Linear B is a syllabic script that was used for writing Mycenaean Greek, the earliest attested form of Greek. The script predates the Greek alphabet by several centuries. The oldest Mycenaean writing dates to about 1450 BC. It is descended from the older Linear A, an undeciphered earlier script used for writing the Minoan language, as is the later Cypriot syllabary, which also recorded Greek. Linear B, found mainly in the palace archives at Knossos, Cydonia, Pylos, Thebes and Mycenae, disappeared with the fall of Mycenaean civilization. The succeeding period, known as the Greek Dark Ages, provides no evidence of the use of writing.
Linear B consists of around 87 syllabic signs and over 100 ideographic signs. These ideograms or "signifying" signs symbolize objects or commodities. They have no phonetic value and are never used as word signs in writing a sentence.
The application of Linear B appears to have been confined to administrative contexts. In all the thousands of clay tablets, a relatively small number of different "hands" have been detected: 45 in Pylos (west coast of the Peloponnese, in southern Greece) and 66 in Knossos (Crete). From this fact it could be thought that the script was used only by a guild of professional scribes who served the central palaces. Once the palaces were destroyed, the script disappeared.
Linear B has roughly 200 signs, divided into syllabic signs with phonetic values and ideograms with semantic values. The representations and naming of these signs has been standardized by a series of international colloquia starting with the first in Paris in 1956. After the third meeting in 1961 at the Wingspread Conference Center in Racine, Wisconsin, a standard proposed primarily by Emmett L. Bennett, Jr. (1918–2011), became known as the Wingspread Convention, which was adopted by a new organization, the Comité International Permanent des Études Mycéniennes (CIPEM), affiliated in 1970 by the fifth colloquium with UNESCO. Colloquia continue: the 13th occurred in 2010 in Paris.
Many of the signs are identical or similar to those in Linear A; however, Linear A encoded is still an unknown language and we cannot be sure that similar signs had similar phonetic values.