I thank the authors for the reply to my comments, for adding an abstract and for clarifying the scope of the paper. I am only adding two additional comments based on the changes you have made and your reply.

Thank you for the added section on monothetic vs polythetic groups, which I think is very useful. I do not have access to the figures you mention - which unless I am mistaken are not included in the zenodo file.
You seem to advocate a switch from polythetic to monothetic groups, which is a discussion already very well present in recent archaeological literature (without being explicitly addressed in these terms, which is why this section of the paper is particularly useful).  However, I believe it would be useful as well to include a discussion of why polythetic groups were adopted in the first place (why were they considered more "natural" p. 28) and of the impact of the use of polythetic vs monothetic groups on the type of questions asked by archaeologists? Both aspects (why polythetic/monothetic groups are used in the first place and the consequences of their uses for comparative analyses) are intermingled and could perhaps be treated separately. Perhaps using the information on the section on "Metaphysics of Classification in Biology" here could also clarify these points. 

Your reply stating that "We maintain that the search for conventional types is a dead
end." Or, in the text, p. 20 "Conventional kinds are, of course, extremely important for the cultural systems who implemented them, but their importance dies with the people who created, used, and believed in them." These strong statements seem to be directly related to the emic vs etic debate, as well as to the theoretical approches used. The article is already quite long as is but I believe it would have benefited from more explanations regarding this.

I’ve read the revised version and, in my opinion, the authors have responded adequately and appropriately to all my comments in version #1. I, therefore, recommend publication.

Overall, I found this a really interesting and compelling paper, and I recommend its publication.

Although I do not necessarily agree with all the points made in this paper, I still find that you've made a compelling argument, and I see the value in considering both the historical/metaphysical roots of the classification/categorization problems and the parallels between similar issues in Biology and Archaeology, something that I would agree most archaeologists who have previously dealt with these issues have largely overlooked.

Almost all my more detailed comments, in fact, are purely editorial in nature. I have made detailed editorial suggestions, mostly for clarity, grammar and style. I would also suggest making changes to the abstract (abstracts should always summarize the paper, including its conclusions, not just introduce the topic), and you might consider a certain amount of reorganization and shortening of the manuscript. As I am quite interested in this topic, I read it eagerly, but some readers might be put off by the fact that the manuscript is rather lengthy and arguably the narrative wanders a bit, especially in the second half, despite the excellent use of subheadings to clarify the structure.

My much pickier and minor editorial comments are marked up in tracking on the attached version of the manuscript.

I paginated the pdf to make it easier to comment. I highly recommend doing this, and numbering the lines in the whole document. I should mention that I am a paleoanthropologist and an archaeologist, not a philosopher or an historian of science.

___________________________________________________________________

The manuscript consists of 3 parts: (1) a discussion of metaphysics and classification in philosophy of science (pp. 1-7), (2) how these terms and concepts are used in biology (pp. 8-13), and (3) how (2) can be applied to archaeology (pp. 14-25).

(1) is quite sophisticated but replete with terms and concepts unfamiliar to most archaeologists or defined differently from archaeological usage. The authors note that epistemology is largely ignored by archaeologists, which is true (and also true of biologists). The text reads much longer and more complex than it has to be, assuming that the target audiences are biologists and archaeologists. It could benefit from stringent copy editing. It needs more headings to break it up into manageable chunks, and some figures summarizing the main points made in the text. It appears to have been written for philosophers or historians of science. It is the strongest part of the essay.

(2) is more comprehensible to someone like me, an archaeologist with a background in human paleontology. This means the logic of inference is more readily understood, as are the concepts and terms (although “The species-as-sets approach can be tentatively compared to species as mereological sums – the grouping of objects under the parthood relation ­– which are also defined by their extensions, in the same way that sets are” (pg. 10) must be amongst the most jargon-ridden sentences in the English language!) I’ve read some of the authors cited (e.g., Richards, Ghiselin, Boyd, Gould, Ereshefsky) because I taught a graduate seminar on evolutionary epistemology. Based on that, (2) basically made sense. It however relies rather too heavily on Richards and ignores Hahlweg, Hooker, Casti, Hull, etc.

(3) is the weakest of the three sections. It offers only a very dated and selective construal of anglophone archaeology, and fails to even mention the francophone research tradition that dominated Stone Age archaeology world-wide for more than a century. It was only in the 1970s that anglophone archaeology shed culture history, now seen as strictly empiricist (. . “the facts speak for themselves”), limited in its ability to explain anything of interest and shot through with all kinds of simplistic, essentialist, variety-minimizing assumptions about ‘culture’, the nature of change, and hunter-gatherer adaptations. This shift in the conceptual foundation of the discipline took place when I was in graduate school. It was wrought by Binford, Flannery and their intellectual progeny (Longacre, Hill, Dibble, Schiffer, Freeman, etc.). Dunnell was important, as they note, but his impact on the field was limited by his book, Systematics in Prehistory (1971), quite possibly the most boring text ever written. After that he was largely ignored. This part of the manuscript should be thoroughly updated and revised. There are few citations and most of them date in the 1940s-1960s. Anglophone archaeology has changed radically in the past 20 years, although little of this is evident in this part of the manuscript. Archaeology has no universal metalanguage (i.e., mathematics) like physical science does, nor is it very introspective. Consequently, confusion in concepts and terms is inevitable both within, and between, intellectual traditions.

In sum, the manuscript does not cohere very well. (1) is complicated and abstruse to biologists and archaeologists, but probably relatively clear to historians of science. (2) looks pretty good to me because I have some control over systematics in biology. An archaeologist without a background in neo-Darwinian evolution would have trouble with it.  (3) is a dated, selective view of archaeology that bears little resemblance to modern practice in the anglophone research tradition. Francophone researchers have also begun to contemplate the logic of inference underlying the culture historical approach. These changes directly affect how we explain pattern in the past.

***

Here are some in-text notes that I made while reading the manuscript (only up to page 5), for your possible interest:

pg. 2 – there are 7-8 definitions of ‘species’ in human paleontology [HP] (see Kimbel & Martin, 2013), point being that ‘species’ depends on how they are defined and whether or not there is a consensus in the discipline.

pg. 2 – essentialism no longer a part of HP research, although it figured prominently in the 19th and early 20th centuries. However, cladistic methodologies, common in HP after the early 1990s, do tend to minimize variation. This is because of the conviction that speciation only, or mainly, occurs because of genetic isolation (i.e., cladogenesis); anagenesis no longer taken seriously by many card-carrying cladists.

pg. 2 – the French archaeological research tradition dominated the field from the 19th century until the early 1960s. It relied upon archaeological index fossils (e.g., Aurignacian blades, lamelle Dufour, etc.) to identify what they thought were prehistoric ‘cultures’ analogous to the bands, tribes, linguistic groups known to us from ethnography – a viewpoint that’s absurd from an anglophone perspective (see Binford & Sabloff 1983, Clark 1993, Barton & Clark 2022). Between c. 1900 and the early 1960s, the French approach also dominated US archaeology (called ‘culture history’ in the anglophone world). CH depended on variety-minimizing essentialist perspectives. After 1960 and continuing to the present, essentialism no longer part of anglophone archaeology, having been replaced by various processual conceptual frameworks.

pg. 3 – There are many definitions of science. My favorite is “a collection of methods for evaluating the empirical sufficiency of knowledge claims about the world of sense experience” (Clark 2007: 60).

pg. 3 – I doubt any classification can be ‘free from theory’ (bias), whether it be explicit or not, and ‘pure observation’ is impossible.

pg. 4 – any generalization arrived at strictly by induction can be refuted, but if a deductive element is ‘built into’ it, it is amenable to testing – so Dupré is correct.

pg. 4 – as to Wittgenstein, in physical science there is a universal metalanguage (mathematics) but biology, although it has a powerful, overarching metaphysical paradigm (neo-Darwinian evolution), and archaeology (which doesn’t) both lack a metalanguage that would allow for universal communication, hypothesis formation and testing.

pg. 5 – re Richards, the conceptual framework of biology is evolution. Any explanation that runs counter to the basic tenets of NDE would be rejected by all biologists.

pg. 5 – numerical taxonomy (Dupré) has fallen out of favor in HP and in archaeology, although I’ve never understood why. Methodological fads permeate both fields.

pg. 5 – see Rose (1998) for excellent discussion of reductionism in biology.

pp. 5,6 – many life scientists define a species as a polythetic set of genetic or biochemical polymorphisms.

pg. 5 – strictly speaking, taxonomy is the study of classification, a distinction usually ignored by archaeologists.

_____________________________

Barton, C.M. & Clark, G.A. (2021). From artifacts to cultures: Technology, society,

and knowledge in the Upper Paleolithic. Journal of Paleolithic Archaeology 4(1): 1-21.

Binford, L., and Sabloft, J. (1982). Paradigms, systcmatics and archaeology. Journal of Anthropological Research 38(1): 137-153.

Clark, G.A. (1993). Paradigms in science and archaeology. Journal of Archaeological Research 1(3): 203-234.

Clark, G.A. (2007). The flight from science and reason: evolution and creationism in contemporary American life. In Proceedings of the International Science Conference ‘Science in Archaeology’ (B. Harrison, ed.), pp. 60-80. Portland, OR: Institute for Archaeological Studies.

Kimbel, W.H. & Martin, L.B., eds. (2013). Species, species concepts and primate evolution. Springer Science & Business Media, 2013.

Rose, S. (1998). Lifelines–Biology Beyond Determinism. Oxford, Oxford University Press.

The article 'On the Physics and Metaphysics of Classification in Archaeology' by Mercedes Okumura and Astolfo G.M. Araujo gives a welcome perspective on one main issue that has been faced by archaeologists since the emergence of the field: how to classify the archaeological material. 

The article is well structured and well written. It first discusses the theoretical concepts (or absence of) behind classification, before developing in more detail the metaphysics of classification in biology and finally discussing the philosophy of classification in archaeology. The authors depart from the statement that this topic has been little explored in archaeology and call for a raise of awareness of what lies behind any kind of classification in archaeology and for being explicit about their theoretical positions.

In particular, the discussion about classification in archaeology that falls into two different types of kinds, i.e., conventional kinds and arbitrary kinds appears important. I however believe this section should be developed and include references to debates that have marked the history of archaeology. One may think, to cite but a few recent examples, of the extensive discussion on this in Tostevin's 2012 Seeing Lithics (Chapter 3 and the discussion of the emic vs. the etic including in the context of the Bordes-Binford debate), or of the discussion on the differences between chaîne opératoire and reduction sequence in Hussain's 2019 The French-Anglophone Divide in lithic research. The strong criticism that follows, of classifications based on conventional kinds, following Read's definition of an artefact as an intentionaly made object also reads like a simplification of the debate, as I wonder to what extent this restricted definition for artefact is used, even in cases of classifications based on conventional kinds. This section also leaves me a bit confused as whether it means that the authors reject any ontological approach to the past, which does not seem to be the case given the sentence p. 21 "In a way, a pluralistic approach using epistemic kinds might be interesting for Archaeology, especially if conceptualism is applied." 

While the paper does present an interesting discussion of fundamentals in archaeological classification and the fact that the authors do present a convincing case for being explicit in our theoretical approaches to classifications, the paper could be improved by better integrating the existing literature on classification in archaeology, which, even if it has probably not received enough attention by archaeologists, is still very much present in some works.

As minor points, I have noted a few typos (others might be present) in the text:
- p. 7 "will be find"
- p. 9 "can be used accordingly to the subject"
- p. 14 "can be, in principle, look much simpler"