« 2011年8月 | トップページ | 2011年10月 »

2011年9月の3件の記事

「门内有径」の意味

今朝程 (2011/09/17 07:06:49)、キーフレーズ [门内有径 日本語 意味] で、このサイトを訪問された方がいらしたようだ。

まぁ、順当に考えて、これは「門」(「门」は、「門」の簡体字) の「内側」には「小道」があると云うことでしょうね。まとめれば「門の内側には小道がある」。

視点を門の外側に置くなら「門の向こう側には小道がある」と云う訳し方も可能でしょう。

大きなお世話ながら [门内有径] で検索してみると、これは、林語堂 (林语堂, Lin Yutang) が、或るアメリカ人に対して、中国人の理想の暮しを表現したものとして引用したと云う明代の書家・画家 [陳継儒] (陈继儒/陳繼儒) の「小窗幽记」(「醉古堂剑扫」とも。日本新字体ではそれぞれ、「小窗幽記」及び「醉古堂剣掃」) に収めれた (巻六 [集景]) 一節

门内有径,径欲曲;径转有屏:屏欲小;屏进有阶,阶欲平;阶畔有花,花欲鲜;花外有墙,墙欲低;墙内有松,松欲古:松底有石,石欲怪;石面有亭,亭欲朴;亭后有竹,竹欲疏;竹尽有室,室欲幽;室旁有路,路欲分;路合有桥,桥欲危;桥边有树,树欲高;树阴有草,草欲青;草上有渠,渠欲细;渠引有泉,泉欲瀑;泉去有山,山欲深:山下有屋,屋欲方;屋角有圃,圃欲宽;圃中有鹤,鹤欲舞;鹤报有客,客不俗;客至有酒,酒欲不却;酒行有醉,醉欲不归。
--wiki.guoshuang.com : 小窗幽记 陈继儒

がヒットする。

高校の漢文の授業を居眠りしないでいた人物には説明不要だろうが、「径欲曲」の「欲」は、漢文の所謂「助字」で、通常「いまにも・・・しようとしいてる」と云う意味だと説明されるのが普通だ。人口に膾炙する 杜甫の「絶句 江碧鳥逾白 山青花欲然 今春看又過 何日是歸」の「山青花欲然」は「山は木々の青葉で被われ、その花は今にも燃えだすのではないかと思われるほど赤い」と云う訣だ。

しかし、上記引用文では、「欲」が繰り返して使われていて、描写そのものと云うより、言い切らないことで空想であることを暗示し、更に、一種の韻律をもたらす為の修辞的技法になっている。だから、必ずしもイチイチ「いまにも・・・しようとしいてる」と解釈しないほうが良い。

こんな感じだろう:

門の内側には小道があって、曲がっている。小道を曲がると、二之門があって、これは小さい。二之門を進むと、石段があって平坦になっている。石段の傍らには花が咲いていて、鮮やかである。花の向こう側には垣根があって、背が低い。垣根の内側には松の木があって、古色がある。松の根方には石があって、不思議な形をしている。石の傍には亭 (あづまや) があって、飾り気が無い。亭の後には竹林があって、まばらである。竹林を過ぎると小屋があって、ひっそりとしている。小屋の傍には道が通っていて、分かれている。分かれた道が合う所に橋があって、谷底は深い。橋のたもとには樹があって、高い。木陰には草があって、青々としている。草の近くには溝があって、細い。溝は泉に繋がっていて、その泉は水を溢れさせている。泉を過ぎると山があって、深い。山の麓には屋敷があって、四角い。屋敷の隅には菜園があって、広々としている。菜園の中には鶴がいて、踊っている。鶴は来客があったのを知らせたのだが、その客は高雅の士である。来客とあれば酒であり、酒は辞退されない。酒がすすめば酔いを発し、酔いを発せば帰らない

| | コメント (0) | トラックバック (0)
|

Errata and Typographical Remarks on Alan Turing's "On Computable Numbers, with an Application to the Entscheidungsproblem"

INTRODUCTION

I'm writing this article to provide you a list of typos found in Alan Turing's [main paper 1937](1) as well as a few typographical remarks. Punctuative and other trivial errors are excluded from the list. (e.g., “If there is no \alpha\rightarrow\mathfrak{B}.” should read “If there is no \alpha$, $\rightarrow\mathfrak{B}.” in the first right column on p.237.) None of his own peculiar wording, some of which appear to me too informal, are taken in either. (e.g., on p.237, he wrote “= \mathfrak{q}, say” where I would write “=, say, \mathfrak{q}”, and “the first symbol marked \alpha”, “the first symbol marked with \alpha”.)

The first half of the paper shows tables describing how the machine (the Turing Machine, which has a format superficially different from the current versions) computes. Unfortunately, there are no captions nor sections in those tables. With different numbers of lines in columns from table to table, the line number can't be effectively used to locate a letter/word on a page. Therefore, I will quote a piece of text including a concerned point for an enough length when I think it helps you to find quickly where the issue is.

Notes on the mathematical font: The fraktur font used here is a little different from the one seen in the [main paper 1937]. Though, they are so similar as to cause no difficulties, I believe. As for the script font, contrarily, the discrepancy is so great that I wish you cautious not to be misled by the different appearances of the typefaces.

LIST

page 238, rows 6-7 of the table It reads as follows:

\begin{tabular}{lll}
$\mathfrak{cr}(\mathfrak{C},\mathfrak{B},\alpha)\quad$ & $\quad\mathfrak{c}\Big(\mathfrak{re}(\mathfrak{C},\mathfrak{B},\alpha,a),\mathfrak{B},\alpha\Big)\qquad$  &\multirow{2}{2cm}{\rightcolumnnote{43mm}{\vspace{1mm}\quad$\mathfrak{cr}(\mathfrak{B},\alpha)$ differs from $\mathfrak{ce}(\mathfrak{B},\alpha)$ only in that the letters $\alpha$ are not erased.  The $m$-configuration $\mathfrak{cr}(\mathfrak{B},\alpha)$ is taken up when no letters “$a$” are on the tape.}}\\
$\ \mathfrak{cr}(\mathfrak{B},\alpha)$ & $\mathfrak{cr}\Big(\mathfrak{cr}(\mathfrak{B},\alpha), \mathfrak{re}(\mathfrak{B},a,\alpha),\alpha\Big)$ & \end{tabular}
However no typos are seen here, it may puzzle you until you realize that there are two different letters \alpha and a, which look quite alike in the original copy of the [main paper 1937]. Worse, using the Latin letter a deviates from Turing's own general convention of using a small Greek letter for a symbol (cf. p.236, l.1). The table should be rewritten by replacing a with a small Greek letter, say \beta, as follows:

\begin{tabular}{lll}
$\mathfrak{cr}(\mathfrak{C},\mathfrak{B},\alpha)\quad$ & $\quad\mathfrak{c}\Big(\mathfrak{re}(\mathfrak{C},\mathfrak{B},\alpha,\beta),\mathfrak{B},\alpha\Big)\qquad$  &\multirow{2}{35mm}{\rightcolumnnote{43mm}{\vspace{1mm}\quad$\mathfrak{cr}(\mathfrak{B},\alpha)$ differs from $\mathfrak{ce}(\mathfrak{B},\alpha)$ only in that the letters $\alpha$ are not erased.  The $m$-configuration $\mathfrak{cr}(\mathfrak{B},\alpha)$ is taken up when no letters “$\beta$” are on the tape.}}\\
$\ \mathfrak{cr}(\mathfrak{B},\alpha)$ & $\mathfrak{cr}\Big(\mathfrak{cr}(\mathfrak{B},\alpha), \mathfrak{re}(\mathfrak{B},\beta,\alpha),\alpha\Big)$ & \\
\end{tabular}

page 238, row 8 of the table: The row

\begin{tabular}{ll}
 $\hspace{5mm}\mathfrak{cp}(\mathfrak{C},\mathfrak{A},\mathfrak{E},\alpha,\beta)\qquad$ &$\mathfrak{f}^{\prime}\left(\mathfrak{cp}_{1}(\mathfrak{C}_{1}\mathfrak{A},\beta),\mathfrak{f}(\mathfrak{A},\mathfrak{E},\beta),\alpha\right)$
\end{tabular}
should read

\begin{tabular}{ll}
 $\hspace{5mm}\mathfrak{cp}(\mathfrak{C},\mathfrak{A},\mathfrak{E},\alpha,\beta)\qquad$ &$\mathfrak{f}^{\prime}\left(\mathfrak{cp}_{1}(\mathfrak{C},\mathfrak{A},\beta),\mathfrak{f}(\mathfrak{A},\mathfrak{E},\beta),\alpha\right).$
\end{tabular}
.
Here, the subscript “1” after \mathfrak{C} in (\mathfrak{C}_{1}\mathfrak{A},\beta) is an error for a comma “,”.

page 239, rows 1-7 of the table: Two m-functions \mathfrak{q}(\mathfrak{C}) and \mathfrak{q}(\mathfrak{C},\alpha) are defined as follows:

\begin{tabular}{ll}
$
\mathfrak{q}(\mathfrak{C}) \qquad\ 
<br /><br />\begin{cases}
 \text{Any}\  \qquad R \quad & \quad\mathfrak{q}(\mathfrak{C})\\
 \text{None} \qquad R \quad & \quad\mathfrak{q}_{1}(\mathfrak{C})
\end{cases}
\qquad$
 & \hspace{5mm}\rightcolumnnote{40mm}{\vspace{2mm}\quad$\mathfrak{q}(\mathfrak{C},\alpha)$. \quad The machine finds the last symbol of form $\alpha. \quad \rightarrow\mathfrak{C}.$}
\\
\end{tabular}

\begin{tabular}{ll}
$
\mathfrak{q}_{1}(\mathfrak{C}) \qquad
\begin{cases}
 \text{Any}  \qquad\ R \quad & \quad\mathfrak{q}(\mathfrak{C})\\
 \text{None} \qquad \phantom{R} \quad & \quad\mathfrak{C}
\end{cases}
$
 &\\
\end{tabular}

\begin{tabular}{ll}
$\mathfrak{q}(\mathfrak{C},\alpha) \phantom{\text{Any}  R \qquad \quad \mathfrak{q}(\mathfrak{C})}$
 &$\quad\mathfrak{q}\left(\mathfrak{q}_{1}(\mathfrak{C},\alpha)\right)$\\
\end{tabular}

\begin{tabular}{ll}
$
\mathfrak{q}_{1}(\mathfrak{C},\alpha) \quad
<br /><br />\begin{cases}
 \quad\alpha  \phantom{L} \quad & \quad\mathfrak{C}\\
 \text{not}\ \alpha \qquad L \quad & \mathfrak{q}_{1}(\mathfrak{C},\alpha)
\end{cases}
$
 &\\
\end{tabular}
Notwithstanding, neither appears again in the paper. Instead, an m-function \mathfrak{g}(\mathfrak{C},\alpha) is used where \mathfrak{q}(\mathfrak{C},\alpha) should be. We must amend the original text to adopt only one of the two. Hence, I discard \mathfrak{g} with \mathfrak{q} left in, and proceed with my work.

page 244, the table and the note for the m-function \mathfrak{con}(\mathfrak{C,\alpha}): It reads:

\begin{tabular}{lll}
$\mathfrak{con}(\mathfrak{C},\alpha)$&
$
\begin{cases}
 \text{Not}\ A  \quad R,R & \mathfrak{con}(\mathfrak{C},\alpha)\\
 \quad A \quad L,P\alpha,R & \mathfrak{con}_{1}(\mathfrak{C},\alpha)
\end{cases}
$
 &\multirow{2}{3.5cm}{\rightcolumnnote{45mm}{\vspace{-2mm}\quad $\mathfrak{con}(\mathfrak{C},\alpha)$. Starting from an $F$-square, $S$ say, the sequence $C$ of symbols describing a configuration closest on the right of $S$ is marked out with letters $\alpha$. \qquad $\rightarrow{\mathfrak{C}}.$}}
\\
\vspace{-12mm}
$
\mathfrak{con}_{1}(\mathfrak{C},\alpha) $&$
\begin{cases}
 \quad A \quad R,P\alpha,R & \mathfrak{con}_{1}(\mathfrak{C},\alpha)\\
 \quad D \quad R,P\alpha,R & \mathfrak{con}_{2}(\mathfrak{C},\alpha)
\end{cases}
$\\
&&\\

$
\mathfrak{con}_{2}(\mathfrak{C},\alpha) $&$
\begin{cases}
 \quad C \quad R,P\alpha,R & \mathfrak{con}_{2}(\mathfrak{C},\alpha)\\
 \text{Not}\ C \quad R,R & \mathfrak{C}
\end{cases}
$&\rightcolumnnote{45mm}{\vspace{13mm}\quad $\mathfrak{con}(\mathfrak{C},\ )$. In the final configuration the machine is scanning the square which is four squares to the right of the last square of $C$.  $C$ is left unmarked.}
\\
\end{tabular}
In the second column of the table, the letter C refers to a single symbol in the standard description (S.D[.]) of a Turing machine for a specific computation (cf. p.240), while in the right column the letter C represents a sequence of symbols that makes up a configuration of the machine (cf. p.231). To avoid ambiguity, the right column should be rewritten to get rid of the letter C, though I'm afraid I must decline to write out the details.

page 244, the table and the note for the m-configuration \mathfrak{anf}: It reads as follows:

\begin{tabular}{lll}
$\mathfrak{anf}\hspace{25mm}$ &$\mathfrak{g}(\mathfrak{anf}_{1},:)$ &\hspace{5mm}\multirow{2}{3.5cm}{\rightcolumnnote{45mm}{\vspace{1mm}\quad $\mathfrak{anf}$. \quad The machine marks the configuration in the last complete configuration with $y$.\quad $\rightarrow{\mathfrak{kom}}$.}}
\\
$\mathfrak{anf}_{1}\hspace{10mm}$ &$\mathfrak{con}(\mathfrak{kom},y)$&
\\
\end{tabular}
Here \mathfrak{g}(\mathfrak{anf}_{1},:) should read as \mathfrak{q}(\mathfrak{anf}_{1},:).

page 244, the table and the note for the m-configuration \mathfrak{kmp}: It reads as follows:

\begin{tabular}{ll}
$
\mathfrak{kmp} \qquad\ \mathfrak{cpe}\left(\mathfrak{e}(\mathfrak{kom},x,y),\mathfrak{sim},x,y\right)$
 & \hspace{5mm}\rightcolumnnote{43mm}{\vspace{15mm}\quad$\mathfrak{kmp}$. \quad The machine compares the sequences marked $x$ and $y$. It erases all letters $x$ and $y$. \quad $\rightarrow\mathfrak{sim}$ if they are alike.  Otherwise $\rightarrow\mathfrak{kom}.$}
\\
\end{tabular}
In the first, the m-function \mathfrak{e}(\mathfrak{B},\alpha,\beta) is not defined in the paper, and should be given as \mathfrak{e}(\mathfrak{e}(\mathfrak{B},\beta),\alpha) (The definition of \mathfrak{e}(\mathfrak{B},\alpha) is seen on p.237).

Secondly, until the \mathfrak{kmp} process reaches \mathfrak{sim}, it should not end in \mathfrak{kom}, but in \mathfrak{anf}. The interested subroutine repeats a searching loop from \mathfrak{anf} to \mathfrak{kmp} to find a configuration in the S.D. that coincides with the last configuration in the complete configuration under construction. Therefore, \mathfrak{kmp}, which seems to stand for a German word “\mathfrak{Komparation}” (“comparison” in English), goes back to \mathfrak{anf} (“\mathfrak{Anfang}”, “start”) to start another search routine when the comparison shows a disparity. Therefore, the table and its note should be corrected to:

\begin{tabular}{ll}
$
\mathfrak{kmp} \qquad\ \mathfrak{cpe}\left(\mathfrak{e}(\mathfrak{anf},x,y),\mathfrak{sim},x,y\right)$
 & \rightcolumnnote{43mm}{\vspace{17mm}\quad$\mathfrak{kmp}$. \quad The machine compares the sequences marked $x$ and $y$. It erases all letters $x$ and $y$. \quad $\rightarrow\mathfrak{sim}$ if they are alike.  Otherwise $\rightarrow\mathfrak{anf}.$}
\\
\end{tabular}

pages 245-246: There are a few typos:

  • The table for \mathfrak{sim}_{2} (p.245)
    
\begin{tabular}{l}
 $\mathfrak{sim}_{2}
\begin{cases}
 \quad A & \mathfrak{sim}_{3}\\
 \text{not}\ A \qquad R,Pu,R,R,R \quad & \mathfrak{sim}_{2}
\end{cases}$
\end{tabular}
    should read
    
\begin{tabular}{l}
 $\mathfrak{sim}_{2}
<br /><br />\begin{cases}
 \quad A & \mathfrak{sim}_{3}\\
 \text{not}\ A \qquad L,Pu,R,R,R \quad & \mathfrak{sim}_{2}
\end{cases}$
\end{tabular}.
.
  • The table for \mathfrak{mk} (p.245)
    
\begin{tabular}{ll}
 $\quad\mathfrak{mk}$& $\hspace{30mm}\mathfrak{g}(\mathfrak{mk},:)$\\
\end{tabular}
    should read

    
\begin{tabular}{ll}
 $\quad\mathfrak{mk}$& $\hspace{30mm}\mathfrak{q}(\mathfrak{mk}_{1},:)$\\
\end{tabular}.
.
  • The table for \mathfrak{sh}_{2} (p.245)
    
\begin{tabular}{l}
 $\mathfrak{sh}_{2}
<br /><br />\begin{cases}
 \quad D \qquad R,R,R,R \qquad & \mathfrak{sh}_{2}\\
 \text{not}\ D & \mathfrak{inst}
\end{cases}$
\end{tabular}
    should read
    
\begin{tabular}{l}
 $\mathfrak{sh}_{2}
<br /><br />\begin{cases}
 \quad D \qquad R,R,R,R \qquad & \mathfrak{sh}_{3}\\
 \text{not}\ D & \mathfrak{inst}
\end{cases}$
\end{tabular}.
.
  • The table for \mathfrak{inst} (p.246)
    
\begin{tabular}{ll}
 $\quad\mathfrak{inst}$& $\hspace{30mm}\mathfrak{g}\left(\mathfrak{l}(\mathfrak{inst}_{1}),u\right)$\\
\end{tabular}
    should read
    
\begin{tabular}{ll}
 $\quad\mathfrak{inst}$& $\hspace{30mm}\mathfrak{q}\left(\mathfrak{l}(\mathfrak{inst}_{1}),u\right)$\\
\end{tabular}.
.

page 246: The m-function \mathfrak{inst}_{1}(\alpha) is defined as follows:

\begin{tabular}{ll}
 $\mathfrak{inst}_{1}(L)$& $\hspace{15mm}\mathfrak{ce}_{5}(\mathfrak{ov},v,y,x,u,w)$\\
 $\mathfrak{inst}_{1}(R)$& $\hspace{15mm}\mathfrak{ce}_{5}(\mathfrak{ov},v,x,u,y,w)$\\
 $\mathfrak{inst}_{1}(N)$& $\hspace{15mm}\mathfrak{ce}_{5}(\mathfrak{ov},v,x,y,u,w),$\\
\end{tabular}
The m-function \mathfrak{ce}_{5} has not been defined explicitly in the paper, while the definitions of \mathfrak{ce}_{2} and \mathfrak{ce}_{3} are seen on p.239. Naturally, \mathfrak{ce}_{4} and \mathfrak{ce}_{5} should be defined as follows:

\begin{tabular}{ll}
$\mathfrak{ce}_{4}(\mathfrak{B},\alpha,\beta,\gamma,\delta)$ & $\hspace{15mm}\mathfrak{ce}\left(\mathfrak{ce}_{3}(\mathfrak{B},\beta,\gamma,\delta),\alpha\right)$\\
$\mathfrak{ce}_{5}(\mathfrak{B},\alpha,\beta,\gamma,\delta,\epsilon)$ & $\hspace{15mm}\mathfrak{ce}\left(\mathfrak{ce}_{4}(\mathfrak{B},\beta,\gamma,\delta,\epsilon),\alpha\right)$\\
\end{tabular}.
.

page 247, the 4th line from the bottom:H” should read “\mathscr{H}”.

page 252, line 20: The formula

\begin{tabular}{l}
 $(\exists{u})N(u) \mathand (x)\Big(N(x)\rightarrow{}(\exists{y})F(x,y)\Big) \mathand \Big(F(x,y)\rightarrow{N(y)}\Big)$
\end{tabular}
should read

\begin{tabular}{l}
 $(\exists{u})N(u) \mathand (x)\Big[N(x)\rightarrow{}\Big((\exists{y})F(x,y) \mathand (y)\big(F(x,y)\rightarrow{N(y)}\big)\Big)\Big]$\\
\end{tabular}
.

page 254, the 7th line from the bottom: The formula

\begin{tabular}{l}
 $\mathfrak{A}_{\phi} \mathand F^{(N^{\prime})} \rightarrow \left(-H(u^{(n)},u^{(m)}\right)$\\
\end{tabular}
should read

\begin{tabular}{l}
 $\mathfrak{A}_{\phi} \mathand F^{(N^{\prime})} \rightarrow \left(-H(u^{(n)},u^{(m)})\right)$.\\
\end{tabular}
.

page 255, line 1:\alpha_{n}=0” should read “\alpha_{n}=\pm{\infty}”.

page 256, line 3:-G(\alpha) \rightarrow \alpha \geqslant \xi” should read “-G(\alpha) \rightarrow \alpha > \xi”.

page 256, lines 7-11: There is a passage:

Owing to this restriction of Dedekind's theorem, we cannot say that a computable bounded increasing sequence of computable numbers has a computable limit. This may possibly be understood by considering a sequence such as

\hfill -1,\ -\frac{1}{2},\ -\frac{1}{4},\ -\frac{1}{8},\ -\frac{1}{16},\ \frac{1}{2},\ \cdots\ .\hfill
.

No matter of what I can't understand the passage, it refers unambiguously to, among others, what we call today the Specker sequence, that is “a computable, strictly increasing, bounded sequence of rational numbers whose supremum is not a computable real number.” (Wikipedia “Specker sequence”.)

He might argue here in effect a sequence \{\varepsilon_{K}\}_{K\in\pint} defined by

\varepsilon_{K} \equiv -2+\sum_{N=1}^{K}2^{N-R(N)-1}
where R(N) is the number of satisfactory positive integers (cf. p.241 and p.247) not greater than a positive integer N. The sequence is increasing with an upper bound, and has a supremum real number \displaystyle{}\lim_{K  \rightarrow \infty}{\varepsilon_{K}}. He might refer to a fact that every partial sum is rational and thus computable, though the supremum is not computable (cf. p.247).

As far as following Turing's convention in the [main paper 1937], the first positive integers reaching to a few thousands are unsatisfactory. As many first members of the sequence \{\varepsilon_{K}\}_{K\in\pint} are identical with those of a geometric sequence \{-2^{-K}\}_{K\in\pint}, that is

\hfill -1,\ -\frac{1}{2},\ -\frac{1}{4},\ -\frac{1}{8},\ -\frac{1}{16},\ -\frac{1}{32},\ \cdots\ .\hfill
.

page 257, line 9: The formula

\begin{tabular}{l}
  $\hspace{5mm}\mathand \Big[H(w,z) \mathand G(z,t) \mathbin{\mathrm{v}} G(t,z) \rightarrow \Big(-H(w,t)\Big)\Big]$
 \end{tabular}
should read

\begin{tabular}{l}
 $\hspace{5mm}\mathand \Big[H(w,z) \mathand \Big(G(z,t) \mathbin{\mathrm{v}} G(t,z)\Big) \rightarrow \Big(-H(w,t)\Big)\Big]$
 \end{tabular}
.
Composition of Logical conjunction “\&” and sum “\mathrm{v}” are not associative, and thus a pair of parentheses must be added to enclose “G(z,t) \mathbin{\mathrm{v}} G(t,z)”.

page 257, the 2nd line from the bottom:m \neq \eta(u)” should read “m \neq \eta(n)”.

page 257, the last line: The formula

\begin{tabular}{l}
 $\hspace{5mm} \mathfrak{A}_{\eta} \mathand F^{(M^{\prime})} \rightarrow G(u^{\eta((n))},u^{(m)}) \mathbin{\nu} G(u^{(m)},u^{\eta((n))})$\\
\end{tabular}
should read

\begin{tabular}{l}
 $\hspace{5mm} \mathfrak{A}_{\eta} \mathand F^{(M^{\prime})} \rightarrow G(u^{\eta((n))},u^{(m)}) \mathbin{\mathrm{v}} G(u^{(m)},u^{\eta((n))})$.
\end{tabular}
.
In the paper, the small Latin letter “\mathrm{v}” seems a prescribed symbol for the logical sum, and the small Greek letter “\nu” should be replaced with “\mathrm{v}”.

page 258, lines 2-3: The formula

\begin{eqnarray*}
  &&\mathfrak{A}_{\eta} \mathand F^{(M^{\prime})} \rightarrow \Big[\big\{G(u^{(\eta(n))},u^{(m)}) \mathbin{\nu} G(u^{(m)}, u^{(\eta(n))})\hspace{20mm}\\
 &&\hspace{40mm} \mathand H(u^{(n)},u^{(\eta(n))} \big\} \rightarrow \left(-H(u^{(n)},u^{(m)})\right) \Big]
 \end{eqnarray*}
should read

\begin{eqnarray*}
  &&\mathfrak{A}_{\eta} \mathand F^{(M^{\prime})} \rightarrow \Big[\Big\{\Big(G(u^{(\eta(n))},u^{(m)}) \mathbin{\mathrm{v}} G(u^{(m)}, u^{(\eta(n))})\Big)\hspace{20mm}\\
 &&\hspace{40mm} \mathand H(u^{(n)},u^{(\eta(n))}) \Big\} \rightarrow \left(-H(u^{(n)},u^{(m)})\right) \Big]
 \end{eqnarray*}
.

page 258, line 10: “the m-configuration b” should read “the m-configuration \mathfrak{b}”.

page 258, the 6th line from the bottom: The row

\begin{tabular}{ll}
 $\hspace{5mm} \mathfrak{u}_{2}$ & $\hspace{30mm}\mathfrak{re}(\mathfrak{u}_{3},\mathfrak{u}_{3},k,h)$\\
\end{tabular}
should read

\begin{tabular}{ll}
 $\hspace{5mm} \mathfrak{u}_{2}$ & $\hspace{30mm}\mathfrak{re}(\mathfrak{u}_{3},\mathfrak{b},k,h)$\\
\end{tabular}
.
The machine \mathscr{N}^{\prime} repeats the computation of the machine \mathscr{N} to get a figure \phi_{n}(n) for every positive integer n, and must bring its m-configuration back to \mathfrak{b}, the first m-configuration of \mathscr{N}, each time one figure is computed and the symbol k disappears from the complete configuration of \mathscr{N}^{\prime}. (I would warn you that the appearance of the script capital \mathscr{N} used here is quite different from the one seen in the [main paper 1937]).

page 259, the last line:S” should read “S_{l}” to accord with the term “R_{S_{l}}(x,y)” on the one line before.

page 260, lines 7-9: In his [correction paper 1938, p.544](2), Turing changed the definition of \mathrm{Inst}\{q_{i}S_{j}S_{k}Lq_{l}\} from

\begin{eqnarray*}
&&(x,y,x^{\prime},y^{\prime})\Big\{\Big(R_{S_{j}}(x,y) \mathand I(x,y) \mathand K_{q_{i}}(x) \mathand F(x,x^{\prime}) \mathand F(y^{\prime},y)\Big)\\
&&\rightarrow \Big(I(x^{\prime},y^{\prime}) \mathand R_{S_{k}}(x^{\prime},y) \mathand K_{q_{l}}(x^{\prime})\\
&&\hspace{40mm}\mathand (z)\left[F(y^{\prime},z) \mathbin{\mathrm{v}} \left(R_{S_{j}}(x,z) \rightarrow R_{S_{k}}(x^{\prime},z)\right)\right]\Big)\Big\}
\end{eqnarray*}
to

\begin{eqnarray*}
&&(x,y,x^{\prime},y^{\prime})\Big\{\Big(R_{S_{j}}(x,y) \mathand I(x,y) \mathand K_{q_{i}}(x) \mathand F(x,x^{\prime}) \mathand F(y^{\prime},y)\Big)\\
&&\rightarrow \Big(I(x^{\prime},y^{\prime}) \mathand R_{S_{k}}(x^{\prime},y) \mathand K_{q_{l}}(x^{\prime}) \mathand F(y^{\prime},z) \mathbin{\mathrm{v}} \Big[\big(R_{S_{0}}(x,z) \rightarrow R_{S_{0}}(x^{\prime},z)\big)\\
&&\mathand \big(R_{S_{1}}(x,z) \rightarrow R_{S_{1}}(x^{\prime},z)\big) \mathand \ldots \mathand \big(R_{S_{M}}(x,z) \rightarrow R_{S_{M}}(x^{\prime},z)\big)\Big]\Big)\Big\}.
\end{eqnarray*}
However, the new definition should be amended as follows:

\begin{eqnarray*}
&&(x,y,x^{\prime},y^{\prime})\Big\{\Big(R_{S_{j}}(x,y) \mathand I(x,y) \mathand K_{q_{i}}(x) \mathand F(x,x^{\prime}) \mathand F(y^{\prime},y)\Big)\\
&&\rightarrow \Big(I(x^{\prime},y^{\prime}) \mathand R_{S_{k}}(x^{\prime},y) \mathand K_{q_{l}}(x^{\prime}) \mathand (z)\Big(F(y^{\prime},z) \mathbin{\mathrm{v}} \Big[\big(R_{S_{0}}(x,z) \rightarrow R_{S_{0}}(x^{\prime},z)\big)\\
&&\mathand \big(R_{S_{1}}(x,z) \rightarrow R_{S_{1}}(x^{\prime},z)\big) \mathand \ldots \mathand \big(R_{S_{M}}(x,z) \rightarrow R_{S_{M}}(x^{\prime},z)\big)\Big]\Big)\Big)\Big\}
\end{eqnarray*}
to include the universal quantifier (z).

page 260, line 15: As Turing wrote in the [correction paper 1938, p.545, l.14], “logical sum” should read “conjunction”.

page 260, lines 18-21: The definition of \mathrm{Un}(\mathscr{M}) given on lines 18-21 was withdrawn and replaced with the new definition (the [correction paper 1938, p.545]):

 (\exists{u})A(\mathscr{M}) \rightarrow (\exists{s})(\exists{t})R_{S_{1}}(s,t)
where A(\mathscr{M}) is abbreviation for

 Q \mathand (y)R_{S_{0}}(u,y) \mathand I(u,u) \mathand K_{q_{1}}(u) \mathand \mathrm{Des}(\mathscr{M})
and Q is abbreviation for

\begin{eqnarray*}
 &&(x)(\exists{w})(y,z)\Big\{F(x,w) \mathand \Big(F(x,y) \rightarrow G(x,y)\Big) \mathand \Big(F(x,z) \mathand G(z,y) \rightarrow G(x,y)\Big)\\
&&\hspace{5mm}\mathand \Big[G(z,x) \mathbin{\mathrm{v}} \Big(G(x,y) \mathand G(y,z)\Big) \mathbin{\mathrm{v}} \Big(F(x,y) \mathand F(z,y)\Big) \rightarrow \Big(-F(x,z)\Big)\Big]\Big\}
\end{eqnarray*}
.
Here, in my opinion, A(\mathscr{M}) might be changed to “A(\mathscr{M})(u)” because the formula comes along with the free variable u.

page 261, line 10:\mathand (y)F\Big((y,u^{\prime})\mathbin{\mathrm{v}}\ldots” should read “\mathand (y)\Big(F(y,u^{\prime})\mathbin{\mathrm{v}}\ldots”.

page 261, line 29: As Turing wrote in the [correction paper 1938, p.545, l.13], 
 r\big(n,i(n)\big) = a, r\big(n+1,i(n+1)\big) = c, k\big(i(n)\big)=b, \mbox{ and } k\big(i(n+1)\big)=d
should read

 r\big(n,i(n)\big) = b, r\big(n+1,i(n)\big) = d, k(n)=a, \mbox{ and } k(n+1)=c .

page 261, line 33: As Turing wrote in the [correction paper 1938, p.545, l.11],

 \mathrm{Inst}(q_{a}S_{b}S_{d}Lq_{c}) \mathand F^{(n+1)} \rightarrow (CC_{n} \rightarrow CC_{n+1})
should read

 \mathrm{Inst}(q_{a}S_{b}S_{d}Lq_{c}) \mathand Q \mathand F^{(n+1)} \rightarrow (CC_{n} \rightarrow CC_{n+1}).

page 262, line 9:A(\mathscr{M}) \mathand F^{(N)} \rightarrow CC^{N}” should read “A(\mathscr{M}) \mathand F^{(N)} \rightarrow CC_{N}”.

page 263, line 20:1+\phi_{\gamma}(u)” should read “1+\phi_{\gamma}(n)”.

page 264, the 4th line from the bottom:U” should read “U_{\gamma}”.

page 265, the 8th line from the bottom: The definition of “Q

\begin{tabular}{l}
  $\hspace{5mm} \Big\{\{Q\}{W_{\gamma}}\Big\}(N_{s}) \mathbin{\mathrm{conv}} N_{r(z)}$
\end{tabular}
should read

\begin{tabular}{l}
  $\hspace{5mm} \Big\{\{Q\}{W_{\gamma}}\Big\}(N_{s}) \mathbin{\mathrm{conv}} N_{r(s)}$.
\end{tabular}
Namely, the subscript “{}_{r(z)}” should be changed to “{}_{r(s)}”.

REFERENCE

  1. [main paper 1937]
    Turing, Alan.
    “On Computable Numbers, with an Application to the Entscheidungsproblem”
    Proceedings of the London Mathematical Society. ser.2 vol.42: pp.230-265. 1937.
  2. [correction paper 1938]
    Turing, Alan.
    “On Computable Numbers, with an Application to the Entscheidungsproblem. A Correction”
    Proceedings of the London Mathematical Society. ser.2 vol.43: pp.544-546. 1938.

NOTE

1. I have drafted the PDF version of this article.

2. You can read the papers in the book below:

| | コメント (0) | トラックバック (0)
|

ブログテンプレート変更

previous template of nousenew template of nouse本ブログ (nouse) のテンプレートを、2007年4月21日(土) から使用してきた [ノート(2列左)] から [シンプルコレクション2/ブラウン(2列左)] に変更した。記事本体部分の幅をひろくするのが目的である。

対比のため、以前のテンプレートによるイメージを左側に、新しいテンプレートによるイメージを右側に示しておく。

| | コメント (0) | トラックバック (0)
|

« 2011年8月 | トップページ | 2011年10月 »