{"id":"0f701643-d024-42d6-9e2b-2ab47fca3d0f","arxiv_id":"2510.09069","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Nobel Physics discoveries from 1972 to 2024 fit Kuhn's two discovery types, supporting the view that a correct understanding is part of what makes a discovery.","lead":"A philosopher tested whether Nobel Prizes in Physics from 1972 to 2024 back Thomas Kuhn's old claim that discovering a new thing in the world requires both seeing it and correctly understanding what it is. All 33 prize-worthy discoveries of new natural kinds fit Kuhn's two categories, and the author argues that Nobel credit practices under-recognize theorists.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Discovery-type coding is underdetermined: the 19/14 split and the surprise percentages depend on flexible rules about 'guidance', publication priority, and Nobel Committee emphasis.","rationale":"The reader's weakest assumption identifies the univocality problem: discovery type seems to depend on publication timing and Committee emphasis. I agree that this is the central soft spot, but I would sharpen it: the paper's operationalization is not merely 'first publication wins'; it also includes a 'theory did not guide the experiments' exception and an appeal to the Committee's award emphasis. These additional criteria make the classification underdetermined in a way that a simple publication-date rule would not. This is load-bearing because the headline empirical results—the 19/14 tally and the nearly perfect surprise profile (Table 2)—are computed from this classification, and the paper's own inter-coder reliability applies only to the descriptor/revolutionary variables, not to the primary type coding. The concern is not that the author is biased; it is that the decision rules are flexible enough that alternative reasonable codings could change the tallies and weaken the confirmation of H2/H3. I therefore recommend no change to the reader's CONDITIONAL verdict: the paper is acceptable as a proposal, but its central empirical claim stands or falls on the available, concrete re-coding check. The paper does get credit for openly disclosing borderline cases (NP 1985, NP 1987), separating descriptive from normative claims, and providing an OSF appendix; these are real strengths. But the primary classification lacks the same verification standard as the descriptor layer.","tokens_in":14704,"tokens_out":6636,"duration_ms":57330,"concrete_test":"Have three independent coders, blind to the author's classifications, re-code all 33 discoveries using a pre-registered codebook that specifies decision rules for: (a) whether a prior theory counts as 'guiding' the observation; (b) how to handle multiple discoverers with different uses of theory (e.g., positron); and (c) whether NP 1975 should be what-that or that-what when the observation preceded the theory but the award emphasized the theory/confirmation. Report Kappa for the primary type variable, not just the descriptor layer. Then recompute Table 2 (surprise/expected percentages) under the alternative codings. If the that-what/what-that split shifts by more than two cases, or if the surprise percentages cease to separate cleanly (e.g., the what-that surprise rate rises above ~20%), then Hypotheses 2 and 3 are not robust to classification choice. A minimal version: re-code just NP 1","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central empirical claim—that all 33 Nobel natural-kind discoveries sort cleanly into that-what and what-that classes with the predicted surprise profiles—depends on a coding rule that is underdetermined and partly circular. In §3 the author introduces an exception: a theory that predates the observation does not make a discovery what-that if the theory 'so happened not to guide the experiments' (quarks, CMB, 51 Peg). This requires counterfactual judgments about whether a theory guided a community, not a simple date. In the same section, NP 1975 is classified as what-that explicitly because 'the Nobel Committee opted not to give the award to the observations prompting Rainwater's theoretical work' and the prize went to Bohr/Mottelson; the discovery type is thus made to follow the Committee's award emphasis, not any property of the discovery itself. The positron discussion (§2.2) concedes that the same event can be both types depending on who publishes first. With decision rules this flexible, the 19/14 split is not a stable fact about discoveries; it is a function of publication timing, authorial judgments about 'guidance', and Committee emphasis. Since H3's surprise percentages are computed on this split, a re-coding of even a few borderline cases (CMB, quarks, 51 Peg, NP 1975) could change the '100% vs 93%' result. The reported Cohen's Kappa was run only on the descriptor layer ('surprising'/'revolutionary'), not on the primary type classification, and the full codebook is not included in the manuscript (only an OSF link is given). Thus the key classification is not independently verifiable from the paper alone.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper offers an empirical test of a stripped-down Kuhnian account of natural-kind discovery, using Nobel Prizes in Physics from 1972 to 2024. The author classifies 33 prize motivations as either “that-what” discoveries (observation first, conceptualization later) or “what-that” discoveries (predicted by prior theory), and tests five hypotheses: (H1) discovery requires an at least partially correct conceptualization; (H2) all discoveries fall into the two types; (H3) that-what discoveries are described as surprising and what-that as expected; (H4) that-what discoveries are revolutionary; and (H5) that-what discoveries involve greater epistemic uncertainty. The paper reports that H1 is confirmed except for one case (NP 1987), H2 is fully confirmed with 19 that-what and 14 what-that discoveries, H3 is nearly perfectly confirmed (100% vs. 93%), H4 is ambiguous, and H5 is confirmed via longer Nobel award lags for that-what discoveries. The author draws normative conclusions about the Nobel Committee's under-crediting of theoretical contributions.","tokens_in":14943,"tokens_out":6160,"duration_ms":57205,"significance":"If the empirical claims are sound, the paper would provide the first systematic, large-scale evidence for a Kuhnian distinction between two types of natural-kind discovery, and it would ground a practical normative critique of Nobel Prize credit allocation. The paper has several strengths: it works with a well-defined corpus, it makes its full classification available via OSF, it explicitly acknowledges anomalies (NP 1985 quantized Hall effect; NP 1987 high-temperature superconductors), and it reports intercoder agreement for the descriptor coding. However, the central classification and the test of H1 are vulnerable to coding underdetermination and sample selection; the current evidence is therefore suggestive rather than conclusive. The paper would be a valuable contribution if these concerns are addressed with robustness checks and independent coding of the primary type classification.","major_comments":[{"comment":"The primary classification into that-what vs. what-that discoveries is not independently validated and is governed by flexible rules. The paper reports Cohen's Kappa for the descriptor coding ('surprising'/'expected'/'revolutionary') in §4.2, but not for the type classification itself. In §3, a theory that predates the observation is declared not to make a discovery what-that if it 'so happened not to guide the experiments' (quarks, CMB, 51 Peg), and NP 1975 is classified as what-that explicitly because the Nobel Committee opted not to award the observations that prompted Rainwater's work. These rules make the type assignment depend on counterfactual judgments about guidance and on award emphasis, not on a stable property of the discovery. Since the 19/14 split underlies the '100% vs. 93%' result in Table 2, re-coding even a few borderline cases (CMB, quarks, 51 Peg, NP 1975) could chang","section":"§3, §4.1 (Hypothesis 2)"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short version: Schindler has done something nobody else has done - systematically coded 53 years of Nobel Physics citations as instances of Kuhn's two discovery types and tested five hypotheses derived from his own earlier account. The empirical core is the 19 vs 14 split, the 100%/93% surprise profiles, and the 20-vs-7-year award lag. Those numbers are internally consistent, and the paper is honest about its anomalies (NP 1985 quantized Hall, NP 1987 high-Tc). It also openly reports that Hypothesis 4 only partially supports Kuhn and that the first operationalization of Hypothesis 5 came out backwards before being reframed. That transparency is real credit.\n\nWhat's genuinely new: nobody has asked whether Nobel practice fits the that-what/what-that distinction. The paper also separates the descriptive claims from the normative credit-attribution argument, which is useful.\n\nThe soft spots are concentrated in the classification. The primary coding - which discovery is that-what vs what-that - is done by the author, using categories he introduced, and the reported kappa covers only the descriptor layer, not the type classification. The codebook is not in the manuscript; there's an OSF link. Some decision rules are flexible: quarks and the CMB are classified as that-what because the pre-existing theory 'so happened not to guide the experiments,' which requires a counterfactual community judgment, and NP 1975 is classified as what-that explicitly because the Nobel Committee chose to reward the theory rather than the motivating observations. That last one conflicts with the paper's own claim that award emphasis and discovery facts are orthogonal. The positron discussion concedes type can depend on who published first. So the 19/14 split is less stable than the prose suggests. A re-coding of a handful of borderline cases could shift the surprise percentages, though I wouldn't expect it to flip the general pattern.\n\nAlso, Hypothesis 1 is close to trivially true given the corpus: Nobel committees rationally avoid rewarding unresolved identifications. The paper acknowledges this, but it means that the main support for Kuhn's correctness requirement is weaker than 'fully confirmed' implies. The normative conclusion about under-crediting theory is plausible but rests on the same coding.\n\nWho should read this: philosophers of science interested in discovery, and anyone studying Nobel prize patterns. It deserves a serious referee - the empirics are concrete and checkable. But the revision needs to publish the full codebook, report reliability for the primary type coding, and either defend or change the NP 1975 rule. If those are addressed, it's a worthwhile contribution.","headline":"A serious first attempt to test Kuhn's two-type discovery taxonomy against Nobel Physics, with honest reporting but a classification scheme that is less robust than its 'fully confirmed' language suggests.","tokens_in":15619,"tokens_out":2444,"would_cite":true,"duration_ms":22345,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"All 33 natural-kind discoveries recognized by Nobel Prizes in Physics from 1972 to 2024 fall into two classes — observations that are later conceptualized and predictions that are later observed — and all but one received a prize only after","keywords":["scientific discovery","natural kinds","that-what discovery","what-that discovery","Nobel Prize in Physics","credit attribution","theory and experiment","Kuhn"],"falsifier":"A single Nobel Prize in Physics between 1972 and 2024 awarded for a natural-kind discovery that, by the paper's own criteria, lacked any at least partially correct conceptualization before the award (other than the cited high-temperature superconductor case), or a natural-kind discovery that fits neither the that-what nor the what-that pattern, would overturn the central claim.","tokens_in":14424,"feed_emoji":"🏆","tokens_out":5937,"duration_ms":50498,"temperature":0.7,"pith_summary":"This paper asks whether an old philosophical distinction between two kinds of scientific discovery holds up against the most celebrated discoveries in physics. Analyzing the 33 Nobel Prizes in Physics from 1972 to 2024 that recognized new natural kinds or their properties, the author finds that all of them fall into either 'that-what' discoveries — a surprising phenomenon is observed first and only later correctly conceptualized — or 'what-that' discoveries — a theoretically predicted phenomenon is then observed. The author also finds that, with one exception, no prize was awarded before an at least partially correct conceptualization of the discovered phenomenon existed. On this basis, the paper draws a normative conclusion: the Nobel Committee's strong preference for honoring experimental over theoretical contributions systematically under-credits the conceptual half of discovery.","feed_headline":"Two discovery types capture every Nobel-winning physics find","feed_subtitle":"A 53-year review of physics Nobels shows discovery needs both observation and a correct concept—and prizes often honor only one side.","key_machinery":"The central machinery is Kuhn's distinction between 'that-what' and 'what-that' discoveries, paired with the premise that a natural-kind discovery requires a correct but not necessarily complete conceptualization of the observed phenomenon. The analysis classifies each Nobel Prize by recording when the observation ('that') and the correct concept ('what') first appeared, then cross-checks the classification against the Nobel Foundation's own descriptive texts. This machinery also produces two derived measures: the time lag between observation and conceptualization, and the time from completion of the discovery to the award, which together are used to test claims about epistemic uncertainty.","core_discovery":"The paper claims that Kuhn's two-type taxonomy of natural-kind discovery, when stripped of its surrounding revolutionary framework, accurately describes every natural-kind discovery that earned a Nobel Prize in Physics over the past 53 years. A 'that-what' discovery starts with an unexpected observation that requires a new or revised concept; a 'what-that' discovery starts with a theoretical prediction that is later confirmed by observation. Of the 33 qualifying discoveries, 19 were that-what and 14 were what-that. The paper further claims that in all but one case the prize was not awarded before an at least partially correct conceptualization of the discovered phenomenon existed, and that N","pith_inferences":["If the type of a discovery is fixed by who published first rather than by the discovery's intrinsic epistemic structure, then the same event — as with the positron — can be a 'that-what' discovery for one team and a 'what-that' discovery for another; in that case the taxonomy describes credit conventions and publication races more than it describes a stable feature of discovery.","The author's own numbers show that half of what-that discoveries were also described as revolutionary, which is a direct challenge to the clean mapping between discovery type and revolutionary status that originally motivated the distinction.","A natural next test is to apply the same classification to Nobel Prizes in Chemistry or Physiology/Medicine; if the two-type taxonomy holds there, it would strengthen the claim that Kuhn's distinction captures something general about natural-kind discovery rather than a quirk of physics prize practices.","The normative argument could be operationalized as a practical reform: prize committees could be required to name both the 'first observer' and the 'first correct conceptualizer' of a discovery, ensuring that both the experimental and theoretical halves receive credit."],"forward_implications":["If the account is correct, the 19/14 split shows that the most prestigious physics prizes recognize both revolutionary surprises and predicted confirmations in comparable numbers, undermining the idea that Nobel-level discovery is mostly about unexpected breakthroughs.","The finding that that-what discoveries wait a median of 20 years for the prize while what-that discoveries wait 7 supports the claim that epistemic uncertainty about a new phenomenon delays its community acceptance and recognition.","The normative conclusion that a discovery consists of both observation and conceptualization implies that awards honoring only experimentalists or only theoreticians leave half of the discovery uncredited, and this under-crediting appears in both discovery types.","The sole counterexample, the 1987 high-temperature superconductor prize, suggests the correctness requirement can be relaxed when the new phenomenon is an extension of a well-understood kind, narrowing the scope of the claim but not overturning it."],"fun_headline_variants":["Kuhn's two discovery types fit every physics Nobel in 53 years","All physics Nobels since 1971 match Kuhn's two-type discovery split","Two discovery types, one rule: they explain all 33 physics Nobel wins","Physics Nobels: 19 found by surprise, 14 predicted, all fit Kuhn","Kuhn's two-type taxonomy perfectly maps 53 years of physics Nobels"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"That each discovery has a single, stable classification determined by whether the correct concept or the first observation appeared first in print, independent of who later received the prize or of how the Nobel Committee framed the award.","fun_headline_variants_meta":{"raw":{"variants":["Kuhn's two discovery types fit every physics Nobel in 53 years","All physics Nobels since 1971 match Kuhn's two-type discovery split","Two discovery types, one rule: they explain all 33 physics Nobel wins","Physics Nobels: 19 found by surprise, 14 predicted, all fit Kuhn","Kuhn's two-type taxonomy perfectly maps 53 years of physics Nobels"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000211,"raw_usage":{"total_tokens":1180,"prompt_tokens":603,"completion_tokens":577,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":347,"completion_tokens_details":{"reasoning_tokens":473}},"tokens_in":347,"tokens_out":577,"duration_ms":5410,"temperature":1.0,"reasoning_tokens":473,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T10:39:55.991113+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single Nobel Prize in Physics between 1972 and 2024 awarded for a natural-kind discovery that, by the paper's own criteria, lacked any at least partially correct conceptualization before the award (other than the cited high-temperature superconductor case), or a natural-kind discovery that fits neither the that-what nor the what-that pattern, would overturn the central claim.","supporting_citations":[],"review_version":1}