The etymology, or origin, of a word frequently reveals its deepest level of meaning, associating it with the ancient culture from which it came. Historical linguists resort to the concept of linguistic archaeology, uncovering layers of meaning through linguistic change. Latin is undisputedly one of the Italic languages, a group within the Indo-European languages. The fact that forms of castrum also appear in Oscan and Umbrian, two other Italic languages, suggests an origin at least as old as Proto-Italic language.
Julius Pokorny (major Indo-Europeanist of the 20th century) in his encyclopedia of Indo-European roots, traces a probable derivation from *k̂es-, schneiden (“cut”) in *k̂es-tro-m, Schneidewerkzeug (“cutting tool”). The k̂ is not a letter or an IPA symbol. Called the circumflex k, it began as a symbol to transliterate Ka (Cyrillic) into Latin script in the Scientific transliteration of Cyrillic. More recently the latter was incorporated into ISO 9.
The k̂ is one of a group of sounds called on the one hand dorsal consonants because articulated by constricting the back of the tongue, or dorsum, against the roof of the mouth, and on the other are subclassified by the exact point on the roof of the mouth, of interest here mainly the velars, which use the velum, and the palatals, which use the hard palate. The IPA classification offers a different point of view on the palatals, that they are formed from the other dorsals by moving the point of contact up to the palate, a supposed process called palatalization. The IPA symbol for the k̂ is therefore kʲ, where the superscript j represents the supposed palatalization. In more ordinary terms, the k sound can be articulated either up on the palate or down on the velum.
To the etymologists, Proto-Indo-European is conceived to have contained a group of words that began with palatals and another that began with velars. After this proto phase; that is, after the formation of the major language groups, the palatal group changed according to one of two different phonetic destinies, one termed centumization and the other satemization, to create the Centum and satem languages. In centumization the k̂ changed to k, merging the group with the velar group. The Italic languages are centum. *k̂estrom therefore became *kastrom in the Italic languages, but there was a change in meaning as well.
These Italic reflexes based on *kastrom include Oscan castrous (genitive case) and Umbrian castruo, kastruvuf (accusative case). They have the same meaning, says Pokorny, as Latin fundus, an estate, or tract of land. This is not any land, but is a prepared or cultivated tract, such as a farm enclosed by a fence or a wooden or stone wall of some kind. Cornelius Nepos uses Latin castrum in that sense: when Alcibiades deserts to the Persians, Pharnabazus gives him an estate (castrum) worth 500 talents in tax revenues. This is a change of meaning from the reflexes in other languages, which still mean some sort of knife, axe, or spear. Pokorny explains it as ’Lager’ als ‘abgeschnittenes Stück Land’, “a lager, as a cut-off piece of land.”
If this is the civilian interpretation, the military version must be “military reservation,” a piece of land cut off from the common land around it and modified for military use. All castra must be defended by works, often no more than a stockade, for which the soldiers carried stakes, and a ditch. The castra could be prepared under attack within a hollow square or behind a battle line. Considering that the earliest military shelters were tents made of hide or cloth, and all but the most permanent bases housed the men in tents placed in quadrangles and separated by numbered streets, one castrum may well have acquired the connotation of tent.