Mathematics of Language: Errors
  
List of Errors
- Page 83 The sentence Notice that the edge colours 
   only the vertex colours are used to steer the derivation.
   is not grammatical and misleading. It should read instead
   Notice that only the edge colours are used to steer the 
   derivation. (This has been spotted by John Colby.)
- Seite 108, (2.22)On the left hand side it must read  
   A &u2192; a. (This has been spotted by Christian 
   Wurm.)
- Page 295 Exercise 141. I have grossly underestimated 
   the status of Leibniz' Principle in set theory. It comes down
   to the identity of indiscernibles. Set theory with Leibniz' 
   Principle has been studied by Jan Mycielski. The structure 
   theory of such universes is everything but easy, see Ali Enayat: 
   "Leibnizian models of set theory", 
   The Journal of Symbolic Logic 69, 2004, 757 - 789. 
- Page 409. Theorem 5.45 ist false for k ≥ 3. This  
   is the result of Makoto Kanazawa, Gregory M. Kobele, Jens 
   Michaelis, Sylvain Salvati and Ryo Yoshinaka: The Failure of 
   the Strong Pumping Lemma for Multiple Context-Free Languages
   (unpublished manuscript). 
- Page 425-426 Die claim that indexed grammars can only generate 
   PTIME languages has been refuted by Bill Rounds in 1973
   ("Complexity in Recognition in Intermediate-Level Languages", 
   in: Proceedings 14th Annual IEEE Symposium on Switching and 
   Automata Theory, 145 - 158. IEEE Computer Science, 
   Northridge, CA). (Thanks for Hans-Jörg Thiede for 
   spotting this.) 
- Page 470 Definition 6.2 should contain the clause that 
   L is not empty.  
- Page 472 The claim that MSO can differentiate between 
   a class that contains the empty word and this class without the 
   empty word is false. This does hold however for QML. 
   Adding the condition that L is not empty the problem 
   will disappear. (Geoffrey Pullum has noticed this.)
- Page 502 The claim that a transducer can be made 
   deterministic is of course false. The implicit use of the 
   analogous fact for finite automata is not licit. The problem 
   is that while strings have a unique decomposition into letters, 
   pairs of strings do not have such a decomposition. For example
   the pair ⟨a, a⟩ has two decompositions into a product 
   of the pairs ⟨a : ε⟩ or ⟨ε : a⟩. 
- Page 520, Example (6.172) This sentence is not 
   grammatical. It should read 
   John is believed to be responsible for himself.
  
Marcus Kracht
Last changed: Thu April 19, 2012 14:30:00