Revisiting Greek ἄχθος
The etymology of Greek ἄχθος, ἄχθομαι ‘burden, to be loaded, to be vexed’ is uncertain. It has been compared to Hittite ḫatk-ⁱ ‘to shut, to close,’ with derivatives appearing to connote ‘tight, pressing.’ Its analysis as a disyllabic root, *h₂edʰgʰ-e-, is also problematic since PIE roots were, in most cases, monosyllabic. The present paper will consider the possibility that Greek ἄχθος is a compound consisting of two separate PIE roots, one of which is *dʰéĝʰ- ‘earth’ as attested in Greek χθών.
Homer and Hesiod’s use of ἄχθος ‘burden, weight’
The Greek phrase here translated as “… burden on the earth” is ἄχθος ἀρούρης, where ἀρούρης is literally ‘arable land.’2
The phrase here translated as “huge as a man could lift” is ἀνδραχθέσι. LSJ glosses this word as “loading a man, as much as a man can carry.”4
Here “little does its weight burden him” translates ὀλίγον τέ μιν ἄχθος ἐπείγει.
This translates φέρε δ’ ὄβριμον ἄχθος ὕλης ἀζαλέης…
“put too great a load on your wagon” translates ἐπ’ ἄμαξαν ὑπέρβιον ἄχθος ἀείρας.
Homer, Hesiod, and later authors also employed ἄχθος, ἄχθομαι and related forms to indicate the concepts: hated, vexed, pained, or grieved, but these uses are considered to be metaphorical derivations by Chantraine (2009: 143–144). With respect to both nominal and verbal forms of this Greek word, Frisk (1960: 200–201) considers them “nicht sicher erklärt.”
Modern Lexicographers
Risch (1964: 78) proposed an etymological connection between Greek ἄχθομαι and Hittite ḫatk-ⁱ ‘to shut, to close.’ This has been accepted by Puhvel, Beekes, Kloekhorst, and LIV (doubtfully).8 Beekes suggests that the semantic value ‘shut’ in Hittite must have arisen from the original concept, ‘to squeeze.’ The proposed root is * h₂edʰgʰ-, which Kloekhorst calls an “awkward looking form.” The vast majority (perhaps originally all) of PIE roots are monosyllabic and this, along with the semantic divergence, raises questions about the etymological correspondence.9
An Alternative Analysis based on PIE *dʰéĝʰ- ‘earth’
Alternatively, ἄχθος could be considered an ancient compound ἄ-χθος. The second half of this compound would then be parallel to Grk. χθών ‘earth, ground, land,’ a reflex of PIE *dʰéĝʰ-om-, *dʰĝʰ-om- ‘earth, ground, man (as earth worker), slave’ as seen in the following cognates: Hit tēkan ‘earth, ground,’ Ved kṣám- ‘earth, ground,’ Skt kṣa ‘field, protector or cultivator of a field, peasant,’ Lat humus ‘earth,’ homo ‘human being,’ OE guma ‘man, (bride)groom,’ Lith žēmė ‘earth,’ OCS zemlja ‘earth, land,’ Phrygian zemel ‘slave,’ TochA tkaṃ ‘earth, ground,’ Alb dhe (from *dʰgʰ-em-) ‘earth.’10
As is well-known, this root often appears in radical metathesis form (with inversion in the ordering of root consonants) and often is attested in zero grade (as in the Greek and Vedic).
If such a compound form was originally a Greek construction, the initial element of ἄ-χθος would either be the well-known alpha-copulative signifying ‘with, together with’11 or else an assimilated form of Grk. ἄγω ‘drive, lead, carry.’12
If the first, this would then signify ‘with earth,’ and would originally have been applied to commoners or slaves tasked with carrying loads of earth, clay, mud, or stones for construction projects, indicating that they were loaded, weighed down, burdened. From here it is not difficult to imagine the semantic shift to vexed, pained, or grieved, since those emotions would be a natural outcome of being burdened with heavy loads and worked to one’s extreme limits.
If the second case is the correct one, then the compound would be ἄγ-χθ-ος, where the gamma was assimilated to the following chi, signifying: carrying earth → loaded, burdened.
On the other hand, if ἄ-χθος reflects an ancient PIE rather than a later Greek compound, then the two possibilities would be the following:
Plus: PIE *dʰéĝʰ- *dʰgʰ- ‘earth, ground, land’ as attested in the previously listed examples.
yielding:
*sm̥-dʰgʰ- or (in metathesis form) *sm̥-ĝʰdʰ- → Grk ἄ-χθ-ος ‘with earth, weighted, loaded.’
Or, if the first element of the PIE compound was an assimilated form of *h₂eĝ- ‘lead, carry, drive’ (instead of *sm̥- as above), then the components would be:
Plus: PIE *dʰéĝʰ- *dʰgʰ- ‘earth, ground, land’ as described above
yielding (with metathesis):
*h₂eĝ-ĝʰdʰ- (then with assimilation) → *h₂e-ĝʰdʰ- → Grk ἄ-χθ-ος ‘carrying earth, weighted, loaded.’
A third (more remote) possibility is that the initial ἄ in ἄ-χθ-ομαι is a prothetic vowel15 parallel to Grk. χθές ‘yesterday’ and ἐχθές ‘yesterday,’ in which case the meaning of ἄχθομαι would be simply ‘to be earthed → to be weighed down with earth.’
Cognate with Hittite?
If either of the above cases is correct, then it could not be possible for Hittite ḫatk-ⁱ ‘shut, close’ to be cognate with Grk ἄ-χθ-ος for the following reasons:
- No reflex of PIE *h₂eĝ- is found in Hittite.
- The PIE prefix *sm̥ ‘with, together with’ is attested in Hittite in the form =(š)šan, not as the ḫa in ḫatk-ⁱ.
What about the goddess Ἀχθεία ?
Hesychius defines Ἀχθεία as ἡ Δημήτηρ, μυστικῶς ‘Demeter, according to the mysteries.’ Since Demeter is typically taken as “Mother Earth,” this would confirm a link between the word ἄχθος and the concept Earth. The second element of Δημήτηρ is obviously ‘mother.’ The first is thought by some as a pre-Greek word for ‘earth,’ and by others as Illyrian (attested by Alb. dhe ‘earth’) and therefore a cognate of Grk χθών, arising from PIE *dʰĝʰ- ‘earth, ground, land.’16
Additional Evidence from the Sanskrit
As mentioned above, PIE *dʰgʰ-em became Ved kṣám-, Skt kṣā́ ‘earth, ground.’ The initial and final consonants of this zero-grade root (via an intermediate Proto-Indo-Iranian dźh-) fused to become Skt kṣ as in kṣám- and in a number of other derivatives, not all of which are typically viewed as being based on the root *dʰgʰ-. The following are a few such instances (with comments):
These are straightforward attestations of PIE *dʰgʰ-em.
In an earlier article (Haynes 2023: 106–109) I have argued that attestations of PIE *dʰĝʰ-em with the semantic value ‘man’ (Lat homo ‘human being,’ Goth guma ‘man,’ Phrygian zemel ‘slave’), as well as another root of the same consonantal form, *dʰ—ĝʰ- (i.e., *dʰeu̯ĝʰ-, with the Greek attestation τυγχάνω carrying as one of its meanings ‘common, or vulgar men’), refer not to men in general, but rather to men of a lower class, i.e., earth workers, cultivators of the soil, or builders employing earth, stones, mud, or clay. The preceding two Sanskrit forms: kṣa ‘cultivator of a field, peasant’ and kṣoṇá ‘a multitude of men, people (as opposed to the chief)’ clearly fit into the same rubric. These terms do not refer to the rulers or chiefs, but rather to the commoners and slaves. According to the ancient logic: they are low, they are dirty, they are dirt.
KEWA expresses uncertainty as to the etymology of these terms, as does EWAia. I would argue that the derivation from *dʰĝʰ- is very likely, considering the many references to ‘earth, land, soil, field, plot of ground, cultivation, plowing, acquisition of land, possession of land, lordship over lands and countries.
The semantic values of these two Sanskrit terms are directly parallel to those derivations of Grk ἄχθομαι and ἄχθος that have been glossed as: ‘mental oppression, to be hated, vexed, pained, or grieved; burden, trouble; burdensome, oppressive; annoyance, suffering; vexed at being conquered, etc.’23 I would argue that these are compound roots based on PIE *dʰĝʰ- as the first element with some root containing a strong labial consonantal component as the second.
These Skt words are typically ascribed to the PIE roots *tḱ-ei̯- and *te-tḱ-. In that previously mentioned earlier article (Haynes 2023: 107–108) I have argued that these two PIE roots:
Ved kṣéti ‘dwells, lingers,’ Myc ki-ti-je-si = /ktiⁱensi/ ‘to build on, cultivate, or work land,’ Lat pōnō ‘put, place, sit down,’ Grk κτίσις ‘settlement,’ κτίζω ‘people a country and build houses and cities in it,’ Av šiti ‘settlement,’ Arm šēn ‘dwell, build on, farm, town.’25
and
Lith tašýti ‘hew, trim,’ OCS tesati ‘hew,’ Skt tákṣati ‘fashions, creates, carpenters, cuts,’ Grk τέκτων ‘architect,’ τέχνη ‘art, craft, skill, technique,’ Hit taksanzi ‘undertake, prepare, cause, joint,’ OHG dehsa ‘axe.’26
are de-voiced and de-asperated variants of PIE *dʰĝʰ-, showing phonetic development similar to that which occurred in Tocharian. The first shows a clear connection to the earth: ‘settling, building, and dwelling on land; cultivating and farming the ground.’ The second refers fundamentally to the practice of building with earth as the basic material of walls and houses constructed of mud bricks. Later, builders began using wood and other materials, but the term for building (originally working with stones, mud, and clay) was later applied to other materials and techniques.
Examples of this phonetic reduction (loss of voicing and aspiration) were given in that earlier article. The part of the introduction of that article that describes this process is reproduced below by way of example, but many more instances are given there.
Phonetic Reduction
Another type of root mutation could be called reduction. This concept is also familiar, since something very close to it is seen in Tocharian (and to some extent in Hittite) where the rich PIE obstruent inventory has been reduced to include only the simple, unvoiced, unaspirated (lenis) forms.27 It has been suggested that this change may have been due to the influence of a substrate language with a similarly limited range of obstruents.28 This same dynamic can be seen in distinct synonymous PIE roots. Obvious examples are often remarked upon in the standard handbooks. The following are some examples of this transformation:
| *gol(H)u̯os ‘bare, bald’ | *kl̥Hu̯os ‘bald’29 |
| *-dʰro-, *dʰlo- = instr. suffix | *-tro-, *-tlo- = instr. suffix30 |
| *h₂eng- ‘bend’ | *h₂enk- ‘bend’31 |
| *peh₂ĝ- ‘fasten securely’ | *peh₂ḱ- ‘fasten securely’32 |
| *pei̯ĝ- ‘draw, color’ | *pei̯ḱ- ‘draw, color’33 |
| *sredʰ- ‘boil, be agitated, move’ | *sret- ‘boil, be agitated, move noisily’34 |
| *bʰendʰ-ŕros ‘relation’ | *pent-h₂ros ‘father-in-law’35 |
| *h₂eu̯g- ‘increase, become strong’ | *h₂eu̯k-s- ‘grow, become large’36 |
| *greh₂bʰ- ‘hornbeam’ | *karp- ‘hornbeam’37 |
| *gʰabʰ- ‘take, seize’ | *kap- ‘have, hold, seize’38 |
| *pleh₂g- ‘strike, beat’ | *pleh₂k- ‘strike, beat’39 |
| *ḱu̯oi̯dis ‘white’ | *ḱu̯oi̯tós ‘white’40 |
| *sab- ‘sap’ | *sap- ‘sap’41 |
| *stei̯b- ‘make stiff’ | *stei̯p- ‘make stiff’42 |
| *dei̯ĝ- ‘teach, show, indicate’ | *deiḱ- ‘preach, say, index’43 |
| *u̯ei̯bʰ- ‘vibrate, be agitated’ | *u̯ei̯p- ‘move back and forth, vibrate’44 |
| *gʰebʰōl ‘head’ | *kapolo- ‘head’45 |
| * ĝʰr̥d- ‘heart’ | *ḱérd- ‘heart’46 |
Many more examples of this dynamic can be observed once the variation of medial resonants in PIE roots is allowed for. The evidence suggests that an ancient dialectical subset of PIE speakers experienced a phonetic influence similar to that which occurred in Tocharian, and then, during a later period of reunification with a group that had not experienced this linguistic change, the dialects became merged. The result is that, after this merger, synonymous pairs (doublets) coexisted within the basic vocabulary of PIE and these have persisted down into the various daughter languages. These synonyms are now considered separate roots, but they should, it will be argued, be seen as variants of an ancient original.
In their most strict formulation, these phonetic reductions can be summarized as follows:
| d, dʰ | became | t |
| b, bʰ | became | p |
| ĝ, ĝʰ | became | ḱ, |
| g, gʰ | became | k |
| gᵘ, gᵘʰ | became | k or kᵘ |
The semantic parallels of *teḱ-s, *te-tḱ- to the voiced and aspirated form of this root can best be seen by way of comparison to the following, also quoted from that earlier article:
*dʰei̯ĝʰ-, *dʰi̯ĝʰs- ‘form, build, mold mud or clay, knead, smear, plaster; wall of mud’ Skt déhmi ‘spread, fill,’ déhī ‘wall, rampart, dam,’ Goth digan ‘form, fashion, knead, make pottery,’ ON deig ‘dough,’ digr ‘thick,’ NE dough, Lith žiedžiù ‘form from mud,’ TochB tsikale ‘to form,’ Lat fingō, finxī ‘form, shape,’ figūra ‘form, shape, figure,’ fictilis ‘fashion out of clay, made of earth or clay,’ figulus ‘potter,’ Av pairi-daēza- ‘enclosure’ (> NE paradise) Grk τεῖχος, τοῖχος ‘wall, embankment,’ possibly Grk θιγγάνω ‘touch with the hand,’ OIr digen ‘build, firm, solid, hard, strong, fixed.’47
Mallory and Adams (223-224, 371) write, “The underlying semantics of *dheiĝh indicate that it was specifically associated with the working of clay (e.g. Lat fingō ‘fashion,’ Skt déhmi ‘smear, anoint,’ TochAB tsik- ‘fashion [pots, etc.],’ hence the English cognate dough; in Greek and Indo-Iranian it is also associated with building walls, e.g. Av pairi-daēza ‘build a wall around’ … but there are also cognates of more general meaning, e.g. OIr con-utainc ‘builds,’ Lith diežti ‘whip, beat,’ Arm dizanem ‘heap up’.” And in EIEC (629) they write: “The substance from which the walls were made, [earth] came to be applied both to the finished product, e.g., Grk τοῖχος ‘wall,’ Av uz-daēza- ‘wall,’ and clay-like substances, e.g. Germanic dough.”
Another probable example of a reduced form of ἄχθος ‘load, burden, mental oppression; vexed, pained, or grieved; burden, trouble,’ can be seen in a Greek word denoting the same semantic field: ἀγανακτέω ‘be indignant or irritated.’ This is a word of uncertain etymology, with Beekes even doubting its source in Greek at all.48 This word could be analyzed as ἀγαν-α-κτέω, with the first element being the familiar Greek adverb ἄγαν ‘much, too much,’ and the remaining compound α-κτέω being a phonetically reduced version of ἄχθος, ἄχθομαι. The original meaning of the compound would therefore be ‘overloaded, overburdened, literally: too much with earth, or carrying too much earth.’
Conclusion
The preponderance of the evidence suggests that Greek ἄχθος, ἄχθομαι is a compound based on the PIE root *dʰeĝʰ- ‘earth, ground, field, earth worker.’
Abbreviations of Literature
| Adams | Douglas Q. Adams, Dictionary of Tocharian B |
| AHD | American Heritage Dictionary |
| ALEW | Hock, Wolfgang, Altlitauisches etymologisches Wörterbuch |
| CLL | Melchert, Craig, Cuneiform Luvian Lexicon |
| DELG | Chantraine, Pierre, Dictionnaire étymologique de la langue grecque |
| de Vaan | de Vaan, Michiel, Etymological Dictionary of Latin & other Italic Languages |
| EDHIL | Kloekhorst, Alwin, Etymological Dictionary of the Hittite Inherited Lexicon |
| EIEC | Mallory, James P., and Douglas Q. Adams, Encycl. of Indo-European Culture |
| EWAia | Mayrhofer, Manfred, Etymologisches Wörterbuch des Altindoarischen |
| IEW | Pokorny, Julius, Indogermanisches etymologisches Wörterbuch |
| KEWA | Mayrhofer, Manfred, Kurzgefaßtes etymologisches Wörterb. des Altindischen |
| LIV | Rix, Helmut, Lexicon der indogermanischen Verben. 2nd edition |
| LSJ | Liddell, Scott, and Jones, A Greek–English Lexicon |
| Mallory and Adams | Mallory and Adams, The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World |
| NIL | Wodtko, Irslinger, and Schneider, Nomina im Indogermanischen Lexikon |
| OCD | Oxford Classical Dictionary |
| OLD | Oxford Latin Dictionary |
References
Abbreviations of Names of Indo-European Languages
| Alb | Albanian | Corn | Cornish |
| Arm | Armenian | Cymr | Cymric |
| Av | Avestan | Gall | Gallo-Roman |
| Bret | Breton | Gaul | Gaulish |
| Bulg | Bulgarian | Goth | Gothic |
| CLuv | Cuneiform Luvian | Grk | Greek |
| HLuv | Hieroglyphic Luvian | OLat | Old Latin |
| Hit | Hittite | OLith | Old Lithuanian |
| Illyr | Illyrian | ON | Old Norse |
| Khot | Khotanese | OPers | Old Persian |
| Lat | Latin | OPrus | Old Prussian |
| Latv | Latvian | ORus | Old Russian |
| Lith | Lithuanian | OSax | Old Saxon |
| Luv | Luvian | Osc | Oscan |
| Lyc | Lycian | Oss | Ossetic |
| Lyd | Lydian | OSwed | Old Swedish |
| Mcymr | Middle Cymric | OWels | Old Welsh |
| ME | Middle English | Phryg | Phrygian |
| MHG | Middle High German | PIE | Proto-Indo-European |
| MIr | Middle Irish | Pol | Polish |
| MPers | Middle Persian | Rus | Russian |
| MWels | Middle Welsh | SC | Serbo-Croatian |
| Myc | Mycenaean Greek | Skt | Sanskrit |
| NE | New English | Slav | Slavic |
| Norw | Norwegian | Sogd | Sogdian |
| NPers | New Persian | Swed | Swedish |
| NWels | New Welsh | TochA | Tocharian A |
| OAv | Old Avestan | TochB | Tocharian B |
| OCS | Old Church Slavonic | Ukr | Ukrainian |
| OE | Old English | Umb | Umbrian |
| OFris | Old Frisian | Ved | Vedic |
| OHG | Old High German | YAv | Young Avestan |
| OIr | Old Irish |