The village was once a ditch

The ditch, the legion's infrastructure

The ditch was standard technology for a Roman camp — but not to a single template, every evening, everywhere. Polybius describes the Republican army (2nd century BC); Vegetius, a late military treatise (late 4th century AD), gives two ditches according to danger.

“The Greeks, in encamping themselves, consider it of the first importance to adapt the camp to the natural advantages of the ground, firstly, because they shirk the labour of entrenching… Whereas the Romans prefer to undergo the toil of entrenching and other defensive work, for the sake of having one sort of camp which never varies and is familiar to all.”

The division of labour is in 6.34 (the allies dig two sides, the Romans the other two); the Greek comparison is in 6.42. Both quoted from the Loeb edition — describing the Republican army, not one timeless Rome.

The dimensioned ditch — according to danger

1. The light ditch: 9 × 7 feet

   “tumultuaria fossa fit lata pedes nouem et alta pedes VII” — when “nimia necessitas non premit”, when no pressing need weighs: 9 feet broad and 7 deep (≈2.7 × ≈2.1 m).

2. The full ditch: 12 × 9 feet

   “legitima fossa… XII pedes lata sit et alta… pedes nouem” — where “uis acrior inminet”, where a keener enemy threatens: 12 broad, 9 deep (≈3.6 × ≈2.7 m). Both dimensions sit in 1.24; recommendations, not site measurements.

3. Roman feet, not modern ones

   The conversion uses the accepted modern value once, 1 pes = 296 mm. Vegetius gives feet only; the metres are our arithmetic, with the factor in the open.

Method: archaeologists still classify these ditches by profile — “fossa fastigata”, the V-section ditch. The 2024 Ermelo dig (Netherlands) cites Vegetius 1.24 and Pseudo-Hyginus §50 as the ancient authorities for the dimensions.

The ditch at continental scale

In North Africa the Romans took the ditch to frontier scale: the Fossatum Africae — one or more linear defence systems. It is claimed to extend over 750 km or more — a length to treat with care, since the source itself says “claimed”.

The reference is our own arithmetic (haversine) from the two towns' coordinates; the top bar renders the source's claim, with “claimed” glued to the number.

1. The single ancient mention, with an exact number

   Fossatum is named once in pre-20th-century historical literature: the law of 29 April 409 from emperors Honorius and Theodosius II to Gaudentius, vicar of Africa — CTh 7.15.1 — on maintaining the limes and the fossatum, raised, says the text, by “the ancients”.

2. Baradez's map

   The extent only became known after World War II, from aerial photographs: Jean Baradez mapped at least four segments, and his survey remains foundational — the map below is “after Baradez (1949)”.

3. Dating: centuries of work

   “It is not known with certainty when the Fossatum was constructed”: digs along the route give dates from Hadrian (2nd c.) to Constantine (4th c.). Baradez proposed a start after Hadrian's first African visit, in 122; he also postulated a building pulse under Gordian III and abandonment around 430–440. Attributions, not certainties.

The rupture: the 270s

The traditional date for Dacia's official abandonment is 271 — but another view moves the evacuation to 272–273, possibly as late as 275. Read it as “the 270s”: the withdrawal of Roman administration and army from north of the Danube under Aurelian. The fate of the Romanized population became mired in controversy.

“The province of Dacia, which Trajan had formed beyond the Danube, he gave up, despairing, after all Illyricum and Moesia had been depopulated, of being able to retain it. The Roman citizens, removed from the towns and lands of Dacia, he settled in the interior of Moesia, calling Dacia that which now divides the two Moesias.”

1. A new province, in the south

   Aurelian created a new province called Dacia south of the Danube — Dacia Aureliana — with its capital at Serdica. Eutropius places it “in media Moesia”, between the two Moesias; modern descriptions vary from page to page — so this page goes no further than Eutropius's wording.

2. The controversy stands

   The “Controversy over the fate of the Daco-Romans” section sets two theories against each other — continuity of a Daco-Roman population north of the Danube, or a near-total pullback. The camps stay facing each other; the thread followed here is the words', not continuity's.

3. Why it matters for words

   Between the fort digging the ditch and the village inheriting its name lies this rupture — as background, not proof. An inherited word proves linguistic filiation; it does not say by itself where every generation of speakers lived.

“The Latin table” gathers more inherited words — with the same evidence vocabulary: inherited, borrowed, reconstructed, attested.

The one who doubts

The Dicționarul etimologic român (Cioranescu, 1958–1966) files the origin of sat under “Origine îndoielnică” (“doubtful origin”) — after listing the fossum camp (Bogrea, Tiktin, Daicovici, Candrea, Scriban, Rosetti) and the Albanian route (Cihac, Meyer, Densusianu…). His preference: Latin sătum. He gets the floor at length — including the concession that keeps this page on the majority's side.

No verdict beyond that: today's standard etymology is fossatum; its history was not unanimous — and the fsat form remains the constraint every proposal must explain.

1521 is not the beginning of written Romanian — read why in “Neacșu's letter”.

The borrowed word

Romania's urban vocabulary is layered, not replaced piece by piece: cetate from Latin, târg from Slavic, oraș from Hungarian. The words coexist, overlap, specialise — oraș did not enter a void.

1. What we know

   MDA2 records the 1521 attestation and the “mg város” etymology; DEX confirms the borrowing. Hungarian város is tied to vár, “fortress, castle”. That is all.

2. It did not enter a void

   târg — from Slavic trŭgŭ (DEX) — also means “smaller urban settlement”; DER gives sense 4 as “Oraș, orășel”, and MDA2 cites COD. VOR. 100/15. When oraș enters, târg is already there.

3. What we don't know

   When város entered Romanian and by which route. The 1521 attestation is only a “no later than” — the word may be older than the oldest surviving paper.

A real ditch, an uncertain date

In Wallachia and Oltenia, ditched earthen ramparts still show — Brazda lui Novac. The perfect case for this page's method: the ditch physically exists, but dating and function differ by segment and by source. We spread the disagreement and draw no conclusion — contextual archaeology, not etymological proof.

1. The disagreement, briefly

   The English version: a Drobeta–Ploiești vallum, 3 m high, 2 m ditch; datings from the 1st century (Donat) to 322 (Constantine). The Romanian version: a 4th-century Constantinian northern Brazda; a southern Brazda — the Troianul — with different dimensions and datings (2nd–3rd centuries, now Augustus, now Hadrian).

2. The interim verdict

   “Datarea valului de pământ Brazda lui Novac de sud este încă neclară și mult discutată” — the dating of the southern vallum is still unclear and much discussed. Physical survival does not make interpretation easy: not even a ditch hundreds of kilometres long dates itself precisely.

And today's village

Tell the story in one breath — the word “sat” comes from “ditch”, “cetate” from “city”, “oraș” from Hungarian — and add the hedge that makes all the difference: dictionaries say confidently where words come from; when, where and why each meaning turned remains, in parts, reconstructed.

What we don't know (and say so)

• The first Romanian attestations of cetate and oraș. This page cites only MDA2's recorded attestations (COD. VOR. and 1521 respectively) — and 1521 is a “no later than”, not the word's birthday.
• Which medieval charters might document the villa/oppidum/civitas vocabulary. Unconfirmed leads: the Andreanum, the Ioannites' Diploma — no text quoted until verified.
• Whether villages ever dug their own defence ditches (the obște). The claim stays off the page until a primary or archaeological source.
• The exact fsat passage in the Hurmuzaki Psalter (MDA2 cites 462/11). Until the edition is checked, the page cites the reference, not the verse.
• Who is right about continuity. A journey of words has no standing to settle it — and this page does not use the 270s withdrawal as etymological proof.

Quiz: what we know vs what we reconstruct

Six questions about the distinctions that matter: attested versus reconstructed, dictionary versus chronology, word versus thing.

1. question 1 of 6

   MDA2 gives the old variant “fsat” — and DER, though it prefers sătum, concedes it is “difficult to explain”. What does fsat give us?

   • direct evidence of an old form, constraining any etymology
   • a photograph of the ditch-to-village passage
   • the exact year of the meaning change

   A form, not a photograph: fsat preserves an old consonantal stage — which is why DER itself stops at it. This page holds fossatum, with the dissent in view.

2. question 2 of 6

   MDA2 puts the sense “Cort” (“tent”) first; DER puts “Câmp, ogor” (“field”) first. What does sense order in a dictionary prove?

   • the order in which senses appeared
   • only lexicographical organisation — not chronology
   • the age of the manuscripts

   Only organisation: two dictionaries, two orders. Chronology needs dated attestations, not sense numbers.

3. question 3 of 6

   Littré: fossé “du lat. fossatum”. Treccani: fossato “dal lat. tardo fossatum”. What does the pair show?

   • that French and Italian borrowed from Romanian
   • that the same fossatum stayed a ditch in the West and became a village in Romanian
   • that sat comes from French

   Semantic drift made visible: same root, two afterlives — the West kept the ditch, Romanian kept the settlement.

4. question 4 of 6

   DEX: târg, sense 3 — “smaller urban settlement”, from Slavic. Did oraș enter a void?

   • yes — that is why it was borrowed
   • no — târg already covered the urban sense too
   • yes — cetate meant nothing any more

   No: the urban vocabulary was layered — cetate, târg, oraș — and the words specialised through coexistence, not scheduled replacement.

5. question 5 of 6

   The Saalburg photo documentation: “Both the ditches belong to the late period of the fort… reconstructed over the Roman foundations”. What does the image show?

   • the ditch dug every evening by the legions
   • the reconstructed defences of a permanent fort
   • a marching-camp ditch

   A permanent fort, late period, walls rebuilt over Roman foundations — archaeological context for what a fossatum was, not proof of the every-evening ditch.

6. question 6 of 6

   Sat < fossatum is an inherited etymology. Does it say where all the speakers lived in AD 700?

   • yes — it proves continuity
   • no — a word's filiation is not its speakers' map
   • only for Wallachia

   No: an inherited word proves linguistic ancestry, not each generation's address. The 270s are context, not etymological proof.

Sources

Dictionaries first, then the linguistic synthesis, ancient texts, history and measurements. Every engine node and every number points at the table below.

Dictionaries and linguistic syntheses

• DEX online: sat (DEX '09, MDA2, DER, Scriban)
• DEX online: cetate (DEX '09, MDA2, CADE, Scriban, DER)
• DEX online: oraș (DEX '09, MDA2, DER)
• DEX online: târg (DEX '09, MDA2, DER)
• Brian D. Joseph, The Balkan Sprachbund (LSA 2017, PDF, archive)
• Littré: fossé
• Treccani: fossato
• Etymonline: fosse
• Wiktionary: sat (Romanian section)
• Wiktionary: foso (Spanish section)

Ancient texts and archaeology

• Polybius, Histories, book 6 (Loeb)
• Vegetius, Epitoma rei militaris, book 1 (Latin)
• Eutropius 9.15 (Latin)
• Historia Augusta, Aurelian 39 (Thayer)
• Driessen et al., the Ermelo fort (2024, PDF)

History and datings

• Fossatum Africae (EN-wiki)
• Roman Dacia (EN-wiki)
• Dacia Aureliana (EN-wiki)
• Brazda lui Novac (EN-wiki)
• Brazda lui Novac, limes (RO-wiki)
• Hurmuzaki Psalter — range datings (EN-wiki)
• Hurmuzaki Psalter — 1500–1520 dating (RO-wiki)
• Voroneț Codex — 16th century (RO-wiki)
• Roman units — the 296 mm foot (EN-wiki)

Type: Fraunces, Newsreader and Atkinson Hyperlegible, under the SIL Open Font Licence.