San Diego State University logo

Other languages as codes

Code vs.

A cipher is an encryption system which involves letter-to-letter mappings.
A code is any other encryption system.

Deciphering is the common term used for the task of discovering the meanings associated with an unknown writing system, whether it is alphabetic, ideographic, or syllabic.

So the terminology is a little inconsistsent with the terminology of cryptography. So it goes.


  1. Language known, writing system unknown: Egyptian Hierogliphics
  2. Language family known, writing system unknown: Mayan writing
  3. Language unknown, writing system unknown: Linear B, Etruscan writing, Iberian, Linear A

The morals:

  1. Methods of writing decipherment and cryptography essentially the same
  2. Many of the people working on decipherment have at one time or another been cryptographers.
  1. Earliest Egyptian hieroglyphs: 3000 B.C.
  2. Lasted about 3500 years (i.e., longer than our alphabet has yet lasted)
  3. Languages can change quite a bit in that kind of timespan; Old English (c. 800 A.D.) is not a whole lot easier for a modern English speaker to learn than German.

Two parallel writing systems:

  1. Hieroglyphic: Ornamental, complicated, formal, palace walls, monuments, royal tombs, etcetera
  2. Hieratic: More mundane everyday writing. Simpler. Commercial transactions [evolves into demotic c. 600 B.C.]
  3. Fade out in 4th century A.D., due to spread of Christianity and introduction of Coptic, Greek-alphabet-based writing system.
the type

In all cases the first task is to determine the type of writing system:

  1. Phonetic
    1. alphabetic
    2. syllabic
  2. Ideographic
Of course mixed systems are also possible, as we shall see.

Hieroglyphs are mostly phonetic.

For a long time (Roman era-> 17th century) people labored under the misimpression that hieroglyphs were entirely ideographic, even to the extent of publishing false translations:

Kircher: 1652 Oedipus Aegiptiacus
Apries (a Pharoah's name): translated as "The benefits of the divine Osiris are to be procured by means of sacred ceremonies and of the chain of the Genii, in order that the benefits of the Nile may be obtained."

Rosetta Stone: Slab found in 1798 in Rosetta (Nile Delta) by a team of scholars sent in after Napoleon's armies conquered Egypt. [Currently in the British Museum, well worth seeing]

The same text written 3 times, once in Greek, once in Hieroglyphics, once in Demotic: a decree by Egyptian priests enumerating great gifts of Pharoah Ptolemy to Egypt and great gifts of Egyptian priests to Ptolemy, including an annual festival in his honor. [46 1/2 " x 30 " x 12 "" ]

    High Res Version (note character sequences occurring in curved outlines called cartouches)


  1. Damage
    1. Last 26 lines of 54 Greek lines incomplete [ends of lines missing]
    2. First 14 lines of 32 Demotic lines incomplete [beginnings of lines missing: script goes right to left]
    3. Half the hieroglyphics GONE; Last 14 lines incomplete.
  2. Language: Not too many speakers of ancient Egyptian around

Thomas Young (1819):

  1. First hypothesized that at least some symbols had phonetic uses.
  2. Deciphered Ptolemaios and Berenika, pharoah and queen, from the Rosetta stone and another source, because both were enclosed in a loop called a cartouche.
  3. Correctly guessed the phonetic value of several symbols.
  4. One symbol for i occurs in both names, establishing that the symbols had conventionalized phonetic uses.
  5. Did not go on to hypothesize that the writing system was largely phonetic.

Jan Francois Champollion (1824):

  1. Began by extending Young's ideas to other cartouches, assigning sound values to a number of new hieroglyphs, still under the impression that phonetics worked only for names or foreign words.
  2. Decipherings of Ptolemy and Cleopatra shared a number of letters, further establish conventionality of phonetic values.
  3. Decipherment of Ramses:
      Cartouche for Ramses ( Simon Singh )
      X -Y -s - s
      glyph "X" resembles a picture of a sun
      "ra" the Coptic word for "sun"
      Ra - Y - s - s
      Vowels often left out
      Cartouche encloses the name of important person
      RaYsVs => Ramses
    1. Coptic language is the language whose phonetics are being represented
    2. "Rebus" principle: ideograph being borrowed to serve a phonetic function.
    3. Some symbols are ideographic ("sun")
  4. After a lot of work:
    1. Most of the time symbols being used phonetically, some representing combinations of sounds
  5. Nowadays our undertsanding of hieroglyphics pretty good. Can even decrypt encrypted hieroglyphs in pharoahs' tombs.
  • Some ancient scripts solved (Babylonian cuneiform, Kok-Turki runes in Turkey, Grahmi alphabet of India)
  • Many still unsolved: Etruscan, Iberian, Linear A
  • References  

    1. Singh, Simon. 1999. The Code Book. Anchor: New York.
    2. Adkins, Roy and Adkins, Lesley. 2001. The Keys of Egypt HarperCollins: New York.
    3. Collier, Mark and Manley, Bill. 1998. How to Read Egyptian Hieroglyphs British Museum Press: London.
    4. Jean-François Champollion. 2001 Egyptian Diaries Gibson Square Books.
    Linear B  

    Linear B: Writings on a series of clay tablets in absolutely amazing condition discovered in 1900 during Arthur Evans excavation of the palace of Knossos on the island of Crete (Minoan civilization,pre-Hellenic):

    1. Refinement of Linear A, also at Knossos [1750-1450 B.C.]
    2. Linear B tablets date 1450-1375 B.C. (the largest number of tablets found)
    3. 90 distinct characters
    4. Many tablets appear to be inventories
    5. Not clear what language Linear B was written in.
      1. Greek [Various scholars]
      2. A Minoan language, now lost [Arthur Evans]

    Question 1: What kind of writing system does the number of characters suggest? Ideographic or phonetic?

    Question 2: If phonetic, alphabetic or syllabic?

    Linear B

    Alice Kober, mid 40s, Brooklyn College

    Her work focused on words that varied slightly in the forms they appeared in:

    1. Stem/suffix variation
    2. Suggests a highly inflected language (consistent with Greek). Endings for
      1. gender
      2. tense
      3. case
      4. number
    3. Some examples:
          Word 1
        Word 2
        Case 1 25-67-37-57 70-52-41-57
        Case 2 25-67-37-36 70-52-41-36
        Case 3 25-67-05 70-52-12
        Noun inflection in Attic Greek (Case)
        Nom logos theos
        Acc logon theon
        Gen logou theou
    4. Bridging syllable: A syllable in which the consonant belongs to the stem and the vowel belongs to the suffix.
        lo • go • s
      go is a bridge syllable beacuse g is part of the noun stem log and o is part of the the suffix. In a syllabic writing system for an inflecting language, stem final consonants will always get obscured by the bridging syllables:
          Word 1
        Word 2
        Case 1 25-67-37-57 70-52-41-57
        Case 2 25-67-37-36 70-52-41-36
        Case 3 25-67-05 70-52-12
    Kober's speculation: 37, 41, 05, 12 are bridging syllables
    1. Same consonant ending stem:
      1. 37 and 05
      2. 41 and 12
    2. Same vowel beginning case suffix:
      1. 37 and 41
      2. 05 and 12

    Michael Ventris b. 1922. Continued the logic with one refinement (1951-53):

      There must be symbols that stand for vowel only syllables. These should occur frequently at the beginnings of words.
    The second part of this reasoning is open to question, if the language is unknown. It requires that word initial vowel syllables occur more often than word-internal vowel syllables. In Japanese hiragana, word-internal occurrences of -i- and -u- are common, because they are used to mark vowel length! But this assumption turns out to be correct for Greek!

    Looking at word-initial statistic, Ventris decided the symbols we are numbering 61 and 8 must be vowel-only syllables.

    Continuing Kober's logic we get the following table Symbols beginning with the same consonant are in the same row, symbols ending with the same vowel are in the same column:

      Linear B Table
      Consonants Vowels
      1 2 3 4 5
      I         57
      II 40   75   54
      III 39       3
      IV   36      
      V   14     1
      VI 37 5   69  
      VII 41 12     31
      VIII 30 52 24 55 6
      IX 73 15     80
      X   70 44    
      XI 53       76
      XII   2 27    
      XIV     13    
      XV   32 78    
      pure vowels   61     8

    Thus far all of the reasoning is structural. Not a single phonetic value has yet been hypothesized.

    Now a guess about a word with a number of occurrences, which begins with 08 (a vowel). Guess it's an important town in the area: Amnisos.

      A miniso -s
    Notice one guess carries a HUGE amount of information. If correct, we know something about the phonetic values of all the sounds in the same rows and columns as 73, 30, 12, 08.

    Leading to another frequently occurring word which shares sign 12, and also has 52 and 70, two signs in the same vowel column:

      ?o ?oso  
      ko noso s
    Big bonus: 52 and 30 are now supposed to represent no and ni respectively, and they occur in the same consonant column!

    Leading to a guess about a 3rd frequently occurring word:

      ?? ?iso  
      tu liso s

    If the guesses so far are correct, we can now identify sign 05 completely:

      05 -o- column, t- row
      05 = to
    and sign 31
      31 -a- column, s- row
      31 = sa
    This gives two complete words (sign sequences often occurring at the ends of messages):
      Signs Values Greek Word Gloss
      05 12 to-so tossos Total (Masc.)
      05 31 to-sa tossa Total (Fem.)
    Greek guess consistent with idea that these are inventories.

    Ventris found other plausible Greek words. But could these all be borrowings? More and more evidence piled up. John Chadwick, a classicist and cryptographer, joined Ventris (adding a better knowledge of Greek) and they made rapid progress.

    1. Resemblance of some letters to a classical Cypriot script proved not to be accidental [This originally argued against a Greek source for Linear B, the known Cypriot representation of "s" syllables never occurred at the end of a word. But this was misleading. The Linear B convention was to always omit final "s"].
    2. Insights into a version of Greek 500 years earlier than homer, Mycenaean Greek, different in grammar and lexicon.
    3. The mystery of Linear A now sharpened.
    Linear A  

    1. Linear B writing clearly related to Linear A, an earlier set of tablets found at Knossos [1750-1450 B.C.] After about 1450 ALL traces of Linear A vanish.
    2. Linear B tablets date 1450-1375 B.C. (by far the larger number of tablets)

    Two rival civilizations found in the Aegean in this era (archeological evidence of their distinctness decisive):

    1. Mycenean: Based on Greek mainland around Mycenae
    2. Minoan: Based on Crete (where Knossos is)

    What should we conclude about Linear A? And about the abrupt transition to Linear B? Here are the salient facts:

    1. Some language we called EteoCretan (for "True Cretan") was spoken on Crete. We have evidence from other writing systems. is a form of Greek and some reason to think they didn't
    2. Linear A remains undeciphered to this day, despite the best efforts of scholars

    1. Linear B also has logograms (not discussed here)
    2. Linear B has unicode designations (10000-1007F) and there are fonts that can represent it for all you Linear B fans.

    The following are both pretty good reads:

    1. Chadwick, John. 1990. The Decipherment of Linear B Cambridge University Press; Cambridge.
    2. Chadwick, John. 1976. The Mycenean World