*dem-
‘to build’
- Language
- Proto-Anatolian
- Class
- Base
- Grammar
- verb
- ID
- 244
1. Individual Anatolian Languages
1.1 Hieroglyphic Luwian
*/tam-/
‘to build’1.1.1Derivatives
1.2 Cuneiform Luwian
*dam-
‘to build’1.2.1Derivatives
1.3 Lycian A
*tem-/m̃m-
‘to build’1.3.1Derivatives
1.4 Lydian
tam-
‘to build’1.4.1Transmission
This lexeme is attested twice in different inscriptions from Sardis, LW 2 (344(?) BCE) and LW 10.
[D.S.]
1.4.2Forms
| 1sg. pret. ind. act. | tamν | LW 2.6 (Sardis), LW 10.6 (Sardis) |
1.4.3Meaning ‘to build’
The word tamν is convincingly identified as a verbal form by Gusmani 1960a:284. The approximate meaning ‘to build’ is suggested in Gusmani 1964a:208 based on the context in inscription LW 2, which refers to parts of the grave built by the owner. Furthermore, he suggests convincingly an etymological connection with Gk. δέμω and HLuw. /tama-(di)/ ‘to build’.
[D.S.]
1.4.4Stem
Without an attestation of the 3rd person singular, it is difficult to determine the exact stem class of the verb tam-. A verb in -i- can be discarded, because the 1st person singular ending would be -idν, but other stems remain as possibilities. Nevertheless, an athematic root stem is the most likely hypothesis from an etymological perspective; cf. under reconstruction.
[D.S.]
1.5 Hittite
*tam-
‘to build’1.5.1Derivatives
1.5.1.1Uncertain Derivatives
- damnaššara- c. (deities) (HEG:85f., Melchert 2001d, Groddek 2009a:51f., Simon 2011c:350)
2. Reconstruction
2.1Reconstruction ‘to build’
The etymological connection of the Lydian form tamν proposed by Gusmani (Gusmani 1964a:208) has been accepted so far; cf. Melchert 1993a:252, Melchert 1997a:32, Gérard 2005a:101. Its stem formation remains obscure, although an athematic verb is here tentatively posited. The verb is surely accented on the first syllable. Although one expects a nasalized vowel <ã> or <ẽ> for an accent vowel in front of a nasal, there might be some exceptions, for which the reason remains unknown. On the other hand, Eichner 1986a:11 explains the verb as follows: “Enklise einer schwachtonigen Verbalform?”, which is also a possibility. Until further evidence, these hypotheses must be considered very tentative.
The verbal root is here reconstructed without laryngeal, thus *dem- instead of *demH-. The spelling with the sign in Hitt. dam-me-e-da (KUB 4.5 + KBo 12.73 iii 16, LNH), should not be taken as reflecting a real geminate /mm/ and, thus, as reflecting an assimilation of the sequence *-mH-, since the sign in New Hittite has also the value <dax> (Melchert 1991b:126, Kloekhorst 2008c:413, Nikolaev 2010a:58-60).
[D.S.]
3. Proto-Indo-European
3.1Semantic Reconstruction
Because no exact morphological analysis of Lyd. tam- is available (cf. Part I above), no comparanda are listed and, therefore, no semantic reconstruction can be given.
3.1.1Root etymology
The Anatolian as well as most of the non-Anatolian verbal formations derived from the root *dem- recommend the meaning ‘to build’, cf. the IE transmission above. The most prominent exceptions are PToch. *tsǝm- ‘to cause to grow’, on which cf. Winter 1962a:26–28, and PGerm. *tema- ‘to suit, be fitting’. However, Frisk (GEW I:364) doubts the identity of the Germanic and Greek (δέμω ‘to build’, cf. above) verbs because of the semantic differences. Additionally, the Middle Iranian languages attest a further exception to the basic meaning ‘to build’, cf., e.g., (Late) Khot. va-diṃ- ‘to make’. Because of the recent attestation and the trivial semantic development,Cf. Buck 1949a:537–539 referring to Yoshioka 1908a who p. 12f. mentions, e.g., Gk. ποιέω ‘to construct, build; make’.these formations have no value for the semantic reconstruction. Thus, the meaning of PIE *dem- can be approximated as ‘to assemble, make fit, build’.
Based on the data we have today, it is still not possible to determine whether PIE *dem- is a (telic) aorist root or an (atelic) present root, cf. LIV²:115 n. 2 suggesting that the thematic present developed from a prospective subjunctive of a root aorist.
If the connection with PIE *demh2- ‘to (op)press, (late-PIE) tame’, as proposed s.v. PAnat. *dmah2-s- ‘to (op)press’, turns out to be valid, a more basic meaning has to be reconstructed, viz. ‘to press (on), cover vel sim.’. PIE *dem-, therefore, originally may have denoted the procedure of applying clay to a wall (earthen plaster) and, hence, may give a hint at the building techniques of the Proto-Indo-Europeans.
Benveniste 1955a separates the root noun PIE *dom-/dem- ‘house’, the verbal root *dem- ‘build’, and the root PAnat. *dmah2-s- ‘to (op)press’.CÉG 2:156 s.v. δέμω, Bjorvand & Lindeman 2000a:984, and others reject this view.The affiliation of the latter one no doubt is a matter of debate. Benveniste’s argument in favor of segregating the root noun and the verbal root is PIE *dom(o)- denoting the ‘house’ not as a building but as a social unit. However, this view is problematic because, as Buck 1949a:133f. states, the IE words for ‘family’ are either identical with or derived from terms for ‘house’, Lat. domus ‘house; family’, PCelt. *tego-slou̯go- ‘household, family’ (OIr. teglach ‘family’, cf. eDIL s.v. teglach), a compound consisting of PCelt. *tego- ‘house’ and *slou̯go- ‘retinue’, cf. EDPC:376f. Therefore, it can be argued that the meaning ‘house, building’ is primary which allows for maintaining the affiliation of PIE *dom-/dem- ‘house’ and PIE *dem- ‘to build’.
3.2Morphological Reconstruction
As long as the morphological analysis of Lyd. tam- ‘to build’ remains tentative (cf. above), no morphological comparison and reconstruction can be given.
3.2.1Root etymology
The most discussed topic regarding the root in question is its auslaut, cf., e.g., Hackstein 1995a:138f., Nikolaev 2010a, Steer 2015a:145 et passim. Some attested formations point to a seṭ-root, while others involve reconstructing an aniṭ-root. PIE *dm̥‑ped‑o- ‘(house‑)floor’, PIE *dems=poti- ‘lord of the house’ etc. presuppose an aniṭ-root (PIE *dem-),It has to be borne in mind, however, that seṭ-roots can develop into aniṭ-roots secondarily, cf. Greek γέ-γα-μεν with γα- from PIE *g̑n̥[h1]-.while Ved. dāra- (m.), dārā- (f.) ‘wife’If < *dm̥h2-r-o-, cf. Luw. dammara-.and Dor. νεόδμᾱτος ‘newly built’ etc. point to a seṭ-root (PIE *demh2-).The seṭ-root variant, however may be due to the generalization of a suffixal (factitive or collective) *-h2-, cf. Nikolaev 2010a for explanations of the presumed seṭ-reflexes.Thus, some scholars prefer to reconstruct PIE *demh2- (LIV²:114f., GEW I:364, DELG:262, Sihler 1995:104, Lipp 2009a:xv) and some opt for *dem- (GED:150, Adams 1988a:41, EWAia I:697, Nikolaev 2010a) ‘to assemble, build’. Most recently, Nikolaev 2010a treated this subject thoroughly by collecting and discussing the evidence for and against the seṭ-character of PIE *dem(h2)-.Nikolaev 2010a:58 n. 9 does not adopt the theory of laryngeal-induced gemination of resonants in Germanic (cf. Lühr 1976, Müller 2007a:88–92) and, therefore, dismisses the possible phonological evidence from PGerm. *tema- ‘to suit, be fitting’. However, if the sound law PIE *V́RHV > PGerm. *V́RRV is accepted, the Germanic evidence (*tema-, not geminated †temma-) speaks against a seṭ-root PIE *demH-.His findings are convincing to a high degree and suggest an aniṭ-root, PIE *dem-.On the possible connection with the root 2.*demh2- ‘to tame, domesticate’ (LIV²:116f.), cf. PAnat. *dmah2-s- ‘to (op)press’.
[A.O.]
4. Transmission
PIE
PIE root noun *dóm-/*dém- ‘what has been built, house’ (Schindler 1972a:32, 36, Gamkrelidze & Ivanov 1995a I:646, Mallory and Adams 2006:221, Olsen 2010b:47, LIPP II:158–161),Lindner 2011a:40 reconstructs an amphikinetic root noun PIE nom. *dóm-s, gen. *d(m̥)m-és based on Arm. town (< *dṓm), gen. tan (< *dm̥m-és), cf. also Meier-Brügger 2010a:350.cf. PIE gen. sg. *dém-sYAv. gen. sg. nəmo (Vd.15.43) < *dm-és as well as Arm. gen./dat./abl. sg. tan < *dm̥m-és (cf. below) attest a secondary zero grade stem *d(m̥)m-, cf. Schindler 1972a:32, Kellens 1974a:393.in PIE *dems=poti- ‘master of the house’ (cf. below); Ved. gen. pl. damā́m (hapax, RV.10.46.7) (KEWA II:18, EWAia I:697, Grassmann 1996a:577), OAv. loc. sg. dąm (Y.48.7, Y.49.19), YAv. loc. sg. dąm-i (Bartholomae 1904:683f., Kellens 1974a:392–395);Av. loc. sg. dąm(-) points to original long vowel in the PIE loc. sg. *dĒm, cf. Martínez & de Vaan 2014a:13 §7.6.2. Such a lengthened grade is found in other Avestan locatives, too, cf. Reichelt 1909a:169f., Martínez & de Vaan 2014a:44, differently ALF 1996:144.(?)Gk. nom./acc. sg. δῶ n. ‘house’ (Il.+) (GEW I:428f., DELG:304, EDG:363, Latacz 2009a:146 ad Il.1.426), (?)Myc. do-de /dō(n)de/ ‘to the house’ (DMic I:185f.);In Homer, Gk. δῶ is found always at the end of the verse. The only possible nominative form is attested in Od.1.392. Hesiod (Th.933) attests acc. pl. δῶ, which is most likely an ad hoc-usage, cf. GEW I:429. PGk. *dō, however, cannot represent PIE *dōm (< nom. sg. *dom-s, cf. Mayrhofer 1986:176, acc. sg. *dom-m via Stang’s Law, cf. EDG:362, Mayrhofer 2004a:12, Meier-Brügger 2010a:350), because PIE *dōm would result in Gk. †δών, cf. Gk. nom. sg. χθών ‘earth’ < *dhg̑hōm. Schmidt 1889a:222, however, regards it as a sandhi phenomenon. For an alternative solution, cf. Le Feuvre 1997a attributing masculine gender.
Gk. δῶ could also be a local particle cognate with PGerm. *tō ‘to’ (cf. OHG zuo), cf., e.g., DELG:304. Similarly, LIPP II:149 n. 10, 159 n. 5 considers Gk. δῶ and δῶμα to be “Nomina departiculativa”, but cf. the objections by Le Feuvre 1997a:217–220. Cf. also Lejeune 1976a.
Gk. δῶμα n. ‘house, home, temple’ (Il.+) is secondarily generated, possibly from the root noun’s acc. sg. *dō̆m-a (< *dō̆m + -m̥), cf. already Grdr² II.1:136, GEW I:429. As a parallel, cf. Gk. divine name Ζεύς ‘Zeus’ which developed a (nearly) complete paradigm starting from the accusative Ζῆν(α), cf. (?)nom. sg. Ζήν (hapax, A.Supp.162), gen. sg. Ζηνός (Il.+), dat. sg. Ζηνί (Il.+).Arm. n-st. nom./acc. sg. town ‘house’, nom. pl. town-k‘, acc./loc. pl. town-s, (Olsen 1999a:117, 134, EDAIL:618).The nom./acc. sg. town, instr. sg. tamb, and instr. pl. tambk‘ can be lautgesetzlich continuations of the root noun; on the remaining forms gen./dat./loc. sg. tan, abl. sg. tanē, gen./dat./abl. pl. tanc‘, cf. Olsen 1999a:134 with n. 263.On Celtic, cf. Weiss 2017a.
(?)PIE compound *dm̥-ped-o- ‘(house-)floor’ (NIL:529, 537 n. 63), cf. Gk. δάπεδον ‘level surface, floor, ground’ (GEW I:347f., DELG:252, EDG:302, LSJ:369., LfgrE II:219f.);Beekes (EDG:302) is not sure, whether this lexeme belongs here or is a non-IE loan.ON f. topt, tomt ‘building place’, (O)Sw. f. tom(pf)t ‘id.’ (GED:150f., Bjorvand & Lindeman 2000a:949–951)Germanic provides an ō-stem (< *dm̥-ped-eh2-) and secondary i-stem forms, cf. on this NIL:537 n. 63 with reference to Lühr 2000a:314f. However, Bjorvand & Lindeman 2000a:949–951, taking the feminine gender of ON topt, tomt into consideration, suggest an abstract noun *dm̥-ti- “bygging, konstruksjon”.
Cf. also Lith. dim-stis ‘court’ (ALEW I:212), possibly OIr. déis ‘vassalry’, and less likely Lat. domesticus ‘of the household’ (Weiss 2017a:70–77), which show the same first member as PIE *dm̥-ped-. The second compound member is derived from PIE *steh2- ‘to step, position (oneself)’ (LIV²:590f.), hence Lith. dim-stis ‘(Germ.) Haus-stand’. However, we would expect †nimstis in East Baltic (cf. Lith. nãmas ‘house’, Latv. nàms ‘id.’, cf.PIE *dómo-), whereas Old Prussian place names like Dymsteyne, Dymsteines provide anlauting d-. Additionally, Lith. dim-stis is attested exclusively in “Preußisch-Litauen”, which makes it highly probable that this word is borrowed from Old Prussian, cf. ALEW I:212.
On Gk. δάμαρ, δάμαρτος ‘wife’, cf. Luw. dammara-.
(?)PIE collocation *dems=poti- ‘lord of the house’, cf. Ved. dám(=)pati- ‘id.’ (KEWA II:20, EWAia I:699, Grassmann 1996a:577),Cf. also inverted Ved. pátir dán ‘master of the house’ and Ved. śíśur dán ‘child of the house’ (hapax, RV.10.61.20; EWAia I:697). Hence, it can be deduced that Ved. gen. sg. dán ‘of the house’ could still be used independently. However, according to Lindner 2011a:39 n. 19 inverted Ved. pátir dán is artificial.
The dual form Ved. dám=patī ‘husband and wife’ was reanalyzed as a dvandva-compound. Thence, OInd. dam- ‘woman’ was extracted by grammarians, cf. EWAia I:699 s.v. dámpati-.OAv. gen. sg. də̄ṇg(=)patōiš ‘lord of the house’ (hapax, Y.45.11) (Bartholomae 1904:684, Kellens 1974a:392);Cf. also Av. nmānō.pati- ‘master of the house’ (Bartholomae 1904:1093), whose first compound member (nmānō° < *dmāna-) is a derivative of PIIr. *dam(a)- ‘house’ (cf. PIE *dómo-, Nikolaev 2010a:72f.). Therefore, and because of inverted Ved. pátir dán, the collocation PIE *dems=poti- ‘master of the house’ has to be regarded as still reanalyzable even in post-PIE times. Most likely, it is not completely univerbated in late-PIE, cf. the notation “idg. *déms *póti-” in EWAia I:699. However, we would expect Szemerényi’s Law yielding PIE gen. sg. *dēm < *dem-s. The concept of a patriarch being the head of the PIE household is confirmed by collocations like Hitt. parnaš išḫa- (KUB XXXIII 62 II 18) ‘master of the household’, Lat. pater=familias ‘id.’, cf., e.g., Mallory and Adams 2006:207f., Kuz’mina 2007a:59f.Gk. δεσπότης ‘master (of the house), lord’ (Pi.+),Myc. do-po-ta ‘master of the house’ (DMic I:188f., Bartoněk 2003a:183, ) is difficult to explain, cf., e.g., LIPP II:160 with n. 14.
On the substitution of *(dems=)póti- by the ā-stem (δεσ-)πότης, cf. Schwyzer 1939a:451.δέσποινα f. ‘mistress of the house, lady’ (Od.+) (GEW I:370f., DELG:266f., EDG:319, LSJ:381, LfgrE II:255)On Arm. tēr ‘master’ (?)< *ti-ayr (cf. ti-kin ‘mistress’) < *dems=ayr, cf. Olsen 1999a:676f.
(??)PIE *dom-u- ‘house’,Arm. tanow-tēr ‘master of the house’ could be indicative for a u-stem, but it is morphologically and phonologically problematic, cf. Olsen 1999a:674f. with n. 21.
Whether Ved. dámūnā- ‘master (of the house)’ (KEWA II:19f., EWAia I:698f.) and Gk. δμώς, δμωός ‘slave, servant’ (Il.+) (< *dm-ōu̯-, GEW I: 402f.) are secondary derviatives of this u-stem is a matter of debate.cf. (?)Lat. domus o/u-st. f. (but cf. PIE *dómo-), PSlav. *dȏmŭ- m. u-stem ‘house’, cf. OCS domŭ, Russ. dom, Cz. dům, SCr. dȏm (ESSJa V:72f., EDSIL:112, 113)
PIE them. pres. *dém-e/o- ‘to assemble, build’,Pace LIV²:114, HLuw. /tama(di)-/ does not continue PIE “*démh2-e-”, cf. PAnat. *domáH-i̯e/o-.cf. Gk. δέμω ‘to build, construct’ (GEW I:364, DELG:261f., CÉG 7 (2002):117f., EDG:314f., LSJ:378); PGerm. *tema- ‘to suit, be fitting’< *‘to join, (ModHG) sich fügen’, cf. Heidermanns 1993:594f. s.v. -tǣmi-. Cf. also (?)PGerm. *tōma- ‘disposable, available’ (Heidermanns 1993:598).(EDPG:513, GED:150f., Kluge & Seebold 2011a:1010); (?)TochB inf. tsämtsi (THT 240 a2, tsamtsi (THT 31 a5) (Malzahn 2010a:984–987, Adams 2013a:804, Hackstein 1995a:136–140, Peyrot 2013a:843)TochB 3rd sg. mid. tsāmtär does not exist, cf. Peyrot 2013a:843 n. 1033.
The anlaut ts- (instead of †ś-) is intraparadigmatically leveled from TochB Kaus. tsäm- pres. VIII, tsemtsa s-pret. (III) ‘to cause to grow, increase, promote’. The alternation 〈ä〉 ~ 〈a〉 seen in the infinitives is due to different chronological layers the texts belong to (archaic-II vs. classical), cf. Peyrot 2008a:39.
(?)PIE s-aor. *dēm-s-/*dem-s-, cf. Gk. 3rd pl. ind. s-aor. ἔδειμαν ‘built’ (Il.+)LIV²:116 n. 13 considers it to be secondary.(GEW I:364, DELG:261f., EDG:314f., LSJ:378, LfgrE II:246); TochB pret. III tsemtsa ‘caused to grow’ (Malzahn 2010a:984–987, Adams 2013a:804, Hackstein 1995a:136–140, Peyrot 2013a:843)Cf. LIV²:116 n. 14 and, on the epenthetic -t-, Pinault 2008:449.
Middle Iranian
(Late) Khot. va-diṃ- ‘to make’, pa-dīm- ‘to make, build’, Khwar. (’)w-z’m (f.) ‘building’, (’)nd’m- ‘to be solid, steady’, (’)nd’my- ‘to make solid, steady’According to Benzing 1983a:66, Khwar. (’)ndm- ‘to be solid, steady’ goes back to Iran. *ham-dam-. A seṭ-root is, pace Cheung 2007a:55 (“*damH1”), not overtly reflected in Iranian.
We have to reckon with a high amount of contamination of the roots Iran. 2.*dam- ‘to tame, compel, (Russ.) укрощать, приручать, принуждать’ and Iran. 3.*dam- ‘to create, make, (Russ.) создавать; строить’, cf. ESIJ II:321–324, esp. 323.(Cheung 2007a:55 s.v. *damH1, Benzing 1983a:65, 66, 415, 416, 660, Bailey 1979:374)
Greek
Gk. μεσόδμη ‘tie-beam’ (Od.+), Att. μεσόμνη (IG II2.1668, 347/6 BCE [6x], SEG 34.122.35, ?333/2 BCE)On the phonology, cf. Schwyzer 1939a:208.(GEW II:213f., DELG:262, 688, EDG:935, LSJ:1106, LfgrE III:156)There are various possibilities how to explain Gk. μεσόδμη: (1) < *medhi̯o-dm-eh2- ← *medhi̯o-dm-o- ‘belonging to the middle of the house’ (cf. GEW II:213f.) or (2) *medhi̯o-dm̥h2- ‘middle structure’ (cf. EDG:935), (3) *medhi̯o-dmeh2- ‘built in the middle’ (cf. DELG:262, LfgrE III:156), and (4) *medhi̯o-dmeh2-, representing the weak stem of an h2-formation, ‘middle of the construction; construction in the middle’. However, laryngeal loss in compounds makes the first solution most likely. It could be recent, though, cf. Dor. νεό-δμᾱ-τος ‘newly built’ etc.
Gk. nom./acc. sg. δέμας ‘bodily frame, body’ (Il.+), dat. sg. δέμαϊ (Pi.Pae.6.80) (GEW I:364, DELG:261f., EDG:315, Nikolaev 2010a:64–68, LSJ:378, LfgrE II:244f.)
Gk. perf. δέδμημαι (Il.+) ‘is built’, Dor. 3rd pl. perf. δέδμανθ’ (Theoc.15.120), Myc. part. fut. de-me-o-te /demehontes/According to LIV²:115, the Mycenaean form continues a PIE desiderative, cf. similiarly Nikolaev 2010a:76–78.(GEW I:364, DELG:261f., EDG:314f., Nikolaev 2010a:81–87, LSJ:378, LfgrE II:246)
Gk. °δμητος ‘built’, cf., e.g., ἐύδμητος ‘well-built’, νεό-δμητος (MAMA VII.157.8, undated, IG XIV.2508.8, undated), Dor. νεόδμᾱτος ‘new-built’Cf. also νεοδμής (h.Ap.231+), νεόδμητος (E.+), νεόδμᾱτος ‘newly tamed’ which could have served as a model, as Nikolaev 2010a:80 n. 94, 81 n. 100 hints at. Analogical influence by Gk. νέμω, perf. *nenmē- (> νενέμη-) is possible, cf. Nikolaev 2010a:84–87.(GEW I:364, DELG:262, EDG:315, Nikolaev 2010a:78–81, LSJ:1168, LfgrE II:770)
Gk. δμῶς ‘slave (taken at war)’ (GEW I:402f., DELG:289f., EDG:343, LSJ:441, LfgrE II:322–324)
Celtic
(?)MIr. dét ‘disposition, habitude’, MW pl. deint ‘temper, character’
eDIL s.v. 3 détCf. Nikolaev 2010a:60–63 (< *dm̥-to-), but Schumacher apud LIV2+ s.v. 1.*demh2- oder *dem-? with doubts.
For an overview on more Celtic forms, cf. Weiss 2017a.
Germanic
(?)PGerm. adj. *tōma- ‘disposable’, cf. ON tómr ‘empty’, OE tōm, OS tōm ‘id.’If this adjective belongs here, the meaning can be associated with PGerm. *tema- ‘to suit, be fitting’, esp. in contexts with containers, cf. Heidermanns 1993:598. Differently, Bjorvand & Lindeman 2000a:948 rejecting the paraphrased proposal (ibid. 949).
Heidermanns 1993:598f., EDPG:520, Bjorvand & Lindeman 2000a:948f.
Tocharian
TochB Gv. tsäma- pres. III or pres. IV, subj. V, pret. I ‘to grow, increase, come into being’, TochA tsämā- pres. III, subj. V, pret. I ‘id.’
TochB Kaus. tsäm- pres. VIII, subj. ICf. Peyrot 2013a:843.or subj. II,Cf. LIV²:114.pret. III ‘to cause to grow, increase, promote’, TochA Kaus. pres. VIII, subj. IX, pret. II ‘id.’Winter 1962a:57 was the first to connect the Tocharian verb with PIE “*dem-, ±‘(auf)bauen’”.
It is debatable whether PToch. *tsäm- ‘to (cause to) grow’ points to a seṭ-root (*demh2-) or not (*dem-), cf. Hackstein 1995a:139, Malzahn 2010a:986f. Peyrot 2013a:843 does not reconstruct an a-root for Proto-Tocharian.
Malzahn 2010a:984–987, Adams 2013a:804, Hackstein 1995a:136–140, Peyrot 2013a:843, CEToM s.vv. tsämā- (TA), tsämā- (TB)