Martin Haspelmath has been working on a project he’s now put online:
The Grammaticon: Linking grammatical comparative concepts to typological databases
This blogpost introduces a new resource for general-comparative linguistics: the Grammaticon, a collection of hundreds of grammatical comparative concept terms (Haspelmath & Englisch 2026). Version 1.0 has just gone online:
Many of these terms are linked to typological features represented in database collections such as WALS, Grambank, or APiCS. Grammatical terminology is quite variable (and often somewhat confusing), so the Grammaticon offers some guidance: Each term has a standard definition, and definitions are typically linked to other terminological resources (such as Wikipedia), and for many of the typological features, the Grammaticon explains how their technical terms relate to the definitions in the Grammaticon.
The Grammaticon was first conceived of in 2017, and the idea was presented at the ALT conference in Canberra (Haspelmath & Forkel 2017). Version 1.0 is now public, and it is hoped that it will be extended and improved greatly over the coming months and years.
Click through for the FAQs; a sample:
The Grammaticon definitions use ordinary language (no abbreviations or other notational devices) and recognize that some terms cannot be defined – they are treated as “primitives”. Is it an accident that this approach is similar to Anna Wierzbicka‘s NSM approach (Natural Semantic Metalanguage, Wierzbicka 1996)?
It is not an accident – the Grammaticon has been inspired by Igor Mel‘čuk‘s approach to definitions of linguistic terms (e.g. Mel‘čuk 1982), and Mel‘čuk in turn inspired Wierzbicka in the 1960s. For the meanings of ordinary words, Wierzbicka‘s approach is compelling and almost without rivals, and it seems to me that technical terms of grammar are best treated in a similar way.
“It is true that the reality is often continuous (as Democritus famously noted)”
Eh?
https://en.wikipedia.org/wiki/Democritus#Atomism
Heraclitus, maybe?
https://en.wikipedia.org/wiki/Heraclitus#Unity_of_opposites_and_flux
The website is certainly a good idea in principle. Haspelmath kinda putting his money where his mouth is.
Hiphilangsci has a podcast with him which is well worth listening to BTW:
https://hiphilangsci.net/2025/11/01/podcast-episode-51/
(I see that the most recent one is actually with Anna Wierzbicka, but I haven’t heard it yet.)
(I have to say that I’m with Heraclitus rather than Democritus when it comes to alleged linguistic primitives, though: πάντα ῥεῖ καὶ οὐδὲν μένει.)
Yay.
I’m so excited.
He’s been on this for more than a decade now and with every year his decisions and arguments get weirder and weirder. That one time, on academia.edu, I had to *shudders* take the side of a Minimalist to argue against Haspelmath’s redefinition of existentials.
I had to *shudders* take the side of a Minimalist
Fas est et ab hoste doceri.
No relation to the Necronomicon?
I think the notion of linguistic primitives is heuristically useful and so doesn’t necessarily suffer from the notion that continuous rules the roost. If anyone can deal with the conflict it might be Haspelmath, one of the premier morphologists of our time.
I think that these supposed grammatical primitives are not absolute universals, but more in the nature of important statistical tendencies among human languages.
That does not at all preclude their being very useful, and inquiry into just why they are so common, and exactly how, and to what degree, they are based on abstraction away from the actual data, is likely to lead to genuine theoretical insights.
But such “primitives” are not, themselves, fundamentals of Language: that is to mistake means for ends, and to short-circuit genuine research.
“Grammatical primitives” are fine by me so long as they don’t mistake their place as useful aids in research and description and start giving themselves ontological airs.
No relation to the Necronomicon?
with every year his decisions and arguments get weirder and weirder
That will be the endpoint. Indeed, the Endpoint of many things …
The Grammaticon of the Mad Linguist Hasb al-Math will reveal All. Iä! Iä!*
* Linguists will go mad from the revelation that iä is a suppletive rugose verb. Of appalling, mind-destroying primitivity … Man was not meant to see such a Conjugation …
Haspelmath has his approach and keeps going with it mit deutscher Gründlichkeit to see where it takes us. Some shall go mad from the revelation. We’re going to see if that includes him.
@DE: it’s usually the grammatical civilizeds that cause the problems, innit?
Until now …
My feeling is that Haspelmath got stuck with an imperfect idea. His intention is clearly good: there should be clarity on concepts like “word” and “verb” if you are going to compare their properties and behaviors between different languages. However, creating axiomatic definitions doesn’t work. Grammars leak. Edge cases are everywhere, and the closer you look the more you’ll find. No single definition is going to work, and Haspelmath himself is forced to wiggle to explain edge cases, like abstract nouns. That is not to say that comparative linguistics and typology are impossible. Rather, each study needs to define its comparanda precisely. The definition of a “verb” in a comparative semantic study need not be the same as in a study of tense marking or alignment, and so on.
Maybe on further reflection Haspelmath’s project should actually be the Grammatinomicon (Grammatikinomicon?), just to get some νομικός in there.
This prompted me to notice that the pseudonym of the main character in This Immortal means “bold counsel about the law.”
@DE Yes, primitive as a heuristic needn’t imply ontological airs, but I do like the way you’ve framed it. So to push the ontology – primitive might refer to the linguistics/psychology dichotomy, in which the psychologists claim that any linguistic primitives are recruited from extant psychological processes and primitives.
I just listened to the hiphilangsci podcast with Anna Wierzbicka, and I must say that I am no nearer to finding her views on meaning at all persuasive. I rather got the impression that James McElvenny felt the same way, but that may be projection …
But then, as a devotee of St Ludwig, I am probably not a great candidate for Wierzbicka fandom. Others may well differ.
Interesting to hear about her time at MIT. So very much not impressed by Chomsky (but also not very taken with the Generative Semantics crowd active at MIT at that time.)
She discusses at one point Matthew 10:28: “And fear not them which kill the body, but are not able to kill the soul”, and comes up with a historically and exegetically highly dubious account of what Jesus “meant” by this, which also involves an attempt at analysing “soul” in terms of her semantic primes; while conceding (as anyone surely must) that terms like this vary enormously in “meaning” between languages and cultures, she seemed (unless I misheard) to be saying that in all cases the equivalent “word” entailed something that could not be seen, and that was inside a person.
The half-verse in question in the Kusaal version goes:
Da zɔti dabiem nɛ banɛ kʋʋd niŋgbina, ka kʋ nyaŋi kʋ siig la.
“Do not fear those who kill a body, but cannot kill a siig”
In a Kusaasi story, a man hides his siig in the thatch of his hut (so much for it being “inside” him.) His neighbour (a witch) tricks the silly fellow into letting him help in rethatching the hut, sees the siig and runs off with it (hidden in his underwear) …
It might reasonably be objected that siig is a mistranslation, and the translators would have done better to go with win “personal spiritual individuality”, which means something like the Latin genius. But that is not regarded as living inside a person either, or indeed, “living” at all: it just isn’t the kind of thing that can be alive or dead.
Wierzbicka does not have an adequate sample of languages to address even this question, much less to set up any system of supposedly universal semantic primes.
@DE, those are not “grammatical primitives” (primitives of the grammar in the sense of the object of study)! They are descriptive primitives: “productivity”, “deictic visibility”
Inside? Like the 19th-century attempts to identify the soul with the pineal gland? FWIW, I used to imagine the soul on the outside of the body, more like an aura.
@drasvi:
Yes: but I was putting the boot in to semantic primes because Haspelmath explicitly likens his project to come up with cross-linguistically valid grammatical primitives to what Wierzbicka has claimed to have achieved regarding (lexical) semantic primes.
It’s just not good enough to try to shut down further questioning about your supposed fundamental terms by invoking a sort of atomicity and claiming that human thought just cannot go any further, as Wierzbicka does repeatedly in the podcast in response to perfectly reasonable questions from McElvenny.
Doing that is assuming your conclusions as postulates. Apart from the fact that this makes a highly questionable assumption about how languages “mean” (reminiscent of the untenable “picture theory”), you need to demonstrate that your semantic “atoms” correspond exactly one-to-one across (all) languages, and this can’t be done by ad hoc “analyses” of actual concepts which arbitrarily treat all the awkward residues of the analyses as insignificant.
Even if these supposed atoms were real, you’re left with the great difficulty of coming up with an adequate theory of how their meanings combine. This cannot be just some sort of mere juxtaposition. Again, McElvenney implicitly raises this very point in the podcast, in suggesting that the MIT Generative Semantics people were perhaps lukewarm about Wierzbicka’s ideas because their focus was more with the link between meaning and syntax, whereas hers was primarily on meaning and the lexicon. He gets a pretty perfunctory reply (I would say.)
I’ve been reading a bit about Construction Grammar lately, and noticing how the question of just how constructions combine is very much an open one (in fact, differing ideas about exactly how this could work have led to several different schools of Construction Grammar.)
Again, I’d be interested in what other Hatters think of the podcast. I am quite possibly being unfair, as being biased against Wierzbicka’s whole approach.
The Grammatonomicon tells me that a noun is (by definition) “the head of a nominal phrase – that is, referring phrase – that denotes an object”; that a verb is (by definition) “the head of a verbal clause – that is, a clause that denotes an action”, and that a subordinate clause is (by definition) “a clause that is not pragmatically asserted.”
I think I’m missing some fundamental metapoint here. What are these definitions for? They are certainly of no use in writing (or reading) descriptive grammars.
I used to imagine the soul on the outside of the body, more like an aura.
A forgivable childish error. Science has by now shown the soul is composed of ectoplasm, which can occupy the same physical space as ordinary matter, e.g. bodies when alive, walls when ghostly.
Presumably these are meant as prototypical examples of “noun” and “verb” (I’m not sure quite what the deal is with the strange one of “subordinate clause”); that would certainly work. But that is a very peculiar use of “definition.” And it depends on a complex and sophisticated notion of class membership of a Lakoffian or Wittgensteinian kind to get back to anything approaching normal linguistic usage of the terms, while giving no clue as to why on earth, for example, I might want to do that whole family-resemblance thing in order to be able to claim (rightly) that gim in the Kusaal
O gim.
“He is short.”
is an unequivocal verb, and neither an adjective, nor a noun used in a zero-copula construction.
I just can’t see the point of this.
But Haspelmath is neither stupid nor ignorant. What am I missing?
those Grammatonomicon definitions seem to me either entirely circular (“a verb heads a… phrase [headed by a verb]”) and thus pretty useless or identical to far-from-rigorous elementary school definitions (a verbal phrase “denotes an action”) that fall apart as soon as an example like “o gim” rears its low-altitude head.
i’m not sure what i think about the aspiration to standardize definitions, to be honest. i think there’s some merit in (explicitly!) working with shared terminology that doesn’t pretend to that level of rigor, since the ways that “the ‘same’ parts of speech” function in different languages can be very meaningfully different. it seems both more honest and more useful to me to use “verb” (for example) as that kind of term of convenience, without either putting it on a pedestal as an undefineable “primitive” or epicycling until a misleading or unhelpful “standard definition” emerges.
but regardless, anything aspiring to be a standard definition should at least not be circular, and able to see beyond anglophone “common sense”.
I think Haspelmath’s project can be salvaged by relaxing its attitude somewhat. For example:
1. Languages generally have different word classes, distinguishable by their morphological or syntactic distributions.
2. Usually there are words of different classes for describing “things”, like ‘dog’ or ‘rock’, than for describing “actions”, like ‘walk’ or ‘eat’. When there are, we call the one class “nouns”, the other “verbs”.
3. (Clarify noun roots versus inflected nouns vs. noun phrases, etc., if you will.)
So, you can start with a semantically-defined prototype, but move as quickly as you can to a distributional definition, which will take care of the semantic difficulties of abstract nouns or stative verbs or what have you. You’ll still have some things that won’t fit easily — so maybe you shouldn’t include them in your cross-linguistic comparison.
A similar approach works in phonology, if you want to universally define consonants and vowels. A purely phonetic or a purely phonological definition won’t work, but you can have a phonological definition riding on top of a phonetically defined prototype.
I think Haspelmath’s project can be salvaged by relaxing its attitude somewhat.
That’s pretty much my take, but he seems to be an unrelaxing sort of dude.
He comes over as a nice sensible bloke in the hiphilangsci podcast.
I’m mystified by how he’s gone from a fine anarchist belief that e.g. there is no cross-linguistic definition of even “word” that really works*, to a view that seems to be almost Chomskyite in its indifference to real deep-seated incompatibilities between the systems of different languages. As I’ve been saying, it seems so odd that I can’t help but feel I’ve somehow missed what he’s driving at.
* True: but who actually needs one?
When prowling the hiphilangsci site to catch up, I noticed this
https://hiphilangsci.net/2026/04/02/manfred-bierwisch-linguistics-in-germany/
which (among other points of interest) goes into how at one point in the DDR, deviation from the tenets of (linguistic) Structuralism was regarded as extremely suspect politically. I had no idea.
“I think I’m missing some fundamental metapoint here.”
@DE, you are reading the column titled “Croft’s comparative concepts”!!!
(“The Grammaticon links to William Croft’s comparative concepts (Croft 2022), which is a very similar enterprise…“)
(I’m not telling that H’s metapoints are more accessible than Croft’s:))
I’d be interested in what other Hatters think of the podcast.
I’m trying to catch up. Only listened to the Haspelmath podcast so far. I agree he seems a ‘nice chap’. My overriding impression is he just waffled, going round in circles, even contradicting himself. Reminded me strongly of someone who used to run the Deutsche Institut in a University city I lived in: incapable of bringing any sentence to a firm conclusion, for fear of offending someone.
He talks about SVO seeming to be an areal feature in S.E. Asia languages. Before you get that far, don’t you have to establish that for each language: constituent order is semantically significant; there is a default or unmarked constituent order. How do you do that without some cross-linguistic notions of nouniness and verbiness?
The Grammatonomicon tells me that a noun is (by definition) “the head of a nominal phrase – that is, referring phrase – that denotes an object”; that a verb is (by definition) “the head of a verbal clause – that is, a clause that denotes an action”, …
@rozele identical to far-from-rigorous elementary school definitions
Yes that was my immediate reaction. That seems to fall into the trap warned against in Ling 1.01 (and by e.g. Pullum) of confusing formal structure with meaning. What is it for some component of an utterance to ‘denote’ anything? Max Cresswell is (was) admirably clear: the unit of meaning is the sentence, because that can be true or false. Trying to attribute meaning to anything less-than-sentence is going to be hand-waving and/or end in tears. He used sentence-frames binding scoped variables into the positions needed to build a sentence/proposition.
If we can’t identify a ‘nominal phrase’ (at least, maybe not so far as it having a ‘head’) in every language of study, and m.m. a ‘verbal clause’, I don’t see how to establish S and O are similarish constituents; V is a differentish constituent. OTOH I remember a lecture where Dan Everett dismissed these sort of abstract philosophical objections as never encountered by a field linguist. (Specifically wrt Quine’s grue and bleen.)
@Y classes for describing “things”, like ‘dog’ or ‘rock’, than for describing “actions”, like ‘walk’ or ‘eat’.
Hmm. English is famous for nouning verbs and verbing nouns. ‘dog’, ‘rock’ are both. ‘walk’ is both. Even ‘eat’ can be nouned. What kind of a “thing” is wonder? What kind of “action” is disappointment?
podcast with Anna Wierzbicka, and I must say that I am no nearer to finding her views on meaning at all persuasive. I rather got the impression that James McElvenny felt the same way, but that may be projection …
ditto and ditto. I kept wanting to send her away to study ‘How to Do Things with Words’. At the stage of child language acquisition, there might be ‘meaning’ in the background as declarative statements, but that’s a superimposition/interpretation from adults. Miw! (plus gesticulating) is sufficient. And is language in the sense of arbitrariness of the sign.
If ‘soul’ is a linguistic universal, I’m afraid its denotation has evaded me. I know it only as a term of art used in some religions. Even from the little I know about other cultures and practices, I greatly doubt it’s a universal.
If you reduce meaning to lexicon, how do you explain the difference between ‘the cat sat on the dog’ vs ‘the dog sat on the cat’? [I see @DE has already made this point.]
@AntC
I understood the statement about Southeast Asian languages to mean something like
1. We are excluding some languages (e.g., Tagalog)
2. The other languages, by and large, all have coherent classes corresponding to S,O, V and “declarative sentence”.
3. Although not related, the other languages show a preference for a SVO ordering in declarative sentences, in the sense that other orderings are non-declarative or marked (e.g., to focus on the S or O).
Regarding disappoint/wonder, I think the action for disappoint is “deceive when the deceived person has reason not to expect to be deceived by the deceiver”. For wonder, I associate a facial/gestural expression, but maybe I am influenced by religious painting and drama.
I’ve been reading a bit about Construction Grammar lately, and noticing how the question of just how constructions combine is very much an open one …
The approach that makes most sense to me is the line from Montague grammar to Kripke to Cresswell (see above). It’s all couched in philosophical logic, so I’m not sure there’s a Linguist-friendly intro. taster here. Stanford E of P also has an article ‘Compositionality’ that cites Cresswell without any direct discussion.
In the lecture I attended, he was much preoccupied with sentences like ‘three of the cats sat on the mat, and one sat on the dog.’ Wherein, ‘and’ cannot simply be a conjunction of two propositions. Rather, ‘three’ and ‘one’ are _not_ semantically tucked away inside a NP, but quantify at widest scope. The overall sentence carries a presupposition the universe of discourse includes at least four cats. ‘The’, ‘a’, ‘some’, … similarly cannot be sub-components of a NP, but rather scope across both the NP and predicate/verbal clause.
@PP, yes thanks that’s a reasonable summary of the SOV ‘finding’.
The question is how do we point to evidence of which S.E. Asian languages follow the pattern? What is it about Tagalog for which the claim doesn’t hold? Does it have S, O and V but not in SVO order? Or (for example) do its NPs (supposing it has them) carry no differentiation of S vs O? We need some prior notions of NP, S, O, V before we can meaningfully investigate.
other orderings are non-declarative or marked (e.g., to focus on …
So ‘non-declarative’ is a linguistic universal, not limited to IE languages? I think there’s less hope of defining ‘marked’ or ‘focus’ in a universal way.
To be clear: I’m not disputing the finding. I’m looking for evidence that would support concepts like NP, S, O, V applying cross-linguistically.
Shamelessly seizing on “wonder” as a link to something quite unrelated, in order to consult the wisdom of the Hatters …
English has four main-clause types, which differ structurally (I’m not talking about Gricean things like “It’s cold with that window wide open”):
Statements: I like syntax.
Questions: Do you like syntax? Who likes syntax?
Commands: Avoid the works of Noam Chomsky!
Exclamatives: How very misguided is the notion of Universal Grammar! What a chimera it is!
The first three of these are probably universal (I certainly have never come across a language that lacks any of them, though logically a language without a structurally distinct “command” type seems quite possible.)
But Kusaal lacks the “exclamative” type altogether, and it’s not alone: even very in-depth studies of the syntax of e.g. Hausa make no mention of such a thing, and nor do Jeffrey Heath’s many grammars of West African languages, nor Kofi Yakpo’s very detailed grammar of Pichi. (But it’s an understudied topic, so I suppose one should beware of an argumentum ex silentio.)
On the other hand, the exclamative use of interrogative words and structures seems to be extremely common cross-linguistically. It’s not just SAE, but e.g. Turkish, Hebrew, Arabic, Mandarin, and at least one Bantu language:
https://williemyers.com/wp-content/uploads/2023/08/mull-jan-31-exclamatives.pdf
So, my question to Hatters: do you know a language which lacks the exclamative clause type? (What I’m wondering about is whether the presence or absence of this clause type is an areal thing, as it seems it might be in West Africa.)
Welsh, now I think of it, doesn’t use interrogatives for this: it uses mor, a marker of the equative form of the adjective (Welsh has four degrees of comparison, not a mere three):
Mor weddaidd ar y mynyddoedd yw traed yr hwn sydd yn efengylu
“How beautiful upon the mountains are the feet of him that bringeth good tidings”
where mor weddaidd is “equally as comely.”
The Kusaal version just recasts it as a statement:
Banɛ mɔr labasʋŋʋ yit zuoya ni na la, ba ken malisi ti sʋnf pamm!
“Those who bring good news out of the mountains, they walk making our heart very glad.”
@AntC: If ‘soul’ is a linguistic universal, I’m afraid its denotation has evaded me. I know it only as a term of art used in some religions. Even from the little I know about other cultures and practices, I greatly doubt it’s a universal.
Rabbi Zvi Yehuda, the only real Jewish scholar I’ve known, used to say that there was no Hebrew word for “soul”. On the other hand, I wonder whether he was thinking of a Christian or pop-Christian meaning, not of the full range of meanings of the English word.
I greatly doubt it’s a universal
I’m perfectly sure that it isn’t.
Wierzbicka actually doubles down on this near the end of the podcast, though in a fairly informal aside. Even so, it seems quite astonishing. There is abundant ethnographic research on all this, confirming that she’s just wrong. Perhaps her actual view is more nuanced (I hope so.)
She seems to be very involved with trying to explain Christian doctrines in terms of her semantic primes. (E.g. Wierzbicka, Anna. 2025. The Nicene Creed in Minimal English: Why Christianity Needs Universal Human Concepts. Cham: Palgrave Macmillan.) She has maybe succumbed to wishful thinking about its feasability.
AntC,
I agree he seems a ‘nice chap’.
He seems, doesn’t he? So quiet, reserved, reasonable…
He is not. I’ve had a few run-ins with him (mostly online) and I have the receipts. He was cuckoo bananas even before Covid, but Covid turned him into MAGA light.
@David,
I’d be interested in what other Hatters think of the podcast.
H has a podcast? Will the madness never cease???
Covid turned him into MAGA light
How very unfortunate.
(It was the Wierzbicka podcast on hiphilangsci that I was wondering what others thought about, rather than the Haspelmath one. AntC, at least, seems to share my view.)
@drasvi,
(I’m not telling that H’s metapoints are more accessible than Croft’s:))
But Construction Grammar actually makes sense for the description morphosyntax.
@AntC,
Max Cresswell is (was) admirably clear: the unit of meaning is the sentence, because that can be true or false.
I wonder which of the 140 or so definitions of ‘sentence’ he had in mind… I am very skeptical about any claims about language made by a logician, unless they clearly and unambiguously disavow the concept of “logical predicate” and cite von der Gabelentz on the first page.
@rozele,
those Grammatonomicon definitions seem to me either entirely circular … and thus pretty useless
That was precisely my reaction and I said so in a recent paper. Haspelmath read it and responded to it and it’s … something. I will write it up on my blog later this week once this deadline looming over me is over.
@David,
Wierzbicka, Anna. 2025. The Nicene Creed in Minimal English
Good Lord Jesus (no pun intended)… Of all things to apply semantic primes to, this is possibly the worst one, with it layers of answers to controversies.
And sure enough, the thing is not anywhere near what it says on the cover. A case in point, the section “I Believe in One Holy Catholic and Apostolic Church” from p. 97:
The very first line is wrong in terms of theology. And the rest is… How is this *Minimal* English when it has to rely on other evidence (Paul here, the Church Fathers elsewhere)?
Or this, from p. 100, same section:
Yeah, no. None of that is accurate, not even the last line.
Sounds as bad as I feared.
She seems to have entirely misunderstood where the difficulties of cross-cultural communication actually lie. I suppose that’s what happens when you don’t understand the intimate connection between language and culture. It’s not a matter of superficial icing on the cake: it’s something that pervades languages deeply.
There is a whole literature about this from sensible missionaries (and Bible translators) who have actually learnt something from the church’s many grave failures in this area.
I will write it up on my blog later this week once this deadline looming over me is over.
Please link it here!
@AntC: English is famous for nouning verbs and verbing nouns. ‘dog’, ‘rock’ are both. ‘walk’ is both. Even ‘eat’ can be nouned. What kind of a “thing” is wonder? What kind of “action” is disappointment?
I meant dog the animal, not dog the word, and likewise ‘walk’, the thing one does with dog the animal.
So once you establish the distributional nature of words like dog — they can be pluralized, be possessed, serve as verbal arguments, etc. — you jettison the semantics and call every word with those distributional characteristics a noun, without worrying about their meaning.
@DE, are your “exclamatives” the same as rhetorical questions?
Rhetorical questions do seem to shade continuously into exclamatives.
I should note that “disappointment” is of course a noun and thus must be an “object” rather than an “action.” Except of course come to think of it “action” is itself a noun and thus itself an “object.”
are your “exclamatives” the same as rhetorical questions?
No. Exclamatives differ structurally from questions, despite containing interrogative words. They differ in word order and in intonation, and there a quite a few other differences too.
“How pointless is Universal Grammar?”
Question; quite possibly rhetorical here on LH.
“How pointless Universal Grammar is!”
Exclamative.
CGEL treats this on pp 918 ff.
Section 2.3.1 in this oldish paper discusses them briefly too.
https://web.stanford.edu/~zwicky/speech-act-distinctions.pdf
Different languages use different strategies for this, though interrogative words seem to be often shangaied for the process cross-linguistically. English only uses “what” and “how”, as in
“What a fruitless hypothesis Universal Grammar has turned out to be!”
I’ve found that there is quite a literature on this, though it seems to be underresearched overall.
The Myers paper on Kirundi that I linked to is interesting, inasmuch as the exclamative construction can use an interrogative particle, but doesn’t use the interrogative verb form, so (as in Engish) the construction has similarities to a question but is still formally distinct from a question.
This is quite a separate issue from Austin-style illocutionary force and Gricean implicatures, though historically some of these exclamatives may have originated from such things.
The question has been nagging at me for a while in relation to Kusaal, which seems just to lack anything like exclamatives. Trawling through grammars of West African languages, I get the impression that Kusaal is probably typical of the region in this, but few grammars go into the kind of detail that means one can be sure that any exclamative-type constructions haven’t just been overlooked. So it may be an illusion.
They don’t seem to be rare cross-linguistically in general, nor one of those things that SAE has just copied from Latin or Greek.
And English has verbless exclamatives that can’t be interrogative:
“What a fruitless hypothesis!”
“How different, how very different from the home life of our own dear Queen!”
(Also verbless questions that can’t be exclamative. “Why a duck?”)
I think the action for disappoint is “deceive when the deceived person has reason not to expect to be deceived by the deceiver”
“Disappoint” doesn’t suggest deception to me. If I’m told that the weather during the solar eclipse may be clear or cloudy, and it’s cloudy, I’ll be disappointed, and I may even say that the weather disappointed me.
Not in German, where they have finite-verb-last order because they’re arguably “subordinate” clauses whose “main” clauses have gone missing, nor in Old English, where they had finite-verb-last order even more often than subordinate clauses.
Still happens in modern English – how unusual this is!
In German it’s etymologically even the opposite: täuschen “deceive”, enttäuschen “disappoint” – literally “de-deceive”, like “disillusion”.
Like Spanish desilusión. (Ilusión can be a hope or dream.) But I think the German etymology as “undeceive” agrees with pp’s picture, whereas the Spanish one is like mine in not presupposing that there was any deception before the disappointment.
The Creed of course nowhere mentions bishops or priests,* although maybe a text that only explicates the actual words of the Creed as built up from a limited set of so-called primitives wouldn’t be long enough to package as a book? The actual word in the Creed usually Englished as “church” is (in its sing. nom. lemma form) ἐκκλησία, and how hard can it be to define ἐκκλησία as a pseudo-mathematically-described combination of various members of the ancient-Greek list of semantic primitives? (Maybe arbitrarily hard, I assume the critics of the approach will say.)
*Not because they weren’t thought important but because their existence was not particularly controversial at the time and thus was not perceived as needing clarification or defense.
It’s a pretty odd paraphrase all round.
I wonder what she does with “Holy Spirit”?
The current Kusaal Bible version says Siig Kasi, where siig means (at least in traditional culture) “life force”, not “spirit”, for which Kusaasi culture has no good equivalent.
Kasi is an ideophone borrowed from Mampruli: its core meaning is “clean.” The nearest real match for “holy” would probably be kisug “tabu, forbidden”, though one can see why that wouldn’t do in this case. Using “clean” for “holy” is just pure translationese, though. It matches nothing in traditional culture – or, indeed in Christianity, really.
The construction Siig Kasi is not really grammatical: like most “quality” ideophones, kasi can only be predicative, not attributive, and for “holy people” you have to say
nimbanɛ an kasi
person.who.Pl be holy
“people who are holy”
Siig Kasi is the sole exception.
The neologism was apparently controversial, though apparently this was more because kasi was felt to be “foreign” rather than because of the ungrammaticality. Previous Bible versions used Sisʋŋ “good life force.” As a translation, this is obviously hopeless, though Kusaasi who are already Christian and have acquired the Christian concept through actual teaching about it presumably are happy to use it as a handy label regardless of its actual etymology.
The idea that you can short-circuit the complex process of conveying an alien concept to speakers unfamiliar with its whole cultural background by One Weird Trick of expressing the idea in putatively human-universal atom concepts is frankly ludicrous.
Unfortunately, the traditional Christian approach has usually been to supplant the local culture in order to transplant the foreign concept into the resulting imposed Western culture, rather than to translate into the indigenous culture in any meaningful sense. I suspect Wierzbicka’s motive is the noble and praiseworthy one of trying to facilitate a kind of real refactoring of Christian concepts into exotic cultures; it’s a pity that her method is so flawed.
Is Such a nice guy! an exclamative? If not, what is the difference to What a nice guy!?
CGEL classifies clauses like that as “non-exclamative exclamations” (p923), giving examples like
Such strange people inhabit these parts.
She hated it so.
It was such a disaster.
It separates them from “exclamatives” on the grounds that “so” and “such” are not markers of a particular clause type, but can occur in any kind of major clause type.
I may have spoken too soon about the complete absence of exclamatives in Kusaal, incidentally. I’ve found a couple of examples in the Bible translation with a focusing deictic within a question, which look like candidates. e.g. Romans 7:24:
Oi, bɔ sʋnsa’aŋʋ nwa ka m nyɛta!
oh what sorrow.LINKER this and I see.IMPERFECTIVE.CONTENT-QUESTION
which is literally “this is what sorrow that I am seeing?”
I’m not sure that these are formally distinguishable from questions, though; you can embed deictics like these in ordinary questions too:
Fʋ maal bɔɔ la tis mam?
“What is this that you have done to me?”
@dm, jf
I thought AntC wanted an action, and for me the removal of the illusion is an effect (of deception), rather than a primary act (but maybe I am overthinking this). Also for me, deception can also mean betrayal, which may be the point of issue, i.e., in jf’s case, he may also feel the weather has betrayed him. Both betrayal and deception can be and maybe are by default unconscious, i.e. the weather and Marlene Dietrich in “the Blue Angel” are just doing what they do and not trying to deceive or betray anyone. Maybe in those cases the removal of illusion is the primary act, but only from the point of view of La Dietrich or the weather. I also do not know if enttäuschen is semantically close to betrügen in German.
@Y I meant dog the animal, not dog the word, and likewise ‘walk’, the thing one does with dog the animal.
So once you establish the distributional nature of words like dog — they can be pluralized, be possessed, serve as verbal arguments, etc. — you jettison the semantics and call every word with those distributional characteristics a noun, without worrying about their meaning.
But, but you’re presuming the conclusion! How do you tell the appearances of the word ‘dog’-as-noun from the appearances of ‘dog’-as-verb? If in English it was only a few words that appear in both roles, you could establish the rules by avoiding the ‘dog’ example, then come back to it later as an exception. But in case of English, you’d have to conclude well more than half of common words are exceptions.
That’s why I’m also questioning the universality of ‘unmarked’ or ‘default’ constituent order.
Or you establish the rules by avoiding Anglo-Saxon vocab altogether, and studying Latinate/abstract/Greek vocab.
@JWB I should note that “disappointment” is of course a noun and thus must be an “object” rather than an “action.” Except of course come to think of it “action” is itself a noun and thus itself an “object.”
Fair point, and I thought that as I wrote it. But English doesn’t really do pro-verbing. Also using an abstract noun meant I could be neutral between the disappointer vs the disappointee. (I agree there’s not necessarily any
deceptiondeceiving going on.) So at risk of tortured English, I’ll try againWhat kind of doing is disappointing, or being disappointed. (Gerunds still grammatically nouns.)
@JF the only real Jewish scholar I’ve known, used to say that there was no Hebrew word for “soul”. On the other hand, I wonder whether he was thinking of a Christian or pop-Christian meaning, not of the full range of meanings of the English word.
That seems bizarre. I thought Christian ‘soul’ came from (translations of) the OT(?). A quick search finds
Yeah, re full range of meanings: I had visions of a passenger flight having to declare Mayday. The emergency services on the ground always ask how many souls on board? Whilst the pilot wrestles with the controls to land the plane, I imagine the co-pilot wrestling with the passenger list: how many Atheists on board? Do Confucians/Taoists have souls? Are there any comfort animals on board? Do pet octopuses have souls?
Assume arguendo you could correctly analyze a given Greek word in let’s say the Creed or perhaps in the Gospels as some objectively-describable combination of semantic primitives that exist in all languages. But all that would let you do at first is confirm the vague impression that a given let’s say West African language does not at present possess a lexeme corresponding to that chemical formula. It either needs a new lexeme or some sort of agreement to use an existing lexeme in a new and different sense. So you’re not much better than before in guidance as to how to do a translation, other than perhaps proving that good translation may not be possible into the given language as it currently stands.* The chemical formula for the semantics does not tell you what that needed new-or-adapted lexeme should be – the connection between signifier and signified remains largely arbitrary even if you have standardized ways of describing signifieds, so there’s no mechanical way of generating a new signifier that transparently carries the needed meaning.
*The sort of people who tend to do initial Bible translations into languages without the lexical and cultural resources to understand what the Bible is going on about are perhaps prone to resisting the conclusion that their intended goal is impossible. This resistance may lead to weird and unfortunate results when they nonetheless persist in their work.
But, but you’re presuming the conclusion! How do you tell the appearances of the word ‘dog’-as-noun from the appearances of ‘dog’-as-verb? If in English it was only a few words that appear in both roles, you could establish the rules by avoiding the ‘dog’ example, then come back to it later as an exception. But in case of English, you’d have to conclude well more than half of common words are exceptions.
“The dog”, noun, and so other words with “the”. “It dogged him”, verb, and so other words which can do “It X-ed him”. Likewise “I put it in a bag (n.)” and “I bag (v.) it”, etc.
Or, if you will, think of words whose form appears in only one word class, like son or eat.
I suppose the German one assumes that wishful thinking is self-deception.
No; betrügen is “defraud” and “cheat on”.
Yes and no; what the terms mean in different books of the Bible is not always straightforward, and then there’s the whole issue that Paul’s letters introduce the distinction between soul and spirit without explaining it – I like the idea that it’s a reaction to Epicureanism that has gotten way out of hand.
I cannot speak about the whole Hebrew Bible, but it does seem that nefeš which is literally “breath” means “living force” and sometimes person as an individual (something like “living soul”), but not individuality. This concept is most readily expressed as lev (heart, same as in any number of languages (I remember that in Kusaal it is gold bladder, tyvm)). In Leviticus it is explained that this life force (remember, it is breath) is located in the blood. Quite literally, this is why consumption of blood is prohibited. Unless ancient Jews wised up about oxygen, I cannot understand how they imagined it exactly.
Re the OT Hebrew word that Jerry F.’s informant says absolutely definitely doesn’t mean “soul,” here’s one reasonably standard-reference summary: ‘נֶפֶשׁ (nephesh) occurs roughly seven hundred fifty-four times, encompassing the whole range of human and creaturely life: the living being itself, the animating life-force, the inner self with its thoughts, emotions, and desires, and even the life that may be forfeited in death. Context, rather than a single English equivalent, determines whether translators render it as “soul,” “life,” “person,” “creature,” “appetite,” or related ideas.’
That said, the LXX translators (themselves Jews, and not Christians) pretty consistently rendered nephesh as psyche (ψυχή). Which itself when used in the NT is often, but certainly not always, translated as “soul.” The distinction in some places in the NT between ψυχή and πνεῦμα has indeed occasioned a vast literature.
The vigorous dogmatic explication by the 16th-century English reformers of “originall or birth sinne” focuses on a flesh/spirit dichotomy that leaves the soul-as-such out of it, and is especially noteworthy for noting the complexity of translation, viz. “And this infection of nature doth remayne, yea in them that are regenerated, whereby the luste of the fleshe called in Greke fronema sarkos (whiche some do expounde, the wisedome, some sensualitie, some
the desyre of the fleshe) is not subject to the lawe of God.” (That’s φρόνημα σαρκὸς, for those who like their “Greke” untransliterated.)
Kusaal nyɔvʋr, a transparent compound of nyɔɔr “nose” and the adjective vʋr “living”, always means simply “life” as opposed to “death”, and never has any sense of “soul” or “spirit.”
Nyɔɔr by itself can mean “life”, too, as in the proverb
Kɔdiŋ ye, Nyɔɔri an gɛl.
francolin that nose.LINKER be egg
“Francolin says, Life is an egg.”
(Implying that life is the prerequisite for anything else, so “Abi gezunt.”)
and (perhaps needless to say) the concepts of “neshome”, “lev”, etc and other words that could be argued to correspond to “soul” in the Tanakh don’t map directly onto the members of different jewish communities’ varied understandings of those words.
i can’t remember offhand what term is used in the early modern accounts of dibukem (which in some ways act differently in yiddish, sefardi, and mizrahi contexts), but in some of them an individual’s personal essence can not only possess another person’s body (cohabiting with their own essence or temporarily submerging it), but also can get stuck in certain kinds of bodies of water until a suitable human (or iirc non-human) host happens by. and of course in the lurianic tradition parts of the personal essence of past figures can inhere in contemporary people through a kind of kaleidoscopic metempsychosis. neither of those kinds of mobility seem to me to cohere all that well with christian concepts of “soul” – i’m not sure that there’s a common core there that’s more specific than “personal essence”/”distinctness from other people”, which doesn’t seem all that helpful as a semantic prime.
There are three common terms: רוּחַ rûaḥ, נֶפֶשׁ nep̄eš, and נְשָׁמָה nəšāmâ. In later exegesis and especially in the Kabbalah their meanings have been demarcated in detail, and they progress in that order from the merely physical to the spiritual. I don’t know how much the meanings overlap in Biblical Hebrew. Much has been written on this by many, none of which has been read by me.
Add: rûaḥ is lit. ‘wind’. nep̄eš is, I presume, metathesized from the root nšp̄: i.e. ‘exhalation’. nəšāmâ is from nšm, i.e. ‘breath’ or ‘inhalation’.
“Francolin says, Life is an egg.”
Am I to be lectured by a chicken?
Fas est et a pullo doceri.
Note btw that the Wierzbicka notion of semantic primitives does not actually match up with a Swadesh-list sort of concept. .The latter will include “water,” but not “hydrogen” and “oxygen.”
Closer to a peacock than to a chicken, but still…
The Swadesh list was a reasonable guess at which words are the most stable diachronically; it has nothing directly to do with “primitive” or “atomic” meanings.
The “[animal] says” frame is quite common in Kusaal proverbs, and apparently in Africa in general. It’s not always apparent why that particular animal has got nominated.
Another Kusaal one is
Nwaaŋ ye, dikanɛ bɛɛ o nɔɔrin la aan o din.
“Monkey says, the food in his mouth is his.”
(Possession is nine points of the law.)
I suppose this reflects something of the same mindset as underlies Aesop’s fables. (Kusaasi folktales often involve talking animals too, including, of course, the trickster hare Asumbul. One story in my collection features a Hajji hyena, because why not?)
But it’s probably a pan-human thing. Are there any cultures that don’t have stories with talking animals?
Am I to be lectured by a chicken?
Don’t quail at the thought, and don’t grouse about it.
i didn’t realize that nefesh and neshome were that kind of pair! that’s really neat, and i’m sure much has been made of it in the mystical literature. i’ll have to keep my eye out for it in yiddish poetry!
@bulbul He seems, doesn’t he? So quiet, reserved, reasonable…
He is not. I’ve had a few run-ins with him …
Yes I was aware of that, hence the scare quotes.
I am very skeptical about any claims about language made by a logician, unless they clearly and unambiguously disavow the concept of “logical predicate” and cite von der Gabelentz on the first page.
Cresswell is a Philosopher/Logician. He doesn’t make “claims about language”; and certainly not linguistic universals. He is trying to find an underpinning for semantics, particularly how to form the meaning of the whole from the parts. Did you read at the link I provided?
Are you sure it’s not you adopting confrontational positions? As I come across (some of) your previous posts on Gabelentz at the Hattery, they seem rather … curmudgeonly (to use Hat’s term) [**]. What’s wrong with “logical predicate” for a Logician to use in a logic?
(And he doesn’t try to define ‘sentence’: that’s Linguistics’ job.)
Answer when you’ve cleared your deadlines. Or is there a paper you can point to?
[**] I plead guilty to the epithet myself.
The “[animal] says” frame is quite common in Kusaal proverbs, and apparently in Africa in general.
There are pages of them in Mammeri’s Poèmes kabyles anciens, motivated primarily by rhyme, e.g.
(Yes, they are pretty much all this negative. Early 20th century Algeria was by all accounts not a happy place.)
@lameen
I do not view these as negative. The man with the bad wife and the (other or maybe the same) man with the bad children are making the best of what life has served up to them (it should not be ignored that he feels, possibly incorrectly, that the son is his).
The man with wealth is ensuring that his children will not be corrupted by that wealth and/or will be motivated to acquire their own wealth.
Cresswell
I’m actually the proud owner of the work on modal logic he wrote with G E Hughes. Who doesn’t like a good Kripke frame now and again?
The issues he is concerned with are quite different from what I was alluding to.
In Construction Grammar, it’s Constructions (form-meaning pairings) all the way down (and up); lexical form-meaning pairings are not regarded as different in kind from syntactic form-meaning pairings.
But leaves you with an issue about how constructions combine. A naive idea would be that you kind of slot lexical-type constructions into appropriate positions in your syntactic-style constructions, but pretty much the whole point of Construction Grammar is that It’s Not That Simple, and the more you think through the implications, the more knotty the problems get.
Different schools of Construction Grammar have different approaches to this. “Creative Construction Grammar”, for example, asserts that this all happens by “conceptual blending”, a rather unhelpful technical term for what are taken to be a range of human powers of analogy and synthesis not linked exclusively to Language.
I don’t know a lot about this, but it seems to suffer from what quite a lot of Construction Grammar does: it’s very probably right in principle, but applying it in practice rapidly because very hard and you readily get lost in the weeds.
But then, you’d surely expect any adequate theory of human communication to be very complex. Minimalism is for insects …
“Negative” is maybe too vague a word; “pessimistic” is closer.
(I like your optimism, but the man with wealth there is definitely neglecting his kids; all three of them are supposed to be bad examples that you should avoid becoming.)
Here’s another sample:
Thank you, AntC, David M., J.W. Brewer, rozele, and Y for interesting comments on Hebrew words and “soul”.
Note FWIW that the Chabad folks (as well as many other current Anglophone Jewish sources) are happy to treat this traditional prayer as one for the “soul” of the decedent and to use “soul” in the English translation of the Hebrew. That said, this specific prayer in these specific words is apparently not attested before the 17th century (CE). As rozele noted, it’s not like Jewish understandings of the referent of specific Hebrew words have all been entirely stable and uniform across space and time.
https://www.chabad.org/library/article_cdo/aid/367837/jewish/Kel-Maleh-Rachamim.htm
Elegant variation on 子曰 “Confucius says”.
Five outdated 1-£ coins say the Pirahã don’t.
@AntC,
Yes I was aware of that, hence the scare quotes.
Ah, my apologies.
Cresswell is a Philosopher/Logician.
I’m assuming the capitals are significant, though not sure what they signify.
He doesn’t make “claims about language”; and certainly not linguistic universals. He is trying to find an underpinning for semantics
Which is, for lack of a better word, a component of language.
Are you sure it’s not you adopting confrontational positions?
I’m sorry, was anything I said insulting? If so, I apologize.
But to answer your question, yes, I am pretty sure I am adopting confrontational positions, aka disagreeing or critiqung.
What’s wrong with “logical predicate” for a Logician to use in a logic?
Absolutely nothing. The problem begins when they try to apply it to language.
(And he doesn’t try to define ‘sentence’: that’s Linguistics’ job.)
So he works with the concepts of “semantics” and “sentence” and you still insist he makes no claims about language? Also I said “had in mind”, not “define”. He might have been happy outsourcing the definition to linguistics, but he still had to use one.
I love von der Gabelentz, it’s just a shame that his work did not gain much traction in linguistics. Same goes for Anton Marty who is perhaps even more relevant to this discussion.
J.W.Brewer, Jews starting from pre AD times (some) until now (almost all, of a traditional kind) beleive in something like afterlife, world to come etc. There must be then something of a person that inhabits those spaces. Because body is in the ground it cannot be the body. Then whatever it is can be named “soul” in translation to “Christian languages”.
The calim is that there was no such concept in (proto) Judaism before composition of the very late books of the Bible and therefore backreading into, say, Pentateuch the concept of soul under any word is suspect. But to quote Y, “much has been written on this by many, none of which has been read by me.”
There must be then something of a person that inhabits those spaces
Orthodox Christians and Muslims believe that the physical body will be resurrected, so a belief in an afterlife does not necessarily imply belief in disembodied souls.
However, (most) mainstream Christians do actually believe in this; the dissenting position is sometimes rather poetically called “the sleep of the soul”:
https://en.wikipedia.org/wiki/Christian_mortalism
There are differing views on this within Islam, too:
https://en.wikipedia.org/wiki/Islamic_view_of_death
Buddhists, of course, officially disbelieve in souls altogether. (Maybe: it’s complicated …)
@jwb, d.o
Whosoever destroys one soul, it is as though he had destroyed an entire world. And whosoever saves a life, it is as though he had saved an entire world.”[30]
https://en.wikipedia.org/wiki/Hillel_the_Elder
—
[30] = Jerusalem Talmud Sanhedrin 4:9
לפיכך נברא האדם יחידי בעולם ללמד שכל המאבד נפש אחת מעלין עליו כאילו איבד עולם מלא.
וכל המקיים נפש אחת מעלין עליו כאילו קיים עולם מלא.
—
This indicates to me that even in “late” usage, nifsh is not exclusively “immaterial soul” but also “life” as a unit incorporating body and soul. But maybe this is an artefact of this particular translation.
@David E.: Some largeish subset of “traditional” (in a modern context) Jews also believe that there will be a physical resurrection of physical bodies when the Messiah comes – this is said to be the or at least a motivation for traditional hostility to cremation etc. I don’t know if there’s a well-developed view on what may go on in the interim. I am reminded of an interesting quirk in the Book of Job. The Hebrew text that we have from the Masoretes eventually ends “So Job died, being old and full of days.” But the LXX Greek goes on to say (and I don’t think this is standardly claimed to be a later Christian interpolation as opposed to a variant Jewish text?) “and it is written that he will rise again with those whom the Lord raises up.”* In the Byzantine lectionary there are various readings from Job during Holy Week, and that one comes up at vespers on Good Friday.
*γέγραπται δὲ αὐτὸν πάλιν ἀναστήσεσθαι μεθ᾽ ὧν ὁ κύριος ἀνίστησιν.
@Lsmeen: Is tlata a loan or are Berber numerals so close to Semitic?
Because body is in the ground it cannot be the body.
within the very varied field of past and present jewish ideas about the afterlife or lack thereof, there is a longstanding tradition (currently held by, among others, ChaBaD) that the body is indeed present in it. that’s the basis of the german-jewish (ashkenazi, in its proper sense) minhag’s rejection of tattooing* and other body modifications. the problem of burial has been considered, and solved in the simplest way possible: those buried in the holy land arise directly from their graves; those buried elsewhere roll through underground tunnels to jerusalem to re-emerge from the earth there. here’s a longer explanation, complete with a subway map.
.
* which is decidedly a minority position among jewish traditions, though it has become hegemonic through the colonial efforts of the Alliance Isréalite Universelle [sic] and ChaBaD.
As usual for this kind of matter, TDOT has extensive articles on nep̄eš, rûaḥ, and nəšāmâ in the Old Testament, discussing etymology, semantics, and theological significance, among the Hebrews and among neighboring people.
rozele, I was under the impression that tattooing was universally prohibited, following Leviticus 19:28 and its later interpreters (the Mishna, Maimonides, etc.)
@bulbul But to answer your question, yes, I am pretty sure I am adopting confrontational positions, aka disagreeing or critiqung.
We’ve not interacted much at the Hattery. I await your write-up of the contretemps with Haspelmath. So far, on very thin evidence:
* Haspelmath is “nice, sensible” [DE], floundering [AntC, Y]
* Wierzbicka is misguided and downright wrong in places [all]
* bulbul is confrontational [bulbul] and citation-lite [AntC]
I’m all for disagreeing or critiquing; do plenty of it myself; always with supporting evidence. So far on this topic all you’ve provided is your own opinions.
What I said in re Haspelmath was “He comes over as a nice sensible bloke in the hiphilangsci podcast.” As to his intrinsic nicitude and sensibiliousness I have no information. I would be inclined to trust bulbul’s opinion on this, and indeed on many other matters.
Conversely, Anna Wierzbicka may very well not have done herself justice on the podcast (she is 88.) I can think of at least one very well-informed Hatter who has a high opinion of her work.
We can disagree on such points; at least we all share a disdain for Noam Chomsky. It is enough.
I gave some pretty detailed constructive criticism about where I differ with Haspelmath’s program. I have no personal opinions about him.
@Y: yes, that’s the commonly circulated account, but it’s very much a provincial european interpretation (and perhaps even an ashkenazi- or ashkenazi-and-yiddish*-specific one), like many things that are presented as pan-jewish.
maghrebi, kurdish, ethiopian, and other communities have longstanding tattoooing traditions (including facial tattoos), generally shared with their non-jewish neighbors. i don’t know for sure about the communities of eastern north africa, the mashreq, and the arabian peninsula, but there are strong regional tattooing traditions there (egypt, for instance), and i’d assume they’re shared by the jewish communities where they appear. (and i know even less about the various communities of the caucasus-to-central-and-south-asia zone, or the wider sefardi diaspora**)
.
* there isn’t a simpler handle, unfortunately, since the communities east of ashkenaz that are neither sefardi nor karaite don’t adhere to a single minhag (there’s polin, estraykh, and ha-ari (a/k/a sefardi, just to confuse matters))
** including whether maghrebi sefardi communities have shared the region’s tattooing practices, or distinguished themselves from the pre-1492 communities by not adopting them.
Facial, specifically forehead, tattoos are a thing in Ethiopian/Eritrean culture for extremely respectable church-going ladies. Which is a bit initially disconcerting from an American POV where you may be used to normal outlaw-biker tattoos on a variety of surfaces of the body but forehead tattoos go beyond normal limits into full Manson-Family territory. But these are indeed upon closer inspection very respectable churchgoing ladies, including in diaspora in North America. I have a vague suspicion that esp. in Eritrea which is roughly 50/50 Christian/Muslim having a cross tattooed on your forehead may be a socially acceptable passive-aggressive way to signal “no I’m not a Muslim.” But I don’t know if the historical timeline would validate that. The Ethopian/Eritrean churches are of course historically connected with the Copts, who have their own robust tattooing tradition (if you make a pilgrimage to Jerusalem, you often get a commemorative tattoo) that AFAIK tends to avoid the face/forehead.
(FWIW my late first wife had a tattoo, or perhaps palimpsest of tattoos, but no one stopped her from being buried in a Jewish cemetery. Although this may perhaps have not been disclosed to the bureaucratic gatekeepers and I wasn’t the one interfacing with them and for all I know whoever was doing so may have been strategically evasive on the subject based on their own views of what was proper under the circumstances.)
As to his [Haspelmath’s] intrinsic nicitude and sensibiliousness I have no information. I would be inclined to trust bulbul’s opinion on this,
Haspelmath has posted occasionally at the Log. Those also seemed sensible and reasoned. That does not mean I agree. Reasonable people can reasonably disagree with reasonable reasons. (As @Y has done.)
I’m being especially cautious on this topic, because as a Philosophical Skeptic, I have an immediate suspicion of grandiose schemes like Universal Anything. I’m trying to correct for that bias.
I have no ‘history’ with bulbul, as I pointed out. So I have no basis to trust anything. So far the posts here come across as just another over-opinionated blowhard on the intertubes. In the interests of reason I suggested clearing those current deadlines; coming back with evidence.
I should say that accusing a respected academic of “cuckoo bananas”, “madness” plays exactly into ‘over-opinionated blowhard’. So I have already raised the bar for evidence by several notches.
Apparently “you can’t get buried in a Jewish cemetery if you have tattoos” is an urban myth. Even though tattoos are technically forbidden, in general there are no consequences for them (maybe among some Haredis you’d be shunned? I don’t know.) In fact, quite a few publicly heavily tattooed Israelis are getting religion these days.In my mind, the stereotype be right-wing rapper (m.) and influencer (f.) types.
JWB: I’ve now read of Ethiopian Jews, likewise tattooed with a cross in order to blend in with potentially hostile Christians.
So far the posts here come across as just another over-opinionated blowhard on the intertubes.
I wish you would get out of the habit of talking like that. It’s no way to win friends and influence people. (I know, I know, you don’t want friends, you’re just talking straight from the shoulder, you call them as you see them, you don’t take any guff, you take your liquor straight and your coffee black.)
He was cuckoo bananas even before Covid, …
H has a podcast? Will the madness never cease???
And yet you don’t call out that language? Is it merely because H doesn’t participate here? For all we know he might have read bulbul’s posts and decided he won’t show up to wrestle with the pig. In which case we’ve all lost out.
I call double standards. And I’ve by no means cited all the arrogant, presumptuous language bulbul has used on this thread. If you won’t call it out, I do.
Is tlata a loan or are Berber numerals so close to Semitic?
Kabyle, like most Berber varieties, has borrowed all numbers from 3 up from Arabic. (Those that have kept the older system use forms like kraḍ for 3.)
Do any of the original Berber numerals have plausible cognate candidates elsewhere in Afro-Asiatic?
“Two” (sin, sǝn etc) connects neatly with Semitic and Egyptian. 5-9 (sǝmmus, sḍis, sa, tam, tẓa) look rather Semitic, and are suspected of being an older layer of pre-Punic loanwords. “One” (yiwǝn etc) might speculatively be linked to, say, Egyptian w`, but there’s not much to go on.
There are three common terms: רוּחַ rûaḥ, נֶפֶשׁ nep̄eš, and נְשָׁמָה nəšāmâ.
rûaḥ is lit. ‘wind’. nep̄eš is, I presume, metathesized from the root nšp̄: i.e. ‘exhalation’. nəšāmâ is from nšm, i.e. ‘breath’ or ‘inhalation’.
Yes, neat!
Genesis 2:7 KJV
Other versions put ‘being’ instead of ‘soul’.
According to this 18:30 on, early Judaism didn’t believe in a ‘soul’ in the sense of something persisting beyond your last breath. No afterlife; no resurrection; no eternal damnation; no torturing of ‘souls’; no lake of fire.
There are three common terms: רוּחַ rûaḥ, נֶפֶשׁ nep̄eš, and נְשָׁמָה nəšāmâ.
All with Arabic cognates: rūḥ, nafs, nasamah. The first two are both highly productive; nafs forms reflexives in Classical Arabic (like English “self”) as well as meaning something closer to “soul”, while rūḥ forms reflexives in many dialects, and also shows up in formulae like rūḥ al-qudus “Holy Spirit, rūḥānī “spiritual” (or, in Algerian dialect, “ghost”). nasamah shows up in much more limited contexts, like measuring populations – and for “breeze”.
nep̄eš is, I presume, metathesized from the root nšp̄: i.e. ‘exhalation’.
Arabic has tanaffasa “breathe”.
There’s quite a little family of breath-related roots there, now I think about it: nps, nsp, nsm – to which we could add *npx (Arabic nafaxa “blow, inflate”), *npḥ “waft”, and apparently even *nsb and *npṭ “blow one’s nose”. Vaguely tempting to link the np- to *ʔanp- “nose”, but probably not. This is what makes biliteralism so frustrating: any biliteral etymology for one subset of these will inevitably exclude another.
So yeah: all these Semitic soul-words come down to air – invisible stuff that moves, and that moves within you for as long as you’re alive, and leaves you with your last breath.
Kusaal vʋ’ʋs “breathe” is surely from the same root as vʋe “be alive” and vʋ’ʋg “make/come alive”, and similarly throughout Oti-Volta.
The -s suffix of vʋ’ʋs is pluractional. The verb also means “rest”, as in vʋ’ʋsʋm daar “day of rest.”
No overlap at all with souls, though. Nor, come to think of it, with “wind”, which is expressed by a quite different family of words, which are also used for things like blowing a musical instrument.
rūḥānī “spiritual”
Adopted into Persian, also meaning “cleric” and as the surname Rouhani or Rohani, like the former president of Iran Hassan Rouhani, for anyone else who’s trying to remember where they heard that (though I’m sure many Hatters knew the surname’s origin the the first time they heard it).
No, he was on someone else’s podcast once.
Da haben sich zwei gefunden – you two are actually quite alike and maybe deserve each other, eh? I prefer to sit back, relax, and wait for the full presentation of what Haspelmath is supposedly like. Once I’ve read that, I might know enough to form an opinion.
“Clark”, then.
There’s also Hebrew שׁאף šʔp ‘inhale’. It is also, remarkably, ‘aspire’. The meanings ‘breathe in’ and ‘desire’ of the Hebrew are both attested in the OT, and both meanings of aspīrō are Classical Latin.
The hiphilangsci podcast with Nick Enfield is also about how words actually end up meaning anything:
https://hiphilangsci.net/2026/03/01/podcast-episode-56/
Interesting idea that the signifier/signified link is a kind of hypothesis on the part of the language user, based on how people behave in relation to the word(s) in question. (Don’t worry, he doesn’t go full-on behaviourist on us.)
[In relation to Anna Wierzbicka et al’s analyses in terms of semantic primes, he says something like “that’s kinda like their best guess at the meaning.”]
[Rouhani]
“Clark”, then.
I hadn’t thought of that. It has hyponyms “Deacon”, “Priest”, and “Bishop”.
It seems “Hassan Rouhani” is “Beauclerk“.
Da./Norw. geistlig “prelate” < Ger. geistlich “(lit.) spiritual”. Not a surname, though.
Geistlicher “cleric”; not any particular rank, though.
The Semitic ‘breathing’ words seem to often change meaning between ‘inhale’, ‘exhale, blow’, and just ‘breathe’.
*npš is pan-Semitic, including Akkadian. So if *npš < *nšp, as I suggested, that would be old.
Also, do *nšp, *npḫ work as onomatopoeias without the *p fricativizing to [f] as in later Hebrew and in Arabic? Evidently they do, but to my ear more as ‘blow’ than ‘exhale’.
Obsolete English “ghostly father”, confessor.
—Charles, duc d’Orléans (who by the end of his 25 years as a prisoner of war in England spoke English better than French, it’s said)
Text from https://www.poetryfoundation.org/poems/50409/confession-of-a-stolen-kiss
“rhetorical”
In Russian context I’d rather think of something adults often say to kids (and sometimes to other adults):
Look, quel(le) [beautiful, large, …] [girl, bird, …]!
(however, it can be a borrowed construction)
So is Russian seks “sex” a verb for H.?
(and why it is “roots” rather than “words” or “[verbal, etc.] stems” that he calls so?)
Kabyle, like most Berber varieties, has borrowed all numbers from 3 up from Arabic.
Despite abundant evidence that even small numbers can be borrowed (and often have been), the words for “three” and “four” still feature in lists of supposed cognates between Volta-Congo and Dogon and (parts) of Atlantic, and even Mande.
(They’re shared by Bangime, which even Roger Blench thinks is an isolate.)
Hausa seems to have borrowed its word for “two.” An interesting start for a people now famous in West Africa for their awesome mercantile powers. But then, wouldn’t a long-distance merchant be exactly the sort of person most likely to adopt foreign words for numbers?
Hi everyone, many thanks for discussing my “Grammaticon” project! I don’t know how exactly this blog works, but let me try to respond to one thoughtful comment in particular (posted by Y on April 24th):
My feeling is that Haspelmath got stuck with an imperfect idea. His intention is clearly good: there should be clarity on concepts like “word” and “verb” if you are going to compare their properties and behaviors between different languages. However, creating axiomatic definitions doesn’t work. Grammars leak. Edge cases are everywhere, and the closer you look the more you’ll find. No single definition is going to work, and Haspelmath himself is forced to wiggle to explain edge cases, like abstract nouns. That is not to say that comparative linguistics and typology are impossible. Rather, each study needs to define its comparanda precisely. The definition of a “verb” in a comparative semantic study need not be the same as in a study of tense marking or alignment, and so on.
Indeed, “grammars leak”, but we cannot compare them without “axiomatic definitions” of the type used in the Grammaticon (though I don’t call them “axiomatic”). For a long time, I thought much like Y: “No single definition is going to work”. But then I pondered what exactly is meant by “work” – because strictly speaking, any definition *works* as long as it can be applied consistently. Y’s comment about “each study needing to define its comparanda precisely” is right on target, and this is what linguists who engage in broad comparisons do on a regular basis. HOWEVER: In order to define their project-specific comparative concepts, they typically make use of a wide range of other terms that are generally thought to be understood uniformly across the discipline. We are aware of problems here and there, but we do not think of terms such as “copula”, “demonstrative” or “ablative” as requiring project-specific definitions. So the Grammaticon definitions can be thought of as providing a default if someone does not want to invest much into providing their own definition using more basic terms. In many cases, it does not really matter which definition is chosen “in a study of tense marking or alignment”, because researchers mostly focus on uncontroversial core cases anyway (“canonical cases”, as Corbett would call them). So it seems to me that most of my definitions can be adopted wholesale without any issues, though a minority will of course raise questions (or even meet with bewilderment) among my colleagues.
(And as for the Covid discussions, I recommend going back to John Ioannidis’s article from March 2020: https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/. This comes from traditional scientific skepticism and has absolutely nothing to do with “MAGA”. Ioannidis is a leading researcher in the science reform movement – https://www.nature.com/articles/s41562-016-0021 – which I have long sympathized with; see also my comments in a recent issue of “Linguistic Typology”, where I cite Ioannidis (2005): https://www.degruyterbrill.com/document/doi/10.1515/lingty-2025-0037/html)
That introduced me to HARKing (“hypothesizing after the results are known”); I knew it as “the Texas sharpshooter fallacy”, illustrated as shooting at a barn and painting a bullseye around the bullet hole.
John Ioannidis used to be a leading figure in a certain kind of scientific research. However, I think that even then, calling what he was involved with “the science reform movement” was rather tendentious. In any case, he proved himself to be an self-important crank when it came to Covid, as Orac clearly showed.
Martin: I changed the quoted paragraph in your comment to italics for clarity; I hope you don’t mind.
I don’t know how exactly this blog works, …
Welcome here Martin. I haven’t figured it out either, as evidenced by my not infrequently getting black marks from our host.
Let me first respond on the question of levels of (scientific) evidence.
Nobody has claimed Ioannidis 2020 was somehow actively contributing to a MAGA viewpoint. What happened was people who already wanted to see COVID as a conspiracy seized upon I’s piece as evidence for their prejudices. If I is going to weigh in to heated public discourse, he should as a minimum have ‘read the room’ better. To the extent some people took his piece as support for
continuing tonot going out of their way to avoid spread the virus at its most virulent and unvaccinable phase [**], he has deaths on his hands. He has also in my view retrospectively trashed the reputation he had built up for “a certain kind of scientific research”. (Like Brett I’m dubious this amounted to “science reform”.) I suggest you stop citing I’s 2005 papers: it’ll have the effect of undermining your claims/make you seem to be following uncritical groupthink. [***]is damning for anybody whose reputation is built on scientific probity.
I note I is not qualified in Public Health. He should have just shut the fuck up. In Public Health, you need to consider not only the costs of action, but also the costs of inaction. As such the evidentiary testing is for ‘balance of probability’, not ‘beyond a reasonable doubt’. You don’t have the luxury of ‘scientific skepticism’.
Ioannidis has to a limited extent walked back that 2020 piece. Too little too late. He above all should know the dangers of hubris in in sticking to poorly supported scientific claims. What would somewhat rehabilitate his reputation is to acknowledge poor methodology and his unjustifiable language (“fiasco”).
[**] I’m afraid the revision of early 2020 history is well under way. Now that the virus has morphed to be less lethal, and vaccination has achieved ‘herd immunity’ in at least some countries, politicians (even some who were there at the time) have started criticising the response as an over-reaction/as too much of a stress on the economy/as producing nothing but inflation (yes it did, but not “nothing but”).
[***] I also don’t see how the double-blind placebo-controlled experimental methodology that I’s 2005 papers are talking about has any bearing on research in Language Typology. As if we could concoct or go find a language just like this one but lacking ablatives.
… as Orac clearly showed.
@Brett’s reference — see the Who is Orac? link (presumably, and not the Jana at cite 48 here).
It’s true that we cannot test typological hypotheses in the way we can test hypotheses in biology or medicine (by randomized controlled trials), but the science reform movement is just as relevant for claims about cross-linguistic patterns. Linguists are very prone to overgeneralize – not only Chomskyans, who have often generalized from a few languages (or just one) to all languages, but also anti-Chomskyans. Sadly, the only typological paper ever to have been published in “Nature” made strong claims on the basis of just four language families: https://www.nature.com/articles/nature09923. Fourteen years later, the authors asked similar questions again, but this time used a much larger database, reversing their earlier overall conclusion: https://www.nature.com/articles/s41562-025-02325-z. (Interestingly, the new results are more in line with what typologists such as Greenberg and Dryer have been saying for decades), When making inferences from observational data, one is inherently limited because causality cannot be established with any certainty, but there are still big differences between conservative approaches (of the Ioannidis type) and approaches that draw strong conclusions on the basis of limited evidence.
Martin, thanks for your reply. The explanation of the purpose of the Grammaticon (a default definition, for comparative studies) makes it clearer, but still: take for example “A noun is a root that denotes an object.” Why not flesh it out a bit, so one doesn’t need to think too much about whether “vacuum” or “possibility” or even “George” denote objects?
I think my idea of a syntactic/distributional definition riding on clear semantically-defined prototypes would answer nicely, but maybe therre are others which would do as well.
Yes, “a syntactic/distributional definition riding on clear semantically-defined prototypes” is what has been used for syntactic functions (A, S, and P, as discussed in my 2011 paper), but it doesn’t work for noun/verb/adjective, because the syntactic/distributional properties are too diverse. In contrast, A, S and P can be defined in terms of flagging and indexing, which occur quite regularly across languages. And indeed, one wonders whether “vacuum” or “possibility” denote objects. However, given the notion of “shared-core definition” (see §5 of my 2021 paper: https://ling.auf.net/lingbuzz/005489), we can work with comparative-concept definitions where the English words “vacuum” and “possibility” ARE NOT nouns in the comparative sense (= in the Grammaticon sense). They are English Nouns, but not nouns for general linguistics. This is unintuitive, but only because we have the mistaken intuition that language-particular categories should map straightforwardly onto general concepts.
Linguists are very prone to overgeneralize …
My take on the discussion here so far is that everyone is being cautious about generalising even a little bit. A Linguist I am not, but my intuition is that trying to apply ‘noun’ cross-linguistically is already over-generalising — that is, unless the definition is circular or so vacuous as to allow that in English nearly all ‘content’ words are/can be nouns.
conservative approaches (of the Ioannidis type)
Here’s the rub. Go back and read the 2005 paper more critically. Ioannidis is _not_ conservative at all. He does not claim merely Findings go beyond the claimed statistics. He claims they are _False_ [**]. That is, he claims the statistics contradict the Findings. I had hoped the “false” was inserted gratuitously by his publicist, trying to ‘ginger up’ the subject. (“Most Research Findings are not proven” would be pretty feeble, after all. And that would be the generous interpretation of its appearance in the Abstract, that you quote in your 2025 piece.) Alas, “false” appears so frequently, it must be attributed as deliberate.
Why “Why Most Published Research Findings are False”
is Falseis not epistemologically conservative.[**] A jury finding the accused ‘not guilty’ is not at all the same thing as declaring the accused ‘innocent’. It means merely the prosecution has not made its case beyond a reasonable doubt.
I completely agree that saying that “Why Most Published Research Findings are False” is not conservative. What I meant was that it’s good to be conservative in not assuming that a research finding is correct just because one is persuaded by the evidence for it. Too many scientists (including generative linguists) build new theories on top of old theories which are built on top of assumptions plus very little evidence. I also think that it’s a big problem that nowadays a lot of speculation by credentialed experts feeds into politics, in a way reminiscent of credentialed astrology feeding into politics in earlier centuries. But I don’t want to discuss politics further on this site (I now deciphered the pseudonym “bulbul”, and I recall some very unpleasant interactions with him a few years ago).
Martin, I don’t think I understand you. If, within a given language, the word designating ‘possibility’ has the same distributional properties as the word designating the object ‘tooth’, why shouldn’t the two be equally serviceable for cross-linguistic comparison regarding “nouns”? Why can’t your definition be amended to something like “A noun is a root that denotes an object (or a root that distributes like one)”?
English words “vacuum” and “possibility” ARE NOT nouns in the comparative sense (= in the Grammaticon sense). They are English Nouns, but not nouns for general linguistics
Is there any human language in which all nouns refer to physical objects?
Does (to pick at random) Kusaal have an equivalent for “vacuum” or ‘nothingness’/’emptiness’/’absence of stuff’ or similar? Is it a noun — to the extent ‘noun’ can be applied to Kusaal vocab?
I can see treating “possibility” as a noun cross-linguistically would be problematic. There’s an abstract -ible (adjective in English) hiding right inside it. But then posse isn’t English. (Arguably the whole structure is a foreign imposter.) ‘Maybeness’ is alleged to be a word, but seems jocular to me. [all of wikti’s quotes are very recent]
I feel this methodology is like Archimedes asking for a place to stand and a solid fulcrum so that he can move the earth with a long lever. Comparative Linguistics is all out of independent places to stand.
Does (to pick at random) Kusaal have an equivalent for “vacuum” or ‘nothingness’/’emptiness’/’absence of stuff’ or similar? Is it a noun — to the extent ‘noun’ can be applied to Kusaal vocab?
Yep. Ka’alim “nonexistence.” (An unequivocal noun.)
Almost any verb can make a gerund, a noun expressing the action or process or state that the verb describes.
Random non-gerund abstract nouns from the same noun class: daalim “masculinity”, biilim “childhood”, tiraannim “neighbourliness.” Others: nɔŋ “poverty”, gɛɛnmis “madness”, sʋnsa’aŋ “sorrow.”
There’s a robust morphological distinction between nominals and verbs in Kusaal, and the associated semantics are in no way unusual; they differ in some respects from English, so you have quality verbs like gim “be short”, and what would be adverbs in English are a subtype of noun in Kusaal syntactically. But neither phenomenon is particularly unusual cross-linguistically, and certainly it would do grave damage to the clarity of any grammatical description to declare that only nouns referring to physical objects were “really” nouns at all. To do that would be every bit as valuable in Kusaal as in English or Nahuatl or Mandarin: i.e. not in the least.
I have mentioned before that linzug “therefore” can be a verb subject in Kusaal, as in e.g. (from the Bible translation)
Ka linzugʋ da kɛ ka o lieb buudi bɛdegʋ yaab.
“Therefore he became the ancestor of many nations.”
(Literally, “And ‘therefore’ made that he became the ancestor of many nations.”)
But language-internally, this isn’t odd at all. Linzug certainly does mean “therefore”, but Kusaal construes it as a clause-level adverb, not a clause linker, and (as I was saying) adverbs are a subtype of noun in Kusaal. No reason why they can’t be verb subjects if the semantics permit.
(Linzug is analysable: it’s lin “this” with the postposition zug “on, onto, on account of.” The postposition, in turn, is a specialised use of the unequivocally nouny noun zug “head.”)
My question wasn’t totally rhetorical. I don’t actually know of any language in which all nouns refer to physical objects (certainly not Kusaal, nor any Oti-Volta language.) But maybe there is some weird exception out there somewhere which justifies decreeing that a real cross-linguistic definition of “noun” should exclude anything which doesn’t refer to a physical object.
There certainly seem to be weirder languages out there than Kusaal (though perhaps this is partly in the ear of the listener.)
I can well imagine that there may be languages that don’t have a “word” for “death”, for example, but need instead to use some sort of periphrasis involving the word (verb?) for “die.” It occurs to me that this sort of project (to set up a sort of “universal grammar”, in which every word for “death” would be declared to be a kind of abbreviation for such a periphrasis) seems to have a lot in common conceptually with Anna Wierzbicka’s attempts at reducing everything to semantic primes.
“(Physical) object” is a culture-bound notion, incidentally. It’s not an intuitively obvious universal human given.
Is a Kusaasi person’s siig an “object”, for example? If not, why not? Because Western linguists don’t think it is?
I feel this methodology is like Archimedes asking for a place to stand and a solid fulcrum so that he can move the earth with a long lever.
I like that analogy.
I can see treating “possibility” as a noun cross-linguistically would be problematic. There’s an abstract -ible (adjective in English) hiding right inside it.
That’s a different issue. “Luck”, then.
it’s good to be conservative in not assuming that a research finding is correct just because one is persuaded by the evidence for it
well, that kinda rejects the entire scientific method, in favor of, um, what exactly?
the basic alternatives to being persuaded by evidence, historically, are (1) assertions of divine revelation (whether framed explicitly as such or as an insistence that evidence contradicting an axiomatically true theory cannot possibly be accepted), (2) handwaving, or what’s now talked about as “vibes”* (which is similar, but lacks an articulated underlying structure), (3) sleight of hand, usually involving changing the definition of “evidence”, “a research finding”, or “persuaded”.
(the straw-man sentence that follows the line i quoted is, i think, an example of (3), since it substitutes “theories which are built on top of assumptions plus very little evidence” for being “persuaded by the evidence”; the chomskyites it refers to, however, are firmly committed to (1).)
me, i’ll stick with evidence, thanks.
.
* which philosopher-of-popular-music robin james has been theorizing in some brilliant ways for a forthcoming book, with a lot of material on her blog.
Sorry for butting in but I was also struk by “it’s good to be conservative in not assuming that a research finding is correct just because one is persuaded by the evidence for it”. Either someone thinks that the research finding is correct or that it is incorrect, or they are not sure. But in any case, if that someone is persuaded it means that they think that the finding is correct. I mean, there is probably some evidence out there that now is the month of October. However, there is also an overwhelming evidence to the contrary. It’s not true that every piece of evidence has to be definitive either way.
If I understand Martin Haspelmath correctly, his definitions are not additive, but subtractive. For example, the definition of “noun” doesn’t try to identify every noun in every language. If someone wants to argue that luck or siig are nouns in their respective languages, let them. But cross-linguistically only objects are included in the core definition of “noun”. This raises some questions: 1) does each of 6000-7000 languages have a grammatical category to which most words for objects belong and 2) are there any objects the words for which belong to such a class in any language. If the answers are “yes” and “yes” than it is a meaningful definition. If not, then, well, maybe not.