Placement. Late in the methodology book, after the negative-space apparatus is established and before the abandoned-project material. It functions as the generalization test: a method derived from one case earns its place only if it deforms predictably across others.
Chapter thesis. The protocol’s layers hold fixed positions but variable content, variable count, and variable separation. Generalizing it is therefore not a matter of applying the Catalan template five times. It is a matter of identifying which parameter each case varies, and what the method yields once it does. Two of the five cases break the protocol outright, and those are the two that teach the most.
§1. The protocol in general form
Strip the Catalonia-specific content and state it abstractly.
- Layer 1: scholarship in the minoritized language. Dense where subject and language are congruent; structurally non-exporting; blind to whatever is internally contentious enough to distort its own field.
- Layer 2: scholarship in the dominant state language. Contains the technical and legal machinery. Nearly unread by Anglophones. Frames the subject as a problem of the state rather than of the community.
- Layer 3: scholarship in English. Event-driven, lagging, framed by whichever Anglophone discipline claimed the subject first.
The claim being generalized is that these three layers say different things, that reading any one alone produces a patterned distortion, and that the richest negative space sits where an object is invisible at all three simultaneously.
§2. The parameters
Name these early, because the case studies are organized by which one each varies.
Layer separation. The distance between Layer 2 and Layer 3. Where they are the same language, the protocol degenerates.
Layer 2 multiplicity. How many state languages the community is divided among. One for Catalan, two for Basque, three for Occitan, four for Sámi.
Layer 1 coherence. Whether the minoritized language constitutes one scholarly community or several.
Middle-layer stability. Whether Layer 2 is itself a dominant language or one losing its own scholarly domain to English.
Layer 2 content. Constitutional law in Catalonia, philology in Occitania, ethnography and Arctic policy in Sápmi. What Layer 2 contains is not fixed and identifying it is a diagnostic step, not an assumption.
Assessment competence. Carried forward from the sovereignty paper.
Withholding. Whether some of Layer 1 is deliberately unpublished. Introduced by the Sámi case and unique to it among the five.
§3. Basque — bifurcated Layer 2, and the isolate condition
What it varies: Layer 2 multiplicity, and mutual intelligibility.
Basque is divided between two states with incompatible legal orders. Spain grants co-officiality in the Basque Autonomous Community and partial recognition in Navarre; France grants nothing. The protocol therefore has two Layer 2s, one Spanish and one French, which disagree not about Basque but about whether Basque is a legal object at all. A researcher reading only the Spanish Layer 2 sees a competence dispute. A researcher reading only the French one sees an absence.
The isolate condition matters more than it appears. Catalan leaks to Spanish partly because a Spanish reader can work through Catalan text with effort; Romance-to-Romance permeability makes the border porous in both directions. Basque is opaque to everyone outside it. This seals Layer 1 in a way that simultaneously protects it, since Basque scholarship must be its own domain or nothing, and renders it maximally invisible, since no outsider stumbles into it. Euskaltzaindia, UZEI, and the Basque-medium provision at UPV/EHU constitute real infrastructure that essentially nobody outside reads.
The finding this case supplies for the sovereignty argument: the Basque Country holds the strongest sub-state fiscal competence in Western Europe through the Concierto Económico and the foral system. It can fund whatever it chooses. It still cannot make a Basque-language publication count toward a sexenio. Fiscal autonomy and assessment competence are separable, and it is the second that matters. This is the cleanest possible demonstration that money is not the lever.
§4. Welsh — the degenerate case
What it varies: layer separation, set to zero.
Layer 2 and Layer 3 are the same language. Everything about Wales not written in Welsh is immediately legible to any Anglophone. The protocol collapses to two layers, and its distinctive yield, the excavation of a hidden Layer 2 counterliterature, vanishes entirely because nothing is hidden.
This is worth a full case rather than a footnote for two reasons.
First, it establishes the method’s scope condition. The protocol repays effort in proportion to layer separation. Where separation is zero, three-layer reading is just ordinary bilingual reading and claiming otherwise would be method for its own sake.
Second, it produces a different failure mode worth naming: proximity-induced neglect. Wales is over-covered in English relative to its size, because it sits inside the Anglophone world. Welsh-language scholarship is nonetheless under-read even by people writing about Wales, precisely because the English-language material is sufficient-seeming. Distance-induced invisibility, the Norwegian and Basque pattern, and proximity-induced neglect, the Welsh pattern, produce similar-looking gaps by opposite mechanisms, and the remedies differ. Distance is fixed by translation and indexing. Proximity is fixed only by convincing researchers that the accessible layer is incomplete, which is harder.
The Coleg Cymraeg Cenedlaethol is heavy teaching-side investment without assessment competence, since the REF is UK-wide. The sovereignty thesis predicts Wales patterns with Catalonia rather than Flanders, and this chapter should say so as a falsifiable claim.
§5. Frisian — the leaky middle
What it varies: middle-layer stability.
Frisian sits under Dutch, and Dutch is itself a medium-sized language losing scholarly ground to English. The chain is Frisian → Dutch → English with the middle term undergoing, one level up, the process the chapter is describing one level down.
The consequence for method is concrete. Layer 2 excavation assumes Layer 2 holds material the Anglophone layer lacks. Where Layer 2 is itself dissolving into English, that assumption weakens over time, and the Frisian researcher of 2040 may find that the Dutch-language layer has thinned to the point of yielding nothing. The protocol has a shelf life that varies by case, and this is where to say so.
Two further notes. The Netherlands runs a comparatively qualitative research assessment culture rather than a metric-driven one, which makes it a useful control against the Spanish design. And the Fryske Akademy hosts the Mercator centre, which is the European clearinghouse for exactly this literature, so the smallest of the five cases is where the field keeps its own records. Worth a sentence for the irony and a paragraph for the practical point that Mercator is where a researcher should start on any of the five.
§6. Sámi — maximal fragmentation and the withheld layer
What it varies: everything, plus one parameter the other cases do not have.
Layer 2 is four languages across four states. Layer 1 is itself fragmented, since North, Lule, South, Inari, and Skolt Sámi are not mutually intelligible and do not constitute one scholarly readership. Sámi parliaments exist in three states with limited and unequal competence. Sámi Allaskuvla at Kautokeino is the institutional anchor.
The sharpest test in the chapter. Norway is the state that built the publication indicator to protect Norwegian-language scholarship. Does that protection extend to Sámi-language output, or does it protect Norwegian specifically? This is empirically checkable and the answer determines whether assessment protection is a linguistic good or a national one. My expectation is the latter: states build these instruments to defend the state language against English, and a minoritized language inside the state receives incidental benefit at best. If that holds, it means the most protective assessment regime in Europe reproduces toward Sámi the structure Catalonia complains of, which is the same finding the Aran paper reaches by a different route and confirms it across an independent case.
The fourth layer. Sámi introduces something the Romance and Germanic cases do not: knowledge deliberately withheld. Indigenous research ethics protocols, community consent requirements, and traditional knowledge that is not published as a matter of choice constitute a body of material that is absent from all three layers for reasons that have nothing to do with publication economics.
This deserves its own section in the chapter, because it is a genuine boundary on the negative-space framework generally. The framework treats absence as a symptom to be explained and, usually, remedied. Chosen absence is not a symptom, and the researcher’s correct response is to note the boundary and stop, not to excavate. A methodology that cannot distinguish an economic gap from a refusal will eventually mistake the second for the first, and that failure is worse than leaving the gap unfilled. Say this plainly.
§7. Occitan — one language, three states, three positions
What it varies: the state, holding the language constant. The best natural experiment in the set, and the reason it goes last.
Occitan spans France, the Val d’Aran in Spain, and the Occitan valleys of Piedmont in Italy. The legal positions are radically different. France recognizes nothing, with Article 2 of the constitution declaring French the language of the Republic and the European Charter for Regional or Minority Languages signed but unratified. Italy recognizes Occitan through Law 482/1999. Spain makes Aranese co-official within Catalonia. Same language, three legal orders, three completely different Layer 2s.
This lets the chapter isolate the state variable cleanly, which none of the other cases permit.
It also demonstrates the variable content point most vividly. Layer 2 for Occitan is not constitutional law, because there is no constitutional dispute in France to generate one. It is philology and medieval literary studies, a large French-language scholarly tradition about Occitan produced by people with no stake in its survival as a living medium. Layer 2 occupies the same structural position as Catalonia’s constitutional literature and contains something entirely different, and a researcher who went looking for a legal counterliterature would conclude, wrongly, that Layer 2 was empty.
The chapter closes here because Aran returns the argument to the Catalan suite. The recursive finding from Paper 7 reappears as one of three positions the same language occupies, which reframes it: Aran is not an embarrassing anomaly in Catalan language policy but the best-off of the three Occitan territories. That reframing is worth the whole chapter.
§8. The diagnostic sequence
The practical payoff. A researcher approaching an unfamiliar small-language subject runs this in order.
- Count the layers. Identify every state language the community is divided among. If Layer 2 equals Layer 3, stop; the protocol has nothing extra to offer.
- Locate each layer’s index. Not the international aggregators. Dialnet for Spanish, RACO for Catalan, the national CRIS systems for the Nordic countries, Mercator for the minority-language literature generally.
- Determine what Layer 2 contains before assuming. Constitutional law, philology, ethnography, and policy studies all occupy that position in different cases.
- Run the spike test on Layer 3. Plot English-language output by year and find the crisis clusters. The troughs are where the gaps are.
- Check Layer 1 coherence. Is this one scholarly community or several presenting as one?
- Look for objects invisible at every layer at once. These are the strongest findings and the rarest.
- Ask whether any absence is chosen. If so, note the boundary and go no further.
§9. Scope and failure modes
Close honestly. The protocol yields most where separation and multiplicity are high and where the subject has never had a crisis event pulling it into English. It yields nothing in the Welsh configuration. It has a shelf life in the Frisian configuration. It has an ethical stopping point in the Sámi configuration. And it carries a standing risk worth naming: a method built to find gaps will find them whether or not they are there, and the discipline against that is the mapping step, which is why three of the six Catalan papers do nothing but establish what is present before anything is claimed absent.
