📄 _'ādām_ as 'Skin' and 'Earth': An Examination of Some Proposed Meanings in Biblical Hebrew

Summary

Hess tests whether three of the five senses that the Koehler–Baumgartner Hebrew lexicon assigns to the noun 'ādām are well founded. Beyond the uncontested senses (humankind and the personal name Adam), the lexicon lists "skin/hide/leather", "ground/earth/steppe", and a place name. Hess revisits every verse cited for the doubtful senses and compares the rendering behind each entry with the existing commentary tradition and the ancient versions.

The "leather" sense rests on G. R. Driver's Arabic-based reading of Hosea 11:4. The "earth/steppe" sense rests on six or seven passages argued by M. Dahood. In each case Hess finds that the ordinary sense "humanity", or an accepted emendation, fits the context at least as well. Only the toponym holds up, and its occurrence in Joshua 3:16 is the sole secure one.

His conclusion is that difficult syntax does not license a new dictionary sense, and that cognate-language parallels alone should not add meanings to the Biblical Hebrew lexicon. The study shows how sense inventories can grow through weakly supported special pleading.

Key points

• The Koehler–Baumgartner entry has five listings for 'ādām: humanity, skin/leather (Hosea 11:4), Adam, earth/steppe, and the site linked to Tell ed-Dimiye.

• In Hosea 11:4, Driver's "bands of leather" reading draws on Arabic cognates. Later commentaries, including Wolff's updated edition, read the phrase as "human cords" instead.

• Dahood's "earth" sense was argued for these passages:

  • Zechariah 9:1
  • Proverbs 30:14b
  • Genesis 16:12
  • Job 36:28b
  • Jeremiah 32:20
  • Zechariah 13:5

  For each one, Hess gives the alternatives: "humanity", the place name, or the emendations of Wellhausen and Ginsberg.

• The place name is confirmed by Egyptian sources (Shishak's campaign list) and by Hebrew sources. Joshua 3:16 is certain, and other proposed occurrences are conjectural.

• Hess's verdict is that two of the five lexicon senses are at best hypothetical for Biblical Hebrew. Other forms from the 'dm root already supply the meaning "land".

Interesting ideas

• I1. Before a dictionary accepts a new sense, the established sense must be shown to fail in context, not merely shown to be less exciting; parsimony keeps the sense inventory honest. — For Mova (world): When building the WordNet/FrameNet-based ontology, block new senses that rest on one hard example unless every existing sense demonstrably fails the frame.
• I2. Hess tests candidate senses by what the surrounding text cares about: in Jeremiah 32:20 the passage concerns peoples, not places, so a "steppe" reading clashes with it. — For Mova (frames): Role and sense choice could include a discourse-topic check (human vs. location), scoring whether a reading fits the entity types already active in the frame.
• I3. He notes that swapping one noun for another leaves the verse's grammatical difficulty untouched, so a "fix" that does not touch the real problem is no fix. — For Mova (learn): Self-induced repair rules should be accepted only if they reduce the specific error they target, verified with negative controls rather than by plausibility.
• I4. Evidence from a related language (Arabic, Ugaritic) suggests a meaning but does not prove that speakers of the target language used it. — For Mova (back): During translation through the meta-world, cross-lingual cognate similarity should only propose candidate senses, never assert them, without in-language usage evidence.
• I5. Hess weighs the agreement of many independent commentators and ancient versions against a single innovative proposal. — For Mova (learn): This mirrors silver data from several witness parsers, where an outlier analysis needs extra evidence before it overrides a broad consensus.