Whom Do University-Based Geoscientists Recognize as Important to the Study of Portuguese Geology? A Purposive Expert Nomination Survey across Seven Chronological and Thematic Domains

1. Introduction

Which researchers a scientific community regards as important to the development of its field is a question with both historical and practical relevance. The answers people give shape how a discipline is taught, commemorated and explained to non-specialist audiences, and they contribute to a community’s collective identity [1,2]. In this article we treat such answers as one observable trace of disciplinary memory, that is, of the figures that members of a community name as significant when they are asked. Three levels need to be kept apart. The first is present-day recognition among a defined group of experts, which is what a nomination survey measures. The second is the canon or memory of a discipline, to which present-day recognition is one contributor among others, such as textbooks, commemorations and institutional histories. The third is historical influence on the production of knowledge, which requires other kinds of evidence. The present study provides evidence at the first level. It uses the second and third as questions to which that evidence can contribute.

This positioning follows recent proposals to study the history of the Earth sciences as a history of knowledge, attentive to the practices, actors and socio-cultural contexts through which knowledge is produced and shared [3]. Collective-memory theory offers a starting point. Halbwachs argued that memory is always shaped by the social groups to which individuals belong [4], and Zerubavel described mnemonic communities that determine which figures remain visible as generations succeed one another [5]. Applied to a scientific community, these ideas suggest that the recognition of predecessors is socially mediated, but they do not predict which figures will be recalled. Studies of scientific recognition show that credit is distributed unequally and cumulatively [1,6], and Kuhn observed that textbooks, through which new generations learn a field, give a selective and retrospective account of its history [7]. Raj’s work on the circulation of knowledge is a reminder that knowledge about a territory is built through exchanges between actors and places of different origins [8], which matters here because foreign researchers appear among the nominees.

The geosciences in Portugal have a documented history of at least two centuries. The state geological commissions and services of the nineteenth and twentieth centuries, in which Carlos Ribeiro and Nery Delgado were leading figures, carried out the geological survey of the mainland territory [9,10]. Researchers who came from abroad also left a lasting mark: Paul Choffat, who came from Switzerland, on Mesozoic stratigraphy, and Georges Zbyszewski on the Cenozoic and on the geology of the Atlantic islands [11,12]. These contributions delineate chronologically differentiated layers in the constitution of Portuguese geology. Whether present-day geoscientists carry a memory of these layers, and of which figures, is an empirical question. The university-based geoscientific community is small enough (230 individuals in the invitation pool used here) for a near-census of that group to be feasible. It is not the whole community of people who work on Portuguese geology, which also includes geoscientists employed outside public universities.

To our knowledge there is no published systematic account of whom present-day geoscientists recognize as important to the study of Portuguese geology. Existing histories cover specific periods and figures and follow the editorial criteria of a small number of authors [9,10,11,12]. They do not aim to document how the current community itself recalls its predecessors. Citation-based indicators, of which the h-index is the best-known example [13], are of limited help for this purpose. They depend on indexed databases and on citing behaviour [14], and they are hard to apply to figures whose work predates digital indexing. More generally, the science-of-science literature shows that visibility depends on structural factors such as collaboration networks and cumulative advantage [15,16,17], and the Leiden Manifesto argues that indicators should support, not replace, expert assessment and should take field-specific practices into account [18]. Reports on research assessment reach a similar conclusion about the continuing role of expert judgement [19]. Reynaud, writing about observation in the geosciences, remarked that what a researcher sees is essentially what he looks for (“Que voit-il [le chercheur]? Essentiellement ce qu’il cherche!”) [20] (p. 17). A method that looks for citations can only find citable recognition. A method that asks members of a community to name figures records something different, and that difference is what this article explores.

Nomination surveys ask members of a defined group to name individuals according to a stated criterion, and so record recognition as it exists among those respondents. They can include historical figures regardless of their presence in bibliographic databases and can cover several thematic and chronological domains in a uniform way. Structured elicitation of expert knowledge is an established practice in other fields, for instance in conservation science [21] and in public-policy analysis [22], although those applications usually elicit quantitative or probabilistic judgements rather than names, and they document the cognitive heuristics and overconfidence to which expert judgement is subject [22]. The design used here is related to the Delphi method [23,24,25] but is not a Delphi study. A Delphi study iterates over several rounds with controlled feedback and defines consensus in advance through explicit criteria [26]. The present survey has one round, no feedback and no prior consensus criterion. We therefore describe it as a purposive, single-round expert nomination survey, and we speak of recognition and of the convergence of nominations rather than of consensus.

This article reports the design and descriptive results of the survey carried out within the project Geology of Portugal through Faces (GeoFaces). Its objectives are: (1) to describe the survey design, the invitation pool, the response and the data processing in enough detail for the article to stand on its own; (2) to present, for each domain, descriptive statistics and the ranking by number of nominations; (3) to describe the biographical corpus that GeoFaces defines from these rankings and the operational criterion used, without claiming that the criterion identifies a consensus; (4) to read the patterns of nomination against the historiography of Portuguese geology, taking as main cases the persistence of nineteenth-century figures and the differences between domains and between recognition profiles; and (5) to state what the evidence does and does not support. Three other pieces of work are distinct from this one. The development and validation of the questionnaire are reported in a separate manuscript under review [27]. The associations between respondents’ profiles and nomination patterns are analysed in a second article. A broader theoretical interpretation, within the sociology of disciplinary memory and the history of Portuguese geology, is planned for a third. The historical interpretation offered here is limited to what the present survey and the cited historiography can support.

2. Materials and Methods

2.1. Study Design and Instrument

The survey is a purposive, single-round expert nomination survey developed within GeoFaces at the University of Aveiro (Portugal). The questionnaire had two sections. The first collected sociodemographic and academic-professional information: sex, nationality, doctoral degree, habilitation title, institutional affiliation and type of employment, academic category, years of research and teaching experience, disciplinary areas of activity (selected from a predefined list) and geochronological terrains of activity (selected from a predefined list of five). The second section presented seven chronological and thematic domains. For each domain, respondents first stated whether they had sufficient knowledge to nominate geoscientists (binary response). Those who answered affirmatively were invited to nominate up to ten geoscientists whom they considered most relevant to the development of geological knowledge in that domain, in decreasing order of importance (1 = most important). The questionnaire was administered in Portuguese. For the first domain, the item read (original wording): “Para o Enquadramento Geotectónico da Península Ibérica e da Europa, indique até 10 geocientistas que considera mais relevantes. Ordene-os por ordem decrescente de importância (1 = mais relevante).” In English: “For the Geotectonic Framework of the Iberian Peninsula and Europe, name up to 10 geoscientists whom you consider most relevant. Rank them in decreasing order of importance (1 = most relevant).” The items for the other six domains had the same wording, with the domain replaced. No list of names was suggested, and nominations were entirely free. This choice followed piloting, in which an earlier closed-list version showed ceiling effects and framing biases [27]. For the geoscientist placed first in each domain, respondents were also invited to state in free text (maximum three lines) the contributions that justified the choice. The final question of the questionnaire invited open comments.

The seven domains are those adopted in the book Uma Geologia de Portugal. Contributos da Geoeducação para a Formação de Professores [28]: (1) Geotectonic Framework of the Iberian Peninsula and Europe; (2) Precambrian; (3) Lower Paleozoic; (4) Upper Paleozoic; (5) Mesozoic; (6) Cenozoic; and (7) Meso-Cenozoic of the Atlantic Islands. The study is limited to geological research on mainland Portugal and on the Atlantic archipelagos of the Azores and Madeira. Geoscientists whose contributions concerned only other regions, such as the former Portuguese overseas territories, were outside this scope and were not eligible for nomination, regardless of nationality.

The division into seven domains combines chronostratigraphic, thematic and practical criteria. The chronostratigraphic domains follow the major divisions of geological time represented in Portuguese territory and in the Atlantic islands, as set out in the international chronostratigraphic scale [29] and in the reference geological cartography [30]. The Geotectonic Framework was treated as a separate domain because it is an interpretive framework that cuts across all periods and involves a community of specialists with its own identity. The Meso-Cenozoic of the Atlantic Islands was separated from the mainland Mesozoic and Cenozoic because it corresponds to a distinct geological context (intraplate oceanic volcanism, no Mesozoic sedimentary record on the islands) and to a group of researchers whose specialization, methods and bibliography are largely independent of those of mainland specialists. This structure has consequences for the results, because it can exclude geoscientists whose work spans several domains, as some respondents observed (Section 4.9).

The instrument went through two iterative cycles of piloting. Its development and validation are reported elsewhere [27]. The information needed to interpret the results (item wording, response options, absence of suggested names, scope and domain definitions) is given here so that this article can be read independently. The survey was administered electronically through the LimeSurvey platform over eight weeks (April and May 2026). Participation was voluntary, and responses were anonymous at the level of the individual respondent.

2.2. Invitation Pool and Respondents

The survey was sent to a purposive pool of 230 geoscientists: researchers and teaching staff of departments of geosciences, earth sciences or geology of Portuguese public universities (10 institutions). The pool was compiled by searching the institutional websites of Portuguese public universities for departments of geosciences, earth sciences, geology or similarly named units. The pool therefore does not include geoscientists who work outside these university departments, for instance in private institutions, geological surveys or companies. In this article, “respondents” and “the panel” refer to the participating university-based geoscientists, and statements about the wider geoscientific community are made only as hypotheses.

A total of 129 responses were received (56.1% of the pool). Of these, 45 (19.6% of the pool; 34.9% of the responses) were complete, meaning that the respondent reached the end of the questionnaire and submitted it, and 84 (36.5% of the pool) were partial. Most partial responses (n = 71; 84.5%) contained no nomination: 43 progressed through the first section as far as the items on areas of activity, 10 abandoned it earlier, and 18 provided no profile data. Thirteen partial respondents (15.5%) provided at least one nomination, one of which was not interpretable (the single character “X”), leaving 12 valid partial responses. To keep the analysis homogeneous, only the 45 complete responses were used to construct the rankings. Of these 45 respondents, 39 nominated geoscientists in at least one domain, and 6 declared insufficient knowledge in all seven domains although they completed the rest of the questionnaire. Profile data from the partial responses are used in the non-response comparison in Section 6.

Of the 45 complete respondents, 57.8% (n = 26) were male and 40.0% (n = 18) were female, and one participant (2.2%) preferred not to answer. Regarding nationality, 95.6% (n = 43) were Portuguese and 4.4% (n = 2) had another nationality (one Spanish and one Italian). Forty-four held a doctoral degree (97.8%), 42 of them (93.3%) in geology, and ten held a habilitation title[1] (22.2%). Thirty-eight were in the university teaching career (84.4%), three in the scientific research career, and four had another type of affiliation. The mean research experience was 29.2 years (SD = 11.9; range 0-43; n = 43, as two respondents did not answer this item) and the mean teaching experience was 27.8 years (SD = 14.0; range 1-47). Respondents could select several disciplinary areas of activity: 24 (53.3%) reported Environment and Society / Environmental Geology; 13 (28.9%) External Geodynamics (Stratigraphy, Sedimentology and Paleontology); 8 (17.8%) Internal Geodynamics (Structural Geology, Tectonics and Geophysics); 16 (35.6%) Rock Materials (Geochemistry, Mineralogy and Petrology); 4 (8.9%) Education and History of the Geosciences; 4 (8.9%) Geomorphology; 17 (37.8%) Resources, Engineering and Energy; and 8 (17.8%) other areas.

Table 1 shows the distribution of the 45 respondents across the ten universities, with the response rate by institution. Response rates varied widely. Among universities with at least one respondent, they ranged from 39.1% (NOVA University Lisbon) to 7.8% (University of Lisbon), the latter despite being the institution with the largest number of invitees (n = 77). University of Minho had a rate of 8.3%, and University of Madeira, with two invitees, had no respondents. This unevenness means that the panel does not reproduce the institutional composition of the pool.

The instrument did not prevent self-nomination, and because responses were anonymous it is not possible to determine whether or to what extent it occurred among the invited geoscientists. The name “Soares de Andrade” appears twice among the nominees (one nomination in the Geotectonic Framework and one in the Precambrian domain), well below the threshold of the operational criterion. This is the co-author António Soares de Andrade. None of the authors of this article, including António Soares de Andrade, belonged to the invitation pool or completed the questionnaire, so these nominations cannot be self-nominations.

2.3. Data Processing: Name Normalization and Computation

Nominations were free-text entries, which produced considerable orthographic variation in the names. Three main types of variation were found: (a) spelling variants of the same name (for example “Rogério Rocha”, “Rogério Bordalo da Rocha” and “Rogério Bordalo Rocha”); (b) abbreviated or partial forms (for example “Zby” for Georges Zbyszewski and “Telles Antunes” for Miguel Telles Antunes); and (c) typographical or memory errors (for example “Berkley” for “Berkeley” and “Ruget” for “Ruget-Perrot”).

Before any quantitative analysis, names were normalized within each domain in three steps: (i) identification of all unique name strings per domain; (ii) construction of a domain-specific normalization map assigning variants to a canonical name; and (iii) validation of all harmonization decisions by the principal investigator. In ambiguous cases, the most complete form documented in the Portuguese geoscientific literature was adopted. Entries in which two names had been typed without a separator were split, and one name divided by a comma (“Christiane Ruget, Perrot”) was rejoined. After these corrections no list exceeded ten entries. Duplicates within a list (the same geoscientist named twice by the same respondent) were reduced to a single entry, keeping the first occurrence. The number of variants merged per domain ranged from 7 (Upper Paleozoic) to 28 (Mesozoic).

Nomination frequencies and percentages were calculated with the assistance of an AI assistant (Claude Opus 5, Anthropic), which executed data-processing code (Python 3.12, pandas library) on the respondent database supplied by the first author in Excel format. The authors defined the inclusion criteria and reviewed and validated all outputs. All tables and figures in this version were cross-checked against the corrected nomination database (domain totals, numbers of distinct nominees, nomination frequencies and corpus sizes) and are consistent with the Supplementary Material.

2.4. Rankings and the Operational Criterion for the Biographical Corpus

For each domain, the ranking counts the total number of nominations received by each geoscientist, irrespective of position on the list. Geoscientists with the same number of nominations share a rank. The position given by each respondent (1st to 10th) was recorded separately and is reported as complementary descriptive information (P1 to P3 in the ranking tables).

The GeoFaces project will produce biographical texts for a limited set of geoscientists in each domain. To define this set, the project adopted an operational criterion: inclusion of all geoscientists with three or more nominations in a domain, provided that the threshold falls at a break in the frequency distribution immediately above a broad tie at two nominations. In the Geotectonic Framework, the domain with the most respondents (n = 32), the threshold was set at five nominations. Three nominations would correspond to 9.4% of respondents in that domain, whereas in the other domains three nominations correspond to between 15% (Mesozoic) and 25% (Upper Paleozoic) of respondents, and five nominations correspond to 15.6% in the Geotectonic Framework.

This criterion is a decision of the project about the size and composition of its biographical corpus. It is not a statistical test, and it does not establish that nominees above the threshold are more important than those below it, or that their recognition is unlikely to be due to chance. The study has no null model for the frequency of free-text nominations of named individuals, so a claim of that kind cannot be tested. Placing the threshold above a tie avoids separating individuals with equal support, but any threshold separates people who differ by one nomination. To show how far the corpus depends on the criterion, Section 3.10 reports its size under alternative thresholds.

2.5. Analysis of the Justifications for First Place

Respondents who placed a geoscientist first in a domain were invited to justify the choice in free text. Each justification is linked to the geoscientist actually placed first by that respondent, which is not always the geoscientist with the most nominations in the domain (this occurs in the Cenozoic). The analysis was qualitative and interpretive and was carried out by a single analyst, without a second coder or a measure of inter-rater agreement. The excerpts quoted in Section 3.2 to 3.8 illustrate the reasons most often given for the leading figures and are not an exhaustive coding of the free-text material. This limitation is taken up in Section 6.

2.6. Scope of Inference and Panel Size

The design should not be judged by the criteria that apply to probability samples. Respondents were not drawn at random from a defined population, response was partial, and the panel is uneven across institutions (Table 1). The results therefore describe the nominations of the respondents who took part. They do not estimate the distribution of opinion in the pool of 230 geoscientists, still less in the wider geoscientific community. Standard errors and significance tests are not reported, because no inferential claim is made.

The literature on Delphi-type panels shows no agreement on an adequate panel size. Published panels range from a handful of participants to more than a thousand [31], and commonly cited recommendations for Delphi panels are smaller than the present panel, for instance 10 to 18 participants [32]. Panel size is therefore not, on its own, an argument for the validity of the results. What matters is the relation between the claims and the panel. In the Upper Paleozoic, with 12 respondents, a single nomination corresponds to 8.3 percentage points of respondents, against 3.1 in the Geotectonic Framework (n = 32), so the rankings in the smaller domains are correspondingly coarse.

As a check on the stability of the upper part of the rankings, the analysis was repeated after adding the 12 valid partial responses to the 45 complete ones (Section 6). This check shows how far the top of the rankings depends on the exclusion of partial responses. It is not a test of validity or of representativeness.

3. Results

3.1. Overview

Table 2 summarizes the seven domains. The 39 respondents who made at least one nomination produced 818 nominations in 126 domain lists. These correspond to 340 nominee-domain pairs and 212 distinct geoscientists after normalization. The number of respondents who nominated in a domain ranged from 12 in the Upper Paleozoic (26.7% of the 45 complete responses) to 32 in the Geotectonic Framework (71.1%). The mean number of names per list was 6.5 (SD = 2.8), from 5.1 in the Atlantic Islands to 7.3 in the Upper Paleozoic. Longer lists in the domain with the fewest respondents may indicate that those who chose to answer had more specific knowledge, but the survey does not allow this to be tested.

The distribution of nominations is strongly asymmetric in every domain. The proportion of nominees named only once ranged from 32.5% in the Mesozoic (13 of 40) to 63.5% in the Lower Paleozoic (33 of 52). In each domain a small group of names receives a large share of the nominations, and a long tail of names is nominated once. Table 3 shows the share of nominations absorbed by the biographical corpus and by the three most nominated geoscientists.

The corpus absorbs a larger share of nominations in the Mesozoic (77.2%), the Atlantic Islands (63.6%) and the Cenozoic (60.4%) than in the other four domains (48.6% to 54.8%), which means that nominations are concentrated on a more limited set of names in those three domains. Because the size of the corpus is itself a product of the threshold and of the number of respondents, these percentages are descriptive and should not be compared as measures of agreement.

3.2. Geotectonic Framework of the Iberian Peninsula and Europe

3.2.1. Descriptive Statistics

The Geotectonic Framework had the largest number of respondents (32; 71.1% of the complete responses), who made 222 nominations of 83 distinct geoscientists. The mean list length was 6.9 (SD = 3.0; minimum 1; maximum 10). All 32 respondents (100%) filled the first position, 25 (78.1%) the fifth and 10 (31.2%) the tenth. Forty-five of the 83 nominees (54.2%) were nominated once, 16 (19.3%) twice and 22 (26.5%) three or more times. The corpus comprises the 12 geoscientists with five or more nominations, who account for 108 of the 222 nominations (48.6%). The reason for using five rather than three is given in Section 2.4.

3.2.2. Ranking

Table 4 presents the ranking, with the percentage of the domain’s respondents who nominated each geoscientist and the number of times each was placed 1st, 2nd or 3rd (P1, P2, P3).

António Ribeiro is the most nominated geoscientist, with 28 nominations (87.5% of respondents) and 16 first places, the largest number of first places received by any nominee in any domain. Rui Dias follows with 17 (53.1%). The two have co-authored work on the Variscan Ibero-Armorican arc [33], and the survey cannot tell whether respondents nominated them together for that reason. Carlos Ribeiro, José Ramón Martínez Catalán and Fernando Noronha are tied for third place (8 nominations each), followed by Nery Delgado (7), Paul Choffat and José Tomás de Oliveira (6 each) and four nominees with 5 (Manuel Julivert, Carlos Teixeira, António Mateus and José Carlos Kullberg). Carlos Ribeiro, Nery Delgado and Paul Choffat, all born in the nineteenth century, were nominated by between 18.8% and 25.0% of respondents in this domain.

3.2.3. Justifications for First Place

Of the 16 respondents who placed António Ribeiro first, 9 (56.3%) gave a justification. These refer mainly to his work on the geodynamics and plate tectonics of the Iberian Peninsula, with recurring mentions of a “geodynamic model for the Iberian Peninsula”, of “knowledge of Variscan tectonics” and of the creation of scientific infrastructure (“created the Laboratory of Tectonophysics and Experimental Tectonics in 1992”). One respondent placed him ex aequo with Rui Dias for the “geotectonic models and the 2 published volumes of Geology of Portugal”.

3.3. Precambrian

3.3.1. Descriptive Statistics

The Precambrian had 15 respondents (33.3% of the complete responses), who made 91 nominations of 45 distinct geoscientists (mean list length 6.1; SD = 3.2; minimum 1; maximum 10). Nine lists reached the fifth position, against 15 for the first, a steeper drop than in the Geotectonic Framework. Twenty-five of the 45 nominees (55.6%) were nominated once. The corpus comprises the 10 geoscientists with three or more nominations, who account for 46 of the 91 nominations (50.5%).

3.3.2. Ranking

Table 5 presents the ranking of the Precambrian.

António Ribeiro leads with 10 nominations (66.7%), four of them in first place. Carlos Teixeira, Rui Dias and Francisco Gonçalves follow with 5 nominations each (33.3%). Nery Delgado, Manuel Bernardo de Sousa and Fernando Noronha have 4 each (26.7%), and António Sequeira, Luís Carlos Gama Pereira and Manuel Francisco Pereira close the corpus with 3 each. Delgado received two first places from his four nominations.

3.3.3. Justifications for First Place

Two of António Ribeiro’s four first places came with a justification. They refer to his “comprehensive geological vision of Portuguese geology” and to his “fundamental contribution to understanding the Variscan Orogeny in Portugal”, themes already seen in the Geotectonic Framework (Section 3.2.3).

3.4. Lower Paleozoic

3.4.1. Descriptive Statistics

The Lower Paleozoic had 16 respondents (35.6%), who made 104 nominations of 52 distinct geoscientists (mean list length 6.5; SD = 2.6; minimum 2; maximum 10). The number of nominations by position was similar up to the fourth (14 to 16) and fell from the seventh. Thirty-three of the 52 nominees (63.5%) were nominated once, 7 (13.5%) twice and 12 (23.1%) three or more times. These 12 form the corpus and account for 57 of the 104 nominations (54.8%).

3.4.2. Ranking

Table 6 presents the ranking of the Lower Paleozoic.

Two geoscientists share first place, Nery Delgado and António Ribeiro, with 8 nominations each (50.0%). Delgado was placed first more often (6 times, against 3 for Ribeiro). Rui Dias and José Tomás de Oliveira follow with 6 each. Carlos Teixeira, Carrington da Costa, Fernando Noronha, José Manuel Piçarra and Artur Sá have 4 each, and Ícaro Dias da Silva, Carlos Ribeiro and Paul Choffat close the corpus with 3 each. Delgado’s position is discussed in Section 4.3.

3.4.3. Justifications for First Place

Of Nery Delgado’s 6 first places, 5 (83.3%) came with a justification, the same rate as in the Upper Paleozoic. The justifications refer to his founding role in Portuguese Paleozoic stratigraphy. He is described as having “established the stratigraphic foundations of Portuguese Paleozoic geology” and as “the greatest historical reference for the Portuguese Paleozoic”, with explicit mention of his “monumental monographs on the Silurian [...] and the Cambrian” and of his “detailed study of fossils (trilobites, graptolites)”. One respondent specified his mapping work “in Valongo and Buçaco”.

3.5. Upper Paleozoic

3.5.1. Descriptive Statistics

The Upper Paleozoic had the fewest respondents (12; 26.7%), who made 87 nominations of 43 distinct geoscientists. It also had the highest mean list length (7.3; SD = 2.8; minimum 4; maximum 10). The number of nominations by position was 12 up to the fourth and fell from the fifth. Twenty-six of the 43 nominees (60.5%) were nominated once, 8 (18.6%) twice and 9 (20.9%) three or more times. These 9 form the corpus and account for 45 of the 87 nominations (51.7%).

3.5.2. Ranking

Table 7 presents the ranking of the Upper Paleozoic.

António Ribeiro leads with 8 nominations (66.7%) and 6 first places, followed by José Tomás de Oliveira (7; 58.3%) and Rui Dias (6; 50.0%). Nery Delgado and Zélia Pereira have 5 each (41.7%), Carlos Ribeiro and Carlos Teixeira 4 each (33.3%), and Fernando Noronha and José Brandão Silva close the corpus with 3 each (25.0%).

3.5.3. Justifications for First Place

Five of António Ribeiro’s six first places came with a justification. They describe him as a “key figure in the tectonics and geodynamics of the Iberian Variscan Chain” with “decisive contributions to the interpretation of the Variscan Orogeny”, and one respondent added a more specific mention of “Upper Paleozoic stratigraphy”.

3.6. Mesozoic

3.6.1. Descriptive Statistics

The Mesozoic had 20 respondents (44.4%), who made 136 nominations of 40 distinct geoscientists (mean list length 6.8; SD = 2.7; minimum 2; maximum 10). The number of nominations by position was 19 to 20 up to the third and decreased from the fourth. Thirteen of the 40 nominees (32.5%) were nominated once, 9 (22.5%) twice and 18 (45.0%) three or more times. The Mesozoic has the largest corpus (18 geoscientists), which accounts for 105 of the 136 nominations (77.2%).

3.6.2. Ranking

Table 8 presents the ranking of the Mesozoic.

Paul Choffat is the most nominated geoscientist (11 nominations; 55.0%), and nine of those were first places, the largest concentration of first places on a single nominee in the survey. Ana Cristina Azerêdo and Miguel Magalhães Ramalho follow with 9 each, then Luís Vítor Duarte and Georges Zbyszewski with 8 each. José Carlos Kullberg, Rui Pena dos Reis and Rogério Bordalo da Rocha have 7 each, and ten further nominees have between 6 and 3 nominations. Among these are three researchers whom the authors identify as French (Jacques Rey, Christian Palain and Christiane Ruget-Perrot), discussed in Section 4.6.

3.6.3. Justifications for First Place

Seven of Paul Choffat’s nine first places (77.8%) came with a justification, a rate exceeded only in the Lower and Upper Paleozoic (83.3% each). The dominant theme is that of a founding figure. He is repeatedly described as the “father of Portuguese Mesozoic stratigraphy” (“father” appears in two independent responses), as the author of “monumental work of surgical precision in the biostratigraphy and cartography of the Jurassic and Cretaceous”, and as having “defined most of the chronostratigraphic units [...] that are still generically valid and recognizable in the field today”. One respondent also noted his “initial contributions to mineralogy”.

3.7. Cenozoic

3.7.1. Descriptive Statistics

The Cenozoic had 16 respondents (35.6%), who made 101 nominations of 43 distinct geoscientists (mean list length 6.3; SD = 2.6; minimum 3; maximum 10). The number of nominations by position was 16 up to the third and fell to 8 from the sixth. Twenty-two of the 43 nominees (51.2%) were nominated once, 9 (20.9%) twice and 12 (27.9%) three or more times. These 12 form the corpus and account for 61 of the 101 nominations (60.4%).

3.7.2. Ranking

Table 9 presents the ranking of the Cenozoic.

João Pais is the most nominated geoscientist (10 nominations; 62.5%) but was placed first only once. Pedro Proença Cunha follows (8; 50.0%) and stands out for his 5 second places. Miguel Telles Antunes has 7 (43.8%). António Galopim de Carvalho and Georges Zbyszewski have 6 each (37.5%), and Galopim de Carvalho was placed first 4 times in 6 nominations, the most first places of any Cenozoic nominee. João Cabral, Mário Cachão and Nuno Pimentel have 4 each, and Carlos Teixeira, Berkeley Cotter, Carlos Marques da Silva and Paulo Legoinha have 3 each.

3.7.3. Justifications for First Place

The geoscientist most often placed first in the Cenozoic is therefore not the most nominated one (Section 4.8). Three of Galopim de Carvalho’s four first places came with a justification. They refer to his “tectono-sedimentary study of the Cenozoic, particularly in the Tagus basin”, to the creation of the “school of Sedimentology in Portugal”, and to the “valorization of geological heritage [and] outreach of the geosciences”, a theme that does not appear in the justifications given for the leading figures of the other domains.

3.8. Meso-Cenozoic of the Atlantic Islands

3.8.1. Descriptive Statistics

The Atlantic Islands had 15 respondents (33.3%), who made 77 nominations of 34 distinct geoscientists, the smallest number of distinct nominees of any domain. The mean list length was 5.1 (SD = 2.0; minimum 3; maximum 10), the lowest of the seven domains. The number of nominations dropped sharply from the sixth position (4 at the sixth, against 15 in each of the first three). Twenty of the 34 nominees (58.8%) were nominated once, 4 (11.8%) twice and 10 (29.4%) three or more times. These 10 form the corpus and account for 49 of the 77 nominations (63.6%), the second-highest share after the Mesozoic.

3.8.2. Ranking

Table 10 presents the ranking of the Meso-Cenozoic of the Atlantic Islands.

José Madeira (8 nominations; 53.3%; 4 first places) and Victor Hugo Forjaz (6; 40.0%) lead, followed by five nominees with 5 each (33.3%): Georges Zbyszewski, João Carlos Nunes, António Brum da Silveira, António Serralheiro and João Luís Gaspar. Zbyszewski and Nunes were each placed first twice, whereas Brum da Silveira, Serralheiro and Gaspar were never placed first. João Mata (4), Ricardo Ramalho and Frederico Machado (3 each) complete the corpus. Machado’s three nominations include one first and two second places.

3.8.3. Justifications for First Place

Only one of José Madeira’s four first places came with a justification. The domain has few respondents, which limits the number of justifications, and the qualitative material for this domain is correspondingly thin.

3.9. Presence of Geoscientists Across Domains

Of the 53 individuals in the corpora, 40 appear in the corpus of one domain and 13 in the corpora of more than one (Table 11). Carlos Teixeira appears in six, Nery Delgado in five, and António Ribeiro, Rui Dias and Fernando Noronha in four each; the latter three appear in the same four domains (Geotectonic Framework, Precambrian, Lower Paleozoic and Upper Paleozoic). Georges Zbyszewski appears in three (Mesozoic, Cenozoic and Atlantic Islands) and is the only member of the Atlantic Islands corpus who also appears elsewhere.

3.10. Sensitivity of the Corpus to the Nomination Threshold

Table 12 shows how many geoscientists each domain would include at thresholds of two, three and five nominations, next to the corpus actually defined by the operational criterion. Raising the threshold to five nominations in every domain would reduce the number of domain-level memberships from 83 to 47, and lowering it to two would raise it to 156. Applying the three-nomination threshold in the Geotectonic Framework as well would increase its corpus from 12 to 22 (and the total to 93). The corpus is therefore sensitive to the criterion. The criterion sets the scale of the GeoFaces biographical programme, and the boundary it draws should not be read as a natural division between recognized and unrecognized geoscientists.

4. Discussion

4.1. What the Nominations Show and What They Do Not

The survey shows whom 39 university-based geoscientists named when they were asked, in each of seven domains, to identify the geoscientists most relevant to the development of geological knowledge. It does not measure scientific influence, and it does not establish a disciplinary canon. A nomination can reflect the nominee’s contribution, but also the nominee’s visibility, proximity to the respondent, presence in teaching or commemoration, the wording of the item and the composition of the panel. One respondent (ID100) put this bluntly, writing that the results “ended up saying more about who answered than about the reality they seek to analyze” (Section 4.9). We therefore read the results as evidence of recognition among a defined panel. Where we go on to ask what this recognition may tell us about disciplinary memory, we present the interpretation as a hypothesis and say what further evidence would test it.

4.2. Concentration, Long Tails and Recognition Profiles

In every domain a small group of names attracts a large share of the nominations (48.6% to 77.2% for the corpus), while between 32.5% and 63.5% of nominees are named once. Skewed distributions of this kind are a familiar feature of scientific productivity and communication [34]. The long tail can be read in two ways. It may reflect the diversity of respondents’ specific knowledge, with specialists naming figures known in their own sub-networks, or it may reflect idiosyncratic choices. The survey cannot distinguish the two. The corpus should not be read as a list of the “best” or “most important” geoscientists. It contains the geoscientists who were named by enough respondents to meet the project’s operational threshold.

António Ribeiro is the most nominated geoscientist in the Geotectonic Framework, the Precambrian and the Upper Paleozoic, and shares first place with Nery Delgado in the Lower Paleozoic. His centrality can be described, heuristically, as that of an obligatory passage point [35] for those who work on the Iberian Variscan belt, in the sense that respondents from several specializations name him. The survey records how often he was named, not why, and the justifications given for first place refer to his work on Variscan tectonics and geodynamics (Section 3.2.3 to 3.5.3). Paul Choffat in the Mesozoic and João Pais in the Cenozoic lead their domains with different profiles: nine of Choffat’s eleven nominations were first places, against one of Pais’s ten. Several explanations are possible. The two domains differ in their number of respondents (20 and 16) and in the specializations of those respondents. The contrast may also partly reflect the difference between a figure of a past generation (Choffat, 1849-1919) and a figure who is a contemporary of many respondents and may be evaluated more variably by peers. The survey does not allow a choice among these explanations. Position also depends on respondents having followed the instruction to rank, and at least one respondent indicated that the list was unordered (Section 6).

4.3. The Persistence of the Founding Generation: The Case of Nery Delgado

The persistence of Nery Delgado across five of the seven domains is the clearest historical pattern in the data. He is nominated by 7 of 32 respondents in the Geotectonic Framework (21.9%), 4 of 15 in the Precambrian (26.7%), 8 of 16 in the Lower Paleozoic (50.0%), 5 of 12 in the Upper Paleozoic (41.7%) and 3 of 20 in the Mesozoic (15.0%), and he does not appear among the listed nominees of the Cenozoic or the Atlantic Islands. His recognition is strongest in the domain where his own work was concentrated: he is the joint most nominated geoscientist in the Lower Paleozoic and was placed first there more often than anyone else (6 times, against 3 for António Ribeiro). Together, Carlos Ribeiro (1813-1882), Nery Delgado (1835-1908) and Paul Choffat (1849-1919) received 75 of the 818 nominations (9.2%), and they hold 11 of the 83 domain-level memberships in the corpora.

The historiography helps to read this pattern. Delgado worked from 1857 as assistant in the Second Geological Commission of Portugal and directed the survey from the death of Carlos Ribeiro in 1882 until his own death in 1908 [36]. With Carlos Ribeiro he co-authored the first and second editions (1876 and 1878) of the Geological Map of Portugal at 1:500 000, and with Paul Choffat its third edition (1899), and his work on the stratigraphy and paleontology of the Lower Paleozoic is a recognized part of his output [37]. His monograph on the bilobites and other fossils of the quartzites at the base of the Portuguese Silurian is an example [38]. The justifications given by respondents for placing him first in the Lower Paleozoic (Section 3.4.3) refer to this kind of work: monographs on the Silurian and the Cambrian, fossils, mapping. Respondents’ reasons and the historiographic account of his work therefore point in the same direction. This is consistent with recognition anchored in identifiable bodies of work, but it does not show that the historiography or the work itself caused the nominations.

The survey shows that Delgado’s name is available to a substantial share of present-day specialists. It does not show how it was transmitted. Several channels are plausible, and each could be examined with other evidence. The first is institutional commemoration, of which the centenary exhibition and volume of 2008 at the Museu Geológico are an example [36,37]. The second is the continued use of his monographs and stratigraphic terms in the specialist literature, which we did not analyse. The third is teaching, including the place of these figures in university curricula and textbooks [7], which the survey did not ask about. The fourth is the wording of the item, which asks for relevance to “the development of geological knowledge” in a domain and may itself invite the nomination of founders. These channels are not mutually exclusive and we do not rank them. What the data do show is that recognition of the founding generation in this panel differs between domains rather than being uniform, and that the same is true of Paul Choffat, whose eleven Mesozoic nominations contrast with three in the Lower Paleozoic and six in the Geotectonic Framework. Choffat’s life and work are the subject of a biographical volume [12], and his collaboration with Delgado on the 1899 map edition is documented [37]. Whether teaching, commemoration or citation explain the pattern is a question for further historical work and for the companion articles.

4.4. Geoscientists Present in Several Domains

Table 11 shows that a minority of the corpus (13 of 53 individuals) appears in more than one domain. Presence in several corpora may reflect the breadth of a career, the breadth of respondents’ own specializations, or the fact that domains defined by geological time or by framework overlap in the work of some geoscientists. The last point was raised by a respondent (ID100) who observed that the chronostratigraphic division may disadvantage specialists whose activity spans several terrains. That António Ribeiro, Rui Dias and Fernando Noronha appear in the same four older domains may mean that respondents who nominated in one of these domains often named the same people in the others. Respondent-level data, analysed in the companion article, can examine this directly.

4.5. Women in the Corpora

Three women are included in the corpora, out of 53 individuals: Ana Cristina Azerêdo (9 nominations in the Mesozoic), Zélia Pereira (5 in the Upper Paleozoic) and Christiane Ruget-Perrot (3 in the Mesozoic). This is 5.7% of the corpus, or about one woman for every seventeen men. The figure is consistent with a field in which women’s access to scientific institutions and research careers was historically limited, but that reading needs historical evidence on women’s participation in Portuguese geosciences, which this study does not provide. The design also cannot separate any such inheritance from present-day patterns of recognition, for example differences in visibility, in the domains or topics that respondents associate with women, or in how nominations are made. The panel was not predominantly male (40.0% of respondents were women), so its composition by sex does not obviously explain the result. Whether nominations differ by respondent sex is a question for the companion article on respondent profiles. One respondent (ID58) observed that gender inequality in professional recognition was neglected in the survey. The GeoFaces biographical materials should state that the corpus reflects present-day nominations and is not a complete record of women’s contributions.

4.6. Foreign Researchers in the Corpora

Foreign-born researchers appear in several corpora. Paul Choffat, who was Swiss-born, appears in three (Geotectonic Framework, Lower Paleozoic and Mesozoic). The Spanish researchers José Ramón Martínez Catalán and Manuel Julivert appear in the Geotectonic Framework (8 and 5 nominations), and the Mesozoic corpus includes three researchers whom the authors identify as French, Jacques Rey, Christian Palain and Christiane Ruget-Perrot (4, 4 and 3 nominations). Their nomination, in a free-response format with no list of names, shows that they remain visible to respondents. It does not establish why respondents recognized them. Contributions to knowledge of Portuguese geology, collaboration with Portuguese colleagues, teaching, and personal acquaintance are all possibilities, and only the justifications for Choffat’s first places (Section 3.6.3) address the reasons directly.

The pattern is compatible with accounts that stress the circulation of knowledge across borders [8] and with the view of the scientific field as a space in which peers confer recognition on one another [39], but the survey neither tests nor requires either. A methodological point follows. Rankings of the geoscientists “most relevant” to Portuguese geology can be read as rankings of Portuguese scientists or as rankings of contributors to knowledge of Portuguese geology, and the item as worded leaves this open. The presence of foreign researchers shows that respondents did not restrict their nominations by nationality. It should not be read as a map of historical collaboration.

4.7. The Atlantic Islands

The Atlantic Islands domain has the lowest mean list length (5.1), the fewest distinct nominees (34) and a high concentration of nominations in the corpus (63.6%), lower only than in the Mesozoic (77.2%). These features are compatible with a small sub-community whose members name a limited set of figures, but they may equally reflect the small number of respondents (15). The corpus combines one geoscientist who is also recognized in the mainland domains (Zbyszewski) with nominees who appear only here. Whether generational differences shape the nominations in this domain can be examined with respondent-level data in the companion article.

4.8. First-Place Justifications and Total Nominations

Justifications were given for 32 of the 49 first places received by the most frequently first-placed nominee of each domain (Geotectonic Framework 9 of 16, Precambrian 2 of 4, Lower Paleozoic 5 of 6, Upper Paleozoic 5 of 6, Mesozoic 7 of 9, Cenozoic 3 of 4, Atlantic Islands 1 of 4). In the excerpts quoted in Section 3, the vocabulary is largely one of foundation and reference (“father of Portuguese Mesozoic stratigraphy”, “established the stratigraphic foundations”, “the greatest historical reference”, “key figure”) rather than one of productivity or bibliometric impact. Because the excerpts come from a small number of respondents and were interpreted by a single analyst, they illustrate the reasons offered and do not amount to a systematic coding. Justification rates ranged from 25% (Atlantic Islands) to 83.3% (Lower and Upper Paleozoic), and with such small numbers we draw no conclusion about how they relate to the concentration of nominations.

The Cenozoic shows that the two indicators can identify different nominees. João Pais is the most nominated geoscientist (10 nominations) but António Galopim de Carvalho received the most first places (4). The justifications for Galopim de Carvalho mention Tagus-basin sedimentology, the creation of a school of sedimentology in Portugal, and the valorization of geological heritage and outreach. Total nominations capture the breadth of recognition across respondents, whereas first places capture the name that comes first to mind for a subset of them. The two indicators describe partly distinct aspects of recognition, and neither should be equated with importance.

4.9. Respondents’ Comments on the Limits of the Method

Ten of the 45 respondents (22.2%) left comments in the final question of the questionnaire. Several show awareness that nominations depend on the respondent’s position. Respondent ID106 wrote that “each of our views on each geoscientist is, first and foremost, limited by our professional context and by physical proximity”, and that “even regarding visibility, my opinion is partial, as it is limited to the opinions provided to me only by the people I know directly”. Respondent ID108 abstained from some questions “for considering that I lack the knowledge indispensable to a fair assessment with a high degree of confidence”, and respondent ID15 noted that a long training period abroad tends to dilute the national references available for nomination.

Other comments concern the design. Respondent ID100 asked “Importance in what sense? There are many types of importance...”, argued that the chronostratigraphic division introduced a structural bias by excluding researchers whose activity spans several terrains (for example, igneous petrologists not associated with a particular era), and concluded that “the results also ended up saying more about who answered than about the reality they seek to analyze”. Respondent ID133 noted a possible conflation between “the great figures of scientists who dedicated themselves to Geology” and “the geologists who had a great influence on our careers”, which implies that some nominations may reflect personal formative influence rather than historical significance. Other respondents suggested a separate treatment of the Quaternary (ID19) and the inclusion of practical applicability as a criterion of value (ID144). Respondent ID58 observed that gender inequality in professional recognition was “neglected in the present survey”.

These comments come from within the panel and support the limitations discussed in Section 6: the possible overrepresentation of close academic and institutional networks, the ambiguity of the criterion of relevance, and the sensitivity of the rankings to the domain structure.

5. Conclusions

This article reports a purposive, single-round nomination survey of university-based geoscientists in Portugal and describes whom the 39 respondents who made nominations named as most relevant to the development of geological knowledge in seven chronological and thematic domains. The conclusions apply to this panel.

1. Nominations are concentrated on a small number of names in every domain. António Ribeiro is the most nominated geoscientist in four domains (tied with Nery Delgado in the Lower Paleozoic), Paul Choffat in the Mesozoic, João Pais in the Cenozoic and José Madeira in the Atlantic Islands. Between 32.5% and 63.5% of the nominees in each domain were named only once.

2. Recognition of the founding generation persists but differs between domains. Nery Delgado is nominated in five domains and most strongly in the Lower Paleozoic, where the historiography places a central part of his work. Carlos Ribeiro, Delgado and Paul Choffat together received 9.2% of all nominations. The survey does not show how this recognition was transmitted, and Section 4.3 identifies the evidence that would be needed to examine it.

3. The operational criterion of GeoFaces defines a biographical corpus of 53 individuals (83 domain-level memberships). The size of the corpus depends on the criterion, from 47 to 156 memberships at thresholds of five and two nominations. The corpus is a working basis for biographical work, not a statistical determination of importance.

4. Three of the 53 individuals in the corpus are women, and foreign researchers appear in several corpora. The design shows present-day visibility. It does not show the historical causes of these patterns or the reasons why respondents nominated particular people.

5. Adding the partial responses removes no one from the corpora and leaves every domain’s leading nominee in first place. This indicates that the top of the rankings is stable, not that the panel is representative of the wider community.

The nominations provide a documented baseline of present-day recognition against which the historiography of Portuguese geology can be compared, and the survey can be repeated with other panels, including geoscientists employed outside public universities. Two further steps are already planned: the analysis of associations between respondents’ profiles and their nominations, and an interpretation within the sociology of disciplinary memory and the history of Portuguese geology. Evidence on transmission, such as curricula, commemorations and citation of the nominees’ work, would strengthen the historical reading offered here.

6. Limitations

Coverage and panel composition. The invitation pool consisted of geoscientists at public universities, and 45 of the 230 invitees (19.6%) completed the questionnaire. Response rates by institution ranged from 0% to 39.1%. The University of Lisbon had the largest number of invitees (77) but only six respondents (7.8%), whereas respondents affiliated with institutions in the Lisbon region represent 33.3% of the panel (n = 15), followed by Évora (17.8%), the North (17.8%) and Coimbra (15.6%). Respondents’ choices may therefore reflect institutional proximity to particular research traditions, a question that the companion article will examine by relating each respondent’s institution to the institutions associated with the figures nominated. The panel is also senior (mean research experience 29.2 years; SD = 11.9), so the rankings may reflect the memory of this generation more than that of early-career researchers, who may have had less exposure to the founding figures or may recognize different contemporary references. The present sample cannot test this. Researchers in applied or industrial contexts may be underrepresented.

Non-response. The complete respondents (n = 45) and the partial respondents (n = 84, with profile data for 56 to 66 of them depending on the item) do not differ substantially in sex (58% male among complete respondents; 64% among partial) or nationality (both over 95% Portuguese). Partial respondents have slightly fewer years of research experience (26.1 against 29.2 years) and a somewhat lower doctoral rate (83% against 93%), which suggests that abandonment is associated with less academic experience. General discussions of non-response in web surveys point to similar patterns [40].

Sensitivity to partial responses. The analysis was repeated with the 12 valid partial responses added to the 45 complete ones (the 13th response with a nomination was not interpretable). The leading nominee of every domain remains in first place. In the Lower Paleozoic, the tie between Nery Delgado and António Ribeiro is resolved in favour of Ribeiro (10 against 9 nominations), and in the Mesozoic José Carlos Kullberg rises to second place (11 nominations). The number of respondents per domain increases by between 27% and 42%, and no geoscientist of the original corpora is displaced. Some additional individuals would enter the corpora (between 1 and 8 per domain, with 8 in the Geotectonic Framework, where several nominees stood one or two nominations below the threshold), which illustrates the sensitivity of the corpus to the criterion (Section 3.10).

Item and domain structure. The criterion of relevance to the development of geological knowledge was not defined for respondents, and some understood it in different senses (Section 4.9). The seven-domain structure can exclude geoscientists whose work spans several domains. Both features affect what the nominations measure.

Measurement. Name normalization was done manually and validated by the principal investigator. Counting nominations without weighting for position is a simplification, because the same frequency with different position profiles may reflect different kinds of recognition (compare Choffat in the Mesozoic with Pais in the Cenozoic). Not all respondents followed the instruction to rank in decreasing order of importance, and at least one explicitly indicated an unordered list, which introduces local uncertainty in the assignment of first place in that case.

Qualitative material. The analysis of first-place justifications is interpretive and illustrative. It was carried out by a single analyst without a second coder or a measure of inter-rater agreement. Not all respondents who placed a geoscientist first gave a justification (rates ranged from 25% in the Atlantic Islands to 83.3% in the Lower and Upper Paleozoic), so the excerpts should be read as illustrations and not as a representative sample of respondents’ motives.

Authors and self-nomination. Self-nomination could not be prevented or detected, and the relation between the authors and the invitation pool, the panel and the nominees is stated in Section 2.2.

Inference. Because the panel is purposive and the response partial, statistical representativeness relative to the pool or to the wider community cannot be assumed. This applies with particular force to the domains with fewer respondents, such as the Upper Paleozoic (n = 12), where each nomination corresponds to 8.3 percentage points of respondents.