**4. Discussion**

The present study adds important findings to the recent longitudinal description of crucial vectors characterizing the division of Catalonian society into two confronted communities, unionists and secessionists, through the eruption of an unexpected and mostly not ye<sup>t</sup> fully explained secession crisis within a rich and advanced region of Southern Europe (Oller et al. 2019a, 2019b).

Figures 1 and 2 offer a useful complement to our main longitudinal portrait by showing maps of the geographical distribution of support to secessionist and unionist forces using recent electoral results. They display that heavily populated coastal areas of Catalonia were less enthusiastic about secession. This was also reflected on the score (y-axis value) of the unionism index (Figure 3), where large municipalities correspond mainly to towns situated on the coastal conurbations of the region. These big conurbations have the highest proportions of citizens from migrant origins, either from other Spanish regions or from abroad (Barceló 2014; Guntermann et al. 2018; Hierro 2015; Lepic 2017; OEC Group 2017; Rodon and Guinjoan 2018).

The economic findings presented here offer a systematic scenario. Figure 4 shows that there was a persistent difference of family incomes between the two main segments of Catalonian citizenry, those whose family language is Catalan vs. those whose family language is Spanish; the former enjoying a higher median income across all the period. The income of both groups fluctuated with the development of the lengthy repercussions of the 2008 financial crises, but the distance of incomes between both groups persisted and even tended to increase throughout the period. Figure 5 shows the evolution of percentages of support for secession in a (hypothetical) referendum of self-determination differentiating two levels of income (above and below 3000 euros) and combining that with family/mother language segmentation, Catalan vs. Spanish. The graph illustrates that, regardless of their level of income, citizenry with family/mother language Catalan gave substantially more support to secession than the

Spanish language segment. The di fference was so large that it appears to reveal that the sharp division on the preference for secession depends mainly on an ethnolinguistic cleavage.

Results displayed at Figures 6–8 present longitudinal analyses of variations on di fferent national identity feelings throughout the whole period 2006–2019. Figure 6 focuses on the evolution of national identity feeling "only Catalan" di fferentiating between the abovementioned two levels of income. The escalation of the identity feeling "only Catalan" was maximum for the wealthier segmen<sup>t</sup> of those with family/mother language Catalan. At the start of the period, however, there were no distinctions on that restrictive national identity within such a Catalan speaking citizenry segment. For citizens with family/mother language Spanish, there was also a very slight trend of increasing "only Catalan" identity, and again the wealthier ones taking the same lead, though a bit later. Figure 7 displays the evolution of di fferences on the national identity "as Catalan as Spanish" using the same type of segmentations. There was an erosion of this dual identity on both citizenry idiomatic segments, with the erosion being higher for the wealthier subgroups. To be noted is that, for the wealthier pertaining to the family/mother language Spanish group, the erosion of this dual identity was more intense, whereas the poorer remained more stable. Figure 8 displays the variations on national identity "only Spanish" within both citizenry lingustic segments di fferentiating again for levels of income, though these were very thin strata (low percentages) within the population. For the family/mother language Catalan fraction there was no variation at all, as percentages of that identity for this citizenry segmen<sup>t</sup> were close to zero. The family/mother language Spanish segmen<sup>t</sup> presented large fluctuations which were likely due to the very small sizes.

Finally, Figure 9 introduced two subjective measures of economic resistance/endurance and perception of current personal economic situation. Two findings were relevant: One, that the intensity of support for secession varied significantly with these measures (the degree of secession support aligning positively and strikingly with both measures); and, second, support for secession attained maximums when economic resistance was stronger and economic perception was clearly optimistic. In all, these results cohere with previous (partial) findings reported by other authors on the relation between secession support and economic wellbeing (Llaneras 2017; Coll et al. 2018; Piketty 2019) and extend their relevance by showing consistent di fferences across the whole period of the secessionist campaign. They also unveiled suggestive co-variations between the main ethno-linguistic cleavage at the region and economic segmentations, but gave more relevance to the idiomatic than to economic distinctions.

Before the dawn of the secessionist surge, (Boylan 2015) had already shown using CEO surveys 2011–2013, that national identity (being Catalan native or assimilated) was a much better predictor of desire for secession, than perceived grievances coming from an unfair fiscal treatment or other economic–political factors. Cohering with that, (Miley 2007) established the operation of divergent modes of national identification across the main segments of Catalonian society that rested on an ethno-cultural gap. Departing from CIS<sup>5</sup> surveys and other social data, he challenged the depiction of Catalonian nationalism as a form of "civic nationalism". He described a cleavage that distinguished the self-identification of two citizenry segments: "Native, Catalan speaking" citizens and their Spanish-speaking neighbours with immigrant origins from other regions of Spain. "Mother tongue" had, in fact, the strongest impact upon identity feelings as predominantly Catalan vs. mostly Spanish or mixed "CatSpanish". In subsequent studies, (Miley 2013) showed that there was also a gap between political preferences of these citizenry segments and those implemented by their representatives in the Regional Parliament: Language and education policies, particularly, were inconsistent with preferences of Spanish-speaking citizens. He identified, moreover, two mechanisms that blocked their representation in the region's institutions: (1) A clear under-representation of those citizens within Parliament; and (2) a partial assimilation of some Spanish-speaking politicians into the attitudes of

<sup>5</sup> CIS (Centro de Investigaciones Sociológicas, http://www.cis.es/cis/opencms/ES/index.html).

Catalan-speaking rulers. He concluded that the social bases of support for Catalan nationalism were "overwhelmingly ethnic" and that the movement was an elite-led, "top down" project.

The present findings add likelihood to that depiction, since we were able to show that ethno-linguistic distinctions were more powerful than economic segmentations in describing variations on both national identity feelings and polarized profiles on the issue of secession, across the entire period of the secessionist campaign. In previous though partial and transversal studies, (Barceló 2014) had already shown that idiomatic and neighbourhood contexts were crucial to keeping distinctive national identity boundaries in Catalonia, and (Hierro 2015) showed that both parents' identities and neighbourhood composition were able to counteract the e ffects of compulsory schooling, mainly or exclusively in Catalan language, at inducing changes in national identities in Catalonia.

The author (Piketty 2019) recently discussed data on Catalonian economic segmentations in relation to preferences in favour of, or against secession, which are also fully coincident with the present longitudinal findings. After contrasting findings obtained from di fferent rent strata or educational levels, he concluded that Catalonian claims for secession should be conceived as a form of "fiscal egoism" born within a rich European region. A phenomenon, by the way, that he and others sugges<sup>t</sup> might appear as well on a variety of segregation tensions that have erupted in other countries, within the European Union (Bourne 2014; Gri ffiths et al. 2015; Muro and Vlaskamp 2016; Miley and Garvía 2019; Piketty 2019).
