A Grammar of Proto-Germanic

Winfred P. Lehmann

Jonathan Slocum, ed.

VI. SEMANTICS AND CULTURE

6.1. The Culture of the Speakers of Proto-Germanic

The semantic system of a language is closely connected with the culture of its speakers. Segments of the language, like words for the manner of living and for the kinship system, correspond to their way of life. We then are fortunate if we have accounts of the culture of speakers of proto-languages, even though the semantic system must be determined on the basis of the language.

Julius Caesar has included such an account of the Germans around 55 B.C. in Book 6 of his Gallic War. Tacitus presents much the same information in his Germania of a century and a half later. From Caesar's account we can conclude that the Germani, as he calls them, were still largely hunter-gatherers in the first century before our era. Their chief activities were hunting expeditions and military pursuits. When they were at peace they had no overall ruler, but only chiefs of smaller entities that carried out justice. Like many simple societies they welcomed guests, sharing their food and housing with them.

6.1.1. Religion

Our knowledge of their religion or religions is based most directly on information gleaned from the accounts of Caesar and Tacitus. According to Tacitus in his brief section 9, they worshipped especially Mercury, his Latin for Wodan/Odin, Hercules for Thor, and Mars for Tiu. But they did not believe it worthy to enclose them in temples or to fashion images of them like men; rather, they dedicated woods and clearings to them. Caesar had stated that there were no priests, but the statement is assumed to result from his observation that there was no such official rank, because Tacitus refers to priests who functioned in several activities. In section 40 he states that a priest alone may touch the sacred wagon that transports Nerthus, i.e. mother earth, among the Lombards and various other tribes. In section 43 he states that a priest in women's clothing is in charge of a wood that is an ancient shrine. And in section 45 he remarks that the tribe of Aestii venerate the mother of the gods — matrem deum venerantur - and also display the sign of their cult with images of boars; these are to protect them from all dangers.

The equations with Roman gods are maintained in our names of the days of the week: Tuesday honoring Tiu, Wednesday Wodan, Thursday Thor, and Friday, with the goddess Freyja representing Venus. No divine figure is represented in Sunday and Monday, and no one replacing Saturn in Saturday. It has been proposed that these continue beliefs of the Indo-European period, with Odin and Tiu representing sovereignty, Thor representing physical strength, and Freyja substituting for Nerthus representing fertility (Polomé 1989:73-82). Further inferences have been based on Scandinavian rock carvings from the Bronze Age, which imply reverence for the sun, or even a sky-god like Zeus or Jupiter, cf. the Trundholm depiction of a chariot drawing the sun, or other depictions along with ships and human figures.

Unfortunately, there is no description of religious practices in the literary remains of Gothic or the West Germanic languages, and those in the North Germanic languages are late. References in the Eddic poems and other texts may provide only partial retentions from the Proto-Germanic period, and our knowledge of religion and religious practices among Germanic peoples is accordingly scanty.

6.1.2. Economic and Personal Practices

According to Caesar the Germans observed strict social rules. They did not have intercourse before their twentieth year. Many of them regarded sexual abstinence as contributing to stature and strength. Yet their clothing did not cover them well and they associated freely, as in bathing together in the rivers, so that sexual activities might have readily been encouraged. Tacitus is somewhat more specific on their clothing in Germania 17, describing the general costume as a cloak held together with a clasp or even a thorn; the cloak might consist of the skins of animals. Only the richest had underclothing, and this was tightly bound so that it revealed every member. Women were similarly clothed, often however with a linen cloak decorated with purple stripes. Their garments had no sleeves, so that their arms and the neighboring parts of their breasts were exposed.

They did not stress horticulture. The bulk of their food consisted of milk, cheese and meat; Tacitus included apples. Individuals did not own property. Land was assigned every year by magistrates and chiefs when the gentes (tribes) and cognationes (clans) assembled; it then was passed on to others in the next year so that the plebes (common people) would be equal in wealth. Caesar gives further reasons for the practice, including the aim of avoiding encouragement of a preference among the people for agriculture rather than for warfare.

Their settlements were on the edge of the Hercynian forest, which extended far to the east; Caesar states that they had found no German who had gone to the end of it; moreover, he assumed that the end could not be reached in sixty days. He went on to describe three animals that inhabited the area and differed from animals known elsewhere. One, an ox shaped like a stag, is assumed to be the reindeer; two others he called alces, presumably elk, and uri, aurochs. The Germans hunted these for food. They also collected the horns of the aurochs and encased the edges with silver to produce elegant drinking vessels.

For reconstructing the culture of the Proto-Germanic speakers we then must recognize that, even before the time of the first written materials in the Germanic languages, the culture and accordingly the semantic system had undergone changes by influence from other cultures, first Celtic, then Latin, and to some extent Greek through missionaries. Evidence for such a cultural change that may be determined from borrowing is the word for iron, which has various representatives in the Germanic dialects: Go. eisarn, ON járn, ísarn, OE īsern, īren, OHG isarn, īsan, īser. Because of the various forms it is assumed that these were taken from Celtic, as in the Gaulish name Isarno-duru, at various times in the first millennium B.C. Among early borrowings from Latin is the term for donkey, Go. asilus, OE esol, OHG esil, probably from asellus, the diminutive of asinus. Some religious terms may have been introduced before the time of writing and then maintained in the later texts. They often differ from dialect to dialect, so we assume that most of the religious terms and also borrowings from Greek and Latin were introduced into the individual dialects rather than into Proto-Germanic.

6.2. The Kinship System and Family Structure

The kinship terminology that was inherited from Proto-Indo-European indicates that the family was patrilineate, that is, the system was of the Omaha type. The terms for the nuclear family are well attested, as illustrated here with representatives from Gothic, Old Norse and Old English. They indicate that the family system reflected in the terms of the parent language was maintained through the Germanic period to the early dialects. But additional terms, especially in Gothic, suggest that modifications were being introduced in the system.

As demonstrated by Go. heiwa-frauja 'master of the family', the father was head of the family: Go. fadar, ON faðir, OE fæder, cf. Skt pitár-. The reflex of the Indo-European term for the mother is not attested in Gothic, though it is the term in the other dialects: ON mōðr, OE mōdor, cf. Skt mātār-. The term for son is well attested in all the dialects: Go. sunus, ON sunr, OE sunu, cf. Skt sūnú-; and similarly the term for daughter: Go. dauhtar, ON dōttir, OE dohtor, cf. Skt duhitár-, as well as the term for brother, except in Gothic: ON brōðir, OE brōðor, cf. Skt bhrāta-, and for sister: Go. swistar, ON syster, OE sweostor, cf. Skt svasar-.

Terms for father's brother, comparable to the term for 'father', are attested in West Germanic dialects: OE fædera, OHG fatureo, cf. Skt pítṛvyas 'uncle' and OE faðu 'aunt'. The terms for nephew and niece, though not attested in Gothic, are well attested in the other dialects: ON nefe, OE nefa, OHG nevo, cf. Skt nápāt 'nephew' and ON nipt, OE nift, OHG nift, cf. Skt naptī 'niece'. Terms have also been maintained for father-in-law: OSwed. svēr, OE swehor, OHG swehur, cf. Skt Śvaśuras, and for mother-in-law: Go. swaihra, ON sværa, OE sweger, OHG svigur, cf. Skt śvaśruṣ. Reflexes of the Proto-Indo-European term *awos for grandfather are attested, though as in Lat. avus 'grandfather, uncle' also with other meanings, as in ON afi 'grandfather', OE ēam, OHG ōheim < WGmc awa-haima- 'uncle'; cf. also Gothic dative singular awon 'grandmother'.

In addition to these terms, a number of parallel terms are attested that have raised questions regarding their origin and also possible modifications in the family structure during the Germanic period. Among these is the word for mother that replaced the Indo-European term in Gothic: aiþei, attested also in ON eiða, MHG eide. It may well be related to the word for oath, Go. aiþs, ON eiðr, OE āþ 'oath', which has been assumed to be borrowed from Celtic, as in OIr. ōeth. Among assumed reasons for the new word for mother is the introduction of a legal view of the relationship in marriage, possibly through Celtic influence.

While such an explanation is theoretical, terms for master and mistress of the household were introduced beside the relationship terms, as in Go. gardawaldands 'master over the household', heiwa-frauja 'master of the family', in addition to the simple terms: OHG hī(w)o 'husband', hī(w)a 'spouse'; ON hȳski, OHG hīwiski 'household, family'. Among other terms based on PGmc hīwa- 'member of a family', are OE hī-red, MHG hī-rat, NHG Heirat 'marriage'. Unlike the kinship terms, which refer to lineal relationships, these terms have a legal basis, from which we may conclude that marriage had become formalized in Germanic society.

Terms associated with aiþei with an implied relationship through marriage include Go. megs 'son-in-law' and, in the other dialects, a male related through marriage: ON māgr, OHG māg. The implication has increased interest because, when Mary addressed her son Jesus in Luke 2:48, rather than sunus she used the term magus, the term also used for a servant, as in ON mǫgr, OE mago, OS magu, OHG maga-. The cognates for girl have a similar connotation, as in Go. mawi, ON mær, OE meowle. The terms have derivatives, such as Go. magula 'little son', Go. mawilo 'little/dear girl', and also a compound in the Beowulf 2931 gomela iōmēowlan 'aged woman'. It is assumed that these terms were introduced into Germanic from Celtic with the aim of distinguishing the sons and daughters of the immediate family from those of servants. They have no ready Indo-European etymology. Among suggestions on their origin is an unidentified matriarchal society because of the implications noted for the first terms exemplified in this paragraph.

Whatever the source of the additional terms, we can conclude that the family rather than a larger political unit was the basic social group. The basic ties of an individual were to his family, in which he was completely acceptable: Go. frijonds, ON frændr, OE frēond, OHG friunt. A member who had not transgressed against the rules of the household was in good standing: ON hýrr, OE hēore, MHG gehiure, Go. *-hiuri 'friendly'. And transgressors owed some form of recompense: ON bōt, OE bōt, OHG buoz 'compensation'. If the transgressor failed to provide recompense, a sanction was imposed on him, the most serious of which was expulsion from the family. The expelled member was regarded as an outlaw or a wolf: ON vargr, cf. OE wearh, OHG warg 'villain', and he might be freely killed.

6.3. The Household

Tacitus in chapter 16 of his Germania informs us that the Germans did not form villages but rather lived in isolated homesteads. The terminology for dwellings bears this out. The general term ON heimr, OE hām, OHG heim means house or home. It is found only in the accusative plural in Gothic, where haimos in Matthew 9:35 translates the Greek word for village, and in Mark 5:14 the word for country; its two compounds in Gothic, both also in the plural, are translated home: afhaimjai 'away from home' and anahaimjaim 'at home'. Derived from the same root PIE key- 'lie', as is also Go. heiwa- cited above, it has no direct etymon in the parent language, but it is related to OIr. cōim 'dear', Latvian sàime 'family'. Similarly Go. þaurp 'land, lived-on property' is the translation for Gk agrós 'land', much like ON þorp 'farm, estate'; only in the later West Germanic texts does its cognate mean 'village' as in OE þorp, OHG dorf. Go. weihs 'village' similarly translates agrós. To translate the Gk pólis 'city' a word of uncertain origin is used, Go. baurgs; it is also used to translate Gk báris 'tower', comparable in this use to ON borg 'height, wall, castle, city', and OE, OHG burg 'fortified place, castle, city'.

Tacitus' account of isolated houses rather than villages is also supported by the term ON tūn, OE tūn 'farmstead' that in OHG zūn has maintained the earlier meaning 'fence, hedge'. Similarly, Go. gards 'house' with the related verb *bi-gaírdan 'gird' apparently refers to a building on fenced property, as supported by ON garðr 'hedge, garden, court' and OE geard, OS gard 'enclosure'. A new formation in Old Norse, Old English and Old High German, hūs, cf. Go. gud-hūsa 'temple', is obscure in origin, but if derived from PIE kewH-, kū- 'cover' it also indicates a humble building. This statement is supported by the meanings of ON salr 'house, room', but 'ground, soil' in the Vǫluspá; cf. also OE sæl, OHG sal 'dwelling, room'. Other terms support Tacitus' report on underground quarters covered with dung for refuge in the winter as well as for a storehouse: ON kofi, OE cofa 'hollow in a rock, room', MHG kobe 'stall', which are cognate with Gk gúpē 'cave, hut'. The relatively large number of terms seems to suggest that new designations were introduced as the type of dwelling was modified from the early one-room house, probably wattled, lacking windows and furniture. When references were made to villages, a term borrowed from Lat. vīcus was introduced, Go. weihs, OE wīc, OHG wīch. Similarly, as noted above, a designation for city was adapted from a term meaning 'tower'.

6.4. Construction

The indication of simple structures is paralleled by the terms for building. Some of these are based on the word for timber: PGmc tem(b)ra-, ON timbra, OE timber, OHG zimber 'building, material', which in turn is based on the Indo-European root *dem- 'join, construct'. Denominative verbs are attested in Go. timbrjan 'build, strengthen', OE timbrian, OHG zimbaren. But the terms for the builder vary, as in Go. timrja, ON trēsmiðr, OE trēowyrhta, OHG zimbarman, suggesting that the process of more complex building developed independently in each of the dialects. Moreover, the absence of a reflex of the Indo-European word for 'carpenter' and 'artisan', as in Skt takṣan, Gk téktōn, we may assume that there was no technical specialist for building but that every householder built his own structures.

Terms for components of buildings also suggest simple construction. The term for wall, Go. -waddjus, ON veggr is based on the reflex of PIE wey- 'turn, bend as in wattling'; from it we may assume that woven reeds were at one time used to produce walls. The term for roof, ON þak, OE þæc, OHG dah suggests a thatch covering. And the door may have been made of wickerwork, as indicated by Go. haurds, ON hurð 'lattice (door)', cf. OHG hurt 'wickerwork'. Although reflexes of the Indo-European word for door, *dhwer-, are found in OE duru and OHG turi, other terms indicate that the door was primitive, such as OE geat 'gate', cf. ON gat 'hole' and ON hlið 'gate', OE hlid, OHG hlit 'cover'. Terms for lock and key vary in the dialects, so that we assume they were introduced only late, like the terms for window: Go. augadauro, ON vindauga, OE eagduru, OHG augtora. Similarly, the words for devices to eliminate smoke vary, as in ON ljōri 'opening in the roof (for light)', later reykberi, ME chimney, OHG scorenstein. A word for 'hearth, fireplace' is not attested in Gothic, but the Old Norse term arinn is cognate with Lat. āra 'altar'; OE heorþ and OHG herd are cognate with Go. haúrja 'burning coals', ON hyrr 'fire'. These support the Roman writers who described the houses of the Germans as simple rectangular, box-like structures, made of wood rather than bricks or stone. There were apparently several such structures on a homestead, as stated above, and it may have been enclosed with a hedge, making it distinct from other homesteads.

The simple dwellings probably had few furnishings, among them seats and tables, for as Tacitus reports in Germania 22: "each has a separate seat and his own table," cf. Go. sitls, OE setl, OHG sedhal 'seat' and Go. stōls, ON stōll, OE stōl, OHG stool 'chair'. The early word for table: Go. biuþs, ON bjōð 'table, bowl', OE bēod, OHG piot, was later replaced to some extent by ON diskr, OE disc, OHG tisc, from Lat. discus. The various terms for bed, as well as their origins, suggest little more than a dug-out place: Go. ligrs, OE leger, OHG legar 'couch, grave', cf. Gk lékhos, and badi, OE bedd, OHG betti and also ON rekkja, sæing.

6.5. Occupations

As we have noted above concerning the words for 'builder', terms for specialized occupations are late, developed in the dialects rather than in Proto-Germanic. At one time a man might be referred to as a worker in the fields, Go. waúrstwja, ON akrmaðr, OE æcerman, OHG accharman. In another reference he might be referred to as a settled landowner: ON bōndi, OE gebūr, OHG gibūro.

The progression to specialization may be illustrated with the word 'smith'; created in Germanic, possibly from reflexes of the Indo-European root smey- 'work with tools', its basic meaning may have been 'producer' — cf. the Gothic verb *ga-smiþon 'produce': Go. -smiþa, ON smiðr, OE smið, OHG smid. It is attested in Gothic only in the word áiza-smiþa 'coppersmith'. As a simplex in Old Norse it means 'worker in wood or metal', but it is also an element of the compound ljōðasmiðr 'song-smith', as of compounds in OE wīgsmið 'battle-smith' and OHG urteilsmid 'judgment-smith'. Such terms were apparently created to suit a given context but may not have survived to later stages of the language.

Terms for tradesman are late. The general term for 'merchant' postdates contacts with Latin: ON kaupmaðr, OE cēapman, OHG koufman. The term for 'shoemaker' is similarly late: ON skōari, OE skohere; the Old High German term sūtāri is based on Lat. sūtor 'shoemaker'. Other terms are based on Celtic, such as those for 'physician': Go. lēkeis, OE læce, OHG lāchi and the refashioned ON læknare, cf. OIr. liaig 'physician'. Celtic was the source for terms referring to wider authority than that of the father, such as Go. reiki, ON rīke, OE rīce, OHG rīhhi for a man in authority, and similarly for his sphere of influence, as in Gothic dative singular reikistin, ON rīke, OE rīce, OHG rīhhe, cf. OIr. rī, rīg 'king'. Similarly, the terms for servant, Go. ambahti, OE ambeht, OHG ambaht, ON ambātt 'maid', cf. Gaul. ambactus 'servant'. We may also assume that legal arrangements for inheritance were late; the term for it, as well as those for 'heir', is apparently based on Celtic, cf. OIr. orbe: Go. arbi 'inheritance', ON Runic arbija, OE ierfe, OHG arbi; ON arfr, OE eafora, OHG erpo 'heir'. As noted before, the Germans were influenced by the Celts as early as 800 B.C.; if Proto-Germanic terms had existed previously for occupations and situations treated here, we have no direct evidence.

6.6. The Economy

6.6.1. The Numerals

The numerals that can be reconstructed for Proto-Germanic provide some insight into the economy of the period. As the lists below indicate, the cardinal numerals from one to ten can readily be reconstructed. Those for eleven and twelve, as illustrated by the Gothic forms ainlif and twalif, are comparable to those in Lithuanian: venúolika and dvýlika, literally 'one/two left over'; they are clearly innovations in these two dialects. The remainder of the teen numerals as well as those to sixty may well only be innovations in Germanic, e.g. Gothic fimftaíhun 'fifteen' and fimf tigjus 'fifty'. This pattern for the decades was continued in Old Norse, e.g. ellefo tiger 'one hundred and ten'. But for one hundred and twenty, Old Norse has the form hundraþ, making use of hund, which is the basis of the numerals in the other dialects, e.g. Go. taíhuntēhund, OE hund tēontig, OHG zehanzo 'one hundred and twenty'. We may conclude that the Germanic speakers maintained the simple economy of the Indo-European culture for some time, but gradually expanded it, leading also to expansion of the numeral system.

To illustrate the conservatism of the lower numerals, forms from other dialects are given, as for 'four' and 'five', to provide evidence on the position of the accent, which enabled application of Verner's Law. For Germanic, only forms of the masculine are given for the numerals 'two' and 'three'. Details on the development of the individual forms are left to the grammar of the dialects.

The ordinal numerals are for the most part based on the cardinals, and consequently provide no additional information on Germanic culture.

6.6.2. The Way of Life of the Germanic Peoples

In keeping with the statements of Roman overseers about the life of the Germans, they had terms for the common animals that were continued from Proto-Indo-European, as representative terms in the dialects and their Indo-European etymon indicate, e.g.:

• ON kȳr, OE cū, cf. cow, OHG kuo, PIE gwōus 'cow';
• ON ær, OE ēowu, cf. ewe, OHG ou(wi), PIE owys 'sheep';
• ON sȳr, OE sū, cf. sow, OHG sū, PIE sūs 'pig';
• Go. swein, ON svīn, OE swīn, cf. swine, OHG swīn, PIE swīnos 'pig', cf. sūs;
• Go. aihʷa-, ON jór, OE eoh, OHG ehu-, PIE ékwos 'horse';
• Go. gaits, ON geit, OE gāt, cf. goat, OHG geiz, PIE ghaydos 'goat';
• Go. hunds, ON hundr, OE hund, cf. hound, OHG hunt, PIE kwōn 'dog'.

But the presence of these names may not indicate that herds or flocks were kept, even though the word for property was maintained: Go. faíhu, ON fē, OE feoh, cf. fee, OHG fihu. Evidence to the contrary may be preserved in the different forms for a herd of sheep that are attested in the various dialects, for these suggest independent developments in the several dialects: Go. awēþi, OE ēowde, OHG owiti, ewit. It is also curious that these words and others have a final dental, which may have a collective meaning; among the others is the word for dog, and also those for deer: ON hiǫrtr, OE heorot, OHG hiruz and for horned animal: OE hrīðer, OHG hrind.

Among the most interesting of the words with such a dental suffix is that for salt: Go. salt, ON salt, OE sealt, the suffix is lacking in other dialects, cf. Gk háls, halós, Lat. sal, Toch. A sāle, OCS sol. A cognate is lacking in Indo-Iranian, suggesting that the term was taken into these languages after the speakers of Indo-Iranian had separated from them. Archeologists have determined that the last millennium before our era was a time of economic expansion centering on development of metals such as iron and on the use of salt. The source of the dental suffix has not been identified, though it has been tentatively ascribed to Illyrian on the basis of the tribal name Soudinoi, which is connected with *sūs 'pig'. But little is known of this group other than its location in the Soúdēta, that is the Harz mountains of southern Germany and the neighboring area.

As metals came to be used, so apparently was the lost wax process of casting, in which a wax model is melted out from a mould to be replaced by bronze. Terms for bees, wax, and honey that have no secure Indo-European etymon were added to the Germanic vocabulary at this time. These include the word for bee: ON bȳ, OE bēo, OHG bīo; the word for swarm of bees: OE ymbe, OHG imbi; the word for wax: ON vax, OE weax, OHG wahs; the word for honey: ON hunang, OE hunigk OHG hona(n)g. Gothic has a word for honey, miliþ, that has been derived from Indo-European, although the original word was apparently *medhu, as in Skt mádhu 'honey' and Gk méthu 'wine' (Lehmann 1986:255-256); the Gothic word has parallels in Gk méli and Lat. mel as well as adjectival reflexes in the other Germanic dialects, so that the modified form has been assumed to have been introduced into the western dialects. Efforts have been made to determine Indo-European sources for these, such as Pokorny's relation of the word for wax to the Indo-European root *weg- 'weave'; but it is the only derivation from the root with an -s- formant, and the meaning is difficult to justify with regard to that of the root. Whether or not such efforts to determine their sources are accepted, the Germanic words were most likely added in the first millennium before our era. They then reflect the gradual shift from an economy based on a hunting-gathering culture to one of increasing settlements with domestic animals and more complex technology.

6.7. The Plant World

In view of the close association of the Germans with nature, a large set of names for plants and their products may be expected. Among the several names for tree is ON tré, OE trēo(w), and OS treo; the Gothic cognate is attested in weina-triu 'vine' and suggests that the word may have implied a smaller tree, or even a branch of wood as by its cognate Gk dóru. The Gothic word for tree, bagms, has varied cognates in the other dialects, such as ON baðmr, OE bēam, OHG boum; these have been accounted for in various ways, such as modification of the -g- to ð in Old Norse and its loss in the other dialects. A further word, ON viðr, OE widu, wudu, OHG witu has been explained as indicating trees that provide a boundary, supported by assumption of its derivation from PIE widh- 'separate'.

The word for root is well attested, as by Go. waurts, ON ort, OE wyrt, OHG wurz. The Gothic word for branch asts has cognates only in OS and OHG ast. The Old Norse word for straw, halmr, has cognates in OE healm and in OHG halm.

In view of this relatively large set of words referring to trees, and of their attestation in many dialects, we assume that earlier forms of their names existed already in Proto-Germanic and that its speakers were closely involved with trees and their products. The trees in the areas inhabited by Germanic speakers as well as those of Proto-Indo-European have been admirably documented by Friedrich (1970). The names will be briefly noted here.

Among the earliest to become widespread after the recession of the glaciers is the birch: ON bjǫrk, OE beorc, OHG birka. The name is also attested in other dialects, as in Skt bhūrja-, and in Lithuanian terms such as bìrštva 'birch forest'.

Another early tree is the pine, to which a number of names were given, probably for different varieties. A term that is well attested is ON fura, fūra, OE furhwudu, OS furia, OHG fur(a)ha, so that an etymon is assumed for Proto-Germanic. If the first syllable of Go. faírguni 'mountain' is related, as by some scholars, the term also has a representative in Gothic, but the source of faír- is among the most disputed in the older language (cf. Lehmann 1986:104-105 and Friedrich 1970:133-140). A cognate is assumed in the first syllable of OHG fereh-eih, the word for a specific kind of oak tree. Other cognates are assumed in the Latin word for oak: quercus and in the first syllable of the Celtic name of the Hercynian forest, which then may have been an oak forest.

A general Germanic word for the oak tree is attested in ON eik, OE āk, OHG eih(ha); it is assumed to have been applied to a variety of the oak tree in the mountains but then generalized.

One term for the maple is based on the root *kel-; it has various suffixes, such as -n- in OHG līn-boum, ON hlynr and OE hlyn. A different term is found in Danish ær corresponding to Lat. acer, and to an adjectival formation in OHG ahorn.

The word for aspen is attested in OE æspe, OHG aspa, which is assumed to be metathesized from the Indo-European form that has reflexes in OPruss. abse and Lett. apse.

One term for elm is OE wīce, cf. NLG wīke; it is also attested in western Indo-European dialects and has an -n- infix in Lith. vìnkšna. A second term is attested in Italic and Celtic, e.g. OIr. lem, as well as in Germanic, where it has a variety of vowels, as in ON almr, OE ulmtrēow, and MHG ilm.

Several names were given to the willow, one based on its sheen: ON selja, OE sealh, OHG salaha and its adjectival cognate salo 'gray'. The OHG falawa is related to OHG falo 'fallow'. ON vīðir, OE wīþig, OHG wīda on the other hand are based on the root *wey- 'weave', reflecting one of the uses of its branches.

These are the names for the earliest trees that in Specht's view came to be introduced after the receding of the glaciers, so that there are terms for them in representative Indo-European dialects. To them he adds terms for the apple and nut trees. Those for apple are attested in Crimean Gothic apel, ON epli, OE æppel, OHG apful, cf. OIr. aball; the more specific terms for the apple tree are attested in ON apuldr, OE apuldr, OHG apholtra. The apple has further interest because Tacitus in his Germania stated that the food of the Germans consisted primarily of wild apples — agrestia poma — fresh venison, and curdled milk.

The terms for nut and the tree: ON hnot, OE hnute, OHG hnuz have been the subject of dispute: a likely cognate, though with a different suffix, is found in Lat. nux. Citing Middle Irish cnú, Specht assumed that Latin and Germanic have the same root with different suffixes. In any event, etyma for the terms cannot be found in Proto-Indo-European.

Two other trees, the beech and the yew, as well as their names are clearly later, as indicated by their forms as well as by archeological evidence. The names for the beech belong to the -o/ā- stems, Go. bōka, ON bōk, OE bōc, OHG buocha; often considered to provide evidence that the home of the Indo-Europeans was in North Europe, because of its distribution west of a line extending from the Baltic Sea to the Black Sea, it no longer is credited for such an argument because cognates refer to different trees, e.g. Gk phēgós 'oak'. The names for the yew are reflexes of PGmc īwa-: ON ȳr, OE īw, OHG ēow. Cognates in other dialects refer to different trees, e.g. Russian iva 'willow', so that the name cannot go back to Proto-Indo-European for the yew; it is assumed that the basis may have been a color term because it is the term for alders in Lithuanian.

While cognates of terms for trees can at least be noted in adjacent dialects, those for the cultivated plants have characteristic forms in Germanic. Oats is generally assumed to be the earliest cultivated grain. The term in Old High German and Old Saxon, gersta, differs in vocalism from Lat. hordeum and even more so from Gk krī́thē; it also refers to barley. We may conclude that there was no continuity among the Germanic speakers from the period of Proto-Indo-European in sowing and reaping the grain. The current German word, Hafer, is a recent borrowing from Low German; it contrasts with OHG habaro. The Old English word barley, OFris. ber, is cognate with ON barr 'grain' and also with the Go. adjective bareinans ('prepared of barley'; acc.pl.); the Latin cognate far refers to spelt, Russian bor to millet. The word for wheat is attested in all the Germanic dialects: Go. hʷaiteis, ON hveiti, OE hwǣte, OHG (h)weizi; it was apparently coined in Proto-Germanic to distinguish the whitish grain from the others. In noting the variations in the references of the names of grains, we may assume that they existed in late Proto-Germanic, and that the speakers were acquainted with the plants at the time, but as Caesar and Tacitus imply they did not cultivate them.

Words for flower (Go. blōma, ON blōmi, OHG bluoma, but not attested in Old English), seed (OHG sāmo), and bud (OHG kīmo) were apparently coined in late Proto-Germanic or even in the early dialects from verbs, cf. OE blōwan, OE sāen, Gothic keinan. Another word for seed, ON sāð, OE sǣd, OHG sāt, is found also in Go. manna-seþs 'mankind'. Earlier forms of bud, and cognates, are attested only in the medieval period. Like the words for grains, they seem to have been coined either in late Proto-Germanic or the dialects.

Similarly, words for small plants are unclear in etymology, such as Gothic gras, Old Norse gras, Old English græs, Old High German gras. The same is true of the words for moss, Old Norse mōsi, Old English mos; cf. also Old High German mīos, Old English mēos. Similarly the etymology of the word for reed is unclear: Gothic ráus, Old English reyrr, also hrēod, Old High German rōr.

The variations among these words and difficulties with their etymologies support the characterization of Caesar and Tacitus that agriculture was not practiced by the Germanic speakers even after the beginning of our era. Only when a characteristic word like that for wheat, or when specific trees were identified for definite uses, do we have reliable evidence that the term in question may be credited to Proto-Germanic.

6.8. Nature

The terms for the prominent items of nature directly reflect those of Proto-Indo-European, although like the word for sun they may have undergone various changes; these may be examined in etymological dictionaries. The word for the sun is represented by Go. sauil, ON sōl, OE sygil, Runic sugil, which have cognates in many dialects, e.g. Lat. sōl. An alternate form, Go. sunno, ON sunna, OE sunne, OHG sunna is assumed to reflect an earlier l/n stem.

The Gothic word for moon, mēna, is comparable to Gk mḗnē, but that in the other dialects varies in form: ON māni, OE mōna, OHG māno. A derivative in Proto-Indo-European from the root *mē- 'measure', it was also used in the early languages as the basis of the word for month; but in Germanic an extended form took over this meaning, e.g. Go. mēnōþs, ON mānaðr, OE mōnaþ, OHG mānōd. The word for star also continues the Indo-European base, but with an added suffix, as in Go. stairno, ON stjarna, OE steorra, OHG stern, sterr, cf. Avestan star-; the -rr- forms are accounted for by assimilation.

The word for light is based on PIE *leuk-, but differs in suffixation in the Germanic dialects as also elsewhere, as in the forms: Go. liuhaþ, ON ljōs, OE leoht, OHG lioht, but also ON ljōmi, OE lēoma. There is one word for day throughout Germanic: Go. dags, ON dagr, OE dæg, OHG tag; it then can be reconstructed for Proto-Germanic as daga-, but has no cognates elsewhere. It is probably a development from PIE dhegʷh- 'burn'.

The word for darkness, Go. riqiz, ON røkkr has cognates in other Indo-European dialects, e.g. Skt rájas 'dust', but not in the West Germanic dialects, which have words of uncertain origin like OE deorc and OHG tunkal. The word for night is found in all the Germanic dialects: Go. nahts, ON nátt, OE niht, OHG naht, as well as widely elsewhere, usually without the t-suffix in the nominative, e.g. Skt nak, Lat. nox.

The word for year is found in all the Germanic dialects: Go. jēr, ON ār, OE gēar, OHG jār, as well as in other dialects, but with ō-grade, as in Gk hôros. But the words for the seasons are less general. That for winter is newly introduced in Germanic from a root indicating the wet season: Go. wintrus, ON vetr, OE winter, OHG wintar. The word for summer is attested in ON sumar, OE summer, and OHG sumar; it has a cognate in Armenian amar̃n. But words for the other seasons are not general. For example, OE hærfest and OHG herbist are based on a root indicating 'reap', while NE fall refers to another aspect of the season. By contrast the English word spring indicates the emerging of plants, while NHG Frühling (early season) has much the same connotation.

Two words for 'water' are found in the Indo-European dialects: Hittite watar, an r/n stem with variant nominative forms, e.g. Go. watō, ON vatn, OE wæter, OHG wazzar; and Latin aqua with cognates elsewhere. The words for fire are highly interesting, as noted in section 3.A.1, because two are attested in the dialects: one originally indicating active fire, as in Lat. ignis, Skt Agnís '(god of) fire'; and the other a neuter indicating fire as a state, as in Hittite pahhur, Gk pũr, and Go. fon, ON funi, OE fỹr, OHG fiur. As contrasting terms, they are residues of the time when Indo-European was an Active/Stative language. The word for snow is also general, although strongly modified in Greek, accusative singular nípha, Lat. nix from the reconstructed PIE snoigʷhos: Go. snáiws, ON snær, OE snāw, OHG snēo.

Still other words for natural objects might be cited, like those for stone and sand, but those noted here provide excellent examples of names that have survived from Proto-Indo-European, though often with modifications in the forms generalized in the various dialects. Nonetheless the variation in inflection among them is obvious. These have led to the view that the Indo-Europeans viewed themselves as the center of all objects in space (Specht, 1947:10, 334 ff.). Such a position is unusual for a linguist, if of interest. We would be more inclined to the view that these words are maintained from the time when there was no inflection so that they consisted of roots or bases. Then suffixes may have been added in accordance with classes, such as -r to indicate non-humans. And in time inflections were introduced to mark syntactic relationships, as in Proto-Indo-European.

6.9. Words for Transportation

The words for transportation have been of great interest for their implications on the spread of the Indo-Europeans. They center about the wagon and the horse. Because extended travel was difficult without wagons with wheels, the terms for wheel have been thoroughly examined for possible indications of the time of the spread. One of these is the PIE kʷekʷlós as in Gk kúklos, Skt cakrás, ON hvēl, OE hwēol, hweowol, based on PIE kʷel- 'drive'. Another is Lat. rota, Ir. roth, ON rath, OHG rad, based on PIE ret(h)- 'roll', from which Skt rátha- 'wagon' is also derived. While these are based on roots, as illustrated, they inflect according to the thematic declension, and accordingly it is difficult to state that they are early.

The same is true of the words for wagon, ON vagn, OE wægn, OHG wagan, and Skt vahana 'vehicle, which are based on the root PIE wegh- 'move' and of the words for yoke, Go. juk, ON ok, OE geoc, OHG joh, cf. Skt yugá-, Lat. jugum, which are based on PIE yeug-. Other words for parts of the wagon have been transferred from items or processes, such as the nave of a wheel from navel and the axle, originally the designation of the shoulder. The terminology for the primary means of transportation then reflects a late Indo-European period. In his massive study of the names of wheeled vehicles and of the horse, Ivanov places it at the beginning of the third millennium B.C. (1999:167-236).

The name of the horse he associates with its domestication. It was long hunted for food, but on the basis of archeological finds it is assumed that domestication should not be dated before the early third millennium. The widespread term in the Indo-European dialects, as in Gk híppos, Lat. equus, is attested in ON iōr, OE eoh, and in the name of a thistle in Go. aihʷa-tundi. Names like horse or NHG Pferd that were later introduced do not concern us here. The presence of reflexes of the standard Indo-European name, however, indicates that it was taken into Proto-Germanic at an early stage, at a time when its speakers were still located in the assumed center in southern Russia and the Ukraine.

With their implications concerning the settlement of the Germanic peoples in northern Europe, the names for transportation and its means accordingly have special importance in providing clues on Germanic culture and its development.

6.10. Conclusions on the Bearing of the Semantic Structure of PGmc
for Evidence on the Culture of the Speakers

From the semantic system that we can reconstruct for Proto-Germanic, we assume that the speakers maintained an advanced form of a culture stressing the hunter-gatherer lifestyle. The terminology for the household suggests simple wooden dwellings. These consisted of a habitation for humans as well as sites for storage; any structure for cattle may have been incorporated in the human household. They may well have been surrounded by some sort of fences or hedges.

The households were apparently inhabited by individual families, with alignment in clans and tribes but without any more general political units. The terms for occupations indicate prevalence for hunting. But the homesteads also kept domesticated animals, certainly cattle and sheep, and possibly also pigs. Horses were probably also included, though transportation with wagons may have relied heavily on oxen. Among plants and trees, oats may have been cultivated, but the people relied on wild apples for fruit, as Tacitus reports, and other foods such as nuts may also have been gathered.

As the speakers came in contact with other cultures they adopted advanced forms of weapons and, later, modifications in their religion. The adoption of iron as they came in contact with Celtic speakers must have provided greatly improved weapons and tools. The identification of their native gods with those of the Romans, as still apparent in names of the week like Wednesday corresponding to Mercury's day, implies modifications in worship. By the time of the translation of the Bible into Gothic, their culture and the lexicon to represent it no longer reflected the culture and lexicon that we reconstruct for the Proto-Germanic period.