Quickly add a free MyWikiBiz directory listing!

Functional Logic : Higher Order Propositions

Idea

If functions of type $X \to \mathbb{B}$ are propositions about things in X, then functions of type $(X \to \mathbb{B}) \to \mathbb{B}$ are propositions about propositions about things in X, the first of a series of higher order propositions based on X.

Higher Order Propositional Expressions

By way of equipping the discussion with a modicum of concrete material, let's begin with a consideration of higher order propositions and logical operators that stem from the ordinary propositions on 1 and 2 variables.

Note on notation. The discussion that follows uses minimal negation operations, expressed as parenthesized tuples of the form $\texttt{(} e_1 \texttt{,} \ldots \texttt{,} e_k \texttt{)},$ and logical conjunctions, expressed as concatenated tuples of the form $e_1 ~ \ldots ~ e_k,$ as the sole expression-forming operations of a calculus for boolean-valued functions or propositions. The expressions of this calculus parse into data structures whose underlying graphs are called cacti by graph theorists. Hence the name cactus language for this dialect of propositional calculus.

Higher Order Propositions and Logical Operators (n = 1)

A higher order proposition is a proposition about propositions. If the original order of propositions consists of maps of the form $f : X \to \mathbb{B},$ then the next higher order of propositions consists of maps of the form $m : (X \to \mathbb{B}) \to \mathbb{B}.$ It is often useful to think of a higher order proposition as a measure on propositions.

For example, consider the case where $X = \mathbb{B}.$ Then there are exactly four propositions $f : \mathbb{B} \to \mathbb{B},$ and exactly sixteen higher order propositions that are based on this set, all taking the form $m : (\mathbb{B} \to \mathbb{B}) \to \mathbb{B}.$

Table 1 lists the 16 measures of the form $m : (\mathbb{B} \to \mathbb{B}) \to \mathbb{B}.$

 x: $1 ~ 0$ f m0 m1 m2 m3 m4 m5 m6 m7 m8 m9 m10 m11 m12 m13 m14 m15 f0 $0 ~ 0$ $\texttt{(~)}$ 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 f1 $0 ~ 1$ $\texttt{(} x \texttt{)}$ 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 f2 $1 ~ 0$ x 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 f3 $1 ~ 1$ $\texttt{((~))}$ 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1

The contents of Table 1 are organized as follows. Columns 1 and 2 form a truth table for the four propositions of the form $f : \mathbb{B} \to \mathbb{B},$ with the row leaders in Column 1 displaying the names of the functions fi for i = 0 to 3, while the entries in Column 2 give the values of each function for the argument values that are listed in the corresponding column head. Column 3 displays a more or less canonical expression for the proposition in question. The last sixteen columns are topped by a collection of conventional names for the measures mj as j ranges from 0 to 15, where the entries in the body of the Table record the values assigned to each fi by each mj.

Table 2 presents a sample of interpretive categories for higher order propositions of type $(\mathbb{B}^1 \to \mathbb{B}) \to \mathbb{B},$ but it's best to put off discussing this Table further until we get beyond the 1-dimensional case. These lower dimensional cases tend to be extremely condensed or degenerate in their structures, and the pattern of what's going on here can be set in higher relief as soon as we get even two logical variables to play off each other.

 Measure Happening Exactness Existence Linearity Uniformity Information m0 Nothing happens m1 Just false Nothing exists m2 Just not x m3 Nothing is x m4 Just x m5 Everything is x f is linear m6 f is not uniform f is informed m7 Not just true m8 Just true m9 f is uniform f is not informed m10 Something is not x f is not linear m11 Not just x m12 Something is x m13 Not just not x m14 Not just false Something exists m15 Anything happens

Higher Order Propositions and Logical Operators (n = 2)

By way of reviewing notation and preparing to extend it to higher order universes of discourse, let's first consider the universe of discourse $X^\circ = [x_1, x_2] = [u, v]$ that is based on just two logical features or boolean variables u and v.

The universe of discourse $X^\circ$ consists of two parts, a set of points and a set of propositions.

The points of $X^\circ$ form a space notated as follows:

 $\begin{matrix} X & = & \lang u, v \rang & = & \{ (u, v) \} & \cong & \mathbb{B}^2 \end{matrix}$

Each point in X may be indicated by means of a singular proposition, that is, a proposition that describes it uniquely. This form of representation leads to the following enumeration of points:

 $\begin{array}{*{11}{c}} X & = & \{ \quad \texttt{(} u \texttt{)(} v \texttt{)} & , & \texttt{(} u \texttt{)} ~ v & , & u ~ \texttt{(} v \texttt{)} & , & u ~ v \quad \} & \cong & \mathbb{B}^2 \end{array}$

Each point in X may also be described by means of its coordinates, that is, by the ordered pair of values in $\mathbb{B}$ that the coordinate propositions u and v take on that point. This form of representation leads to the following enumeration of points:

 $\begin{array}{*{11}{c}} X & = & \{ \quad (0, 0) & , & (0, 1) & , & (1, 0) & , & (1, 1) \quad \} & \cong & \mathbb{B}^2 \end{array}$

The propositions of $X^\circ$ form a space notated as follows:

 $\begin{array}{*{7}{c}} X^\uparrow & = & (X \to \mathbb{B}) & = & \{ f : X \to \mathbb{B} \} & \cong & (\mathbb{B}^2 \to \mathbb{B}) \end{array}$

As always, it is convenient to overlook the finer marks of distinction between isomorphic structures, so long as one is aware of their presence and knows when it is critical to call on them again.

The next higher order universe of discourse that is built on $X^\circ$ is $X^{\circ 2} = [X^\circ] = [[u, v]],$ which may be developed in the following way. The propositions of $X^\circ$ become the points of $X^{\circ 2}$, and the mappings of the type $m : (X \to \mathbb{B}) \to \mathbb{B}$ become the propositions of $X^{\circ 2}.$ In addition, it is convenient to equip the discussion with a selected set of higher order operators on propositions, all of which have the form $w : (\mathbb{B}^2 \to \mathbb{B})^k \to \mathbb{B}.$

To save a few words in the remainder of this discussion, let us use the terms measure and qualifier to refer to all types of higher order propositions and operators. To describe the present setting in picturesque terms, the propositions of [u,v] may be regarded as a gallery of sixteen venn diagrams, while the measures $m : (X \to \mathbb{B}) \to \mathbb{B}$ are analogous to a body of judges or a panel of critical viewers, each of whom evaluates each of the pictures as a whole and reports the ones that find favor or not. In this way, each judge mj partitions the gallery of pictures into two aesthetic portions, the pictures $m_j^{-1}(0)$ that mj dislikes and the pictures $m_j^{-1}(1)$ that mj likes.

There are 216 = 65536 measures of the form $m : (\mathbb{B}^2 \to \mathbb{B}) \to \mathbb{B}.$ Table 3 shows the first 24 of these measures in the same style of higher order truth table used above. The column headed mj shows the values of the measure mj on each of the propositions $f_i : \mathbb{B}^2 \to \mathbb{B}$ for i = 0 to 15, with blank entries in the Table being optional for values of zero. Let us refer to the arrangement of measures that continues according to this plan as their standard ordering. In this scheme of things, the index j of the measure mj is the decimal equivalent of the bit string in the corresponding column of the Table, reading the binary digits in order from bottom to top.

 $\begin{matrix}u\!:\\v\!:\end{matrix}$ $\begin{matrix}1100\\1010\end{matrix}$ f $\underset{0}{m}$ $\underset{1}{m}$ $\underset{2}{m}$ $\underset{3}{m}$ $\underset{4}{m}$ $\underset{5}{m}$ $\underset{6}{m}$ $\underset{7}{m}$ $\underset{8}{m}$ $\underset{9}{m}$ $\underset{10}{m}$ $\underset{11}{m}$ $\underset{12}{m}$ $\underset{13}{m}$ $\underset{14}{m}$ $\underset{15}{m}$ $\underset{16}{m}$ $\underset{17}{m}$ $\underset{18}{m}$ $\underset{19}{m}$ $\underset{20}{m}$ $\underset{21}{m}$ $\underset{22}{m}$ $\underset{23}{m}$ f0 0000 $\texttt{(~)}$ 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 f1 0001 $\texttt{(} u \texttt{)(} v \texttt{)}$ 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 f2 0010 $\texttt{(} u\texttt{)} ~ v$ 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 f3 0011 $\texttt{(} u \texttt{)}$ 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 f4 0100 $u ~ \texttt{(} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 f5 0101 $\texttt{(} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f6 0110 $\texttt{(} u \texttt{,} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f7 0111 $\texttt{(} u ~ v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f8 1000 $u ~ v$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f9 1001 $\texttt{((} u \texttt{,} v \texttt{))}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f10 1010 v 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f11 1011 $\texttt{(} u ~ \texttt{(} v \texttt{))}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f12 1100 u 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f13 1101 $\texttt{((} u \texttt{)} ~ v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f14 1110 $\texttt{((} u \texttt{)(} v \texttt{))}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 f15 1111 $\texttt{((~))}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Umpire Operators

We now examine measures at the high end of the standard ordering. Instrumental to this purpose we define a couple of higher order operators, $\Upsilon : (\lang u, v \rang \to \mathbb{B})^2 \to \mathbb{B}$ and $\Upsilon_1 : (\lang u, v \rang \to \mathbb{B}) \to \mathbb{B},$ referred to as the relative and absolute umpire operators, respectively. If either of these operators is defined in terms of more primitive notions then the remaining operator can be defined in terms of the one first established.

Given an ordered pair of propositions $e, f : \lang u, v \rang \to \mathbb{B}$ as arguments, the relative umpire operator reports the value 1 if the first implies the second, otherwise it reports the value 0.

 $\Upsilon (e, f) = 1 \qquad \operatorname{if~and~only~if} \qquad e \Rightarrow f$

Expressing it another way:

 $\Upsilon (e, f) = 1 \qquad \iff \qquad \texttt{(} e \texttt{(} f \texttt{))} = 1$

In writing this, however, it is important to notice that the 1 appearing on the left side and the 1 appearing on the right side of the logical equivalence have different meanings. Filling in the details, we have:

 $\Upsilon (e, f) = 1 \in \mathbb{B} \qquad \iff \qquad \texttt{(} e \texttt{(} f \texttt{))} = 1 : \lang u, v \rang \to \mathbb{B}$

Writing types as subscripts and using the fact that $X = \lang u, v \rang,$ it is possible to express this a little more succinctly as follows:

 $\Upsilon (e, f) = 1_\mathbb{B} \qquad \iff \qquad \texttt{(} e \texttt{(} f \texttt{))} = 1_{X \to \mathbb{B}}$

Finally, it is often convenient to write the first argument as a subscript, hence Υe(f) = Υ(e,f).

As a special application of the relative umpire operator, we next define the absolute umpire operator, also known as the umpire measure. In the present setting this is a higher order proposition $\Upsilon_1 : (\mathbb{B}^2 \to \mathbb{B}) \to \mathbb{B}$ that is defined by the equation Υ1(f) = Υ(1,f). Here, the subscript 1 on the left and the argument 1 on the right both refer to the constant proposition $1 : \mathbb{B}^2 \to \mathbb{B}.$ In most contexts where Υ1 is actually applied the subscript 1 is safely omitted, since the number of arguments indicates which type of operator is intended. Thus, we have the following identities and equivalents:

 $\Upsilon f = \Upsilon_1 (f) = 1_\mathbb{B} \qquad \iff \qquad \texttt{(} 1 \texttt{(} f \texttt{))} = \mathbf{1} \qquad \iff \qquad f = 1_{\mathbb{B}^2 \to \mathbb{B}}$

The umpire measure is defined at the level of boolean functions, but can also be understood in terms of its implied judgments at the syntactic level. Interpreted this way, Υ1 recognizes theorems of the propositional calculus over [u,v], giving a score of 1 to tautologies and a score of 0 to everything else, regarding all contingent statements as no better than falsehoods.

One remark in passing for those who might prefer an alternative definition. If we had originally taken Υ to mean the absolute measure, then the relative version could have been defined as $\Upsilon_e f = \Upsilon \texttt{(} e \texttt{(} f \texttt{))}.$

Measure for Measure

Define two families of measures:

 $\alpha_i, \beta_i : (\mathbb{B}^2 \to \mathbb{B}) \to \mathbb{B}, ~ i = 0 ~ \text{to} ~ 15,$

by means of the following equations:

 $\begin{matrix} \alpha_i f & = & \Upsilon (f_i, f) & = & \Upsilon (f_i \Rightarrow f), \\[6pt] \beta_i f & = & \Upsilon (f, f_i) & = & \Upsilon (f \Rightarrow f_i). \end{matrix}$

The values of the sixteen αi on each of the sixteen boolean functions $f : \mathbb{B}^2 \to \mathbb{B}$ are shown in Table 4. Expressed in terms of the implication ordering on the sixteen functions, αif = 1 says that f is above or identical to fi in the implication lattice, that is, $\ge f_i$ in the implication ordering.

 $\begin{matrix}u\!:\\v\!:\end{matrix}$ $\begin{matrix}1100\\1010\end{matrix}$ f α15 α14 α13 α12 α11 α10 α9 α8 α7 α6 α5 α4 α3 α2 α1 α0 f0 0000 $\texttt{(~)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 f1 0001 $\texttt{(} u \texttt{)(} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 f2 0010 $\texttt{(} u\texttt{)} ~ v$ 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 f3 0011 $\texttt{(} u \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 f4 0100 $u ~ \texttt{(} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 f5 0101 $\texttt{(} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 f6 0110 $\texttt{(} u \texttt{,} v \texttt{)}$ 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 f7 0111 $\texttt{(} u ~ v \texttt{)}$ 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 f8 1000 $u ~ v$ 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 f9 1001 $\texttt{((} u \texttt{,} v \texttt{))}$ 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 f10 1010 v 0 0 0 0 0 1 0 1 0 0 0 0 0 1 0 1 f11 1011 $\texttt{(} u ~ \texttt{(} v \texttt{))}$ 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 f12 1100 u 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 1 f13 1101 $\texttt{((} u \texttt{)} ~ v \texttt{)}$ 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 f14 1110 $\texttt{((} u \texttt{)(} v \texttt{))}$ 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 f15 1111 $\texttt{((~))}$ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

The values of the sixteen βi on each of the sixteen boolean functions $f : \mathbb{B}^2 \to \mathbb{B}$ are shown in Table 5. Expressed in terms of the implication ordering on the sixteen functions, βif = 1 says that f is below or identical to fi in the implication lattice, that is, $\le f_i$ in the implication ordering.

 $\begin{matrix}u\!:\\v\!:\end{matrix}$ $\begin{matrix}1100\\1010\end{matrix}$ f β0 β1 β2 β3 β4 β5 β6 β7 β8 β9 β10 β11 β12 β13 β14 β15 f0 0000 $\texttt{(~)}$ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 f1 0001 $\texttt{(} u \texttt{)(} v \texttt{)}$ 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 f2 0010 $\texttt{(} u\texttt{)} ~ v$ 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 f3 0011 $\texttt{(} u \texttt{)}$ 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 1 f4 0100 $u ~ \texttt{(} v \texttt{)}$ 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 f5 0101 $\texttt{(} v \texttt{)}$ 0 0 0 0 0 1 0 1 0 0 0 0 0 1 0 1 f6 0110 $\texttt{(} u \texttt{,} v \texttt{)}$ 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 f7 0111 $\texttt{(} u ~ v \texttt{)}$ 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 f8 1000 $u ~ v$ 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 f9 1001 $\texttt{((} u \texttt{,} v \texttt{))}$ 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 f10 1010 v 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 f11 1011 $\texttt{(} u ~ \texttt{(} v \texttt{))}$ 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 f12 1100 u 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 f13 1101 $\texttt{((} u \texttt{)} ~ v \texttt{)}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 f14 1110 $\texttt{((} u \texttt{)(} v \texttt{))}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 f15 1111 $\texttt{((~))}$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

Applied to a given proposition f, the qualifiers αi and βi tell whether f is above fi or below fi, respectively, in the implication ordering. By way of example, let us trace the effects of several such measures, namely, those that occupy the limiting positions of the Tables.

 $\begin{array}{*{8}{r}} \alpha_{0} f = 1 & \operatorname{iff} & f_{0} \Rightarrow f & \operatorname{iff} & 0 \Rightarrow f, & \operatorname{hence} & \alpha_{0} f = 1 & \operatorname{for~all} ~ f. \\ \alpha_{15} f = 1 & \operatorname{iff} & f_{15} \Rightarrow f & \operatorname{iff} & 1 \Rightarrow f, & \operatorname{hence} & \alpha_{15} f = 1 & \operatorname{iff} ~ f = 1. \\ \beta_{0} f = 1 & \operatorname{iff} & f \Rightarrow f_{0} & \operatorname{iff} & f \Rightarrow 0, & \operatorname{hence} & \beta_{0} f = 1 & \operatorname{iff} ~ f = 0. \\ \beta_{15} f = 1 & \operatorname{iff} & f \Rightarrow f_{15} & \operatorname{iff} & f \Rightarrow 1, & \operatorname{hence} & \beta_{15} f = 1 & \operatorname{for~all} ~ f. \end{array}$

Thus, α0 = β15 is a totally indiscriminate measure, one that accepts all propositions $f : \mathbb{B}^2 \to \mathbb{B},$ whereas α15 and β0 are measures that value the constant propositions $1 : \mathbb{B}^2 \to \mathbb{B}$ and $0 : \mathbb{B}^2 \to \mathbb{B},$ respectively, above all others.

Finally, in conformity with the use of fiber notation to indicate sets of models, it is natural to use notations like the following to denote sets of propositions that satisfy the umpires in question.

 [ | αi | ] = $\alpha_{i}^{-1}(1),$ [ | βi | ] = $\beta_{i}^{-1}(1),$ [ | Υp | ] = $\Upsilon_{p}^{-1}(1).$

Functional Conception of Quantification Theory

Up till now quantification theory has been based on the assumption of individual variables ranging over universal collections of perfectly determinate elements. Merely to write down quantified formulas like $\forall_{x \in X} f(x)$ and $\exists_{x \in X} f(x)$ involves a subscription to such notions, as shown by the membership relations invoked in their indices. Reflected on pragmatic and constructive principles, these ideas begin to appear as problematic hypotheses whose warrants are not beyond question, projects of exhaustive determination that overreach the powers of finite information and control to manage. Therefore, it is worth considering how we might shift the medium of quantification theory closer to familiar ground, toward the predicates themselves that represent our continuing acquaintance with phenomena.

Extending the Existential Interpretation to Quantificational Logic

The forms commonly viewed as quantified propositions may be viewed again as propositions about propositions, indeed, there is every reason to regard higher order propositions as the genus of quantification under which the more familiar species appear.

Let us return to the 2-dimensional case $X^\circ = [u, v]$. In order to provide a bridge between propositions and quantifications it serves to define a set of qualifiers $\ell_{ij} : (\mathbb{B}^2 \to \mathbb{B}) \to \mathbb{B}$ that have the following characters:

 $\begin{array}{*{11}{l}} \ell_{00} f & = & \ell_{\texttt{(} u \texttt{)(} v \texttt{)}} f & = & \alpha_{1} f & = & \Upsilon_{\texttt{(} u \texttt{)(} v \texttt{)}} f & = & \Upsilon_{\texttt{(} u \texttt{)(} v \texttt{)} ~ \Rightarrow f} & = & f ~ \operatorname{likes} ~ \texttt{(} u \texttt{)(} v \texttt{)} \\ \ell_{01} f & = & \ell_{\texttt{(} u \texttt{)} ~ v} f & = & \alpha_{2} f & = & \Upsilon_{\texttt{(} u \texttt{)} ~ v} f & = & \Upsilon_{\texttt{(} u \texttt{)} ~ v ~ \Rightarrow f} & = & f ~ \operatorname{likes} ~ \texttt{(} u \texttt{)} ~ v \\ \ell_{10} f & = & \ell_{u ~ \texttt{(} v \texttt{)}} f & = & \alpha_{4} f & = & \Upsilon_{u ~ \texttt{(} v \texttt{)}} f & = & \Upsilon_{u ~ \texttt{(} v \texttt{)} ~ \Rightarrow f} & = & f ~ \operatorname{likes} ~ u ~ \texttt{(} v \texttt{)} \\ \ell_{11} f & = & \ell_{u ~ v} f & = & \alpha_{8} f & = & \Upsilon_{u ~ v} f & = & \Upsilon_{u ~ v ~ \Rightarrow f} & = & f ~ \operatorname{likes} ~ u ~ v \end{array}$

Intuitively, the $\ell_{ij}$ operators may be thought of as qualifying propositions according to the elements of the universe of discourse that each proposition positively values. Taken together, these measures provide us with the means to express many useful observations about the propositions in $X^\circ = [u, v],$ and so they mediate a subtext $[\ell_{00}, \ell_{01}, \ell_{10}, \ell_{11}]$ that takes place within the higher order universe of discourse $X^{\circ 2} = [X^\circ] = [[u, v]].$ Figure 6 summarizes the action of the $\ell_{ij}$ operators on the fi within $X^{\circ 2}.$

 $\text{Figure 6.} ~~ \text{Higher Order Universe of Discourse} ~ [\ell_{00}, \ell_{01}, \ell_{10}, \ell_{11}] \subseteq [[u, v]]$

Application of Higher Order Propositions to Quantification Theory

Our excursion into the vastening landscape of higher order propositions has finally come round to the stage where we can bring its returns to bear on opening up new perspectives for quantificational logic.

It's hard to tell if it makes any difference from a purely formal point of view, but it serves intuition to devise a slightly different interpretation for the two-valued space that we use as the target of our basic indicator functions. Therefore, let us declare the type of existential-valued functions $f : \mathbb{B}^k \to \mathbb{E},$ where $\mathbb{E} = \{ -e, +e \} = \{ \operatorname{empty}, \operatorname{exist} \}$ is a pair of values that indicate whether or not anything exists in the cells of the underlying universe of discourse. As usual, let's not be too fussy about the coding of these functions, reverting to binary codes whenever the intended interpretation is clear enough.

With this interpretation in mind we note the following correspondences between classical quantifications and higher order indicator functions:

 $\begin{array}{clcl} \mathrm{A} & \mathrm{Universal~Affirmative} & \mathrm{All} ~ u ~ \mathrm{is} ~ v & \mathrm{Indicator~of} ~ u \texttt{(} v \texttt{)} = 0 \\ \mathrm{E} & \mathrm{Universal~Negative} & \mathrm{All} ~ u ~ \mathrm{is} ~ \texttt{(} v \texttt{)} & \mathrm{Indicator~of} ~ u \cdot v = 0 \\ \mathrm{I} & \mathrm{Particular~Affirmative} & \mathrm{Some} ~ u ~ \mathrm{is} ~ v & \mathrm{Indicator~of} ~ u \cdot v = 1 \\ \mathrm{O} & \mathrm{Particular~Negative} & \mathrm{Some} ~ u ~ \mathrm{is} ~ \texttt{(} v \texttt{)} & \mathrm{Indicator~of} ~ u \texttt{(} v \texttt{)} = 1 \end{array}$

The following Tables develop these ideas in more detail.

 $\begin{matrix}u\!:\\v\!:\end{matrix}$ $\begin{matrix}1100\\1010\end{matrix}$ f $\begin{smallmatrix} \texttt{(} \ell_{11} \texttt{)} \\ \mathrm{No} ~ u \\ \mathrm{is} ~ v \end{smallmatrix}$ $\begin{smallmatrix} \texttt{(} \ell_{10} \texttt{)} \\ \mathrm{No} ~ u \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \texttt{(} \ell_{01} \texttt{)} \\ \mathrm{No} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ v \end{smallmatrix}$ $\begin{smallmatrix} \texttt{(} \ell_{00} \texttt{)} \\ \mathrm{No} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \ell_{00} \\ \mathrm{Some} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \ell_{01} \\ \mathrm{Some} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ v \end{smallmatrix}$ $\begin{smallmatrix} \ell_{10} \\ \mathrm{Some} ~ u \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \ell_{11} \\ \mathrm{Some} ~ u \\ \mathrm{is} ~ v \end{smallmatrix}$ f0 0000 $\texttt{(~)}$ 1 1 1 1 0 0 0 0 f1 0001 $\texttt{(} u \texttt{)(} v \texttt{)}$ 1 1 1 0 1 0 0 0 f2 0010 $\texttt{(} u\texttt{)} ~ v$ 1 1 0 1 0 1 0 0 f3 0011 $\texttt{(} u \texttt{)}$ 1 1 0 0 1 1 0 0 f4 0100 $u ~ \texttt{(} v \texttt{)}$ 1 0 1 1 0 0 1 0 f5 0101 $\texttt{(} v \texttt{)}$ 1 0 1 0 1 0 1 0 f6 0110 $\texttt{(} u \texttt{,} v \texttt{)}$ 1 0 0 1 0 1 1 0 f7 0111 $\texttt{(} u ~ v \texttt{)}$ 1 0 0 0 1 1 1 0 f8 1000 $u ~ v$ 0 1 1 1 0 0 0 1 f9 1001 $\texttt{((} u \texttt{,} v \texttt{))}$ 0 1 1 0 1 0 0 1 f10 1010 v 0 1 0 1 0 1 0 1 f11 1011 $\texttt{(} u ~ \texttt{(} v \texttt{))}$ 0 1 0 0 1 1 0 1 f12 1100 u 0 0 1 1 0 0 1 1 f13 1101 $\texttt{((} u \texttt{)} ~ v \texttt{)}$ 0 0 1 0 1 0 1 1 f14 1110 $\texttt{((} u \texttt{)(} v \texttt{))}$ 0 0 0 1 0 1 1 1 f15 1111 $\texttt{((~))}$ 0 0 0 0 1 1 1 1

 $\begin{matrix}u\!:\\v\!:\end{matrix}$ $\begin{matrix}1100\\1010\end{matrix}$ f $\begin{smallmatrix} \texttt{(} \ell_{11} \texttt{)} \\ \mathrm{No} ~ u \\ \mathrm{is} ~ v \end{smallmatrix}$ $\begin{smallmatrix} \texttt{(} \ell_{10} \texttt{)} \\ \mathrm{No} ~ u \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \texttt{(} \ell_{01} \texttt{)} \\ \mathrm{No} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ v \end{smallmatrix}$ $\begin{smallmatrix} \texttt{(} \ell_{00} \texttt{)} \\ \mathrm{No} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \ell_{00} \\ \mathrm{Some} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \ell_{01} \\ \mathrm{Some} ~ \texttt{(} u \texttt{)} \\ \mathrm{is} ~ v \end{smallmatrix}$ $\begin{smallmatrix} \ell_{10} \\ \mathrm{Some} ~ u \\ \mathrm{is} ~ \texttt{(} v \texttt{)} \end{smallmatrix}$ $\begin{smallmatrix} \ell_{11} \\ \mathrm{Some} ~ u \\ \mathrm{is} ~ v \end{smallmatrix}$ f0 0000 $\texttt{(~)}$ 1 1 1 1 0 0 0 0 f1 0001 $\texttt{(} u \texttt{)(} v \texttt{)}$ 1 1 1 0 1 0 0 0 f2 0010 $\texttt{(} u\texttt{)} ~ v$ 1 1 0 1 0 1 0 0 f4 0100 $u ~ \texttt{(} v \texttt{)}$ 1 0 1 1 0 0 1 0 f8 1000 $u ~ v$ 0 1 1 1 0 0 0 1 f3 0011 $\texttt{(} u \texttt{)}$ 1 1 0 0 1 1 0 0 f12 1100 u 0 0 1 1 0 0 1 1 f6 0110 $\texttt{(} u \texttt{,} v \texttt{)}$ 1 0 0 1 0 1 1 0 f9 1001 $\texttt{((} u \texttt{,} v \texttt{))}$ 0 1 1 0 1 0 0 1 f5 0101 $\texttt{(} v \texttt{)}$ 1 0 1 0 1 0 1 0 f10 1010 v 0 1 0 1 0 1 0 1 f7 0111 $\texttt{(} u ~ v \texttt{)}$ 1 0 0 0 1 1 1 0 f11 1011 $\texttt{(} u ~ \texttt{(} v \texttt{))}$ 0 1 0 0 1 1 0 1 f13 1101 $\texttt{((} u \texttt{)} ~ v \texttt{)}$ 0 0 1 0 1 0 1 1 f14 1110 $\texttt{((} u \texttt{)(} v \texttt{))}$ 0 0 0 1 0 1 1 1 f15 1111 $\texttt{((~))}$ 0 0 0 0 1 1 1 1

 Mnemonic Category $\mathrm{Classical~Form}$ $\mathrm{Alternate~Form}$ $\mathrm{Symmetric~Form}$ Operator $\begin{matrix} \mathrm{E} \\ \mathrm{Exclusive} \end{matrix}$ $\begin{matrix} \mathrm{Universal} \\ \mathrm{Negative} \end{matrix}$ $\mathrm{All} ~ u ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\mathrm{No} ~ u ~ \mathrm{is} ~ v$ $\texttt{(} \ell_{11} \texttt{)}$ $\begin{matrix} \mathrm{A} \\ \mathrm{Absolute} \end{matrix}$ $\begin{matrix} \mathrm{Universal} \\ \mathrm{Affirmative} \end{matrix}$ $\mathrm{All} ~ u ~ \mathrm{is} ~ v$ $\mathrm{No} ~ u ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\texttt{(} \ell_{10} \texttt{)}$ $\mathrm{All} ~ v ~ \mathrm{is} ~ u$ $\mathrm{No} ~ v ~ \mathrm{is} ~ \texttt{(} u \texttt{)}$ $\mathrm{No} ~ \texttt{(} u \texttt{)} ~ \mathrm{is} ~ v$ $\texttt{(} \ell_{01} \texttt{)}$ $\mathrm{All} ~ \texttt{(} v \texttt{)} ~ \mathrm{is} ~ u$ $\mathrm{No} ~ \texttt{(} v \texttt{)} ~ \mathrm{is} ~ \texttt{(} u \texttt{)}$ $\mathrm{No} ~ \texttt{(} u \texttt{)} ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\texttt{(} \ell_{00} \texttt{)}$ $\mathrm{Some} ~ \texttt{(} u \texttt{)} ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\mathrm{Some} ~ \texttt{(} u \texttt{)} ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\ell_{00}$ $\mathrm{Some} ~ \texttt{(} u \texttt{)} ~ \mathrm{is} ~ v$ $\mathrm{Some} ~ \texttt{(} u \texttt{)} ~ \mathrm{is} ~ v$ $\ell_{01}$ $\begin{matrix} \mathrm{O} \\ \mathrm{Obtrusive} \end{matrix}$ $\begin{matrix} \mathrm{Particular} \\ \mathrm{Negative} \end{matrix}$ $\mathrm{Some} ~ u ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\mathrm{Some} ~ u ~ \mathrm{is} ~ \texttt{(} v \texttt{)}$ $\ell_{10}$ $\begin{matrix} \mathrm{I} \\ \mathrm{Indefinite} \end{matrix}$ $\begin{matrix} \mathrm{Particular} \\ \mathrm{Affirmative} \end{matrix}$ $\mathrm{Some} ~ u ~ \mathrm{is} ~ v$ $\mathrm{Some} ~ u ~ \mathrm{is} ~ v$ $\ell_{11}$

References

• Quine, W.V. (1969/1981), "On the Limits of Decision", Akten des XIV. Internationalen Kongresses für Philosophie, vol. 3 (1969). Reprinted, pp. 156–163 in Quine (ed., 1981), Theories and Things, Harvard University Press, Cambridge, MA.

Appendix : Generalized Umpire Operators

In order to get a handle on the space of higher order propositions and eventually to carry out a functional approach to quantification theory, it serves to construct some specialized tools. Specifically, I define a higher order operator Υ, called the umpire operator, which takes up to three propositions as arguments and returns a single truth value as the result. Formally, this so-called multigrade property of Υ can be expressed as a union of function types, in the following manner:

 $\Upsilon : \bigcup_{\ell = 1, 2, 3} ((\mathbb{B}^k \to \mathbb{B})^\ell \to \mathbb{B}).$

In contexts of application the intended sense can be discerned by the number of arguments that actually appear in the argument list. Often, the first and last arguments appear as indices, the one in the middle being treated as the main argument while the other two arguments serve to modify the sense of the operation in question. Thus, we have the following forms:

 $\Upsilon_p^r q = \Upsilon (p, q, r)$ $\Upsilon_p^r : (\mathbb{B}^k \to \mathbb{B}) \to \mathbb{B}$

The intention of this operator is that we evaluate the proposition q on each model of the proposition p and combine the results according to the method indicated by the connective parameter r. In principle, the index r might specify any connective on as many as 2k arguments, but usually we have in mind a much simpler form of combination, most often either collective products or collective sums. By convention, each of the accessory indices p,r is assigned a default value that is understood to be in force when the corresponding argument place is left blank, specifically, the constant proposition $1 : \mathbb{B}^k \to \mathbb{B}$ for the lower index p, and the continued conjunction or continued product operation $\textstyle\prod$ for the upper index r. Taking the upper default value gives license to the following readings:

 $\Upsilon_p (q) = \Upsilon (p, q) = \Upsilon (p, q, \textstyle\prod).$ $\Upsilon_p = \Upsilon (p, \underline{~~}, \textstyle\prod) : (\mathbb{B}^k \to \mathbb{B}) \to \mathbb{B}.$

This means that Υp(q) = 1 if and only if q holds for all models of p. In propositional terms, this is tantamount to the assertion that $p \Rightarrow q,$ or that $\texttt{(} p \texttt{(} q \texttt{))} = 1.$

Throwing in the lower default value permits the following abbreviations:

 $\Upsilon q = \Upsilon (q) = \Upsilon_1 (q) = \Upsilon (1, q, \textstyle\prod).$ $\Upsilon = \Upsilon (1, \underline{~~}, \textstyle\prod)) : (\mathbb{B}^k\ \to \mathbb{B}) \to \mathbb{B}.$

This means that Υq = 1 if and only if q holds for the whole universe of discourse in question, that is, if and only q is the constantly true proposition $1 : \mathbb{B}^k \to \mathbb{B}.$ The ambiguities of this usage are not a problem so long as we distinguish the context of definition from the context of application and restrict all shorthand notations to the latter.

Document History

Note. The above material is excerpted from a project report on Charles Sanders Peirce's conceptions of inquiry and analogy. Online formatting of the original document and continuation of the initial project are currently in progress under the title Functional Logic : Inquiry and Analogy.

 Author: Jon Awbrey November 1, 1995 Course: Engineering 690, Graduate Project Cont'd from Winter 1995 Supervisors: F. Mili & M.A. Zohdy Oakland University
| Version:  Draft 3.25
| Created:  01 Jan 1995
| Relayed:  01 Nov 1995
| Revised:  24 Dec 2001
| Revised:  12 Mar 2004