Reconstruction

Chapter 3 found the regular correspondences between related languages. Reconstruction takes the next step: it asks what single ancestor sound would, after regular change, give each of the sounds seen today. The result is a form in a language that nobody has recorded. This chapter explains how that is done and how far to trust it.

The Asterisk

Reconstructed forms are prefixed with an asterisk (*) to distinguish them from attested forms. "Asterisks are used by linguists as a conventional mark of reconstructed words" (Wikipedia, Proto-Indo-European language). So *tréyes is not a word anyone wrote down; it is a conclusion. Chapter 1 flagged this convention as my own summary; this chapter attributes it to its sources.

Two Kinds of Reconstruction

A reflex is a known derivative of an earlier form, attested or reconstructed. An attested word from which a proto-language root is reconstructed is a reflex. A reflex predictable from the reconstructed history is "regular", and reflexes of the same source are cognates.

Rules of Thumb

A worked example from the source

Spanish cantar and French chanter suggest that the stop [k] is older than the fricative [ʃ], because stops tend to become fricatives. This is the "most natural development" principle in action, and it is the same direction as the Romance [k] : [ʃ] set in Chapter 3.

Proto-Indo-European

Proto-Indo-European (PIE) is the reconstructed common ancestor of the Indo-European language family. No direct record of it has been found; its features were derived by comparing documented descendant languages. Examples from the page:

The sound correspondence is the one from Chapter 2: Italian piede, padre and pesce against English foot, father and fish show a consistent initial p / f pairing. Grimm's law (1822) and Verner's law (1876) describe such regular changes, and the Neogrammarians held that sound laws apply without exception.

When and where?

• Date: PIE is hypothesised to have been spoken as a single language from about 4500 BCE to 2500 BCE. Other estimates place the period more than a thousand years later.
• Kurgan hypothesis (first proposed by Marija Gimbutas in 1956): the most popular view. It places the language with the Yamnaya culture on the Pontic steppe around 3400 BCE.
• Anatolian hypothesis: PIE spread from Anatolia with agriculture, beginning around 7500-6000 BCE.
• The page calls these two the most widely accepted and the most debated against each other. Colin Renfrew, the Anatolian hypothesis's original proponent, has accepted migrations from the Pontic steppe after 2015 ancient DNA studies. The Armenian hypothesis, the Paleolithic continuity paradigm and the indigenous Aryans theory are not regarded as credible.

I show both main views because the page treats the question as open. Where scholars disagree, this course does not choose.

A reconstruction that was later confirmed

The laryngeal theory was proposed by Ferdinand de Saussure in 1879, based only on internal reconstruction. It explains irregularities in PIE phonology as effects of hypothetical sounds. It gained general acceptance after Jerzy Kuryłowicz found consonantal reflexes of these sounds in Hittite. That is why many PIE forms above contain a subscript h₂.

How reliable is it?

The page says PIE phonology "has been reconstructed in some detail", but that the most widely accepted reconstruction is "not uncontroversial". Some points remain debated, including the existence of /a/ as a separate phoneme and the accent system, which some linguists reconstruct as a tone system instead. It also says the sound laws are reliable enough to support the Neogrammarian hypothesis.

Try It Yourself