specifies that the file is a TAN file containing lexico-morphology data about a class-1 source or a language in general. Root element. identifies which lexical definition is meant. This attribute is essential in cases where a lexicon has multiple entries for lexemes (lexical headwords) that are identical. Because there is no TAN format for lexicons, assertions about lexica will not be validated. points via idrefs to one or more lexicons, persons, organizations, or algorithms. This attribute is weakly inheritable. See main.xml#inheritable_attributes specifies the quantity of tokens that are the subject of the <ana>. This attribute is useful for language-specific TAN-A-lm files, where the total number of tokens that formed the set might be lost. It is also useful for calculating the relative probability of token-to-lm combinations. This attribute is weakly inheritable. See main.xml#inheritable_attributes points to one or more <morphology> IDs This attribute is inheritable. See main.xml#inheritable_attributes contains a one or more assertions about the lexical or morphological properties of one or more tokens. Claims within an <ana> are distributive. That is, every combination of <l> and <m> within an <lm> is claimed to be true for every <tok>. The sequence of consecutive <ana>s is immaterial. names a lexeme, by pointing to the main word entry in the lexicon identified by the element's inherited value(s) of @lexicon. This element should not be used to point to roots, only to lexical headwords. In many languages, especially those that are lightly inflected, the lexeme will be identical to the word token itself. If <l> is omitted, the calculated value of <tok> is to be inferred as its value. The sequence of consecutive <l>s is immaterial. groups lexical or morphological data. Components within <lm> combine with each other and with all sibling <tok>s. That is, every <l> is asserted against every <m> within an <lm> is asserted of every <tok>. The sequence of consecutive <lm>s is immaterial. carries a morphological code that conforms to the rules or patterns defined in the TAN-mor file upon which the data depends. Codes are space-delimited. If a value of <m> violates the rules established by the dependency TAN-mor file, an error will be generated. For more about how codes are built, and how they function, see main.xml#tan-a-lm. The sequence of consecutive <m>s is immaterial. [^\+\s]+(\s+[^\+\s]+)* points to a token in the abstract, no matter where it is found, either in the source (if there is one) or the language (if there is no source). The sequence of consecutive <tok>s is immaterial, except in the context of <group>. specifies an exact value of @val of <tok>s in the file. Collectively, every <tok> should point to at least one <tok-is>. This element is an optional way to improve the efficiency of validation and applications. The sequence of consecutive <tok-is>s is immaterial. specifies one opening string that characterizes the @val of <tok>s in the file. This element is an optional way to improve the efficiency of validation and applications. The sequence of consecutive <tok-is>s is immaterial. names a lexicographical authority. identifies a dependency <TAN-mor> file that defines the parts of speech for a language, the codes for those parts, and the rules for combining them