On the semantics of the Proto-Indo-European roots *mel-, *men-, and *steh 2 : from the external-positional to the internal-cognitive perspective 1

. The present discussion, which is articulated within the theoretical framework of cognitive semantics, aims to reconsider the homonymy between the Indo-European roots *men - ‘to think, to have in mind’ and *men- ‘to delay, linger, remain’. We should instead imagine the existence of a single archetype, whose different semantic values represent the developments of a metaphorical shift from a concrete to an abstract meaning, exactly as in the case of the historical products of IE *steh 2 - ‘to stay’, and * mel - ‘to delay, linger, remain’; ‘to think, take care of’, which convey meanings related to both ‘stasis’ and ‘reflection/knowledge’.

While from a formal point of view there do not seem to be particular difficulties in tracing the words back to a possible common prehistoric antecedent, problems instead emerge in relation to the semantic aspect, as noted by the Author. The words in fact convey fairly disparate meanings.
The formal tracing back of the two Anatolian forms to the common origin *steh 2is supported, as noted above, by fairly plausible, linear explanations (cf. HED 2: 463 and Kloekhorst 2008: 413): as observed by Giusfredi himself (2016) 6 , in fact, the lack of haplology in the suffix -nāna-suggests that ištanāna-may represent a new formation derived from an ancient theme in nasal comparable with forms such as Cretan Doric στανύω 'I collocate, position', Lat. dēstināre 'to stop, bind, secure', OCS stanǫ, Arm. stanam 'I place myself, position myself' (< *stǝ-nā -IEW: 1005) etc. According to the scholar, in essence, we are dealing with an adjectival expansion of the zero-grade of *steh 2 -, from which a neutral substantive in -n-was built (later thematized in -a). The form ištanzan-7 , which instead, as said, means 'soul, mind, spirit' (as well as 'will, intention'), could, in his opinion, be traced back to the Indo-European base *steh 2 -through a previous *sth 2 -ent-i-on-8 .
Despite the apparent semantic divergence with respect to the series just mentioned, nor can we rule out that išta(n)ḫ-may be etymologically connected with the aforementioned forms, especially if it is interpreted -in light of the reasoning proposed by Pozza (2014 and2019) 11 -, as a verb that expresses the act of experiencing/knowing through the senses, within a more general picture of the bodily dimension of cognition 12 . Experience -in this case "gustatory" experience -in fact plays a fundamental role in the structuring of cognitive processes and, consequently, in that of the subsequent associated linguistic manifestation. At this point, the relation of išta(n)ḫ-with the remaining three verbs in the series displays, by virtue of the observations presented by Giusfredi (2016) in relation to ištanāna-'altar', and ištanzan-' 'soul', a further, double, corroboration.
On the one hand, in fact, according to what was already hypothesized by Eichner (1988), the meaning attested in Hittite represents the result of a shift, followed by semantic shrinkage, which, from the generic sense originally conveyed by *steh 2 -'to stand', led to the more specific one of 'den Geschmack, oder den Zustand oder ähnlich feststellen' (ivi: 143), or, again, of 'zu sich nehmen, bei sich behalten' (HEG A-K: 421), up to the documented "kulinarisch-technischer Sonderbedeutung" (Eichner 1988: 143) that the Hittite verb išta(n)ḫ-actually expresses.
On the other hand, if we analyse the problem in the theoretical framework of cognitive linguistics, following, among others, the work of Lakoff -Johnson (1980) and Lakoff (1993: 235 ff.) 13 , one could reinterpret the sense historically conveyed by the verb išta(n)ḫ-'to taste, savour' as that of 'to (re)cognise through the senses, appropriate' and, therefore, 'to come into contact with; know', exactly as in the case of Lat. sapere, in which the sense of taste is connected not only with perceptual experience as such, but also with more general and abstract mental experience (cf. also Sweetser 1990: 36) 14 .
Basically, words indicating position, positioning in space, lingering, staying, etc. seem to be connected -not only formally but also due to semantic contiguity -with words indicating the act of knowing (see also infra, § 2), of coming into contact through the senses and through proximity with the object of knowledge, by virtue of the fact that the notion of 'knowing' in the pure, ahistorical state, freed from historical, religious and cultural implications, seems to be absent from many historical linguistic traditions 15 . Belardi (1976) had already noted that being in a position in 11 Refer directly to the Author's works for the etymological discussion of the forms in question and for the conclusions reached regarding the probable common origin of these words. 12 The reference literature is vast. We will thus merely mention, as representative, Varela -Thompson -Rosch (1993), Lakoff-Johnson (1999), and Gibbs (2005). 13 See the metaphors relating to the interpretation of ideas such as food, perception as reception and as contact between the perceiver and the perceived. Cf. Pozza (2014 and2019) for details. 14 "The sense of taste here is evidently connected not merely with general experience of perception, but with mental experience as well". 15 Cf. Belardi (1976). See in particular the brilliant discussion (ibidem) of words such as Lat. superstitio 'superstition' but, originally 'knowledge of the truth' (*uper-steh 2 -'to be above'), Gr. ἐπιστήμη 'knowledge, skill' (< *epi-steh 2 -'to be above'), OEng. understandan 'to understand, comprehend' (< *n̥ dher-steh 2 -'to be below'), OEng. far-standan 'to defend', but also 'to intend, comprehend, acknowledge' (< *pr̥ -steh 2 -'to be in front of'), OIr. (h)ires(s) 'faith, credo' (< *iriss < *eriss < *peri-steh 2 -'to be around'), am(a)ires, amaras 'unbelief' (prefixed with the privative *n̥ -), the Pahlavi name of Avesta, apastāk 'knowledge, religious knowledge', to be intended as *upa-steh 2 -ka-, 'being near'. relation to the perceived object represents something very similar to our 'to know', something that is expressed as a spatial relationship between two physical entities: the object of knowledge is that on which our attention "lingers". The perceptual metaphor "to experience is to taste "16 -of which the Hittite verb išta(n)ḫ-represents an example -completes the picture presented here and offers support for the hypothesis that it is etymologically correlated with the series derived from *steh 2 -, a root which, as noted, is continued, in historical Indo-European languages, mostly by words indicating knowledge 17 . The connection between the "material" cognitive process (the source domain) and the "mental" cognitive process (the target domain) is identified, in the specific case, with the well-known and widespread metaphor "ideas are food/ to experience is to taste" (among others, Lakoff 1993 andForesti 2002) 18 .
The two roots are generally quoted as different headwords (see LIV, s.v.: 1. *men-'einen Gedanken fassen'; 2. *men-'bleiben, warten'), while however not ruling out some possible identity (cf. already IEW: 729 *men-'bleiben, (sinnend) stillstehen' = men-'denken'?) 23 , based on a relationship that is not only formal but also semantic. In particular, according to a line of reasoning already proposed by Carruba (1986), we could establish a connection not only between Lat. mora and memor 24 , but also between Gr. μένω and μέμονα, according to the parallelism: 'I wait, linger', 'I waited, I lingered', so thus 'I think, I have in mind'. A peculiarity of an abstract nature such as that of 'to think' would therefore be connected -as a logical consequence -with the concrete act of 'stopping', 'delaying', as we will see shortly.
Unlike Bomhard (2004), who considers it necessary to isolate a further root *men-with the sense of 'to desire ardently, with passion' -by virtue of specific meanings conveyed by attestations such as Toch.B mañu 'to desire', Toch.A mnu 'spirit, desire', Skr. manyú-'spirit, ardour, zeal, passion', Gr. μενοινάω 'to desire ardently, crave', OIr. menn-'desire', menme 'desire', OHG minna 'love' etc. -, we consider it more economical here, as well as clearer from a semantic point of view, to include these forms within the same semantic sphere conveyed by the root for 'to think' 25 . The act of thinking repeatedly, with intensity, in fact, presupposes, by metaphorical extension, the concept of desiring, a concept which, moreover, is part of the polysemy of many terms such as Gr. μένος 'vital spirit, force (of spirit), Skr. mánas-'mind, intellect, will', man-'to think' (but also 'to hope, desire') etc.
The same reasoning could be extended -as indeed Rieken effectively does (1999: 51) -to another Indo-European root, *mel-, to which some attribute the meaning of 'to delay, linger, hesitate' (IEW: 720; EDG: 927; Rieken 1999: ibid.), others those syllabic nasal *n̥ (*me-mn̥ -> mema-, as in katta-'down'< *k/k̑m ̥ tm ̥ etc.), see in particular HEG (L-N: 189) and, above all, the recent discussion presented by Melchert (2019). See also, on this subject, the literature cited by Kimball (1999: 252-253). Cf. Dardano (2018: 365 ff.) for the hypothesis that in the case of the Hittite expression (-za) … ZI-ni āppa mema-'to speak from the bottom of (his) soul/mind' we are dealing instead with a loan translation from Akkadian (where no verb for 'to reflect' exists either, but where expressions such as itti libbi-šu qabû 'to speak with his own heart', itti libbi dabābu 'to speak to his own heart', hence 'to ponder, think' are documented). 21 Cf. Sturtevant (1933: 133), Melchert (1984: 100). For semantic aspects, see Pedersen (1938) and Jasanoff (2003: 128 ff.): "the development from 'stand fast' to 'stand firm' and 'refuse' hardly requires comment". 22 The vocalism of the Latin verb is considered problematic by LIV (437: note 6a): perhaps a reduced degree of *m e n-or a formation from a *mon-ē-with delabialization after *m°, or, instead, from a *mn̥ (n)-eh 1 -(for reference bibliography and objections, see LIV and LIV Add.: s.v.). 23 See also Buck (1949: s.v. 'remain, stay, wait'). 24 Despite the doubts of DELL: 396. 25 The etymological dictionaries also reconstruct a third homonymous root *men-with the meaning of 'to rise up, stand out' (IEW: 726), from which Lat. mentum 'chin', mōns 'mount', Av. mati-(< *mn̥ ti-) 'mountain outcrop', OCorn. menit 'mountain', Wel. mynydd id., MIr. moned, monad (found in toponyms) etc. In any case, the reconstruction of this root does not prejudice the reasoning proposed here, since it would be, in this case, a formally homonymous archetype, but semantically distinct. of 'to worry, think, have in mind' (LIV Add.: *mel-'nachedenken, sich sorgen';Serangeli 2016 26 ). Both meanings attributed to the root in question are preserved in Greek, where the verb μέλλω (< *μέλ-jω, with subsequent analogical extension of the form with the geminate to the rest of the verbal paradigm and to the noun derivatives) indicates the idea of 'to linger, be late, delay' (from which, 'stop thinking', according to HED M: 21) as well as that of 'being on the point of', while μέλω/ μέλομαι means 'to take care of, think about; be at heart, be an object of thought'.
In Latin, this stem would seem to be continued by remelīgō 'idler', prō-mellere 'litem promovere', in Old Irish by mall 'slow'. In the Anatolian sector the outcomes are represented, in Hittite, by the noun māl-'thought, mind, spiritual force' 27 and by the denominative verb mala-/malāi-'to have in mind, to meditate; to approve' 28 , in Cuneiform Luwian by the verb mali(ya)-, mal(ā)i-'to consider, think' and by the noun mālī-'idea, thought' (see Serangeli 2016).
On the other hand, as already observed by Puhvel (HED M: 21), the relationship between Hitt. māl-'thought, mind, spiritual force' and Gr. μέλω/μέλομαι 'to take care of, think of, be important to, be the subject of thought' reminds us of that between Gr. μένος 'vital spirit, force (of spirit), courage' and Gr. μένω, μίμνω 'I remain, I stop', or of that (cf. also supra) between Lat. memor 'which has in mind, remembering' and Lat. mora 'pause'.
Therefore, it would seem we may not rule out a possible semantic development, of a metaphorical type, which goes from the concrete to the abstract (cf. infra § 2 for the details), and which led, also in this case, from the meaning of 'to wait, delay' to that of 'to meditate, think', as already anticipated by Rieken (1999: 49-51) 29 . In Greek, both μέλλω 'to linger, delay, play for time' / 'to be on the point of', and μέλω/ μέλομαι 'to take care of, think of; to be important to, to be an object of thought', therefore, could be traced back to a single and common root *mel-30 , able to convey both the meanings subsequently documented historically.
The idea that the act of pausing or lingering generates reflection and thought could therefore be formulated both for *men-and for *mel-, both, therefore, polysemic in origin. The outcomes of these roots, in fact, display both the concrete and abstract meaning. 26 According to Serangeli (2016) one should imagine that this root conveyed the sense of 'to think', since this meaning would be perfectly able to explain both the forms documented by Anatolian and by the Greek verb μέλω/μέλομαι, which then developed the two specific meanings of 'to take care of' and 'to be the object of someone's thoughts'. 27 Cf. Rieken (1999: 49): "Denken, Geist(esstärke), Verstand". For CHD (L-N: 124) it referred to "a quality desirable for men in combat, such as boldness, ferocity, skill", but also "something which is known or recognized". [-la?] ma-a-al UL šakti 'He knows not for himself even a little mal, but courage has been given to him tenfold. The child whom the …-s beget for themselves, you do not know the mal of it [either]' . 28 Cf. Otten, Bronzetafel 20: kuin-za imma DUMU-an ID LAMA malāizzi 'whatever son Kuruntas has in mind'; KBo XVIII 48 Rs. 18: n-at-za mān malasi 'if you agree with it' (cf. HED M: 25-26). 29 Cf. ivi: 51: "ich will noch warten → ich will noch denken". Contra, Puhvel (HED M: 21): "Rieken […] placed the cart before the horse by postulating a reverse semantic development 'wait' > 'think'". According to Melchert (1994: 169), who hypothesizes a proto-Anatolian dissimilation of */n/ in */l/ near a nasal, "Hitt. māl-is 'inner strength' < *'mental force' from a root noun to *men-". 30 The etymologies of μέλω and μέλλω are generally considered obscure, but various scholars think that they may be related (see IEW: 720,EWG: 196;DELG and GEW,s.v., and the bibliography contained therein), despite the different meanings (see EDG: 927; 929).

Final considerations
In conclusion, if we contextualize the hypothesis here proposed within a more general framework of semantic change, it could be observed, as Traugott (1982 and1990) and Traugott -Dasher (2002: 94-96) have shown, that verbs related to the physical sphere frequently evolve into verbs of speech acts or mental state, which is precisely what seems to happen from 'to stand' and 'to linger' to 'to think, to have in mind'. The semantic change here proposed falls within a general trend according to which meanings "based in the external described situation (positional)" change into meanings "based in the internal (evaluative / perceptual / cognitive) described situation" (ivi: 94). This tendency, explicitly described by Traugott (1990: 500) and Traugott -Dasher (2002: 94) as "Tendency I", subsumes many semantic changes from concrete to abstract, most especially from physical to mental (among the examples offered by the scholars, we remind OE felan 'touch' > 'experience mentally', or agan to 'have for' > 'obligation,' 'ought'). Yet, according to Sweetser (1990: 31) "there is a general tendency to borrow concepts and vocabulary from the more accessible physical and social world to refer to less accessible worlds of reasoning, emotion, and conversational structure" 31 . "Tendency III", according to which "meanings tend to become increasingly based in the speaker's subjective belief state/attitude toward the situation" is, according to Traugott (1989: 34-35;1990: 500) and Traugott -Dasher (2002: 94), the dominant one, fed by other tendencies. Concerning this latter kind of semantic shift, the scholars remind Langacker's (1986: 467) words: "Whereas the basic meaning profiles physical motion by an objectively-construed mover, namely the subject, one (unprofiled) facet of the extended meaning is abstract motion by a subjectively-construed mover, specifically the conceptualizer. The pivotal factor in this type of semantic shift is therefore subjectification".
Particularly, the theoretical scheme adopted here provides, in light of cognitive semantics, the semantic link between the source domain, represented by 'stasis' and 'sensory perception', and the target domain, represented by 'knowledge', which manifests itself linguistically through verbs which, by indicating gustatory perception or lingering with respect to the nearby object of knowledge, denote a more general and abstract cognitive activity. In both cases, the connection with the gnoseological sphere thus develops from the experiential domains that envisage either the material incorporation ("embodiment") of the object of knowledge (Hitt. išta(n)ḫ-) or proximity to it -lingering in its observation, we could say -(Lat. superstitio, Gr. ἐπιστήμη, OEng. understandan, far-standan, Lat. maneō : mēns, Gr. μένω, μίμνω : μένος, Toch. mäsk-: mañu etc.).
In fact, most Indo-European languages (see the table below) display outcomes characterized by both the semantic values attributed to the two "traditional" radical 31 See also Traugott (1982: 253), even if specifically focused on grammaticalization processes: if a meaning-shift in the process of grammaticalization occurs within a component, it is more likely to involve "less personal" to "more personal" than the reverse. Sweetser (1990: 31) underlines that Traugott's propositional level (the "less personal") corresponds fairly close to what she defines "sociophysical level", and that Traugott's textual/expressive level (the "more personal") coincides partially with her epistemic level. The same argues Traugott (1990: 500), quoting Sweetser: "Tendency I also subsumes the tendency to use vocabulary from the external (sociophysical) domain in speaking of the internal (emotional and psychological) domain". archetypes of *men-, which would therefore seem to confirm the historical continuation of the polysemy which we consider here to be original. *men-/mon-/mn̥ -'think, have in mind' *men-/mon-/mn̥ -'delay, linger, remain' Latin: mēns, mentis 'mind', meminī 'I remember, reflect', moneō 'I make think, remind, warn', mōnstrāre 'to show' etc.
It will be noted that, except for the Germanic and the Balto-Slavic sectors (which document only outcomes of the root *men-'to think', and instead continue other IE roots for the meaning of 'to delay, wait'), most ancient Indo-European languages testify both of the meanings attributed to the two roots traditionally considered homonymous, which leads us to consider it highly probable that we are actually dealing with a single root capable of "accepting" within it the semantic shifts hypothesized here.
The fact, finally, that one can think that an IE root such as *men-/mon-/mn̥ -was originally polysemic and that, therefore, conveyed both the meaning of 'to think, have in mind' and that of 'to delay, remain', supports the parallel interpretation that sees in *mel-a similar polysemy, showing a further "experiential" connection between (initial) stasis and (subsequent) reflection. It is no coincidence, in fact, that for the latter root the major etymological dictionaries reconstruct either the semantic value of 'to delay' (IEW: 720; EDG: 927) or, alternatively, that of 'to think' (LIV, Add., s.v.), or consider the reconstruction of the main meaning of certain historical outcomes to be problematic (cf. GEW: 203 "Da der konkrete Begriffskern von μέλλω unbekannt bleibt, sind alle Erklärungsversuche hypothetisch"). The same polysemy would seem to occur, as we have seen (cf. § 1.1), also for *steh 2 -, which displays outcomes characterized both by the value of 'to stay' and of 'to think', mostly in the presence of localistic preverbs. The bodily model in fact represents -as already observed in the ethnolinguistic studies of G.R. Cardona 32 -the primary point of reference around which the subsequent abstract concepts would take shape and develop: it is language itself that shows how important spatial references are to us (it is through denomination that the real appropriation of space is perceived), and it is precisely through language that the universal psycho-physical model may be filtered.