About the concept of "geodifferential feature" between linguistic varieties in contact

CHAPTER FIVE

ABOUT THE CONCEPT OF “GEODIFFERENTIAL FEATURE” BETWEEN LINGUISTIC VARIETIES

IN CONTACT1

RAMÓN DE ANDRÉS DÍAZ, FERNANDO ÁLVAREZ-BALBUENA GARCÍA,

MARÍA CUETO FERNÁNDEZ AND XOSÉ MIGUEL SUÁREZ FERNÁNDEZ2

Abstract The Estudio de la Transición Lingüística en la Zona Eo-Navia, Asturias (ETLEN) is a project undertaken by the University of Oviedo (Asturias, Spain), whose objective is the study of the dialectal variation in an area of geolectal contact between two Ibero-Romance language domains: the Galician-Portuguese and the Astur-Leonese, in the westernmost part of Asturias; this is, the Eo-Navia area, which receives its name from the two delimiting rivers on the coast. The ETLEN analyses three aspects: the dialectographic one – a classical dialectal description ‒, the dialectrometric one ‒ the measurement of the difference between geographical points, accordingly to the Salzburg School of Dialectometry ‒, and the horiometric one, a neologism that we employ to define the

1 The work in this chapter has been funded by the Ministry of Science and Innovation of the Government of Spain (ref. MICINN-08-FFI2008-01774/FILO). It has also been possible thanks to a pre-doctoral grant funded by the Programme “Severo Ochoa” for the training in research and teaching (2009-2013), of the Government of the Principality of Asturias. We want to acknowledge the helpful support received from Vital de Andrés Díaz and Andrés Menéndez Blanco. 2 Members of the Seminariu de Filoloxía Asturiana of the University of Oviedo (Asturias, Spain).

Chapter Five

86

measuring of the differences between point to point, according to the assignment of each differential feature to one of the two geolects. The classification of phenomena and differential features is a topic of interest within variational linguistics. From a horiometric perspective, an important factor to consider in the classification of different features is their geographic distribution. In this essay, we will entertain a series of theoretical reflections about concepts of geographical dialectology under the light of the ETLEN research.

1. Language and glottodiversity

Human language shows a multi-level diversity and complexity, from the idiolects to the phyla of the genealogical classification or the types of the typological classification, going through what are known as dialects, languages, families, etc. As a science of language and languages, linguistics has a natural inclination for glottodiversity through different specific disciplines. This diversity implies variation, and variation implies the contrast of features or systems over a commonly shared background. The comparative result moves to a higher level of abstraction from the one in which the features or compared systems are located (Pulgram 1964: 67).

1.1. Differential linguistic contrasts. A general approach

The linguistic contrasts that serve as a foundation for glottodiversity can respond to certain differential types. On one hand, there are inter-systematic contrasts that take into account individual phenomena, provided that they affect the system configuration. On the other hand, there are extra-systematic contrasts; they are the ones that grasp individual phenomena without the need to be analysed as elements integrated within systems. At this general level of analysis, the contrasts can be established with total indifference to kinship or to typological groups. According to this, we would have this general outline:3

3 General models of classification of linguistic contrasts have been considered by different authors. It is worth highlighting the names of Weinreich (1953) or Sala (1998) because of their studies about the interference between languages. A highly developed model with a generativist base and a synchronic perspective is the one offered by Viaplana (1990-1991). See also Veny (1985) and Chambers and Trudgill (1998).

About the Concept of “Geodifferential Feature”

87

differential linguistic contrasts

(I) inter-systematic contrasts4 (II) extra-systematic contrasts (1) paradigmatic contrasts

(2) syntagmatic

contrasts (a) in terms of

the inventory of paradigms5

(b) in terms of the inventory of

paradigmatic units and the

way they relate to each other

in terms of their

arrangement, combination

and unit distribution

Table 5.1: Classification of the differential linguistic contrasts. All these contrasts occur in the three main levels of language: phonetic-phonological, morphosyntactic and lexical-semantic.

1.2. Synchrony and diachrony within linguistic contrasts

Linguistic contrasts can be observed either in synchrony or diachrony.6 The contrasts within synchrony dispense with their evolutionary past

and are valid for any phenomena or system without the intervention (or with total indifference) of genealogical and kinship factors. The synchronic contrast is phenotypic. It allows us to discover facts of linguistic typology.7

The contrasts within diachrony are based on evolution from a common origin; therefore, they are necessarily established between lects placed in the same field of kinship. Accordingly, we are dealing with a genotypic contrast, with evolutionary divergences from a

4 These types of contrasts are in connection with the debate about structural dialectology. See Weinreich (1954), Chambers and Trudgill (1998: 33-37). 5 The distinction between “differences of inventory” and “differences of distribution” is already present in Weinreich (1954: 393-394). 6 “Synchronic dialectology” and “diachronic dialectology” are concepts already debated in Weinreich (1954: 390-392, 395). 7 An example of an atlas within a purely synchronic approach is the Survey of English Dialects (SED) (Orton, Halliday & Barry 1962-1971).

Chapter Five

88

common origin, and, therefore, the phenomena and the features are stated in terms of conservation, maintenance, or either of alteration, loss, etc.8

1.3. The genealogical kinship factor

According to what has been said above, two lects such as the Basque and Castilian languages can be easily contrasted within synchrony. As far as diachrony is concerned, in spite of the genealogical remoteness of both lects, they can also be contrasted through consideration of historically shared elements. Thus, the flood of Latinisms absorbed by Basque from spoken Latin allows us to observe that, from Latin words such as PACE, PĬCE, there is a contrast between the Basque results bake ‘peace’, pike ‘pitch’, with the conservation of [k], and the Castilian paz, pez, with a [θ] as the product of an old palatalization. However, between a form of the Basque potential mode such as daiteke and the corresponding Castilian periphrasis puede ser ‘it may be’, a contrast of the diachronic type cannot be done, since they have absolutely disparate origins (it is, obviously, possible to draw a typological synchronic contrast between the Basque grammatical type ‘synthetic potential’ and the Castilian ‘periphrastic potential’). To this respect, Castilian (Cst) and Catalan (Ctl) may be easily contrasted, either in synchrony or diachrony. In this case, the common origin refers to Vulgar Latin (VL), or maybe to a set of common features in different Middle Age stages. Thus, Cst luna [l]- || Ctl lluna [ʎ]- ‘moon’ can be synchronically understood as a contrast between two lateral consonants at the beginning of the word, but their cognate status also carries a diachronic move: from L-, in the protolanguage (VL), at some moment in its evolution, Catalan introduced an innovation: the palatalization of L- > [ʎ]-. Accordingly, it is said that Castilian conserves or maintains L- whereas Catalan has altered it. In some cases, between related lects, this diachronic perspective might not be relevant. Thus, the contrast of the idiomatic expressions Cst en mi casa || Ctl a casa meva ‘in my house’ has a full typological interest regardless of the divergent evolution they reveal.

8 In the field of structural dialectology, the issue of synchrony and diachrony gave way to an intense theoretical debate about the concept of “diasystem”, since its first formulations, of a synchronic and systematic kind, did not satisfactorily account for similarities and differences with a diachronic origin, and hence reactions such as Moulton’s (1968). See also Veny (1985: 70-71, 179) and Chambers and Trudgill (1998: 35-39).


89

The kinship becomes evident through the cognates, that is to say, signs from different lects that share synchronic similarities of signifiers associated to similarities of signified. Thus, contrasting Romance languages ‒ such as Galician-Portuguese (GP) olho, Astur-Leonese (AL) güeyu, Cst ojo, Aragonese güello, Ctl ull, Occitan uèlh, French oeil, Franco-provençal uely, Romansh egl, Sardinian ogru, Tuscan occhio, Romanian ochi ‒ reveals evident and shared phonic and semantic similarity, explained by a common origin from VL ŎC(Ŭ)LU ‘eye’; concomitantly, these forms also reveal some lines of differentiation.

We just saw above that the contrast Cst luna [l]- || Ctl lluna [ʎ]- can be analysed synchronically, in spite of being established between two cognates (since, we say it again, two kindred lects allow diachronic or synchronic contrasts). The phonic similarity among cognates is the fact that simultaneously allows the quantitatively (and synchronically) detection and measurement of the differences among phonic segments in different lects.

On the other hand, when there is a kinship among lects from a set, it is possible to build a diasystem, which is a theoretical construct with a higher level of abstraction. It encompasses the mentioned lects in a unique supra-system involving a higher synthesis of linguistic diversity.9

1.4. The topicity factor

Linguistic variation can be described and displayed at an abstract level, regardless of the dimension in which such variation can become evident. In this sense, we can give the name of topicity to the fact that linguistic variation is placed on concrete levels or dimensions. By virtue of topicity, the differential features present a topical distribution. This means that they have a material distribution within a geographical territory, a social space or a communicative domain.10 When variation is geographic (diatopic), topicity can be translated as spatiality. In the diastratic and diaphasic axes, topicity is a fact not so easily visualised as in diatopy. Nonetheless, even a concept such as social dialect continuum (Chambers & Trudgill 1998: 7-9; 9 The concept of diasystem is due to Weinreich (1954). As more or less equivalent terms, we may also find, among others, high level sysytem, system of systems, supra-system or geosystem. See Pulgram (1964: 76), Francescato (1964), Berruto (1979: 94-95) and Veny (1985: 178). 10 About the concept of communicative domain, see Fishman (1972: 73-81). Other denominations are setting (Ervin-Tripp 1964: 86-87), frame or field (“Conjunto de lugares sociales en los que se producen ciertos tipos de situaciones y de eventos sociales”, Berruto 1979: 129).

Chapter Five

90

Romaine 2000: 19) evokes topicity and is named the continuum spatial reference.

2. The geographic dialectology

2.1. Geographic dialectology versus social dialectology The linguistic variation that we identify par excellence as dialectal variation is diatopic and geographic. However, the diatopic axis is just one of the axes that shape linguistic diversity. Besides this, there are the diastratic axis, representing linguistic variation in connection with social groups or strata, and the diaphasic axis, representing variation related to communicative situations. From a glottologic point of view, there is not any epistemological obstacle to consider that the diastratic and diaphasic varieties are as dialectological and diasystematic as the diatopic ones. This is something that, one way or the other, has been sufficiently highlighted by different authors.11 López Morales writes it in very clear terms:

La descripción de un grupo de idiolectos sintópica, sinstrática y sincrónicamente es tarea de la dialectología, pues el cometido de esta rama de la lingüística no es otro que el de describir dialectos, aunque éstos sean verticales. El que la dialectología (tradicional) haya estado esencialmente preocupada por describir dialectos en un plano diatópico es factor circunstancial muy ligado a limitaciones de época y de escuela. No es lícito establecer una dicotomía entre dialectología y sociolingüística basados en el carácter horizontal o vertical de los dialectos estudiados: una característica como ésta no puede convertirse en principio definitorio.

López Morales (1989: 24-25) As a matter of fact, the label of social dialectology has been applied to the dialectological survey of diastratic and diaphasic varieties within the Labovian variationism. This “broad” concept of dialectology must not be mistaken for the sociology of language: the focus of the latter is the study of the interrelationships between society and the linguistic system (or diasystem), whereas dialectology focuses on the language itself.

Nonetheless, even though we can assume that dialectology could be diatopic, diastratic and diaphasic at the same level, diatopic dialectology is different from the others for obvious reasons relating to the material

11 Rona (1976) extensively reviews the concept of social dialect. Coseriu (1981: 13) talks about isoglosas “diatópicas”, isoglosas “diastráticas” and isoglosas “diafáticas”. Bibiloni (2000) offers a complete overview of the three axes of dialectal variation.


91

dimension in which it works. That is to say, the spatial distribution of the lectal variation displays a visualisation in the form of isoglosses, bundles, areas, maps, etc., which is analogically harder to conceive in the other axes. However, the concept of differential feature, to be defined further on, can be used – in terms of an abstract concept – either in the diatopic dimension or in the diastratic and the diaphasic ones.

2.2. Horiometry. Its place within geographic dialectology

ETLEN is a dialectological research project that, particularly, deals with linguistic phenomena appearing between geotypes. This kind of dialectology studies transitional or contact areas located between geotypes (geolects, domains), especially the so-called geolectal boundaries between related geotypes, cartographically visualised in isogloss bundles. The ETLEN tackles the northernmost stretch of geolectal contact between Galician-Portuguese and Astur-Leonese geotypes, in the westernmost part of Asturias and the northeasternmost part of Galicia (see Figure 5.1).

Figure 5.1: The squared area signals the survey area of the ETLEN project. One of its three facets is the horiometric study (from Greek hórion ‘limit, boundary’ and métron ‘measurement’), a neologism that we define as “the mathematic-statistical measurement of the geodifferential features in a bundle of isoglosses by means of their ascription to the contiguous geolects”. As we were working in the treatment of features, we realized that the spatial disposition of the geodifferential areas and their corresponding isoglosses is a determinant factor in their classification and quantification. Thus, we also realized that a huge mass of geodifferential features (and their areas) follows some Western/Eastern distribution

Chapter Five

92

patterns, associated to a North/South orientation of the corresponding isoglosses. Moreover, our approach has enabled us to specify some questions about geolectal variation that we will present next.12

2.3. The differentiality in dialectometry and horiometry

Dialectometry has established itself as a dialectological methodology with a mathematic-statistical basis for the treatment and measurement of the differential features in a given territory. The process of determining the differentiality in dialectometry can be summarized as follows:13 The point of departure is a corpus of dialectal data usually extracted

from a linguistic atlas, which has been made by means of a questionnaire based on a collection of individual items. Each corpus exemplifies a particular phenomenon, whether by itself or as part of a group with other items from the questionnaire.

For each item (further on we will examine with more detail this concept), we can observe the variety of collected answers and we can also proceed to an operation of synthesis, reducing the obtained variation to a few types called taxates. The taxates have three characteristics: they are units of linguistic variation able to be statistically processed; they are obtained applying logical criteria arranged by the dialectologist – at least, in the Salzburg School of Dialectometry (S-DM); and they are determined “in abstract terms”, with no need of taking into consideration their geodistribution on the map.

Each taxate of every item is present in certain geographic points on the map, and absent from any other (there cannot be two taxates at the same point). From this, we obtain, for each item, a so-called working map, which is essentially the same type of map as the ones provided by any traditional linguistic atlas: what is displayed is the synthesized spatial distribution of the variants.

12 For anything related to the ETLEN project and its methodological foundations, see: Andrés Díaz, Álvarez-Balbuena García and Suárez Fernández (2007); Andrés Díaz and Álvarez-Balbuena García (2011); Andrés Díaz (2011); Álvarez-Balbuena García, Andrés Díaz; Suárez Fernández and Cueto Fernández (2011); Andrés Díaz (to appear); Cueto Fernández (to appear); and Andrés Díaz, Álvarez-Balbuena García, Cueto Fernández and Suárez Fernández (2012). 13 For anything related to Salzburg Dialectometry, see Goebl 2003a, 2006, 2008a, 2008b, 2010a, 2010b, 2011.


93

A computer program is able to confront all the geographic points on the map, with their diverse taxates, thus finding the mutual differences and similarities in a similarity matrix, where they become mathematical-statistical values. The accumulation of these values provides, as one of its possible results, a map that synthesizes, through a chromatic range, the degree of similarity or difference among the points on the map.

The ETLEN project has a dialectometric orientation in agreement with the S-DM. To this extent, the procedures mainly conform to the ones described above. However, with regard to the mentioned geolectal variation, ETLEN also proposes another type of dialectometric research requiring a different methodology, which we named horiometry, or the surveying and measuring of the linguistic boundary. Relying on general dialectometric procedures, as summarised above, we present below the main characteristics of horiometry, as understood within our current research on the isogloss bundle that runs through the Eo-Navia area. The point of departure is a corpus of dialectal data obtained from an

ad hoc questionnaire, designed and based not on a collection of individual items, but on a collection of contrastive phenomena characterising an western geotype (approximately equivalent to Galician-Portuguese) and an eastern geotype (approximately equivalent to Astur-Leonese). Even though we rely on acknowledged Ibero-Romance types (Catalan, Aragonese, Castilian, Astur-Leonese and Galician-Portuguese) as a point of departure, it is very important to stress the fact that our research is not based on these domains, since all the geotypifications within the ETLEN project result from the effective western/eastern geographical distribution of the geodifferential features under survey.

Each contrastive phenomenon is illustrated by a series of items (words, phrases). The variants of each collected item are susceptible of classification as western, eastern and even axial (from the axis or central area). The western, eastern or axial character of each variant is determined by geodistribution patterns.

For each contrastive phenomenon in each geographic point, we obtain a proportioned statistics of geotypical ascription (western/axial/eastern), in agreement with the characteristics of the items collected at each point.

Additionally, for each locality, we also obtain an accumulation of proportions of each geotype, granting their visualisation through circular diagrams or chromatic ranges on horiometric maps.

Chapter Five

94

Therefore, the S-DM and the horiometric dialectometry deal with the geolectal variation according to different principles: The S-DM classifies the variants according to logical criteria, leaving

aside the geotypes (which, at any rate, will show up a posteriori as a synthetic outcome). Differently, horiometry classifies the variants according to geotypes, that is to say, taking into account certain geodistributions of features allowing their ascription to these geotypes.

In the S-DM, the unit of statistical accumulation is the taxate; in horiometry, the unit of statistical accumulation is the geotypical attribution (in the ETLEN, western / axial / eastern).

As an example contrasting both approaches, we may consider the phenomenon “results of the sequence stressed [ˈkwa] + nasal” in the Eo-Navia area. For the Latin etymons QUANTŌ ~ QUANTUM and QUANDŌ, the variants canto, cando || conto, condo || cuanto, cuando ‘how much; when’ are registered in the area.

The dialectometric analysis would start establishing the taxates or differential and computing units, identifying each taxate with a number (whose value merely functions as identifier). For example:

answers taxatescanto, cando 1conto, condo 2

cuanto, cuando 3

Table 5.2: Dialectometric taxates established with the results from Lat. QUANTŌ ~ QUANTUM and QUANDŌ in the ETLEN survey area. The classification by taxates does not consider facts of geographical or geolectal type (belonging to geolects or concrete domains). Consequently, each geographic point in the net is represented by a taxate, as opposed to other points that display the same or different taxates. Finally, with this example, just like with many other types of phenomena and their corresponding taxates, which are introduced in an array of data, we obtain a map of interpunctual differences. Without previous geolectal consideration, such maps display statistically obtained geotypical structures, which crucially match the known configuration of the different linguistic domains. Let us say, then, that in dialectometry the geotypes come out a posteriori.


95

In the horiometric analysis, from the start, we have in mind geodistributional criteria for the differential features. In the present example, we would check that canto/cando has a western distribution, conto/condo is axial, and cuanto/cuando is common to all geotypes. The resulting geotypological diagram is as follows:

western axial eastern

cuanto, cuando

canto, cando conto, condo

Table 5.3: Geodistribution of the results from Lat. QUANTŌ ~ QUANTUM and QUANDŌ in the ETLEN survey area, for further horiometric quantification. Accordingly, the presence of these forms in each survey point receives a score according to the following four parameters, in such a way that the sum will always be 100:

+ Western – Eastern [= feature of western distribution] – Western + Eastern [= feature of eastern distribution] – Western – Eastern [= feature of axial distribution] + Western + Eastern [= feature of common distribution]

For a great amount of data, the resulting horiometric maps show the proportion of each of the parameters in the forty points of the survey. This allows us to obtain a quantified and highly detailed image of the transition between Galician-Portuguese and Astur-Leonese in the surveyed area.14

14 Some previous papers can be mentioned whose theoretical foundation does not substantially differ from the one we propose in our horiometry. Thus, Reed and Spicer (1952), about an area in the Northwest of Ohio; Speitel (1969), who studies the distribution of lexical isoglosess in both sides of the border between Scotland and England; Recasens (1982, 1985), on the border between the Eastern and Western block of dialectal Catalan, in the province of Tarragona; or the “scale of variability” used by Elizaincín (1996) to measure the typological percentages of features in the confines between Portuguese and Spanish in the fronteiriço of Uruguay.

Chapter Five

96

3. The geolectal variation

3.1. Variation, continuity, contiguity The linguistic family (Romance, Germanic, Slavic, etc.) is a field of geolinguistic continuity, although, of course, it is not the only one: there is also geolinguistic continuity at superior levels and, obviously, at inferior ones. As it is well known, a continuous grading of minimal differences in several directions is produced within the Romania. The minimal differences found among geographically close points – among contiguous locations in a survey net – are the core and the point of departure for the geolectal variety. Besides geolinguistic continuity at superior or inferior levels, this is where the geolectal variation finds its primary expression.

As we move away from a given geographical point, the differences keep piling up quantitatively and progressively. We can see the quantitative accumulation of geolectal differences (a) in a linear direction, as found in the Romance field from Fisterra to Utrantu; or (b) in a ring-shaped/concentric distribution, as if waves of increasing differentiality set off from the initial point of reference, which is very close to what dialectometry displays in relative similarity maps (Goebl 2003a, 2006, 2008a, 2008b, 2010a, 2010b, 2011).

Thus, it turns out that among contiguous points in the continuum there are some minimal differences reducible to minimal geodifferential features. These become the minimal units of the geolectal variety. It is advisable then to deepen into these features and into their nature and characteristics.

3.2. Contrastive phenomenon and geodifferential feature

First, we will pay attention to some concepts:15

A linguistic element is any linguistic characteristic, regardless its geolectal contrastive pertinence.

A linguistic variety is any geotype, regardless its hierarchical taxonomic level, whether is called language, dialect, speech or whatever labels we use.

A contrastive phenomenon (P) is the variable, that is to say, whatever linguistic element susceptible of geolectal variation.

15 These concepts are formulated for the first time in Andrés Díaz et al. (2012).


97

A (geo)differential feature or (geo)variant (V) is each one of the differential manifestations that the phenomenon at issue presents. They are linked to complementary geodistributions (areas) able to be represented cartographically. For example, in the area surveyed by ETLEN, (P) “results of the Latin intervocalic -L-” presents two geovariants or geodifferential features: (V1) “loss of the -L-”, with a western geodistribution; and (V2) “conservation of the -L-”, with an eastern distribution. Another example: in the same area, (P) “placement of clitic pronouns in the verbal infinitive periphrases haber (a, de) + infinitive ‘to have to’, ir + infinitive [future tense], querer + infinitive ‘to want to’ and poder + infinitive ‘can’”, presents two variables, which are two sorts of placements: (V1) “coexistence of types: auxiliary + clitic-infinitive, and auxiliary + infinitive-clitic”, as in quérolo facer ~ quero facelo ‘I want to do it’, with western geodistribution; and (V2) “exclusive type auxiliary + infinitive-clitic”, as in quero facelo ‘idem’, with an eastern geodistribution.

A variation field is the geovariant set (the geodifferential features) ascribed to one phenomenon.

Figure 5.2: Contrastive phenomenon, feature/geovariant and variation field.

Chapter Five

98

3.3. Some characteristics of geodifferentiality

The contrastive phenomena and the geodifferential features display the following characteristics: The contrastive phenomena may refer to (sub)systems or to units of

the system. For example, in the Ibero-Romance field (P) “existence or non-existence of a verbal subsystem of compound tenses with HABĒRE”, as opposed to (P) “existence or non-existence of the voiceless palatal phoneme /ʃ/”.

They are discrete entities (units) and, therefore, subject to individual recognition and accountability.

There are geodifferential features displaying large paradigmatic extension, thus affecting a great number of items. For example, in the Eo-Navia (P) “conservation of the stressed Latin vowels /ɛ/ and /ɔ/, as opposed to its diphthongation” presents two features with their corresponding areas, a (V1) “conservation” with western geodistribution and (V2) “diphthongation” with eastern geodistribution, which is registered in a great amount of items. Other features, however, have a very limited paradigmatic extension; in the Eo-Navia area (P) “presence or non-presence of the falling diphthong -ei- in the reflex of Latin FĒRĬAM” refers to a unique item, in its variants (V1) western feira and (V2) eastern feria ‘market’.

The geodifferential features necessarily fit into a complementary geodistribution in several areas, being also present as discrete entities. An area is each one of the territories that, being complementarily distributed in a geographic space, constitutes the whole geography of the geovariants of a phenomenon.

The geodifferential features are present in all levels of the linguistic system: phonetic-phonological, morphosyntactic and lexical-semantic.

They may have synchronic or diachronic character. They may have paradigmatic or syntagmatic character.

3.4. About the presence of more than one feature in a point

It happens very often that, after the contrastive base of a phenomenon has been determined, and the geodifferential features and their spatial distribution are established, some points display the presence of a single feature, while in others more than one feature coexists. It is quite frequent to deem the coexistence of features or polymorphism as something


99

problematic, because it challenges the notion of isogloss or dialectal boundary.16

It is possible that such coexistence results from some recent importation of a contiguous feature, although that coexistence might also be old. At any rate, we do not believe that such a circumstance would create special problems when dealing with the typification of geolectal phenomena. The crucial is to be able to conceive of a phenomenon (P) as having in principle two variants: (V1) = feature “a” and (V2) = feature “b”, in such a way that there are geographic points that belong to one or the other variant. Let us suppose that we also have points simultaneously displaying both variants. This coexistence would admit two treatments from a logical point of view: either to consider that the simultaneous manifestation of (V1) and (V2) does not constitute a separate type or even to consider a separate type, in other words, a new variant (V3):

(P) (V1) feature “a” (V2) feature “b” (V3) feature “a + b”

Moreover, the consideration of a feature (V3) equivalent to “a + b” may be weighted, namely by determining the dominant presence (either “a” or “b”) and its degree of dominance. If three or more features coexist, such weighting becomes more effective when constructing the typification. Another possibility is a typification according to combined parameters:

16 For example, Coseriu (1985: 136) presents the problem in these terms: “Pero al comparar varios mapas, o al reunir varios hechos en un mapa sintético, se impone otra observación esencial, o sea, que, muy a menudo, los límites entre los varios fenómenos léxicos, gramaticales y fonéticos análogos no coinciden, que cada fenómeno tiene su área y tienen áreas distintas hasta palabras que presentan el mismo fenómeno [...] Ahora bien, la no coincidencia entre las varias “isoglosas” implica una nueva visión del problema de los límites dialectales y de aquel otro de las llamadas leyes fonéticas, pues parece indicar que, simplemente, no existen límites entre los dialectos y que las ‘leyes fonéticas’ tienen una aplicación arbitraria caótica: un cambio ocurre en ciertas palabras y no ocurre en otras que se hallan en la misma situación; así, en Renania, no coinciden las áreas del mismo cambio k > ch en make y en ik”. See also Chambers and Trudgill (1998: 106-109).

Chapter Five

100

(P) + feature “a” – feature “b” – feature “a” + feature “b” + feature “a” + feature “b” – feature “a” – feature “b”

In any case, any solution depends on the decision of the dialectologist, who will chose the most appropriate one for the objectives of his/her research.17

3.5. Determining geodifferentiality

The establishment of the diverse contrastive phenomena and geodifferential features is an operation anchored in the objective observation of the facts. Obviously, this operation inevitably entails a certain degree of conventionality when dealing with the treatment and classification of phenomena, features and concrete items. This is inherent to any taxonomic system (García Mouton 1999: 76, in reference to the limitations of language geography and basilectal studies). This conventionality may have several manifestations as we will see next. The first problem, widely debated and illustrated in different papers

of theoretical dialectology, is that, in the determination of geodifferentiality, many more linguistic data than the ones that traditionally have been considered come into play. In our opinion, the fact that in the determination of certain features and areas may intervene different kinds of data – which is undeniable – cannot prevent the researcher from making a selection from a host of data, and keeping those that he/she determines as more relevant. By saying

17 Different procedures in quantitative dialectology take into account feature coexistence. Thus, Rumpf et al. (2009) suggest the “area-class maps”, on the basis of the idea that one of the coexistent features may obtain more statistical weight, which can be measured and placed on a map. On the other hand, Seiler (2004: 367-399, 380-381, 383) studies the geographic coexistence of features of a phenomenon of verbal periphrases in Swiss German. Rona (1976: 17-19) set out his thesis saying that a feature may be present in “active” or in “passive language”, which is the same as verifying an intermediate area with the coexistence of two features.


101

this, we are not denying the complexity of reality, but just accepting the possibility of choosing its representative aspects.18

The same phenomenon may show different geovariants, from a synchronic or a diachronic perspective. For example, in the Ibero-Romania (P) “kinds of article” allows two perspectives. Synchronically, we can consider three geovariants: (V1) “article with consonantal support /l/”, as Cst el, la; (V2) “vocalic article”, as GP o, a; and (V3) “article with consonantal support /s/”, as Balearic Ctl es, sa. However, diachronically these variants can reduce to two: (V1)

“article coming from ĬLLE / ĬLLUM” as opposed to (V2) “article coming from ĬPSE”. We will come back to this later on.

The same phenomenon may show different geovariants, from a paradigmatic or syntagmatic perspective. For example, the (P) “phonological metaphony (raising) of the stressed vowel by a high final vowel” allows in principle two geovariants: (V1) “lack of phonological metaphony” (furacu ‘hole’, secu ‘dry’) and (V2)

“existence of phonological metaphony” (furecu, sicu). Thus, it may be considered a syntagmatic phenomenon if we take into account only its phonic aspect, because the inventory of the phonemic system has not been altered; but it may be considered a paradigmatic phenomenon if we approach it from the morphological point of view, because the metaphony carries a morphological hypercharacterisation in the morphemes of gender, number and continuity (secu / seca / seco / secos / seques || sicu / seca / seco / secos / seques). It may also happen that the phenomenon concerns only one of the perspectives. For example, in the area of the Iberian Peninsula the (P) “placement of clitic pronouns with regard to the verb”, displaying the geovariants (V1) “basic proclitic position” (Castilian, Aragonese, Catalan) and (V2) “basic enclitic position” (Galician-Portuguese, Astur-Leonese), is only considered from a syntagmatic perspective.

Two issues within the ETLEN project have become relevant (Andrés Díaz et al. 2012): The first one concerns the role (for effects of determination and

classification) ascribed to the item-dependent feature’s geodistribution.

18 The scientific practice abounds in this type of cases. Thus, in physics constant speed is an ideal concept that does not take into account the friction; in geometry, the concepts of point, line, or plane do not exist in the perceived reality, and so on.

Chapter Five

102

The question is: must a corpus of concrete items form a group with respect to one contrastive phenomenon, or to several? To this respect, a simple formulation (or “wording”) of the alleged contrastive phenomenon in linguistic terms may be insufficient. For example: (P1) “maintenance of the diphthongs ei, ou or monophthongization in e, o” presents the geovariants (V1) “maintenance of the decreasing diphthongs ei, ou” (western) and (V2) “monophthongization in e, o” (eastern), whose isogloss comes to mark out the boundary between central and western Astur-Leonese. Accordingly, at first glance, the area studied by ETLEN is entirely included in the area of (V1), as represented in Figure 5.3. The map in Figure 5.4 shows how the conservation of ei is displayed in an item such as veiga ‘meadow’ in all the area.19 However, a different geodistribution appears when the geovariants (V1) and (V2) correspond to a diphthong ei resulting from VL /ɛ/, /e/ + yod. Importantly, this change concerns the Eo-Navia area. The map in Figure 5.5 shows the isogloss drawn by the item < STRĬCTUM, displaying (V1) estreito and (V2) estreto, estrecho ‘narrow’. The same behaviour extends to other items, such as the results of Latin DIRĒCTUM ‘right’, FĬLĬCTUM ‘fern’, INTĔGRUM ‘whole’, LĔCTUM ‘bed’, MATĔRIA ‘(wooden) matter’, etc. Therefore, from a geolectal and horiometric point of view, besides (P1), we must consider (P2) “conservation of the decreasing diphthong -ei- resulting from Latin /ɛ/, /e/ + yod, as opposed to its reduction in /e/”. Thus, these phenomena are considered different because, despite presenting the same geovariants, these clearly display divergent geographic distribution.

A phenomenon may contain sub-phenomena at a hierarchically inferior level of analysis. For example, the (P) “conservation or loss of intervocalic Latin -N-” has two geovariants: (V1) “loss of -N-” (western area) and (V2) “conservation of -N-” (eastern area), but, additionally, the area (V1) displays different subareas. These subareas have to do with different phonetic consequences of the loss of -N-, such as vowel nasalization and its subsequent denasalization, conservation of vowel hiatus or its alteration, etc. Within the ETLEN project, as subdivisions within the western domain, these sub-phenomena, and their corresponding sub-features, do not have

19 The maps in figures 5.4, 5.5, 5.6, 5.7, 5.8 and 5.9 have been obtained with CartoDial (Cartografía Dialectal), a program conceived and registered by the ETLEN project team.


103

horiometric relevance, at a primary level, in the spatial distribution between western/eastern linguistic domains.

Figure 5.3: Map of Asturias. Conservation of ei (in the West) as opposed to its monophthongization in e (in the East), for an example like *UAIKA > veiga || vega ‘meadow’. The box signals the ETLEN surveyed area.

Figure 5.4: Survey area of ETLEN. The conservation of the diphthong ei in the results of *UAIKA ‘meadow’ is general: no isogloss divides the area.

Chapter Five

104

Figure 5.5: Survey area of ETLEN. The results of Lat. STRĬCTUM ‘narrow’ show an area of conservation of ei in the West, as opposed to its simplification in e in the East. As a matter of fact, neither dialectology nor any science works with all the full range of phenomena available in reality; instead, simplification or reduction devices are put into operation. That said, we must highlight the fact that the ETLEN project has initially considered 532 phenomena. This is a number notoriously superior to other similar studies, at least for the Eo-Navia area.


105

3.6. Geodifferential feature, item and isogloss

A geodifferential feature is inductively determined from cases or concrete elements that show the same type of difference and a similar geographic distribution. Each one of these elements (cases, examples) is called an “item”.20

Items illustrate the differentiality and geodistribution in areas that, being alike, allow us to associate features to the same phenomenon (Andrés Díaz et al. 2012). In relation to (P) “results of Latin intervocalic -L-”, we differentiate, in the Eo-Navia area, two differential features: (V1) “loss of -L-”, western; and (V2) “preservation of -L-”, eastern. The items are, in this case, the concrete words displaying this differentiality: the reflexes of GELARE ‘to freeze’, MŎLA ‘millstone’ ‘molar tooth’, PALU ‘stick’, MOLĪNU ‘mill’, FĪLU ‘thread’, CALĔNTE ‘hot’, FOLIŎLA × FRĪXU ‘a sort of crepe’,21 COLŌRE ‘colour’, AUIŎLU ‘little uncle, grandfather’ and others. However, items do not necessarily correspond to words. Thus, with regard to (P) “placement of clitic pronouns in the verbal infinitive periphrases haber (a, de) + infinitive, ir + infinitive, querer + infinitive and poder + infinitive”, we saw that the differential features were (V1) “coexistence of types: auxiliary + clitic-infinitive, and auxiliary + infinitive-clitic” (quérolo facer ~ quero facelo ‘I want to do it’, western); and (V2) “exclusive type auxiliary + infinitive-clitic” (quero facelo ‘idem’, eastern). In this case, the items are the pronominal combination possibilities with the periphrases haber (a, de) + infinitive, ir + infinitive, querer + infinitive, and poder + infinitive.

Under a simplistic, ideal outlook, all the items associated with the same phenomenon formulation would show the same geodifferential features, and each item’s areas would overlap in a perfect way, at least for some of their boundaries.22 The isogloss – a cartographic line that separates points (areas, in fact) displaying different geodifferential features with respect to the same phenomenon – would refer, in that way, to the phenomenon as a whole, but could equally apply to each one of its items. Needless to say, there are phenomena approaching this ideal situation: in

20 Spruit (2006: 8-9) calls the items atomic variables, as opposed to feature variables or features. 21 Etymological formulation that combines the form FOLIŎLA suggested by Corominas and the hypothetic *FRIXEŎLU defended by Meyer-Lübke. 22 That is the idea that used to be sustained by the neogrammarians, but was promptly refuted by language geography. In Dauzat (1922: 55-58) this controversy is presented and several Gallo-Romance examples on how the same phonetic phenomenon is manifested in different isoglosses are provided.

Chapter Five

106

the Eo-Navia area, (P) “conservation of stressed Latin vowels /ɛ/ and /ɔ/, as opposed to its diphthongization” presents (V1) “conservation of stressed /ɛ/ and /ɔ/”, western, and (V2) “diphthongization in ie, ue”, eastern, which display exactly the same geographical limits for a great amount of items – see the map in Figure 5.6.23

Figure 5.6: Survey area of the ETLEN. Isogloss of conservation of stressed Vulgar Latin /ɛ/ vs. its diphthongization in /ie/ (illustrated by a great number of items, among which those resulting from Latin TĔRRAM ‘earth’‘land’).

23 It is pertinent to stress, in any case, that many isogloss associations which show up as one single line on the map respond to a visual simplification. As Chambers and Trudgill (1998: 92) point out in relation to the Rhenish fan: “The essential difference between the area of the fan and the other areas is that the isoglosses are more widely separated at the fan. Nevertheless, they are more or less separated everywhere”.


107

But as it very often happens, the items ascribed to the same phenomenon formulation move away from this regularity. This is an observation which has become a repeated topic in theoretical dialectology presentations.24 In the ETLEN, we realized the existence of at least three different forms how the items move away from that regularity:

a) The items present different geodifferential features repertoires. Let us

consider, for example, (P) “preservation of the hiatus -a’e- resulting from the loss of a non-nasal intervocalic consonant, as opposed to its reduction or to the generation of an antihiatic consonant”. Initially, two of its items are *TRA(G)ERE (for classic TRAHĔRE)25 and CA(D)ERE. However, the full range of their geovariants does not coincide, as we can see in the following comparative table:

*TRA(G)ERE CA(D)ERE

-[aˈe]- traer ‘to bring’ -[aˈe]- caer ‘to fall’ -[a i̍]- traír -[a i̍]- caír — — -[ˈaj]- cair -[a j̍e]-, -[a̝ j̍e]- traier, trayer -[a j̍e]-, -[a̝ j̍e]- caier, cayer -[aˈɰe]- traguer — — -[ˈe]- trer -[ˈe]- quer

Table 5.4: Geovariant repertoire of results from Latin *TRA(G)ERE and CADERE in the survey area of the ETLEN. b) The items present different geodistributions for the same geodifferential

features. That means, the feature -[aˈje]-, -[aˈje̝]- does not conform to the same geographic area in the case of traier, trayer ‘to bring’ as in the case of caier, cayer ‘to fall’, consequently displaying different paths of isoglosses – see maps in Figures 5.7 and 5.8.

24 One of the pioneers of geographic dialectology, Georg Wenker, is often quoted because he was the one who established that the phonetic laws that the neogrammarians postulated did not apply to the German dialects. From then on, this is a fact which has been observed again and again in different contexts. See Chambers and Trudgill (1998: 32-33, 159-165). 25 For the *TRA(G)ERE etymology, see Corominas and Pascual (1983: 577-578), Cunha (1986: 785) and Ferreiro (1999: 344-345).

Chapter Five

108

Figure 5.7: Survey area of the ETLEN. Results from Latin *TRA(G)ERE.


109

Figure 5.8: Survey area of the ETLEN. Results from Latin CADERE.

As far as the Eo-Navia area is concerned, and with respect to the horiometric purpose – or feature ascription to western and eastern geotypes –, the following tables display the different feature geodistribution in each case:

Chapter Five

110

western eastern -aˈe-

-aˈɰe--a i̍-

-ˈe--a j̍e-, -a j̍e-

Table 5.5: Geodistribution of results from Latin *TRA(G)ERE in the survey area of the ETLEN.

western axial eastern -aˈe-

-a i̍- -ˈe- -a j̍e-, -a j̍e̝--ˈaj-

Table 5.6: Geodistribution of results from Latin CADERE in the survey area of the ETLEN.

Accordingly, we can observe, for example, that the form trer (<*TRA(G)ERE) appears in the eastern side, while the form quer (<CA(D)ERE) is exclusively axial.

Therefore, the confluence of both circumstances, (a) and (b), led to the splitting of what looked like the same phenomenon, then considered as two phenomena: (P1) “preservation of the hiatus -a’e- resulting from the loss of a non-nasal intervocalic consonant, as opposed to its reduction or to the generation of an antihiatic consonant, in the case of the verb *TRA(G)ERE” and (P2) “ preservation of the hiatus -a’e- resulting from the loss of a non-nasal intervocalic consonant, as opposed to its reduction or to the generation of an antihiatic consonant, in the case of the verb CA(D)ERE”.

In the same way, in our survey, we account for what at first appears as a generic phenomenon (P) “implosive consonant -/s/ or -/θ/ before /p/ or /k/” with a field of variation always formed by (V1) “implosive consonant -/s/” and (V2) “implosive consonant -/θ/”. Nonetheless, the corresponding isogloss has visibly different trajectories depending on the following items. For that reason, we have considered three different phenomena: (P1) moscar ~ mozcar ‘to notch’ and máscara ~ mázcara ‘mask’, for which (V1) and (V2) share a western / eastern geodistribution; (P2) rasp(i)ar ~ razp(i)ar ‘to scrape’ and casp(i)a ~ cazp(i)a ‘dandruff’, with (V1) common to both geotypes and (V2) axial; and (P3) cascarr(i)a ~ cazcarr(i)a ‘filth, tick’ and isquierda ~ izquierda ‘left’, with common (V1) and eastern (V2).

On the other hand, for (P) “results of intervocalic Latin -L-”, there is not a single isogloss dividing the area (V1) “loss of -L-” and the area (V2)


111

“preservation of -L-” because each item displays its own distribution of geodifferential areas. However, in this case, the different isoglosses are located fairly close to each other, which led us to associate them to the same phenomenon.

The same thing happens with the morphosyntactic phenomena (P1) “placement of clitic pronouns in the verbal infinitive periphrases haber (a, de) + infinitive, ir + infinitive, querer + infinitive and poder + infinitive”, and (P2) “placement of clitic pronouns in the verbal infinitive periphrases tener que + infinitive and haber que + infinitive”.

Figure 5.9: Survey area of the ETLEN. Isoglosses of clitic placement in three verbal periphrases (western area: two clitic positions vs. eastern area: invariably enclitic to the infinitive).

Chapter Five

112

In these cases, there is no single isogloss that divides the areas (V1) “auxiliary + clitic-infinitive ~ auxiliary + infinitive-clitic” and (V2) “exclusive type: auxiliary + infinitive-clitic”. Instead, there are specific isoglosses for each item – in this case, the placement of the clitic in each one of the mentioned periphrases. However, it has been found that the same differential features (concerning clitic placement) are geographically delimited by isoglosses displaying particular item-dependent paths, but within an area of mutual proximity that allows them to be grasped within the same phenomenon; see the map in Figure 5.9, where three of these isoglosses are displayed (Andrés Díaz, to appear). c) Some items do not show differentiality. Indeed, when the examined

phenomenon is made up of differential items, it is likely that some related items do not show differentiality. For example, (P) “results of intervocalic Latin -L-” presents two geovariants: (V1) “loss of -L-”, displaying western geodistribution, and (V2) “preservation of -L-”, which shows eastern geodistribution. There is a series of items that illustrate this phenomenon: GELARE, MŎLA, PALU, MOLĪNU, FĪLU, CALĔNTE, FOLIŎLA × FRĪXU, COLŌRE, AUIŎLU, among others. Nonetheless, items such as GŬLA or UALĒRE do not show differentiality: GP gola ‘throat’‘shirt collar’, AL gola ‘narrowness of terrain’; GP and AL valer ‘to be worth’. Thus, these items cannot be associated with the same phenomenon.

Drawing one single isogloss representative of all the features of a phenomenon is something feasible in some cases: in the Eo-Navia area it happens, for example, with the geovariants of (P) “conservation of stressed Latin vowels /ɛ/ and /ɔ/, as opposed to its diphthongization”. In this case, the isoglosses of the type t[ɛ]rra ‘earth’‘land’ (western) || tierra (eastern) overlapp for a great number of items. Also, for (P) “form of 1st person singular of present indicative and all the present subjunctive of some irregular verbs in -ner, -nir”, the contrast poño, poña – veño, veña – teño, teña (western) || pongo, ponga – vengo, venga – tengo, tenga (eastern) is described by overlapping isoglosses in the case of the verbs PONERE ‘to put’, UENIRE ‘to come’ and TENERE ‘to have’.

However, the most frequent case is the accumulation of items that obtain some degree of divergence in their geographic tracings. In the project ETLEN, the frequent existence of a great disparity of geographic tracings led to a redefinition of the contrastive phenomena. We have to understand the notion of geographic unevenness in close relation to the


113

cartographic scale that is used. Thus, those isoglosses that follow a more or less similar path, though looking uneven at a regional level, are practically seen as one single line at a peninsular level. Accordingly, we call homophenomenic isoglosses those isoglosses that follow a very similar or very close path in the map, so that they can be ascribed to the same geolectal phenomenon; and we call heterophenomenic isoglosses those isoglosses that display a very uneven path, which therefore remain distributed in connection to various phenomena.

From the issues presented above, two conclusions can be drawn: The isogloss is, strictly speaking, the spatial manifestation of an item,

that is to say, the cartographic line that marks a concrete geodistribution associated to an item displaying a differential feature. It is important to highlight that the isogloss is the spatial visualisation of the geodifferential feature associated to the individual item, and not the geodifferential feature itself.

The feature geodistribution is an important criterion to classify the features as pertaining in the same phenomenon or as belonging to different phenomena.

3.7. Geodifferential features and geotypes

As we verified in the ETLEN project, a great amount of geodifferential and homophenomenic features tend to have a binary and complementary distribution in two areas: the western one (Galician-Portuguese) and the eastern one (Astur-Leonese). An isogloss bundle that runs approximately in a North-South direction separates these areas (and, sometimes, some features are exclusive to the contact or axial area). The reccurrence of such distribution of areas shapes two main geolectal domains. As a result, the involved features operate as attributes of those geolectal domains.

Obviously, this understanding of the facts does not contradict the existence of a geolectal continuum in the Ibero-Romania. As usually illustrated by the image of the traveller who moves from Cabo de Fisterra in Galicia to Cap de Creus in Catalonia without noticing abrupt linguistic differences, this continuum undoubtedly exists, but at the same time it displays isogloss discontinuities (areas of isogloss concentration) and areal continuities (areas of isogloss dispersion), respectively, corresponding to the peninsular Romance domains.

Chapter Five

114

Thus, it is possible to carry out the synthesis of massive amounts of features in order to discover configurations at a different level of analysis. Individual features construct quantitative realities; if, in a broad territory, these quantitative realities showed a perfectly gradual territorial distribution without isogloss concentration of any type, then, for sure, it would be highly arbitrary to make partitions.26

But the quantitative study of features and areas (as the dialectometric studies reveal) do not offer that type of results: on the contrary, they show accumulated discontinuities from which we may derive the linguistic geotypes that are grosso modo identifiable with the classic domains; in short, geolectal structures become visible through the accumulative synthesis of individual features, which can be found at a different gnoseologic level.27 The Romanist G. I. Ascoli explained it as follows (italics ours):

Un tipo qualunque, – e sia il tipo di un dialetto, di una lingua, di un complesso di dialetti o di lingue, di piante, di animali, e via dicendo –, un tipo qualunque si ottiene mercè un determinato complesso di caratteri, che viene a distinguerlo dagli altri tipi. […] [M]a il distintivo necessario del

26 The heuristic importance of the isogloss bundles and their characteristics is something that is not unnoticed in dialectology: “Els feixos d’isoglosses són, aleshores, interpretats com a divisòries dialectals d’una certa entitat; i la importància dels feixos, com a determinant de rang en el conjunt de l’estructura jeràrquica que configuren els dialectes en el marc de la llengua corresponent. Enteses les coses així, el feix més gran d’isoglosses manifesta la fissura dialectal més important de la llengua; els altres feixos manifesten la resta de fissures dialectals, cadascuna de les quals s’enquibeix en espais definits per fissures dialectals superiors” (Viaplana 1996: 103, 101-106). The determination of the Romance geotypes in France is in close relationship with the isogloss bundles: “Two facts will become apparent from the following chapters: firstly, there are unmistakably isogloss bundles separating French, Provençal, and Franco-Provençal from each other, and secondly, there are other major isogloss bundles in France, as well as any number of non-bundling isoglosses with independent paths. The former group of isogloss bundles proves that France can indeed be divided into three parts. The other isogloss bundles and independent isoglosses force us to ask whether these divisions are meaningful, and if so, are they the only meaningful or the most meaningful divisions that exist?” (Jochnowitz 1973: 20). 27 This leap from the individual to the synthetic occurs in all the sciences. For example, in climatology the simple detailed registration of daily rainfall in different parts of a country would be an information that exhausts itself. But the mathematical-statistical analysis of a large body of data allows to build a classification of climates.


115

determinato tipo sta appunto nella simultanea presenza o nella particolar combinazione di quei caratteri.28

Ascoli (1876: 387) In short, glottodiversity is the object of a taxonomization in which diverse glottological criteria are used, but ultimately these are based in the real bricks of this classificatory complex: the geodifferential features (Andrés Díaz 2011). Dialectometry and other similar methods suggest that that taxonomization can be improved with the help of quantitative procedures (Aurrekoetxea 2010).

3.8. The weighting of the geodifferential features

Bearing in mind the fact that while geodifferential features are the bricks of the classificatory building of language diversity, in the scientific taxonomy there is an old controversy between those who advocate the differential weighting of features and the supporters of an equal value attribution. The weighting supporters defend that there are some features that, for different reasons, have more weight and value, and are therefore more relevant than others. Those against the weighting defend instead that all the features should have an equal score (Adansonian or isocratic principle, Goebl 1987: 102). In traditional dialectology, the ponderation trend has been more relevant, although always subject to doubts or criticism. In synthesis, there have been three main ponderating positions: Choosing a small group of isoglosses / features as representative, and

leaving aside the majority of other isoglosses / features; Considering only one feature / isogloss as most important or

exclusive (monoisoglottism); Giving more weight to some isoglosses than to others.

We think there are reasons for the ponderation of features, but, until adequate statistical instruments are provided, it remains somewhat intuitive. That is why we believe that, at least in our case, it is better to dispense with ponderation for the moment. In general, the modern quantitative dialectology also operates this way, although it is moving into pondering procedures (Goebl 1987: 96, 102). In our opinion, the quantitative ponderation of features would be admissible only when the statistical-mathematic or logical criteria allowing it will be available 28 In connection to Ascoli’s approach and the evolution of the polemics that came out, see Vàrvaro (1968: 104-107), Jochnowitz (1973: 25-33), and Goebl (2003b).

Chapter Five

116

(Andrés Díaz et al. 2007: 756; Andrés Díaz & Álvarez-Balbuena García 2011: 213-214). These possibilities would be related to the following criteria: Paradigmatic criterion. It is the reliable verification that some

contrastive phenomena affect a statistically superior number of items in the system.

Syntagmatic criterion. It is the reliable verification that some contrastive phenomena have a frequency of appearance statistically superior in the produced texts.

Systematic criterion. It is the reliable verification that the different categories of the linguistic structure have different importance, according to a greater or lesser degree of abstraction.29

Hierarchical criterion. It is the logical estimate of some contrastive phenomena being placed at a superior hierarchical level than others. They contain other sub-phenomena logically depending from them. The detection of phenomena and sub-phenomena (and features and sub-features) in the project ETLEN is done under a geodistribution criterion. As we already pointed out, in our case, this distinction does not have “pondering” effects on the statistics, since the sub-phenomena and the sub-features are not taken into consideration when dealing with horiometric measurements. However, we believe that the hierarchy of phenomena and features through geodistribution are interesting concepts that deserve our attention.

29 Different linguistic categories of isoglosses and features have been proposed by different authors as a good criterion for the hierarchization. Thus, Jochnowitz (1973: 20): “One might decide to weight isoglosses according to linguistic criteria. Thus, a phonemic change could be considered more important than a phonetic change, since its effect on the structure of a dialect is more profound. […] A structured group of phonemic changes, such as a chain shift, can be considered more important than a single phonemic change”; or Rona (1976: 9-10): “Accordingly, a very small questionnaire will allow us to determine the principal isoglosses corresponding to phonological and morphological or syntactic features; and these isoglosses will allow us to determine dialect areas, which then can be individually studied and described”. In fact, he says that with only four isoglosses he could single out the dialectal affiliation of Spanish in Chile. See also Chambers and Trudgill (1998: 96-100).


117

3.9. Hierarchy of phenomena and features by geodistribution

The contrastive phenomena and the geodifferential features can be articulated in hierarchical systems. This happens, for instance, when a geodifferential feature (including its area) allows a subdivision in two new sub-phenomena. They, in turn, branch out in two new sub-features (with their own sub-areas), thus establishing consecutive hierarchical levels that may have developed either synchronically or diachronically.

From a synchronic perspective, a hierarchical system is built with logical criteria. The aforementioned example about the kinds of articles in the Iberian Peninsula can be used to illustrate this point: masculine articles with a consonantal support /l/ (el / lo), with a support /s/ (es) and without consonantal support (o). From here, a hierarchy like this (the symbol S means sub-) can be built:

(P) “kinds of article”

(V1) “with consonantal support”

(SP1) “kind of consonantal support”

(SV1) “with support /l/”: el / lo

(SV2) “with support /s/”: es

(V2) “without cons. support “: o

Table 5.7: Synchronic hierarchy of the lectal variation of (P) “kinds of article” in the Iberian Peninsula. The criteria for this synchronic classification follow a logical pattern. They are also subject to a conventionality very similar to that used by the S-DM taxating for fixing the differential units and their statistical computation.

The hierarchical systems establish at the same time implicative scales, based upon the assumption that a phenomenon has to depend on (or be implicated in) another. This can be easily proved from a diachronic perspective. For example, in the Ibero-Romance area (P) “conservation or loss of intervocalic Latin -N-” produces two geovariants: (V1) “conservation of -N-”; and (V2) “loss of the -N-”, associated to specific geographic areas. This feature generates sub-phenomena at the same time. One of them is (SP1) “loss of -N- and its consequences in the vocalic system”, subdivided at the same time in two sub-features: (SV1) “conservation of nasal vowels” and (SV2) “non conservation of nasal vowels”. Another one is (SP2) “conservation or reduction of hiatus resulting from the loss of -N- intervocalic”, giving way to the features (SV3) “conservation of hiatus” and (SV4) “non-conservation of hiatus”, which establish the (SSP) “non-conservation of hiatus, and further

Chapter Five

118

transformations”, with (SSV1) “formation of diphthong” and (SSV2) “monovocalic reduction”. The dependence can be visualised in the following table: (P) “conservation or loss of intervocalic Latin -N-”

(V1) “conservation of -N-”

(V2) “loss of -N-”

(SP1) “loss of-N- and its consequences in the vocalic system”

(SV1) “conservation of nasal vowels”

(SV2) “non-conservation of nasal vowels”

(SP2) “conservation or reduction of hiatus coming from the loss of the intervocalic -N-”

(SV3) “conservation of hiatus”

(SV4) “non-conservation of hiatus”

(SSP) “non-conservation of hiatus & further transformations”

(SSV1) “formation of diphthong”

(SSV2) “monovocalic reduction”

Table 5.8: Hierarchy of phenomena and geovariants, starting from (P) “conservation or loss of intervocalic Latin -N-”. The implicative character of the previous table is quite obvious when examining, for example, the feature of conservation of nasal vowels: we realize that it also implies the feature of loss of -N-, but not the opposite, something that can be represented in the following implicative scale.

loss of -N- > -Ø- nasal vowels lat. LANAM lect A – – llana lect B + – la lect C + + lã

Table 5.9: Implicative scale of three results of -N- in Latin LANAM ‘wool’. In the ETLEN project, these different hierarchical levels are distinguished for the purposes of horiometric quantification, because they reveal a


119

geolectal hierarchization of the boundaries, and, consequently, of the geotypes and sub-geotypes.

References

Álvarez-Balbuena García, F., R. de Andrés Díaz, X. M. Suárez Fernández & M. Cueto Fernández 2011, La “horiometría” o dialectometría de frontera. IX Congreso Internacional de Lingüística General. 21-23 de junio de 2010. Universidad de Valladolid. Actas del congreso, Valladolid: Universidad de Valladolid – Junta de Castilla y León – Instituto Castellano y Leonés de la Lengua – Ayuntamiento de Valladolid – Diputación de Valladolid. 107-134.

Andrés Díaz, R. de 2011, Fronteras lingüísticas y geotipos, con atención a la zona Eo-Navia. In Ramón de Andrés Díaz (coord.) Lengua, ciencia y fronteras, Uviéu: Ediciones Trabe & Seminariu de Filoloxía Asturiana – Universidá d’Uviéu. 121-152.

Andrés Díaz, R. de (to appear), Tractament horio- i dialectomètric de dues isoglosses a la frontera entre el galaico-portuguès i l’asturlleonès, Actes del XXVI Congrés Internacional de Lingüística i Filologia Romàniques.

Andrés Díaz, R. de & F. Álvarez-Balbuena García 2011, Projecte ETLEN de mesurament de la frontera entre els dominis asturlleonès i galaicoportuguès. In G. Colón Domènech & Ll. Gimeno Betí (eds.), Noves tendències en la dialectologia contemporània, Castelló de la Plana: Universitat Jaume I. 205-232.

Andrés Díaz, R. de, F. Álvarez-Balbuena García, M. Cueto Fernández & X. M. Suárez Fernández 2012, Frontières linguistiques et horiométrie. La transition linguistique de l’interfluve Eo-Navia (Asturies) et le projet ETLEN. In X. A. Álvarez Pérez, E. Carrilho & C. Magro (eds.) Proceedings of the International Symposium on Limits and Areas in Dialectology (LimiAr). Lisbon 2011, Lisboa: CLUL. 1-21. Online:

http://limiar.clul.ul.pt (accessed January 15, 2013). Andrés Díaz, R. de, F. Álvarez-Balbuena García & X. M. Suárez

Fernández 2007, Proxecto ETLEN para o estudio dialectográfico e dialectométrico da zona Eo-Navia, Asturias: fundamentos teóricos e metodolóxicos. In H. González Fernández & M. X. Lama López (eds.) Actas do VII Congreso Internacional de Estudos Galegos. Mulleres en Galicia. Galicia e os outros pobos da Península. Barcelona 28 ó 31 de maio de 2003, vol. 2, Sada: Ediciós do Castro. 749-759.

Ascoli, G. I. 1876, P. Meyer e il franco-provenzale, Archivio Glottologico Italiano 2, 385-395.

Chapter Five

120

Aurrekoetxea, G. 2010, La dialecticidad de los dialectos. In A. M. Cano González (ed.) Homenaxe al Profesor Xosé Lluis García Arias. Lletres Asturianes, Anexu 1, Tomu I, Uviéu: Academia de la Llingua. 53-78.

Berruto, G. 1979, La sociolingüística, México: Nueva Imagen. Bibiloni, G. 2000, Llengua estàndard i variació lingüística, València:

Eliseu Climent. Chambers, J. K. & P. Trudgill 1998, Dialectology, Cambridge: Cambridge

University Press. Corominas, J. & J. A. Pascual 1983, Diccionario Crítico Etimológico

Castellano e Hispánico, vol. V, Madrid: Gredos. Coseriu, E. 1981, Los conceptos de “dialecto”, “nivel” y “estilo de lengua”

y el sentido propio de la dialectología, Lingüística Española Actual 3-4, 1-32.

—. 1985, La geografía lingüística, El hombre y su lenguaje. Estudios de teoría y metodología lingüística, Madrid: Gredos. 103-158.

Cueto Fernández, M. (to appear), Continuidad y geotipificación en la frontera lingüística del gallego y el asturiano, Interlingüística 22.

Cunha, A. G. da 1986, Dicionário Etimológico Nova Fronteira da Língua Portuguesa, Rio de Janeiro: Nova Fronteira.

Dauzat, A. 1922, La géographie linguistique, Paris: Flammarion. Elizaincín, A. 1996, Los estudios lingüísticos en la frontera uruguayo-

brasileña. In J. M. Carrasco González & A. Viudas Camarasa (eds.) Actas del Congreso Internacional Luso-Español de Lengua y Cultura en la Frontera (Cáceres, 1 al 3 de diciembre de 1994), vol. I, Cáceres: Universidad de Extremadura. 267-275.

Ervin-Tripp, S. 1964, An analysis of the interaction of language, topic and listener, American Anthropologist 66/6-2, 86-102.

Ferreiro, M. 1999, Gramática histórica galega. I. Fonética e Morfosintaxe, Santiago de Compostela: Laiovento.

Fishman, J. 1972, The sociology of language, Rowley: Newbury House. Francescato, G. 1964, Dialect borders and linguistic systems. In H. G.

Hunt (ed.) Proceedings of the Ninth International Congress of Linguists, The Hague: Mouton. 109-114.

García Mouton, P. 1999, Dialectología y geografía lingüística. In M. Alvar (dir.) Manual de dialectología hispánica. El español de España, Barcelona: Ariel. 63-77.

Goebl, H. 1987, Encore un coup d’oeil dialectométrique sur les Tableaux phonétiques des patois suisses romands (TPPSR). Deux analyses interponctuelles: parquet polygonal et treillis triangulaire, Vox Romanica 46, 91-125.


121

—. 2003a, Regards dialectométriques sur les données de l’Atlas linguistique de la France (ALF): relations quantitatives et structures de profondeur, Estudis Romànics 25, 59-120.

—. 2003b, Graziadio Isaia Ascoli, Carlo Battisti e il ladino. Breve controstoria di una pietra dello scandalo della linguistica a cavallo tra Otto e Novecento. In A. Trampus & U. Kindl (eds.) I linguaggi e la storia, Bologna: Il Mulino. 273-298.

—. 2006, Recent Advances in Salzburg Dialectometry, Literary and Linguistic Computing 21/4, 411-435.

—. 2008a, La dialettometrizzazione integrale dell’AIS. Presentazione dei primi risultati, Revue de Linguistique Romane 285-286, 25-113.

—. 2008b, Le laboratoire de dialectométrie de l’Université de Salzbourg, Zeitschrift für Französische Sprache und Literatur 118/1, 35-55.

—. 2010a, Dialectometry: theorethical prerequisites, practical problems, and concrete applications (mainly with examples drawn from the Atlas Linguistique de la France, 1902-1910), Dialectologia, Geolinguistics around the World, Special Issue I, 63-77. Online:

http://www.publicacions.ub.edu/revistes/ejecuta_descarga.asp?codigo=684 (accessed July 2, 2012).

—. 2010b, Introducción a los problemas y métodos según los principios de la escuela dialectometrica de Salzburgo (con ejemplos sacados del Atlante Italo-Svizzero, AIS). In G. Aurrekoetxea & J. L. Ormaetxea (eds.) Tools for Linguistic Variation, Bilbao / Bilbo: Universidad del Pais Vasco – Euskal Herriko Unibertsitatea. 3-39.

—. 2011, Áreas, fronteras, similitudes y distancias: lección breve de geolingüística cuantitativa. In R. de Andrés (ed.) Ciencia, lengua y fronteras, Uviéu: Ediciones Trabe & Seminariu de Filoloxía Asturiana – Universidá d’Uviéu. 11-33.

Jochnowitz, G. 1973, Dialect boundaries and the question of Franco-Provençal, The Hague / Paris: Mouton.

López Morales, H. 1989, Sociolingüística, Madrid: Gredos. Moulton, W. G. 1968, Structural Dialectology, Language 44/3, 451-466. Orton, H., W. J. Halliday & M. V. Barry (eds.) 1962-1971, Survey of

English Dialects: The Basic Material, vols.1-4, Leeds: E. J. Arnold & Son Ltd.

Pulgram, E. 1964, Structural comparison, diasystems and dialectology, Linguistics 4, 66-82.

Recasens i Vives, D. 1982, La parla del Camp de Tarragona (assaig de síntesi), Tarragona: Òmnium Cultural Tarragonès.

Chapter Five

122

—. 1985, Estudi lingüístic sobre la parla del Camp de Tarragona, Montserrat: Curial edicions catalanes – Publicacions de l’Abadia de Montserrat.

Reed, D. W. & J. L. Spicer 1952, Correlation methods of comparing idiolects in a transition area, Language 28/3, 348-359.

Romaine, S. 2000, Language in Society. An Introduction to Sociolinguistics, Oxford: Oxford University Press.

Rona, J. P. 1976, The social dimension of dialectology, International Journal of the Sociology of Language 9, 7-22.

Rumpf, J., S. Pickl, S. Elspaß, W. König & V. Schmidt 2009, Structural analysis of dialect maps using methods from spatial statistics, Zeitschrift für Dialektologie und Linguistik 76/3, 280-308.

Sala, M. 1998, Lenguas en contacto, Madrid: Gredos. Seiler, G. 2004, On three types of dialect variation and their implications

for linguistic theory. Evidence from verb clusters in Swiss German dialects. In B. Kortmann (ed.) Dialectology meets Typology. Dialect Grammar from a Cross-Linguistic Perspective, Berlin / New York: Mouton de Gruyter. 367-399.

Speitel, H.-H. 1969, An areal typology of isoglosses near the Scottish-English border, Zeitschrift für Dialekotologie und Linguistik 36, 49-66.

Spruit, M. R. 2006, Measuring Syntactic Variation in Dutch Dialects. In J. Nerbonne & W. Kretzschmar, Jr. (eds.) Progress in Dialectometry, special issue of Literary and Linguistic Computing. Selected proceedings of a workshop at Methods in Dialectology XII, Moncton, Aug. 5, 2005, Groningen: University of Groningen and University of Georgia. 1-15.

Vàrvaro, A. 1968, Storia, problemi e metodi della linguistica romanza, Napoli: Liguori.

Veny, J. 1985, Introducció a la dialectologia catalana, Barcelona: Enciclopèdia Catalana.

—. 1992, Fronteras y áreas dialectales. In Actas del Congreso Internacional de Dialectología (Bilbao 21-25, X, 1991), Bilbao: Euskaltzaindia – Real Academia de la Lengua Vasca. 197-245.

Viaplana Lleonart, J. 1990-1991, Comparació interdialectal i llengua, Estudis de Llengua i Literatura 4, 215-241.

—. 1996, Dialectologia, València: Universitat de València. Weinreich, U. 1953, Languages in contact: findings and problems, New

York: Linguistic Circle of New York. —. 1954, Is a structural dialectology possible?, Word 10, 388-400.

About the concept of "geodifferential feature" between linguistic varieties in contact

Documents