, these locally used terms tend to be avoided, with the linguistic uniformization of schooling as a vehicle of standard words. Thereby, it is important to observe the vitality of the Madeiran traditional lexicon and find ascertain whether it is detected only in rural areas or also in the urban municipality of Funchal. Thus, we can observe the existence or not of lexical and semantic differentiation between localities and, considering more traditional words and meanings, which geographical areas are more conservative and innovative. Mateus et al. (2003), on the use of language and variation, state that it is easy to see that the variation extends to the whole set of speakers of the same language, although such a process is triggered by factors as distinct as belonging to a given generation, insertion in a particular socio-professional group, prolonged experience in a given region, or contact with other communities, etc., claiming that all variation seems to be based on a general principle of frequency of use. In this study, we test mainly vocabulary related to daily rural life, not common in the urban centres, considering that rural areas are conservative and taking the innovative character of cities into account. However, we must consider that Funchal, the capital of the Madeira archipelago, has peripherical localities that were rural areas until recently, and we need to consider their evolution from rural isolated areas in the periphery of Funchal city to new urban areas, as well as the influence of the internal migration of population from Madeira's rural areas to Funchal peripheral areas. These are the called rurban areas (Bortoni-Ricardo 2004), characterized as transition zones with a linguistic continuum between the rural and urban varieties, preserving features of their linguistic variation, namely the lexical and semantic repertoire. As explained by Carrilho, Magro and Álvarez (2013), linguistic facts are inextricably linked to a human community, in each time and territory. The internal variation results mainly from social mobility and access to formal education. Such contacts with standard language or centres of prestige often elicit attitudes of adoption, imitation or rejection, among others. The authors stress the change in dialects over time and the relation between dialectal variation and human settlements.
In this sense, we refer to the project Tesouro do Léxico Patrimonial Galego e Português (Thesaurus of the Galician and Portuguese Heritage Lexicon, from now on Tesouro), presented by Álvarez Pérez and Sousa (2013), which functions as a dialectal lexical portal. It offers lexicographical materials, the dialectal lexicon and its variation, with the indication of semantic fields linked to traditional, regional or local cultures from Galicia, Portugal (including vocabulary from Madeira and the Azores) and Brazil, with the purpose of contributing to the study of inherited material and non-material traditions. This project aims to allow comparative analysis and studies of the distribution and spread of lexical items in the countries considered, providing an electronic edition that describes the heritage vocabulary, semantic fields and geographical specification or location of dialect vocabulary and of variant forms. For Madeira, Tesouro uses the dialectal materials from degree monographs such as Cintra (2008), excluding the vocabularies from nonacademic authors that do not distinguish regional from popular forms and do not indicate collection date or location of the dialectal forms in Madeira archipelago. These lists of what they called popular Madeiran vocabulary, for example Silva (1950), Sousa (1950), Caldeira (1993Caldeira ( [1961) and Pestana (1970), document some forms and meanings of the lexicon studied in this article, which is why we mention them. Concerning the lexical and semantic variation in Madeiran spoken Portuguese, in recent years, it has been the topic of a considerable number of researches, mainly in University of Madeira master's theses and other academic studies: Santos (2007), Rebelo (2007), Figueiredo (2004, Santos (2013), Bazenga (2014Bazenga ( , 2015, Teixeira (2015), Rebelo and Nunes (2016), Nunes (2014Nunes ( , 2017Nunes ( , 2019aNunes ( , 2019b, Andrade, Bazenga, Rebelo and Nunes (2018). We also refer to them when they present terms that are part of this study.

Cognitive Semantics
To explain and better understand the lexical and semantic variation of Madeiran regionalisms, we use the notions of cognitive semantics. Cognitive linguistics studies the conceptual, dynamic and encyclopaedic nature of the lexical meaning, through three interconnected perspectives: the meaning in the mind (focusing on the phenomenon of polysemy and the cognitive processes that originate it, like metaphor and metonymy), the meaning in the culture (underlying the cultural specificities of the lexical concepts or word meanings) and the meaning in society (the social meanings of the lectal variation). About the meaning in the culture, Silva (2010a) mentions the historical and cultural specificities of the meaning. He explains that there is the experiential base of the meaning (the embodiment of mind experience) in the human body, which is universal, and, at the same time, that the meaning has historical and cultural origins that are not universal. It means that the embodiment of the mind, cognition, language, and meaning is situated in a historical and sociocultural context, as is the case of spoken Portuguese on Madeira Island. So, cognition is socio-historical and consequently the embodiment of mind or the cognitive notion of embodied meaning implies the sociocultural situatedness of the lexicon, defined as the ways in which individual minds and cognitive processes shape themselves through social and cultural interactions. These concepts are associated with the flexibility, variability, and instability of the meanings, explaining lexical and semantic variation. As a language is a social diasystem and semantic knowledge is differently spread by the members of a linguistic community, linguistic variation is the best manifestation of the social dynamics of the meaning, especially the lectal variation (all kinds of linguistic varieties or lects -national varieties, dialects, sociolects, etc.). So, the standard word may not be equivalent to the regionalism because it does not comprise the same cognitive reality or knowledge of the word, which is why lexical and semantic variation coexist in one linguistic space or territory. We can study these socio-cognitive dimensions of linguistic variation using consistent empirical methods, something we intend to do in this paper.
To test some Madeiran regionalisms in the speech of informants from different geographical areas of Madeira Island, we applied a semasiological questionnaire, according to the distinction between semasiology and onomasiology established in lexical semantics by Baldinger (1964). While the semasiological approach takes the word as a departure point to analyse its senses or referents and creation of new senses of a lexical item (polysemy that comes from the fact that one word has more than one meaning with a semantic association between them), the onomasiological approach starts from the concept for the different words or expressions that designate it, including the expression of a concept by a new or different lexical item (synonymy, different words with similar senses). We use the semasiological approach and, because of the polysemy of words, we have occurrences of synonymy between the lexical units tested e.g., bisalho and busico (meaning 'little'), burquilha and maltrapichado ('scruffy'), sobressi and trapichado or trapichento ('crazy'). Silva (2010a) explains that denotational synonyms (terms that denominate the same concept or referent) are evidence of regional, social, stylistic and pragmatic-discursive differences that bring about the existence and competitiveness of varieties of a language and, as we can see with these examples, also within the same variety. As the linguistic meaning is dynamic and flexible, it explains the existence of the semasiological phenomenon of polysemy. When a word is polysemic a synonym can comprise only some of its senses, partial synonymy. The author states that synonymic words generally are dialectal variants (as is the case with semilha) and archaisms (in the case of burquilha) or neologisms (in cases of trapichado and trapichento from the Madeiran word trapiche). Regarding semasiology, there are different cognitive processes in the origin of the various senses of a word (polysemy) and their relationships: metaphor and metonymy (a transfer of senses resulting from an association respectively by similarity and contiguity, called figurative senses, for example, respectively joeira, a round object to winnow wheat, compared to voluminous hair, and arriota 'problem' or 'confusion' also carrying the meaning of 'party', probably because it can cause problems); generalization (a semantic extension of the word, like in cieiro from 'dirty skin' to 'dirtiness' in general); and specialization (a restriction of the sense, as in asservado from a 'quiet person' to a 'careful person'). Silva (2010b) writes that the phenomenon of polysemy or meaning differentiation is probably the hardest problem and the one with greatest theoretical and methodological implications. One of the most problematic areas in semantic analysis is to know how many meanings a word has, indicating the semantic flexibility, nuances and adaptations of the word that occur in specific contexts, the inevitable variability and change. This means instability of the polysemy and of the flexibility of meaning, which leads to no stable differentiation of senses. He explains that the flexibility of meaning comes from the fact that it represents the world, and this is a changing reality, so different situations give rise to nuances and cases of deviations in the lexical meanings. It means that the system is flexible enough to adapt to new circumstances; and the structural stability enables the system to function because it maintains its general organization during the same time. It is the apparently contradictory double effect of flexibility (adaptation of a category to new contexts) and stability (interpretation of new circumstances based on previous knowledge) that explains the lexical and semantic variation within a speech community and among different varieties of a language, revealing the natural objectives of communication: expressiveness and efficiency. Chaves (2013) states that words may vary in the richness and precision of their meaning, because one important characteristic of the sense of the words is their vague semanticlexical nature. He explains that the set of semantic features that defines one word we call sense, and the sense of a word or expression can be understood as the definition that the word has in the mental lexicon of a speaker and that distinguishes it from other words. The majority of words have a vague sense and this semantic propriety is called vagueness, allowing some variation in the words' sense and explaining their flexibility. Thereby, the specification of the word can be different from one speaker to another and the vaguer the sense of a word, the more comprehensive is its extension, related to knowledge of the world and personal experience. As regards polysemy, the author claims that words are mainly ambiguous i.e., diverse senses correspond to one lexical form and the greater its frequency in use, the more this happens. Although, in our study of Madeiran regionalisms we can see that the best known words which are used by the informants are the least polysemic and most stable, as is the case of dentinho 'appetizer' and semilha 'potato'. Silva (2013) certifies that onomasiological change can occur by word formation (derivation, composition), deformation (phonetic), semasiological extension of existing words, creation (neologism) and borrowing. Therefore, linguistic contact with other languages is a means of lexical variation or differentiation between dialects, because speakers are influenced especially in the lexicon, as is the case of some words from Spanish and English that are Madeiran lexical loans (as mentioned before, respectively semilha and bambote). Andrade, Bazenga, Rebelo and Nunes (2018) say that language contact happened in Madeira because there was an ancient English community on the island, and because of the contact of local population with English ships in Funchal through the trade in regional products. Concerning contact with Spanish, authors indicate the proximity of the Canaries, the historical presence of Galician individuals in Madeira, and the migration of local population mainly to Venezuela, who returned to the island. Silva (2015) shows that lexical loans have impact on diachronic lexical variation with implications on onomasiological homogeneity within and across different varieties of a language (onomasiological variation between semantically equivalent words or denotational synonyms of which lexical loans are part). He refers to the distinction between strict loan and cultural loan, or superfluous and necessary loans. The last one designates a new object or concept as the only lexicalization of one concept, as is the case of the Madeiran regionalism bombote (for trading regional products in a small boat), while the other duplicate existent terms competing with them, as is the case of semilha (instead of the Portuguese word batata 'potato'). As we can see, lexical loans, synonymy and polysemy occur in languages and in dialects, with even more semantic and lexical variants originating in relation to the standard variety.
The criteria taken into consideration for choosing the regional vocabulary for this study was the popular expressiveness of the words used mainly in rural areas, testing knowledge of lexical items in individuals from Funchal and comparing them with others from different rural locations on the island. Therefore, the selection was made considering the traditional or heritage lexicon of Madeira to ascertain its vitality. The majority of these regionalisms tend to be less known and/or to present different or unexpected meanings when compared to those documented in Madeiran vocabularies, and some words are not even registered. Through the application of the questionnaire in August and September 2018 to collect the corpus of analysis in different geographical areas of the island, we intended to confirm the variability and flexibility of the lexical units and their meanings in Portuguese spoken in Madeira, considering four levels of knowledge of the tested words: known with the expected meaning(s), other meaning(s), unknown, and with the standard meaning, in a comparative perspective.
We do not follow the main premise of variationist sociolinguistics, the quantitative model of the linguistic variation, regarding the ideal sample number of informants representativeness by location (18 individuals stratified into cells or categories of three men and three women by three age groups and three levels of education). Nevertheless, we have an equal number of informants from each of the social groups defined in sociolinguistic studies as a stratified sample, and we present the frequencies and percentages by locations and by individuals. The data of the questionnaires applied to 30 informants is limited to six respondents by location, allowing the comparison of the five different geographic areas. Despite the limitations of this exploratory study, in terms of informants by location (only six informants instead of 18), this empirical approach seems to be representative of the use of the tested regionalisms in the Portuguese spoken on the island of Madeira.
Although this is a study of spoken language, as our purpose is to study semantic and lexical variation, we applied a written questionnaire filled out by the speakers, except the older individuals, who had the assistance of the researcher to write their answers (graphically recording the phonetic forms used). The questionnaires were applied in the houses of the respondents, in a controlled space and time with the presence of the investigator so that informants did not ask others for answers or consult other means of information, in the case of younger individuals. In this sense, the lexical questionnaire method allows to know the linguistic behaviour of individuals reporting spoken language used in their community i.e., it allows a dialectological, sociolinguistic and variation study, revealing the complexity and systematicity of the language structure.
The study is limited to five inquiry points on Madeira Island: three from the south coast (Câmara de Lobos, Ponta do Sol and Funchal, the capital of the archipelago), one that belongs simultaneously to the south and north coast (Machico), and one from the north coast of the island (Porto Moniz), as we can see in the Madeira and Porto Santo Islands map in Figure 1 below. The study does not include Porto Santo, the smallest populated island of Madeira archipelago.
The sampling procedure consisted of including the selected regional vocabulary in a semasiological questionnaire (starting from the designation to obtain the meaning). The questionnaire presents two different parts. The first part is about the sociocultural characteristics of the informants: sex or gender, age group, education, but also geographical origin inside Madeira Island, including their parents and grandparents, and linguistic contacts with rural or urban areas. The second part consists of the words presented without a linguistic context or semantic field to collect their possible different denotative and non-denotative meanings, and examples of use of the lexical units tested in the five different Madeira Island areas, to confirm the meanings indicated.
The data was collected from 30 respondents, whose answers constitute the basic sample of the corpus, six participants in each survey point: Funchal, Machico, Câmara de Lobos, Ponta do Sol and Porto Moniz, distributed by sex/gender (M -men and W -women), three age groups (A -18-35 years, B -36-55 years and C -56-75 years) and three levels of education (elementary: 5-8 years, intermediate: 9-11 years, high: more than 11 years). The choice of the informants considered the dialectology requirements: local speakers and residents, with few contacts with other regions, but also the social stratification (as mentioned above). Table 1 shows the sociocultural characteristics of the informants.
In data collection, definitions of the lexical items are the fundamental part of this study, including information about local or regional aspects, such as description of tools, their parts, how they are used and related customs that aid understanding of the meanings. The comparison with different lexicographic sources allows analysis of the specificities of the non-standard Madeiran vocabulary and its classification in lexical and semantic regionalisms. We use the Dicionário da Língua Portuguesa Contemporânea/Dictionary of Contemporary Portuguese Language (DCPL) and Dicionário Priberam da Língua Portuguesa/ Priberam Dictionary of Portuguese Language (Priberam) online, when it is a semantic regionalism, for the standard meanings and dialectal senses of the terms in European Portuguese and in Brazilian Portuguese. Then we present the information offered in the Dictionary of Regionalisms and Archaisms (DRA), from Leite de Vasconcelos, and the definitions documented in Madeiran vocabularies, when the words are listed. The Madeiran vocabularies consulted were Silva (1950), Sousa (1950), Caldeira (1993Caldeira ( [1961) and Pestana (1970). We also refer to dissertations and other academic studies that include the lexicon studied. Finally, we show the comparison with the Tesouro database (with variants, headwords, semantic fields, and geographical location provided by multiple sources). The information from Madeira in Tesouro is from linguistic-ethnographical monographs (generally conceived of as supervised academic works about the speech or language variety of a narrowly defined area or areas), namely Macedo (1939), Monteiro (1945), Gonçalves (1956), Rezende (1961) and Nunes (1965). These are the studies of different Source: https://www.marktest.com/wap/a/n/id~14c6.aspx. aspects of regional culture and of the dialect of some localities of Madeira archipelago used by Cintra (2008). Our purpose is to allow a comparison of the regional meanings of the words, when they are documented in the listed vocabularies and in Tesouro, with the answers of the respondents to our questionnaire.
The comparison of the words with the Tesouro documentation allows us to see the circulation of words between different areas to study points of convergence and divergence between Madeira, the Azores, mainland Portugal, Brazil, and Galicia. Phonetic variants provide useful information e.g., tratuário and trotoário as different forms of a lexical loan from French trottoir. Geographical indications about the place where the variant was collected or observations about whether the form is listed in the Portuguese language dictionaries and/or in the Madeiran vocabularies are offered. The mention of the Tesouro is also important because its lexicographic sources provide examples of use for the forms given that represent a sample of the living speech of the locality, as in our study. So, for the lexical units studied we intend to offer information about the collected meanings, and their other historical and sociocultural aspects. Due to the length of the article, this supplementary information about the lexicon is given in the appendix.

Data Description and Analysis
We chose the theoretical frame of the cognitive semantics, a model oriented to meaning and use, to describe the semantic and lectal (regional or local and social) variation and the cultural and historical specificities of some word meanings in contemporary spoken Portuguese on Madeira Island. Our purpose is to show the inevitable flexibility and variability of the meanings of Madeiran regionalisms, based on their use through the analysis of the data collected for each lexical item with its semantic variation in Funchal, comparing it with the other inquiry points, first the rural areas from the south coast (including Machico), and then with that of the north coast of Madeira Island.

Madeiran regionalisms tested with greater vitality
According to the answers given by the informants from the different geographical areas, of the 15 lexical regionalisms, 24 semantic regionalisms and one expression the best known Madeiran regionalisms (by all the speakers), those with more vitality and less variability of meaning are: i) azougar, with the phonetic variants azoigar/azagar (verb, 'to die', used mainly of animals), a semantic regionalism, meaning: dar um mal ('feel unwell/dying'), informant 1, and estalar/morrer ('to die'), informants 2, 3, 4, 5 and 6 in Funchal, as well as informants 1 to 6 in Machico, in Câmara de Lobos and in Porto Moniz, and respondents 2, 3, 5 and 6 in Ponta do Sol. This word has only two correlated meanings: 'to die' and 'to feel unwell/dying'. In this case, there is stability of the word meaning related to the fact that practically all informants in Funchal and in rural areas of Madeira Island know the word with its traditional sense, revealing great vitality. ii) bisalho (noun, 'chick or little chicken'), a semantic regionalism, with the graphic variant bizalho, meaning: pintainho ('chick' or 'little chicken'), informants 1, 3, 4, 5 and 6 in Funchal, respondents 1 to 6 in Machico and in Ponta do Sol, pineco (with the same sense), informants 1, 2, 4, 5 and 6 in Porto Moniz, and 1 and 3 in Câmara de Lobos, but also criança ('child'), speaker 2 in Funchal, informant 3 in Câmara de Lobos, as well as pequeno ('little'), by generalization of the sense, respondents 4, 5 and 6, and miúdo pequeno ('little boy'), speaker 3 in Porto Moniz. This word seems to be a conservative form that has the ancient specific meaning of 'chick' in Madeira, that started as a metaphor to designate 'little child' or 'little boy' as well.
vii) furado (noun, 'tunnel'), a semantic regionalism with one main answer: túnel ('tunnel'), informants 1 to 6 in Funchal, respondents 1, 2, 4 and 6 in Machico, speakers 2 to 6 in Câmara de Lobos, informants 1 to 6 in Ponta do Sol, and respondents 1, 2 and 4 to 6, in Porto Moniz; com buraco/roto ('with a hole/shabby'), informant 3 in Machico, informant 1 in Câmara de Lobos, and informant 3 in Porto Moniz; (tem) bicho ('with a bug'), respondent 5 in Machico. The last two senses of the word are standard in European Portuguese. It is a polysemic word that in Madeira, as a dialectal variant, corresponds to the denotational standard synonym túnel 'tunnel'. The regional meaning revealed great vitality in Funchal, with no semantic variation between the informants. viii) semilha (noun, 'potato'), a lexical regionalism that is a loan from Spanish semilla 'seed', with the same meaning at all inquiry points: batata ('potato'). It exists in Madeira in co-occurrence with the standard term batata and became dominant as a feature of Madeiran lexicon, but not used in the expression batata frita 'French fries'.
Of the lexical units contemplated in this study, these seem to be the most widely used and have more vitality in contemporary Portuguese spoken on the island of Madeira, perhaps the ones with more social prestige or due to their strong expressiveness in the regional culture. As they are spoken and not written words, they present phonetical and graphical variation when written as answers in the semasiological questionnaire e.g., azougar/azoigar/azagar, bisalho/bizalho and busico/buzico.

Madeiran regionalisms with the greatest variability in meaning
The Madeiran regionalisms with the greatest variability in meaning, and greatest polysemy, are characterized by less stability. They seem to be less well known, as shown by the co-occurrence of the regional meaning alongside the standard meaning: i) abicar (verb, 'push' and 'jump'), a semantic regionalism with the dialect form or variant abicar-se (in Madeira). In Funchal, the answers collected on the questionnaire for this lexical item were: jogar-se/atirar-se ('to jump'), informants 2, 3, 5 and 6. We recorded the same meaning in Ponta do Sol, respondents 1, 2, 3, 4, 6, and in Porto Moniz, informants 1, 2, 4. In Funchal cair ('fall') was also documented by respondent 4, as well as in Ponta do Sol, speaker number 5, and in Porto Moniz, informants 3, 4, 5, 6. The same meaning occurs also in Machico, together with jogar ('throw away'), informants 1, 2, 4, synonym of atirar ('throw'), respondent 3. In Machico, speakers 5 and 6 gave the answer empurrar ('push'). In Câmara de Lobos, the word was recorded with the sense of matar-se ('to kill oneself'), informants 4, 5 and 6. As we can verify, the meaning of the word is given through synonyms, revealing that abicar(-se) is an expressive lexical type with a sociocultural function. The polysemic character of the word translates its dynamic and flexible meanings, with a semantic association between them, as it is the case of 'to jump' and 'to kill oneself', 'to fall' and 'to push'.
ii) aboseirado and emboseirado (adjective, 'lazy'), lexical regionalisms derived forms from boseira ('bullshit'), meaning: in Funchal, sentado/deitado ('seated/lying'), respondents 1, 3, 4, 5 and 6, with the same sense in Machico, informants 1, 2, 3 and 6, as well as in Ponta do Sol and Porto Moniz, for all six speakers. In Câmara de Lobos, we also recorded the meaning sentado ('seated'), informant 1. In Funchal, we also documented the meaning acomodado ('accommodated'), informant 2, the same as in Machico, sem pachorra para fazer nada ('with no patience/time for anything'), respondents 4, 5. In Câmara de Lobos, we found a similar sense, malandro ('lazy'), speakers 2, 4, 5 and 6, together with pessoa parada ('inactive person'), informant 3 (an expressive lexical type from a cognitive process of metaphor with 'bullshit'). The meaning of the word with the forms aboseirado/ emboseirado is given mainly by synonyms and by a periphrastic form sem pachorra para fazer nada, equivalent to 'with no patience/time for anything'. The sense pessoa parada 'inactive person' seems to be a specification of the word meaning, and malandro 'lazy' a generalization of the sense. This semantic variation results from the vagueness of the word, the characteristic of the lexicon that allows flexibility of meaning and consequently polysemy.
vi) cieiro (noun, 'dirtiness on the skin'), a semantic regionalism with the definitions: suor acumulado na pele ('accumulated sweat on the skin'), respondent 2, sujidade/sujo ('dirtiness/dirty', generalization of the word sense), speakers 3, 4 and 6 in Funchal, and informants 3, 5 and 6 in Ponta do Sol, as well as informants 2, 3, 4 and 5 in Machico; mal lavado/sujidade no corpo/pele/pessoa encardida/suja ('badly washed/dirtiness on the body/skin/dirty person', the previous meaning), speaker 5 in Funchal, and informants 2 to 6 in Câmara de Lobos, also speakers 2 to 6 in Porto Moniz; pedra para esfregar ('stone to rub', through a cognitive process of contiguity or metonymy), informant 6 in Funchal; in Machico, tarro (synonym for 'dirtiness'), respondents 1 and 6. Other meanings of the word were: lábios gretados ('chapped lips'), the standard meaning, informant 1 in Funchal, and by metonymy, association or analogy, cera de lábios ('lip balm'), respondent 1 in Câmara de Lobos. Curiously, another sense of the word collected was erva ('herb'), speaker 4 in Ponta do Sol, with apparent lack of semantic relationship with previous ones. There is a linguistic continuum in the distribution of the traditional meaning of the word, despite the flexibility of the meanings.
ix) embeiçado (adjective, 'without money'), a semantic regionalism with three different definitions: apaixonado ('in love'), the standard meaning, informants 1 and 2 in Funchal, and respondent 3 in Machico, also speakers 4 to 6 in Câmara de Lobos, and respondent 1 in Ponta do Sol; sem dinheiro/teso ('with no money'), the traditional Madeiran meaning, respondents 3 to 6 in Funchal, and respondents 2, 5 and 6 in Machico, as well as informants 3 to 6 in Câmara de Lobos, informants 2, 3, 5 and 6 in Ponta do Sol, and informants 2 to 6 in Porto Moniz; amuado ('sulky'), informant 1 in Machico and in Porto Moniz; mal-humorado ('grumpy'), informant 1 in Câmara de Lobos (new and unexpected meanings). There is an opposition between the meaning of the word in the rural and more conservative area of the north of the Island, Porto Moniz, presenting the traditional meaning 'penniless', and the other more innovative areas where the standard meaning of the word occurs, 'in love'.
xv) quinau (noun, 'opinion/advice'), a semantic regionalism with the meanings: dar uma opinião/conselho/parecer ('give an opinion/advice'), informants 5 and 6 in Funchal, respondents 3, 4 and 5 in Machico, respondents 1, 5 and 6 in Ponta do Sol, informants 2, 4 and 6 in Porto Moniz; exemplo ('example'), informant 6 in Machico; matar ('to kill'), maybe confusion with the verb quinar (above), respondent 5 in Câmara de Lobos. The traditional meaning ('give an opinion/advice') predominates, along with other senses of the word. xvi) renheta (adjective and/or noun, 'grumpy'), a semantic regionalism that seems to be a phonetic variant of ranheta ('snotty') from ranho ('snot'), defined as: pessoa resmungona/rabugenta ('grumpy person'), informants 1, 3 and 5 in Funchal, and informants 1 to 6 in Machico, as well as respondents 1 and 3 to 6 in Câmara de Lobos, speakers 1 to 6 in Ponta do Sol, and informants 1 to 4 in Porto Moniz; zangado ('angry'), informant 6 in Porto Moniz; discutir/reinar ('discuss'), respondents 4 and 6 in Funchal; pessoa reles ('bad person'), informant 5 in Porto Moniz. The flexibility and vitality of the word on the island of Madeira show its expressiveness and efficiency in relation to the synonyms of the standard variety. xvii) sobressi (adjective, 'non-normal person'), a semantic regionalism corresponding to the meanings: pessoa anormal/sem juízo/que não joga bem da cabeça ('non-normal person/crazy'), informants 3 to 6 in Funchal, and respondents 3 to 6 in Machico, as well as informant 5 in Ponta do Sol, and respondents 2 to 6 in Porto Moniz; desorientado ('lost'), informant 1 in Ponta do Sol; falar sobre uma pessoa ('to talk about someone/bad-mouth'), respondent 3 in Câmara de Lobos. The predominant sense of the word is 'non-normal person/crazy', the traditional meaning. We documented also 'lost', having a semantic relationship with the previous meaning, and an unexpected new sense 'bad-mouth'. xviii) sovado and sovento (adjective, 'sweaty/dirty'), sovado is a semantic regionalism and sovento probably is a phonetic alteration of sebento (from sebo 'sebum'), meaning 'filthy' or 'dirty'. It could also be a new word formed from the verb sovar ('knead'), which, on the island of Madeira, has the meaning of 'get dirty'. The words were defined as: nojento/sujo ('dirty'), informants 3 to 6 in Funchal, and respondents 1 to 4 in Machico, as well as 1, 5 and 6 in Câmara de Lobos, and informants 5 and 6 in Ponta do Sol; pessoa nojenta/que cheira mal ('dirty person/who smells bad'), respondents 5 and 6 in Machico, also 2 to 4 in Câmara de Lobos, and informants 2 to 6 in Porto Moniz; agredido fisicamente ('physically assaulted', standard meaning), respondent 2 in Ponta do Sol. The traditional meaning in Madeira Island is 'dirty' for a thing and for a person. There is convergence and continuity between the responses collected in Funchal and in rural areas. The fact that a respondent gave the standard meaning reveals that the regional meaning is not well known.
xix) terçol (noun, 'youngest son'), a semantic regionalism with the collected meanings: terçolho no olho ('inflammation in the eye'), the standard meaning in European Portuguese, informants 1 and 3 to 6 in Funchal, and 1, 2, 3 and 5 in Machico, as well as respondents 1, 2 and 4 to 6 in Ponta do Sol, and 1, 2 and 4 to 6 in Porto Moniz; filho mais novo ('youngest son'), the regional meaning, informants 3 to 5 in Funchal, and respondents 1 and 6 in Machico, as well as 1 to 6 in Câmara de Lobos, 5 in Ponta do Sol, and informants 3 and 5 in Porto Moniz; criança ('child'), by a cognitive process of generalization from the previous meaning, respondent 4 in Machico. As we can see, responses were split between the standard and regional meanings, which means that the regional sense of the word is losing ground especially among younger speakers. xx) trapichado and trapichento (adjectives, 'messy' and 'crazy'), lexical regionalisms, new derived forms by suffixation from the Madeiran term trapiche ('institution for the mentally ill', 'crazy person', 'alcoholic', and 'confusion'), presenting different meanings: atontalhado/tarado ('crazy'), informant 3 in Funchal, and 4 in Machico; mal vestido/malarranjado ('scruffy/ill-arranged'), respondents 4 and 5 in Funchal, and 6 in Machico, as well as informants 1, 2 and 6 in Porto Moniz; trapalhão ('bungler'), respondent 3 in Machico (all of these with the regional sense); bem vestido ('well-dressed'), informants 2, 5 and 6 in Ponta do Sol; alegre/contente como um bobo ('happy as a crazy person'), respondent 4 in Porto Moniz; jogar à porrada ('give and take a beating'), informant 6 in Funchal; dar um mal ('feel unwell'), respondents 5 and 6 in Câmara de Lobos (new and unexpected meanings); pessoa carregada com as coisas que leva ('person loaded with a lot of things'), a confusion with the word atrapichado (from atrapichar), respondent 5 in Porto Moniz. The word reveals great variability, from 'crazy', 'ill-arranged' and 'bungler' to 'well-dressed', with the opposite sense, and 'feel unwell'. The meaning 'give and take beating' is related to the sense of 'crazy' person.

Madeiran regionalisms with the least vitality
Some of the lesser known Madeiran regionalisms studied are archaisms and seem to be dying, disappearing, or losing vitality, such as: i) arriota (noun, 'confusion' or 'problem'), a semantic regionalism (from the standard word arriosca with some phonetic alterations), with the assigned meanings: in Funchal, sempre rindo ('always laughing'), informant 6; in Machico festa ('party'), respondents 1 and 6; in Câmara de Lobos aldrabão ('crook'), speaker 4; in Ponta do Sol, there was no answer; in Porto Moniz, descer ('to descend'), informant 6. This word was revealed to be less known (it was not recognized in one location) and showed not only variability but also instability of meaning. The sense aldrabão 'crook' is related to the standard meaning 'cunning manoeuvre to trick someone'. We can see that the old meaning 'confusion' and/or 'problem', previously documented by old people in rural Madeira, is been replaced by new meanings. Some of them maybe because of confusion with other words or expressions, like 'always laughing', 'party', and 'to descend' (probably by confusion with arriar 'drop').
ii) bambote (noun and/or adjective, 'small boat' and 'loafer', but also 'rascal'), a lexical regionalism that is a loan from the English language bumboat ('little boat to carry goods'), with the documented meanings: barco de pesca ('fishing boat'), informant 3, vendedor a bordo dos barcos ('salesperson on board'), speaker 4, and homem fino ('elegant man', specialization of the meaning of the word, because these men had money to buy new clothes), respondent 6, in Funchal; comprador/vendedor de tudo ('buyer/seller of everything', generalization of the sense), speaker 3, in Machico; canalha ('scoundrel'), informant 5, in Funchal; bêbado ('drunk'), respondent 5, and bandido ('bandit'), informant 6, in Machico; cair ('fall') respondent 6, in Porto Moniz. There were no answers in Câmara de Lobos and in Ponta do Sol. We can see the evolution of the word meaning, with the disappearance of the referent, from the name of the fishing boat used by the bamboteiro to sell products on ships, that was a 'elegant man', to the generalization or semantic extension of the sense of the term to 'buyer/seller of everything', and then, due to the fact that these men worked only when there were ships in the port of Funchal, they were seen as rascals, bêbado 'drunk' and bandido 'bandit', specialization of the sense, losing its relationship with the original meaning, and having probably also a semantic association with cair 'fall (down)'. The word is a clear example of the sociocultural situatedness of the meaning, existing in Madeira because it is related to the socio-historical reality of the profession bamboteiro. This form and bambote are respectively phonetic changes of bomboteiro and bombote. In Figure 2, we reproduce a photograph from the Vicentes Museum to illustrate the object and concept of bambote and bamboteiro described and defined above.
iii) burquilha (noun and/or adjective, 'rude person' and 'dullard'), a lexical regionalism with the phonetic variant broquilha, meaning: pessoa da serra que não sabe falar e mal vestida ('person that doesn't know to speak or dress'), informants 5 and 6, in Funchal; pessoa do Porto da Cruz ('person from Porto da Cruz'), respondents 1, 2 and 6, and pessoa sem modos ('badly behaved person'), speaker 3, in Machico; pessoa mal vestida ('scruffy'), informants 2, 4 and 6, pessoa que não sabe fazer as coisas ('person that doesn't know how to do things'), respondent 3, and pessoa mal-encarada ('bad person'), speaker 5, in Porto Moniz. Other meanings documented were: pessoa tonta ('foolish person'), informants 2 and 3, and maricas ('sissy'), speaker 4, in Funchal. No answers were found in Câmara de Lobos and Ponta do Sol. The meanings 'person that doesn't know to speak or dress' and 'person from Porto da Cruz' are synonyms because this locality is from the north of Madeira island and people from there were considered rustic people. So, it is the more conservative meaning of this ancient lexical form and the polysemy of the word show an evolution of its sense to the more recent meanings: by generalization, 'person that doesn't know how to do things', and by specialization 'bad person', 'foolish person', and 'sissy'. The word seems to exist only in Madeira and was not recognized in two locations. iv) corsa (noun, 'wooden board without wheels for dragging transport'), a semantic regionalism with the graphic variant corça, presenting several responses that correspond to one main meaning: pedaços de pau para carregar lenha e mercadorias ('sticks to load firewood and goods'), informants 4, 5 and 6 in Funchal; carro de madeira/carro de pau ('wooden cart'), speakers 3 and 6, tábua de arrasto ('dragged board'), respondent 5 in Machico; ajuda de corça ('transport aid'), informant 3 in Câmara de Lobos; transporte de madeira usado para transportar coisas ('wooden carrier used to transport goods'), respondents 1 to 6 in Porto Moniz. In Ponta do Sol there were no answers, which may show that the word is losing vitality. The word is known by all the informants in Porto Moniz. It is a rural word related to agricultural work. This can be explained by the fact that Porto Moniz is in the north coast of Madeira Island and is the most isolated area and the most rural community of our research field, because in the other geographical areas only the older speakers know this regionalism. The conservative form and meaning of the word show the existence of a linguistic continuum between rural Madeira and urban Funchal. This can be explained by the fact that, despite being a rural situation, in Funchal there is a similar means of transport for tourists. However, we can see that the word is losing its vitality, not having been recognized in a locality. v) cuscuzeiro (noun, 'ball to cook rice'), a semantic regionalism with different meanings: the old sense is the name of the recipient used for making cuscuz ('couscous'), recipiente de cozer o cuscuz/panela furada ('object to cook couscous, pierced pot'), referred to by informants 3, 5 and 6 in Ponta do Sol, and respondents 3, 4 and 6 in Porto Moniz, places were even today couscous is made; and fabricante ou vendedor de cuscuz ('couscous manufacturer or seller'), speaker 4, also in Porto Moniz. Then, we collected the meaning of recipiente de cozer arroz ('object to cook rice', inside a Portuguese stew), speakers 5 and 6 in Funchal, and informants 3 and 6 in Machico, through a cognitive process of metonymy of function. It means that, in areas where there is no tradition of making couscous, the word denominates a ball to cook rice. Finally, the new sense documented was pessoa coscuvilheira/vida alheia ('person that talks about others' lives/gossipy person'), informants 1, 3 and 4 in Funchal, speakers 1 and 3 in Machico, respondent 4 in Câmara de Lobos and in Ponta do Sol, and informants 1, 2 and 5 in Porto Moniz, as a synonym for bilhardeiro/a, a Madeiran lexical regionalism, probably by confusion with the standard word cusco ('gossipy'). This is a new and unexpected sense of the word. The term, while gaining new life (acquiring a new meaning), mainly in the mouth of speakers that are not familiar with old rural and cooking objects, is losing vitality (19 informants out of 40 did not recognize it).
vi) engamoer and esgamoer (verb, 'starve'), lexical regionalisms with the following meanings: in Funchal, encher-se de comer/comer muito ('eat a lot'), informants 3, 4 and 6, ter fome ('hungry'), respondent 5, the same as in Machico, informants 1, 3, 5 and 6, together with digerir ('digest'), respondent 4, a new and related meaning; the words were not recognized in Câmara de Lobos and in Ponta do Sol. In Porto Moniz, the terms have the sense of animal que está à fome ('hungry animal'), informants 3 and 5, pessoa com muito apetite ('person with a great appetite'), respondent 4, or com muita fome ('very hungry'), informant 6. The words do not have great variability in their meaning, presenting above all the traditional sense. They were unknown in two locations, revealing the loss of their vitality.
vii) papiar and papilar (verb, 'speak in a way that is not understood' or 'talk too much without been understood'), semantic regionalisms identified only in two inquiry points: 'to eat', informant 5 in Funchal, and respondent 6 in Porto Moniz; 'swallow', respondent 6, meaning related to the previous, and 'to speak in a way we cannot understand', informant 4, but also a 'person that talks a lot', informant 5, in Porto Moniz. These words seem to be ancient forms of the Portuguese language, only known with the meaning 'to speak/talk in a way that is not understood' by the older speakers of Porto Moniz. In Funchal, the two oldest informants recognize the regionalism as 'to eat'. The possible correlation between 'to eat' and 'to talk' is the use of the tongue, and the word papilar can be related to papila (taste bud of the tongue). In the other geographical areas, the speakers did not recognize the words.
viii) poita (noun, 'heavy backside'), a semantic regionalism with the definitions: âncora de um barco ('boat anchor'), informant 4 in Funchal, and respondent 6 in Câmara de Lobos, the standard meaning; rabo pesado ('heavy backside'), informant 6 in Funchal, the dialectal meaning of Madeira created by a metaphor through comparison with the 'boat anchor'; andar um pouco ('to walk a little'), respondent 5 in Funchal, sense that can be related to walk with difficulty because of a 'heavy backside'; ponto de arame que prende/segura as estacas da vinha ('wire point used to hold the grapevine'), informants 1 to 6 in Porto Moniz. There were no answers in Machico and in Ponta do Sol. In the north of the island, Porto Moniz, we found a different meaning related to agriculture instead of fishing, identified by all the informants of this area. These meanings are not recorded in the Madeiran vocabularies consulted (V. Appendix).
ix) stique (loan from English stick, noun for 'inebriation'), an innovative form as semantic regionalism, defined as: vergasta ('whip', the literal sense of stick), informant 3 in Funchal; desodorizante ('deodorant'), the standard meaning, a cognitive process of metonymy through the shape of the object, respondent 5 in Funchal; pessoa vaidosa ('vain person'), probably a metaphor by comparison with the way the person walks straight like a stick, informant 3 in Câmara de Lobos; maneira de falar diferente ('speak in a different way'), it can be a semantic extension from the previous sense, respondent 4 in Porto Moniz. There were no answers in Machico and in Ponta do Sol, revealing that the word may be unknown in these areas. None of the meanings corresponds to the regional one documented before and expected, 'inebriation'.
x) tentaréu (adjective and/or noun, 'someone that bothers'), a lexical regionalism with the phonetic variant atentaréu, meaning: diabo/inquieto/que atenta ('devil/restless (person)/that disturbs'), informants 3 to 6 in Funchal, and respondent 1 in Machico; pessoa chata/irritante ('boring/annoying person'), informants 3 to 6 in Machico, also respondent 3 in Câmara de Lobos, 5 in Ponta do Sol, and 2 to 6 in Porto Moniz, probably a generalization of the previous meanings. 14 informants did not answer the question, revealing that the word is not well known as an old term that is losing vitality.
xi) tertilho and entretilho (noun, 'toy' or 'something to entertain a child'), lexical regionalisms probably from the verb entreter ('entertain'), with the definitions: brinquedo/coisa que entretém ('toy'), informants 3 to 5 in Funchal, the traditional meaning; pião ('spinning top'), respondent 6 in Funchal, a specialization in one specific toy; distração/algo com que se ocupar ('distraction/something to keep busy'), informant 3 in Machico, 2 in Ponta do Sol, and 1 to 4 in Porto Moniz, by a process of generalization; pessoa entretida/alegre ('funny person'), respondents 5 and 6 in Porto Moniz, probably by contiguity or metonymy of the object to the person; there was no answer in Câmara de Lobos. The words are known mainly in Porto Moniz, but also in Funchal. xii) tratuário (noun, 'sidewalk'), a lexical regionalism that is a regional French loan from trottoir, with the phonetic variant trotoário, meaning: passeio/calçada ('promenade/sidewalk'), informants 3 to 6 in Funchal, and respondents 1, 3, 4 and 6 in Machico, as well as informant 2 in Ponta do Sol, and respondents 3 and 4 in Porto Moniz. The word tratuário with its phonetical variant trotoário showed great stability of meaning, but it was recognized only by 11 informants. The continuity in its geographic distribution was interrupted in Câmara de Lobos where there were no answers.

Discussion of quantitative data
After analysing the qualitative outcomes of the inquiry of this survey, with the description of the various responses of the informants in the different locations, we present the quantitative data for geographical and social variables.
In Tables 2, 3, 4, 5 and 6, we show the frequencies and percentages of the informants' responses to the dialect lexicon tested per geographical area: known vocabulary with the expected meanings, with other meanings, unknown meanings, and standard meaning words (in the case of the semantic regionalisms). The results of the responses are also presented per informant: 1 and 2 are highly educated young women and men, respectively; 3 and 4 are adults, female and male, with an intermediate level of education; 5 and 6 are older speakers, female and male, with lower education (elementary school and illiterate).
As we can see in the results of the tables above, in Funchal, adults and older informants (3, 4, 5 and 6) are the ones who know the tested Madeiran regionalisms best, and the oldest man is the only one who identifies the entire vocabulary (with 0% unknown words). In Machico, the oldest man recognized 31 words (77.5%), while the oldest woman knew only 23 (57.5%), the same number as the adult man, and the adult woman identified 27 (67.5%). In Câmara de Lobos, the oldest woman recognized 22 (55%) of the 40 Madeiran regionalisms, while the man identified 19 (47.5%), the adult man 18 (45%)  and the younger woman 17 (42.5%). In Ponta do Sol, the results of the oldest respondents are respectively 24 (60%) of the lexical units known to the woman and 21 (52.5%) to the man, followed by the adult woman with 17 (42.5%) of known words. In Porto Moniz, the adult man was the one who recognized most vocabulary with the traditional meaning, 32 (80%), as well as the older woman with 31 (77.5%). The youngest man shows the same result as the oldest man, 29 (72.5%). These results indicate that the generational differences between young individuals, adults, and the elderly, in the knowledge of regionalisms, are greater especially in Funchal and less evident in Porto Moniz.
As for the regionalisms identified with other meanings, in Funchal it was the oldest man who showed the highest number of this type of responses, 6 (15%), while the standard meaning was used mainly by young people, 3 (7.5%). In Machico, the largest number of   other meanings corresponds to the youngest female informant, 4 (10%), and the adult woman was the one who indicated the most uses of the standard meaning, also in 4 words (10%). In Câmara de Lobos, the two adult informants presented 4 (10%) of other meanings and it was the youngest woman who indicated more standard meanings, 3 (7.5%). In Ponta do Sol, the adults and young people are those who have most other meanings, 3 (7.5%), and the standard meaning is used mainly by the youngest, but also by the oldest man, also 3 words (7.5%). In Porto Moniz, the oldest man was the one who indicated more other meanings, 5 (12.5%), and the highest record of standard meanings was given by the adult woman and the oldest man, both with 4 (10%).
Regarding extralinguistic variables, when comparing the sample results of the geographical variation of Funchal, Machico, Câmara de Lobos, Ponta do Sol and Porto Moniz, as urban and rural areas and with sociocultural variation (sex/gender, age group, and schooling), we can observe the influence of these factors on the knowledge and use of the non-standard Madeiran vocabulary. However, social outcomes of the informants by geographical area showed that the level of education and rural vs. urban opposition was not as important as we expected.

Geographical variables
In Table 7 below, we present the total results of the 30 informants by location. If we compare the total results of the vocabulary known by the 30 informants, Porto Moniz is the area with the most answers with traditional Madeiran meanings (67.9%), followed by Funchal (with 64.1%), and Machico (59.5%). The locations with the lowest percentages are Câmara de Lobos (42.5%) and Ponta do Sol (44.5%) on the south coast. In the middle is Machico, located simultaneously on the north and south coasts of Madeira, being more conservative than the others.
Independently of more urban or rural communities, the geographical areas where the words are most known are Funchal and Porto Moniz, which are the most separate and the most urban and rural, respectively. This proximity between results of Porto Moniz (the most distant rural area) and Funchal perhaps can be explained by the fact that Madeira Island was isolated from the mainland and it was a rural society until thirty years ago, even in Funchal (Carita 2013). Concerning the city of Funchal, another important fact was the internal migration from the rural areas of Madeira Island to the urban area (Nunes 2014) throughout the 20th century. This population settled mainly in the peripheral areas of the city and can explain the diversity of conservative and innovative linguistic forms in the same territory.
In Nunes (2017), a study of several regionalisms in urban Funchal and rural areas of Madeira Island reveals that, although many of the Madeiran regionalisms are of rural origin, as they are related to agricultural activity, they have some vitality in Funchal and are recognized by older informants, probably because peripheral localities of the municipality are represented in the sample; until recently they were rural, as well as being areas that received migrants from rural areas of Madeira Island. The author concludes that the diatopic factor proved to be relevant in the case of the oldest terms that are archaisms, as opposed to the most usual or current ones, which we can observe also in the results of this paper, mainly in the south coast areas of Ponta do Sol and Câmara de Lobos where the oldest words were not recognized. So, these two locations are more innovative areas than Machico which showed better results in the recognition of the traditional meanings of the words. Perhaps this is because Machico was included in the north zone of the island's division (since its settlement in the 15th century), together with Porto Moniz. Belonging simultaneously to the north and south coast, it is more conservative than the other southern areas. Nunes (2019a), in a sample of 1st year students at the University of Madeira, concludes that in Funchal 65% of the answers obtained for the meaning of the words tested correspond to the expected concepts, while in the other geographic areas studied the value is 78%. The informants residing in Funchal exhibit 29% different meanings from the regional ones, mainly from standard Portuguese, revealing ignorance of some semantic regionalisms in Madeira, especially when they are archaisms, while in respondents from other municipalities, the author obtained 12% of meanings which are different from the traditional ones expected. It should also be mentioned that in Funchal 94% of the tested lexicon was recognized by the informants, 6% of which was unknown. In rural areas, 90% of the words are known and 10% unknown. These percentages are explained by the fact that some of the tested archaisms exist only in some municipalities in the west and south of the island, as well as loans from Venezuela, being recognized by some informants in Funchal. Nunes (2019b), about a sample of 3 rd -year students at the University of Madeira, from Funchal and from various rural areas of the island, also concludes that there are no significant differences in the knowledge of the vocabulary tested by locations or by gender, with a linguistic and sociocultural continuity between the different geographic areas studied. This proximity between urban and rural varieties can be explained by the fact that the informants were students who attended the cultural studies course and had had the opportunity to deepen their knowledge of the sociocultural and linguistic situation of the region, in contrast to 1 st -year students. However, we found no significant differences between Funchal and the rural locations in the results obtained in the two studies.
Another interesting result of this study of the lexical and semantic variation of 40 Madeiran regionalisms is the localisms that have to do with differences between the geographical areas: their main socioeconomic activities and implications for the meanings of the words, like poita, which in the northern area of Porto Moniz means 'wire to tie vines', a different sense from the other locations. In Ponta do Sol, there is a particularity in the meanings of the words atrapichar ('dress up'), and trapichado/trapichento ('well dressed'), that is the opposite of the senses registered in the other areas: 'bungler', and 'scruffy' or 'ill-arranged'. Old people of Ponta do Sol told us that when they go to Funchal, they try to be well dressed and even perfumed with ancient cologne. This can be explained by the fact that rural people in the town were known as burquilha/broquilha, 'someone that doesn't know how to dress, behave or speak' like civilized urbanites.

Social variables
We can ascertain that age group is a very important external variable and is associated with a higher level of education of young people and adults, more so in the city than in rural areas. In general, older speakers know many more regionalisms than younger people, and middle-aged adults know less or as many words as older informants.
The knowledge and the use of Madeiran regionalisms as non-standard variants tend to disappear with the standard European Portuguese imposed by the scholarly norm. This leads to the reduction of the lexical and semantic features that constitute the contemporary Portuguese spoken on the island of Madeira and of its sociocultural variation. However, it is interesting to observe that new meanings emerge from old words, giving them new life with the efficiency and expressiveness characteristic of the Madeiran variety. So, it is not linear that the level of education (increase of urbanization and access to school) leads to unfamiliarity in the knowledge and use of regionalisms. Although young people with a high level of education tend to follow the standard Portuguese language (Teixeira 2015), we verify that in Machico young women know as many regionalisms as old women and that of the three female respondents, the adult women are those that know the highest number of the tested words. In Porto Moniz, young men know as many words as old men, and the adult man with an intermediate level of education is the one that knows more regionalisms.
The age group variable is decisive in almost all researches on social variation of language to see if linguistic changes are taking place. In this study, the use of Madeiran regionalisms which are generally more frequent in the speech of older people, as opposed to younger people, is also visible in our sample, showing the loss of vitality of some lexical units that seem to be archaisms. Nunes (2014) shows that in Funchal, in general, the difference between young people and other speakers is more accentuated than in other areas, regarding the knowledge and use of regionalisms, while adults and older people show similar outcomes. This is also true for the results of this study, revealing the influence of sociocultural and geographic factors of the urban or rural origin of the informants on the recognition and use of Madeiran regionalisms. Yet, in this study, these differences are not so evident, with the results obtained in Funchal and Porto Moniz being close.
We can say the same about sex/gender as a social category. There is apparently no relevant statistical difference between the knowledge of regionalisms rates presented above for women and men. Thereby, we can observe that women are a little more conservative than men, if we look at the results of the older informants of Câmara de Lobos, Ponta do Sol and Porto Moniz. However, in Funchal they recognized the same number of words, and in Machico men recognized more regionalisms than women. In Nunes (2017), the factors of social variation also were not as relevant as might be expected, with no major differences between speakers of either gender, although women tend to be more conservative in rural areas, especially when it comes to recognizing archaisms. Nunes (2019a) claims that the sociocultural gender factor is relevant, since, in Funchal, women generally lead in knowledge of the Madeiran regional lexicon, as well as of the new registered meanings. The author concludes that, in Funchal and in other geographical areas, women are both more conservative and innovative than men. About different age groups and different levels of education, some differences are noted, again in relation to archaisms. In rural areas, there was no great distinction between the knowledge of regionalisms by young people, adults and the elderly, perhaps because there is greater intergenerational contact.
Despite the numeric limitation of the sample per locality and informant for this survey, to compare different geographical areas, it seems to be representative of the linguistic actuality of the Madeiran variety. Now it is necessary to enlarge the samples to more speakers per area and to the other rural and urban locations of the island to confirm and validate these results. So, we cannot generalize, for example, that women use more linguistic forms of the prestigious standard language than men and that in rural and in peripheral areas of Funchal men lead the change to prestigious forms, or that people with a higher level of education use more the standard Portuguese language than people with poor schooling. Yet, we can verify that older women in rural areas tend to know more regionalisms and their meanings than men. This linguistic situation can be explained by the traditional role of women in Madeiran society, although, in this study, the oldest man in Machico recognized more words with the regional meaning than the woman, and in Funchal they identified the same number of words. In Câmara de Lobos, Ponta do Sol and Porto Moniz, older women revealed greater knowledge of the vocabulary tested than men of the same age group. About the Funchal results, we are witnessing a recent phenomenon of valorization of regionalisms in the urban variety due to a great divulgation of regional traditions correlated with the development of cultural tourism and popular festivities organized in the island for residents and tourists, which contributes to increased awareness of Madeiran linguistic heritage and identity.
The great socioeconomic and touristic development in Madeira is very recent, which is why the 40-year-old generation continues to have rural memories and references that somehow tend to disappear in the new generations (Santos 2013). These rural language experiences are present in almost all the Madeiran regionalisms which reflect the regional identity, namely the sociocultural elements of the popular or traditional knowledge of this insular population. Thereby, this kind of linguistic research contributes towards discussing the multi-dimensional, geographical, historical, and sociocultural characteristics, and consequently the complexity of the Portuguese spoken in Madeira, in urban Funchal and in rural areas.

Conclusion
As a language and a dialect are heterogeneous and variable systems, we studied some lexical and semantic Madeiran regionalisms, analysing the variation between the standard variety of European Portuguese and the Portuguese spoken on Madeira Island. We can observe that the island is a space of conservation and at the same time innovation and (re)creativity of words and meanings. This occurs mainly in Funchal, the capital of the archipelago of Madeira, which has more linguistic contacts abroad, revealing the variability and expressiveness of lexical and semantic non-standard variants in contemporary Portuguese spoken on the island of Madeira.
The city of Funchal is more open to external contacts, especially with the influx of tourism by air and by sea, which can explain the fact that it is an innovative linguistic area and at the same time conservative, because most of its territory was rural until the last decades of the 20th century, receiving migrants from other rural areas on the island of Madeira. Yet, the semantic variation of the words does not occur only in Funchal but also in rural areas. In some words, there is more variation of meanings in Funchal than in the other locations, for example in the regionalism bambote (phonetic variant of bombote, a loan from English): 'fishing boat', 'salesperson on board', 'scoundrel', and 'elegant man'. Perhaps because it was a local reference of Funchal, where men and boys sold Madeiran products to tourists on ships that docked in the bay, from the 19 th century onwards, a situation that no longer exists.
The focus of our study was the analysis of the different meanings of Madeiran non-standard vocabulary, some not yet registered in regional vocabularies, as trapichado and trapichento ('messy/crazy', recent forms derived from the Madeiran word trapiche, 'institution for the mentally ill, crazy person, alcoholic and confusion'). The meanings were collected in the semasiological questionnaire that was used to test the knowledge of lexical and semantic regionalisms among the informants. In despite of the social stratification of the informants by gender, age group and level of education, also their geographical origin from the city of Funchal and from rural areas of Madeira Island taking into account, in general, the data show a continuity of the linguistic forms and senses of the words in the different areas and sociocultural groups. The answers given by the respondents present great flexibility and consequently variability of meanings, except for words like dentinho 'appetizer', semilha 'potato' and tratuário 'sidewalk', which showed great stability, probably due to the high frequency of knowledge and use.
The diversity of meanings and the complexity of the linguistic systems show the importance of cognitive semantics, together with lexical dialectology and sociolinguistics, to better understand the dialectal convergence and divergence between different word meanings in the sociocultural groups and geographical areas studied. In this case, the results show the proximity between Funchal and Porto Moniz, although this latter location belongs to the more conservative north coast. The city of Funchal proved to be both a conservative and innovative area. The collection of new meanings and senses of the words and their forms also reveals that there are some regionalisms not documented in the dictionaries of Portuguese language, in the glossaries of degree monographs, and in master's degree theses about the Madeiran variety, as well as in published vocabularies, showing the necessity of this type of studies.
The majority of the 39 words and one expression tested in our questionnaire are polysemic, with different senses for which it is possible to establish one correlation through one or more semantic common features, generally through cognitive processes of transfer or analogy by metaphor or metonymy, of semantic extension or generalization, and of specialization of the word meaning. Some of them are new and unexpected meanings that seem to appear by language (re)creativity or are lexical innovative forms like trapichado and trapichento, along with the regional loans by language contact, mainly with the English and Spanish language, but also from the French language like tratuário/trotoário.
For the semantic Madeiran regionalisms like embeiçado 'penniless', furado 'tunnel', joeira 'paper kite', and terçol 'youngest son', some informants, especially the youngest from Funchal, indicate the standard instead of the regional meaning, showing that they do not know, use or remember it (nevertheless, we know that some regionalisms are situational used by younger individuals unconsciously, in informal and familiar contexts). The connexion between the mind of the speaker and the sociocultural context, shown by the cognitive semantics, happens in the pragmatic dimension of the linguistic use, on lexical and semantic choices between the existing language alternatives or co-occurrences of different meanings, in this case words with their dialectal or standard senses. Thus, the answers to the questionnaire mainly highlighted the expressiveness and efficiency of the lexical and semantic Madeiran variants and their sociocultural situatedness: the differentiation of the regional and the standard varieties more than the internal social and geographic (rural vs. urban) varieties, projecting an image of linguistic unity and identity.
With our study of 24 semantic regionalisms, 15 lexical regionalisms and one expression, mainly in the semantic field of humans (physical, psychological and behavioural aspects), we observed that the majority of these Madeiran regionalisms are conservative forms with traditional meanings, used in rural areas as well as urban, but also gaining new senses and forms. The studied vocabulary was revealed to have an expressive function in the socio-historical and cultural identity of Madeiran variety, reinforcing the heterogeneity of European Portuguese and the complexity of contemporary spoken Portuguese on Madeira Island. The qualitative analysis together with the quantitative approach allow changes to be recognized that are occurring in the knowledge and use of Madeiran lexicon, the lectal (geographical and sociocultural) variation.
In conclusion, the data collected confirm the existence of great lexical and semantic variability (polysemy and synonymy) in the variety of Portuguese spoken in Madeira. We systematized the lexicon studied in regionalisms with little variability, with great variability and with little or less vitality, with most of the vocabulary having great variability. This can reveal instability of words and their meanings i.e., that a linguistic change is taking place, which can mean new life for some regionalisms or their disappearance. In the case of semantic regionalisms, we found the co-occurrence of the traditional meaning on Madeira Island with the standard meaning of the variety of European Portuguese. Young speakers generally only know the standard meaning of these words, for example dar um amorzinho, 'give way a little/give a little help', giving the literal sense 'give a little love', and embeiçado, 'with no money', using only the standard meaning 'in love'.
For some words like poita we found a significative difference in the special meaning of the word in the north of the island, in the more isolated locality of Porto Moniz, with a sense related to agriculture, and more conservative in the meanings of some words like cachada (with the unanimous meaning of 'cheek', instead of 'buttock', the main answer in the other locations). Through the comparative analysis of the lexicon with the Tesouro, we discovered that some words exist and have the same meaning on Madeira Island and in the Azores: asservado (with a similar meaning and form assorvado, a 'peaceful person' on Graciosa Island), cieiro ('dirt on the skin', also with the same meaning on Graciosa Island), esgamoer ('starve', with the same meaning on S. Miguel Island), poita (with the non-standard meaning 'heavy backside', common to Graciosa Island), and common to Brazil renheta ('grumpy') V. Appendix. This is an exploratory and limited study in the number of informants and of areas of Madeira Island included, and more vocabulary items need to be tested, including in the small island of Porto Santo, to validate these results. With this study, the aim is to contribute to a better knowledge of the dialect lexicon of Madeira, for which much work remains to be done.