Linear B is a type of writing system used to record the Mycenaean Greek language, which is the earliest known form of the Greek language. This script was used hundreds of years before the Greek alphabet, with the oldest examples dating to about 1450 BC. Linear B was developed from an earlier script called Linear A, which has not been fully understood. It is similar to the later Cypriot syllabary, which also recorded Greek. Linear B was mostly found in palace records at places like Knossos, Kydonia, Pylos, Thebes, and Mycenae. It stopped being used after the Mycenaean civilization fell during the Late Bronze Age collapse. The time that followed, called the Greek Dark Ages, has no evidence of writing being used.
Linear B was deciphered in 1952 by Michael Ventris, an English architect and self-taught linguist, who built on the work of Alice Kober, an American classicist. It is the only script from the Bronze Age Aegean region that has been fully understood. Other scripts, like Linear A, Cypro-Minoan, and Cretan hieroglyphic, remain unreadable.
Linear B includes about 87 symbols that represent sounds and over 100 symbols that represent objects or items. These object symbols, called ideograms, do not represent sounds and are not used to spell words.
The use of Linear B was mostly limited to administrative tasks, such as keeping records in Mycenaean palace locations. Thousands of clay tablets were written by a small number of scribes: 45 in Pylos (on the west coast of the Peloponnese in southern Greece) and 66 in Knossos (on the island of Crete). Linear B symbols also appeared more often on trade items like amphorae. After the palaces were destroyed, the script was no longer used.
Script
Linear B has about 200 signs, which are divided into two groups: syllabic signs that represent sounds and ideograms that represent meanings. The way these signs are written and named was made official through a series of international meetings that began in Paris in 1956. After the third meeting in 1961 at the Wingspread Conference Center in Wisconsin, a standard created mainly by Emmett L. Bennett, Jr. became known as the Wingspread Convention. This standard was later adopted by the Comité International Permanent des Études Mycéniennes (CIPEM), an organization that became part of UNESCO in 1970. These meetings continue today; the 13th meeting happened in Paris in 2010.
Many signs in Linear B look similar to those in Linear A, but Linear A uses a language that has not yet been understood. It is unclear if the same signs in Linear A had the same sound values as in Linear B.
During the study of Linear B, Michael Ventris and John Chadwick created a grid that shows the sound values for syllabic signs. (Note: The letter "q" in the grid represents sounds like [ɡʷ, kʷ, kʷʰ], not the uvular sound shown in the IPA. The letter "j" represents the same sound as the English letter "y" in words like "yes" and "yoke.")
On the grid, the first column lists initial consonants, and the top row lists vowels. Next to each sign is a transcription of the sound it might represent, along with a number from Bennett (marked with an asterisk, as Ventris and Chadwick used). If the sound of a sign is uncertain, the number alone is used to identify it. Signs on tablets and sealings often look different from each other and from the grid. Scientists are still studying why these differences happen and what they might mean.
In addition to the grid, the first edition of Documents in Mycenaean Greek included signs called "homophones" because they seemed to sound like other syllables. For example, pa 2 and pa 3 were thought to sound like pa. However, later editions of the book removed these labels, and pa 2 was changed to qa. It is now believed that there are no true homophones.
Some sign values are still unknown because there is not enough evidence about them. For example, 34 and 35 are mirror images of each other, but it is not certain if this visual similarity means they also have similar sounds.
CIPEM now decides which signs are "confirmed" and how to officially represent different sign categories. In published texts, signs with unconfirmed values are written as numbers with an asterisk (e.g., *64). CIPEM also assigns these numbers. If a new sign is discovered or a sign is unclear, it is written as a bullet point in brackets: [•].
The signs in Linear B are not exact representations of sounds. Each sign can be used to show about 70 different sound combinations, depending on rules and conventions. The grid shows a system of signs that represent single syllables (V/CV). Some special values tested the limits of this system, but Chadwick concluded that syllabic signs can still be considered monosyllabic.
Exceptions to this rule may include two diphthongs, 𐁀 (ai) and 𐁁 (au), which represent two vowels combined into one sound. For example, 𐁀 𐀓 𐀠 𐀴 𐀍 (ai-ku-pi-ti-jo) represents Aiguptios (Αἰγύπτιος, "Egyptian"), and 𐁁 𐀐 𐀷 (au-ke-wa) represents Augewās (Αὐγείας, "Augeas"). However, diphthongs are sometimes treated as two separate syllables.
Some signs, like 𐁌 (Twe), 𐁍 (two), and 𐁃 (dwe), may begin with labialized consonants rather than two consonants. Others, like 𐁈 (rya) and 𐁊 (ryo), begin with palatalized consonants.
The one sign Chadwick considers an exception to the monosyllabic rule is 𐁇 (pte), which he thinks developed from pje, as in kleptei < klep-jei.
Linear B does not always show the difference between voiced and unvoiced stop consonants or between aspirated and unaspirated stops, even though these differences are important in Mycenaean Greek. An exception is the dental series, where syllables starting with voiced dental stops are written differently from those starting with voiceless stops. For example, pa-te is patēr (πατήρ), pa-si is phāsi (φησί), and do-ra is dōra (δῶρον, "gift").
In some cases, aspiration may be marked, but this is optional. For example, pu-te is phutēr ("planter"), and phu-te-re is phutēres ("planters"). Initial /h/ is only marked when followed by a vowel, and this is rare.
The q-series represents syllables starting with labialized velar consonants, which were common in Mycenaean Greek but later disappeared from classical Greek. These consonants came from Proto-Indo-European, assimilation, and borrowed words. For example, a-pi-qo-ro is amphiq oloi (ἀμφίπολοι), and qa-si-re-u is g asileus (βασιλεύς, "basileus").
The j-series represents the semivowel equivalent to the English "y" and is used at the beginning of words or after syllables ending in "i." For example, -a-jo is -αῖος (-aios). The w-series represents semivowels used at the beginning of words or after syllables ending in "u." For example, tu-ri-so is Tulisos (Τυλισός).
The r-series includes both /r/ and /l/ sounds. For example, ti-ri-po is tripos (τρίπος, "triple").
Some consonants are not written but are understood to be present. For example, word-initial "s" and "w" before a consonant are not written, as in pe-ma for sperma (σπέρμα, "seed"). Syllable-final -l, -m, -n, -r, and -s are also not written, and only word-final velars are written fully. For example, a-to-ro-qo is anthrōq os (ἄνθρωπος, "human being").
Clusters of two or three consonants are sometimes not written but are understood to be present. For example, in some cases, consonants like "s" or "w" before a consonant are omitted, but the rules for this are clear.
Archives
Inscriptions in Linear B have been discovered on tablets, stirrup jars, and other objects. Two types of tablets have been found: "palm-leaf" shaped tablets, with writing along the long side, and larger "page" shaped tablets, divided into multiple lines. "Palm-leaf" tablets usually record one transaction, while "page" tablets summarize many transactions.
Inscriptions are organized and sorted based on where they were found during excavations.
Another 170 inscriptions in Linear B have been found on various vessels, bringing the total number of known inscriptions to about 6,058. Most vessel markings come from Inscribed Stirrup Jars, with many discovered at Thebes, Mycenae, Tiryns, and Khania.
For many years, scholars have worked to piece together broken tablet fragments. This helps make the tablets and their information more complete while reducing the total number of tablets.
The oldest Linear B tablets are likely those from the Room of Chariot Tablets at Knossos, dating to the second half of the 15th century BC. The Kafkania pebble, though from an earlier time, is not real. The earliest mainland inscription is a clay tablet found at Iklaina, dating to between 1400 and 1350 BC.
Tablets are grouped into series based on the most common symbol used. For example, the E series includes symbols related to grain (gra). The X series is for tablets without symbols, the W series for labels and sealings, and Z for inscribed objects other than tablets. If a group within a series can be identified by shape, vocabulary, format, or writing style, it is labeled with a lowercase letter (e.g., the Er series).
An amber seal with Linear B symbols was found in 2000 near Kranzberg, southern Germany. Its authenticity is still debated.
The Aegean region contributed many early Greek words related to daily life, such as terms for tools and common items. The timeline and spread of Cretan hieroglyphs, Linear A, and Linear B—three distinct writing systems used in Bronze Age Crete, the Aegean islands, and mainland Greece—are summarized as follows:
Major Linear B archives are linked to stages of Late Minoan and Helladic pottery.
Sixteen tablets found at the Megaron in Pylos are believed to date to LHIIIA.
The Knossos archive was dated by Arthur Evans to around 1400 BC, when a fire destroyed the palace and preserved the clay tablets. He placed this event in the LM II period. This view changed in 1939 when Carl Blegen excavated ancient Pylos and found Linear B tablets. These were burned during a fire that destroyed Pylos around 1200 BC, at the end of LHIIIB. After Linear B was deciphered by Michael Ventris in 1952, questions about Evans’s dating were raised. Blegen noted similarities between 1200 BC Pylos and 1400 BC Knossos, suggesting Evans’s dating might need reevaluation.
Investigations later found issues with Evans’s records. His assistant, Duncan Mackenzie, had detailed notes that did not match Evans’s reports. Disagreements also existed about the tablet locations and layers of soil. These findings were published in On the Knossos Tablets, including works by Leonard Robert Palmer and John Boardman, representing Blegen’s and Evans’s views. This debate became known as the "Palmer–Boardman dispute." No final resolution has been reached.
Major cities and palaces used Linear B to record the distribution of goods such as wool, sheep, and grain, often given to religious groups. Some tablets also mention military matters.
Like cuneiform tablets, many Linear B tablets were preserved when fires destroyed the buildings they were stored in, as the heat baked the clay.
Discovery and decipherment
The ancient Greeks could not read Linear B, but some ancient writers mentioned its symbols. For example, Plutarch wrote about Spartan king Agesilaus II (r. 400–360 BC) sending a bronze tablet with "many letters marvellously old, for nothing could be made of them" to Egyptian priests, hoping they could understand them.
In 1886, British archaeologist Arthur Evans, who worked at the Ashmolean Museum, received a sealstone from Crete. Greville Chester gave it to him, and the stone had writing that Evans believed was Mycenaean. Earlier, Heinrich Schliemann had seen similar symbols but did not clearly recognize them as writing. He wrote in his book about Mycenae that he had found only three or four examples of signs that looked like inscriptions. In 1893, Evans bought more sealstones in Athens and confirmed they came from Crete. The next year, he noticed the same script on other items in the Ashmolean Museum. In 1894, he traveled to Crete to search for the script. At Knossos, he saw a double axe symbol on a wall and thought this might be the source of the writing. Later, he found stones from ruins being used as amulets by Cretan women, called "milk-stones," believed to help with breast milk production.
Starting in 1894, Evans wrote articles about the script in The Journal of Hellenic Studies. His first article was titled "Primitive Pictographs and a Prae-Phoenician Script from Crete." In these writings, he described two types of writing: "pictographic writing" (symbols that looked like pictures) and "a linear system of writing" (symbols that were more geometric). He did not clearly define these terms, which caused confusion later. In 1898, he explained that the linear symbols were simple and less likely to change, while the pictographic ones were more complex. He also said the linear symbols likely came from early, simple drawings.
At the same time, Evans worked to buy the land at Knossos. He started the Cretan Exploration Fund with his own money and by 1896 had purchased one-fourth of Kephala Hill, where the ruins were. However, he could not get permission to dig from the Ottoman government. He returned to Britain. In 1897, the Christian people of Crete rebelled against the Ottoman Empire. By 1898, the last Ottoman soldiers left the island, and Evans and his friends completed buying the site. A new Cretan Republic was formed in 1899. After getting permission to dig, work began on 23 March 1900.
In his report to the British School at Athens, Evans wrote that on 5 April 1900, workers found a large collection of Linear B tablets in a wooden box inside a broken terracotta bathtub. More tablets were found later, including over 350 in a room called the "Room of the Chariot Tablets." The tablets were small, with writing in about 70 characters. Evans noted that some symbols seemed to have meanings related to ideas or objects.
By 2 June 1900, the excavation for that year ended. Evans reported that many tablets were broken due to rain, crumbling, or being discarded by workers who did not recognize their value. In a September 1900 report to the Royal Anthropological Institute, he began using terms like "palace of Knossos" and "palace of Minos." He described the script as "linear" and said it was mostly syllabic (representing sounds). He also mentioned some symbols had ideographic or determinative meanings (representing ideas or objects).
After 1900, Evans focused on excavating Knossos and studying the tablets. He planned a major work called Scripta Minoa to describe Cretan scripts. Before publishing the first volume in 1909, he included the Phaistos Disc, which had been discovered in 1908, and some unpublished tablets from Hagia Triada. He divided the linear script into "Class A" and "Class B." The Knossos tablets were Class B, while Class A came from other sites. He described differences between the two scripts as variations in the shapes and arrangements of the symbols.
The first volume of Scripta Minoa covered "Hieroglyphic and Primitive Linear Classes," including early scripts, pictorial writing, and the Phaistos Disc. Evans planned more volumes on Linear A and Linear B tablets but ran out of time. He became involved in war and politics in the Balkans and later focused on finishing his book Palace of Minos, published in 1935. He died in 1941, shortly after Nazi forces invaded Crete.
The Knossos tablets remained in a museum in Irakleion, Crete, where many were lost. Unpublished notes and materials from Scripta Minoa II were created by Clarendon Press. In 1939, Carl Blegen discovered the Pylos Tablets, increasing pressure to complete Scripta Minoa II. After Evans’s death, Alice Kober and Emmett L. Bennett, Jr. helped finish the work. The second volume was published in 1952, just before it was discovered that Linear B wrote an early form of Greek.
Unicode
Linear B was included in the Unicode system in April 2003 when version 4.0 was released. The Unicode block for the Linear B Syllabary covers the range from U+10000 to U+1007F. The Unicode block for the Linear B Ideograms covers the range from U+10080 to U+100FF. The Unicode block for the related Aegean Numbers covers the range from U+10100 to U+1013F. Many fonts support the characters used in Linear B.