Révision 177

CSL17/arithmetic.tex (revision 177)
42 42

  
43 43

  
44 44
\anupam{Collection principles for prenexing? Otherwise need to add closure under sharply bounded quantifiers.}
45
\begin{definition}\label{def:polynomialinduction}
46
[Polynomial induction]
47
The \emph{polynomial induction} axiom schema, $\pind$, consists of the following axioms,
48
\[
49
A(0) 
50
\cimp (\forall x^{\normal} . ( A(x) \cimp A(\succ{0} x) ) )
51
\cimp  (\forall x^{\normal} . ( A(x) \cimp A(\succ{1} x) ) ) 
52
\cimp  \forall x^{\normal} . A(x)
53
\]
54
for each formula $A(x)$.
45 55

  
56
For a class $\Xi$ of formulae, $\cax{\Xi}{\pind}$ denotes the set of induction axioms when $A(x) \in \Xi$. 
46 57

  
58
%We write $I\Xi$ to denote the theory consisting of $\basic$ and $\cax{\Xi}{\ind}$.
59
\end{definition}
60

  
61

  
47 62
\begin{definition}
48 63
Define the theory $\arith^i$ consisting of the following axioms:
49 64
\begin{itemize}
50 65
	\item $\basic$;
51 66
	\item $\cpind{\Sigma^\safe_i } $:
52 67
\end{itemize}
53
and an inference rule:
68
and an inference rule, called $\rais$:
54 69
\[
55 70
 \dfrac{\forall \vec x^\normal . \exists  y^\safe .  A }{ \forall \vec x^\normal .\exists y^\normal . A}
56 71
\]
......
174 189
%The distinction between modal and nonmodal formulae in $(R)$ induces condition 1
175 190
 Conditions 2 and 3 are standard requirements for nonlogical rules, independently of the logical setting, cf.\ \cite{Beckmann11}. Condition 2 reflects the intuitive idea that, in our nonlogical rules, we often need a notion of \textit{bound} variables in the active formulas (typically for induction rules), for which we rely on eigenvariables. Condition 3 is needed for our proof system to admit elimination of cuts on quantified formulas.
176 191

  
192
%\begin{definition}
193
%[Polynomial induction]
194
%The \emph{polynomial induction} axiom schema, $\pind$, consists of the following axioms,
195
%\[
196
%A(0) 
197
%\cimp (\forall x^{\normal} . ( A(x) \cimp A(\succ{0} x) ) )
198
%\cimp  (\forall x^{\normal} . ( A(x) \cimp A(\succ{1} x) ) ) 
199
%\cimp  \forall x^{\normal} . A(x)
200
%\]
201
%for each formula $A(x)$.
202
%
203
%For a class $\Xi$ of formulae, $\cax{\Xi}{\pind}$ denotes the set of induction axioms when $A(x) \in \Xi$. 
204
%
205
%We write $I\Xi$ to denote the theory consisting of $\basic$ and $\cax{\Xi}{\ind}$.
206
%\end{definition}
177 207

  
208
As an example any axiom can be represented by such a nonlogical rule $(R)$, with no premise ($I=\emptyset$), $\Delta'$ equal to the axiom and $\Gamma=\Sigma'=\Pi$. For instance the axiom $\pind$ of Def. \ref{def:polynomialinduction}.
209

  
210
Actually  $\pind$ is equivalent to the following rule:
211
\begin{equation}
212
\label{eqn:ind-rule}
213
\small
214
\vliinf{\pind}{}{ \normal(t) , \Gamma , A(0) \seqar A(t), \Delta }{ \normal(a) , \Gamma, A(a) \seqar A(\succ{0} a) , \Delta }{ \normal(a) , \Gamma, A(a) \seqar A(\succ{1} a) , \Delta  }
215
\end{equation}
216
where $I=2$ and  in all cases, $t$ varies over arbitrary terms and the eigenvariable $a$ does not occur in the lower sequent of the $\pind$ rule.
217

  
218
Similarly the $\rais$ inference rule of Def. \ref{} is represented by the rule:
219
 \[
220
 \begin{array}{cc}
221
    \vlinf{\rais}{}{ \Gamma , \Sigma' \seqar \Delta' , \Pi  }{ \{\Gamma , \Sigma_i \seqar \Delta_i , \Pi \}_{i \in I} }
222
\end{array}
223
\]
178 224
\subsection{Free-cut free normal form of proofs}
179 225
\todo{State theorem, with references (Takeuti, Cook-Nguyen) and present the important corollaries for this work.}
180 226

  
CSL17/ph-macros.tex (revision 177)
87 87
	\newcommand{\cut}{\it{cut}}
88 88
	\newcommand{\wk}{\mathit{wk}}
89 89
	\newcommand{\cntr}{\mathit{cntr}}
90
	\newcommand{\rais}{\mathit{raise}}
91
 
90 92
	
91 93
% theories
92 94
	\newcommand{\theory}{\mathcal T}
93 95
	\newcommand{\system}{\mathcal S}
94 96
	
97
	%induction
98
%	\newcommand{\ind}{\mathit{PIND}}
99
%	\newcommand{\pind}{\mathit{PIND}}
100
	\newcommand{\cax}[2]{#1\text{-}#2}
101

  
102
	\newcommand{\sigone}{\Sigma^{\word^+}_1 }
103
	\newcommand{\sigzer}{\Sigma^{\word^+}_0}
104
	\newcommand{\bharith}{\mathcal A^1_2}
105
	%\newcommand{\arith}{I\sigone}
106
	
107
	

Formats disponibles : Unified diff