Abstract
We present a Revised Projectivity Calculus (denoted RC) that extends the scope of inclusion and exclusion inferences derivable under the Projectivity Calculus (denoted C) developed by Icard (Stud Log 100(4):705–725, 2012). After pointing out the inadequacies of C, we introduce four opposition properties (OPs) which have been studied by Chow (in: Aloni et al (eds) Proceedings of the 18th Amsterdam Colloquium, Springer, Berlin, 2012; Beziau, Georgiorgakis (eds) New dimensions of the square of opposition, Philosophia Verlag GmbH, München, 2017) and are more appropriate for the study of exclusion reasoning. Together with the monotonicity properties (MPs), the OPs will form the basis of RC instead of the additive/multiplicative properties used in C. We also prove some important results of the OPs and their relation with the MPs. We then introduce a set of projectivity signatures together with the associated operations and conditions for valid inferences, and develop RC by inheriting the key features of C. We then show that under RC, we can derive some inferences that are not derivable under C. We finally discuss some properties of RC and point to possible directions of further studies.
Similar content being viewed by others
Notes
Icard (2014) provided an axiomatization for the inclusion and exclusion reasoning and conjectured that the resulting calculus (called C2) is complete (but without proving it). Since C2 does not contain as much detail as the calculus C developed in Icard (2012), the discussion of this paper is mainly based on Icard (2012).
The names and symbols of the relations are adapted from MacCartney (2009), MacCartney and Manning (2009) and Icard (2012). The definitions are taken from Icard (2012). The “exclusive” and “exhaustive” relations are the same as the “CC” (short for “contrary or contradictory”) and “SC” (short for “subcontrary or contradictory”) relations in Chow (2012, 2017). Moreover, in what follows, “iff” is short for “if and only if”. Note that we have chosen to use Δ and ∇ so that the shape of Δ reminds us of the “∧” in the definition of the exclusive property whereas the shape of ∇ reminds us of the “∨” in the definition of the exhaustive property.
The following proposition can be easily proved by invoking the fact that x ≤ y iff x ∧ ¬y = 0 and de Morgan’s Law. The above fact can be found in Keenan and Faltz (1985).
Note that the JOIN operation defined by Icard (2012) is different from the concept of “join” in a partial order. To highlight this difference, the symbol representing this special operation is capitalized in this paper.
Icard (2012) used the terms “monotone” and “antitone” instead of “increasing” and “decreasing”.
What Icard (2012) studied is the “complete” version of the properties.
In this paper, we treat proportional quantifiers more than/less than/at least/at most r of the as equivalent to more than/less than/at least/at most r of.
The truth conditions under GQT of these two exceptive quantifiers are (where s denotes Smith):
ǁall except… Smithǁ(A)(B) is true iff A – B = {s}
ǁno except… Smithǁ(A)(B) is true iff A ∩ B = {s}
The truth conditions under GQT of these two identity comparative quantifiers are:
ǁthe same… as….ǁ(A)(B1)(B2) is true iff A ∩ B1 = A ∩ B2
ǁdifferent… than….ǁ(A)(B1)(B2) is true iff A ∩ B1 ≠ A ∩ B2
For the purpose of discussion in this paper, we assume that any person is either asleep or awake but not both, and so treat asleep and awake as contradictory.
Although the results in Chow (2017) are about generalized quantifiers, these results (together with their proofs) can be readily extended to general functions on Boolean algebras.
The previous version of this paper did not discuss (sufficient) conditions associated with inclusion/exclusion reasoning. I am grateful to an anonymous reviewer for pointing out the need for clarifying these conditions. Without the reviewer’s advice, this paper would not have included this subsection.
These conditions are not parts of the projectivity signatures but are additional information specified in the lexicon (the Appendix to this paper can be seen as a part of this lexicon). Therefore, the discussion in this subsection should be seen as a justification of the conditions specified in the Appendix rather than part of the calculus to be introduced in the next section.
Here we have adapted definitions (86) and (88) in Keenan and Westerståhl (2011): let Q be an arity reducer and B an n-ary predicate, then Q(B) = {(x1,… xn−1): Q({xn: (x1,… xn) ∈ B})}.
In GQT, a proper name can be seen as a generalized quantifier. Such quantifiers are called “Montagovian individuals” in Peters and Westerståhl (2006).
To simplify matters, we treat “are noisy” as a unit.
Note that in Moss (2012), t2(x ← s) means the same thing as t2(s ← x) in this paper.
In GQT, “Not everybody loves or hates everybody” can be represented as not everybody(everybody)(loves ∨ hates).
To facilitate comparison, in what follows we modify the notations for the seven relations and projection used in Icard (2012).
References
Beghelli, F. (1994). Structured quantifiers. In M. Kanazawa & C. Piñón (Eds.), Dynamics, polarity and quantification (pp. 119–143). Stanford: CSLI.
Chow, K. F. (2012). Generalizing monotonicity inferences to opposition inferences. In M. Aloni, et al. (Eds.), Proceedings of the 18th Amsterdam Colloquium (pp. 281–290). Berlin: Springer.
Chow, K. F. (2017). Opposition inferences and generalized quantifiers. In J.-Y. Beziau & S. Georgiorgakis (Eds.), New dimensions of the square of opposition (pp. 155–199). München: Philosophia Verlag GmbH.
Icard, T. F. (2012). Inclusion and exclusion in natural language. Studia Logica,100(4), 705–725.
Icard, T. F. (2014). Higher-order syllogistics. In G. Morrill, et al. (Eds.), Formal grammar (pp. 1–14). Berlin: Springer.
Icard, T. F., & Moss, L. S. (2014). Recent progress on monotonicity. Linguistic Issues in Language Technology,9(7), 167–194.
Keenan, E. L. (2003). Excursions in natural logic. In C. Casadio, et al. (Eds.), Language and grammar: Studies in mathematical linguistics and natural language (pp. 31–52). Stanford: CSLI.
Keenan, E. L. (2008). Further excursions in natural logic: The mid-point theorems. In F. Hamm & S. Kepser (Eds.), Logics for linguistic structures (pp. 87–104). Berlin: Mouton de Gruyter.
Keenan, E. L., & Faltz, L. M. (1985). Boolean semantics for natural language. Dordrecht: Reidel.
Keenan, E. L., & Westerståhl, D. (2011). Generalized quantifiers in linguistics and logic. In J. van Benthem & A. ter Meulen (Eds.), Handbook of logic and language (2nd ed., pp. 859–910). Amsterdam: Elsevier.
MacCartney, B. (2009). Natural language inference. Ph.D. dissertation, Stanford University.
MacCartney, B., & Manning, C.D. (2009). An extended model of natural logic. In Proceedings of the eighth international conference on computational semantics (pp. 140–156).
Moss, L. S. (2012). The soundness of internalized polarity marking. Studia Logica,100(4), 683–704.
Peters, S., & Westerståhl, D. (2006). Quantifiers in language and logic. Oxford: Clarendon Press.
Sánchez Valencia, V. (1991). Studies on natural logic and categorial grammar. Ph.D. dissertation, Universiteit van Amsterdam.
van Benthem, J. (1986). Essays in logical semantics. Dordrecht: Reidel.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix: Signatures and Conditions of Some Functions
Appendix: Signatures and Conditions of Some Functions
This Appendix provides a list of some important functions (including logical operators/generalized quantifiers) and their projectivity signatures and sufficient conditions for valid inferences. For each function, the signature listed in the middle column below represents the strongest MP and/or OP it possesses. For quantifiers with more than one argument, the signature in respect of each argument is specified. Otherwise, the signature in respect of that argument is understood to be ●. For a function with complex signature, different components of the signature may be associated with different conditions. For clarity, the components of a complex signature are listed on separate rows with their corresponding conditions given in the right column (where “–” represents no condition). In what follows, A and B represent the 1st (unary) argument and the 2nd (n-ary) argument of a determiner, while A, B1 and B2 represent the 1st (unary) argument and the 2nd and 3rd (n-ary) arguments of an identity comparative quantifier. For example, the information given below shows that every/all has complex signature “(↓ ∧ ∇ → Δ)” in respect of the 1st argument. Moreover, every/all is decreasing (i.e. ↓) with no additional condition and anti-exhaustive (i.e. ∇ → Δ) given the additional condition that ∀x1…xn−1.{xn: (x1,… xn) ∈ B} ≠ Ư.
Function | Signature | Condition | |
---|---|---|---|
ID | ↑ | – | |
Δ → Δ | – | ||
∇ → ∇ | – | ||
not | ↓ | – | |
Δ → ∇ | – | ||
∇ → Δ | – | ||
Smith, i.e. singular proper name | ↑ | – | |
Δ → Δ | – | ||
∇ → ∇ | – | ||
every/all | 1st | ↓ | – |
∇ → Δ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B} ≠ Ư | ||
2nd | ↑ | – | |
Δ → Δ | A ≠ ∅ | ||
some | 1st | ↑ | – |
∇ → ∇ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B} ≠ ∅ | ||
2nd | ↑ | – | |
∇ → ∇ | A ≠ ∅ | ||
no | 1st | ↓ | – |
∇ → Δ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B} ≠ ∅ | ||
2nd | ↓ | – | |
∇ → Δ | A ≠ ∅ | ||
not every/not all | 1st | ↑ | – |
∇ → ∇ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B} ≠ Ư | ||
2nd | ↓ | – | |
Δ → ∇ | A ≠ ∅ | ||
more than n, at least n | both ↑ | – | |
fewer than n, at most n | both ↓ | – | |
most | 2nd | ↑ | – |
Δ → Δ | – | ||
more than r of (1/2 ≤ r < 1) | 2nd | ↑ | – |
Δ → Δ | – | ||
more than r of (0 < r < 1/2) | 2nd | ↑ | – |
∇ → ∇ | – | ||
at least r of (1/2 < r < 1) | 2nd | ↑ | – |
Δ → Δ | – | ||
at least r of (0 < r ≤ 1/2) | 2nd | ↑ | – |
∇ → ∇ | – | ||
less than r of (1/2 < r < 1) | 2nd | ↓ | – |
Δ → ∇ | – | ||
less than r of (0 < r ≤ 1/2) | 2nd | ↓ | – |
∇ → Δ | – | ||
at most r of (1/2 ≤ r < 1) | 2nd | ↓ | – |
Δ → ∇ | – | ||
at most r of (0 < r < 1/2) | 2nd | ↓ | – |
∇ → Δ | – | ||
exactly r of (1/2 < r < 1) | 2nd Δ → Δ | – | |
exactly r of (0 < r < 1/2) | 2nd ∇ → Δ | – | |
between q and r of (1/2 < q < r < 1) | 2nd Δ → Δ | – | |
between q and r of (0 < q < r < 1/2) | 2nd ∇ → Δ | – | |
more than r or less than q of (1/2 < q < r < 1) | 2nd Δ → ∇ | – | |
more than r or less than q of (0 < q < r < 1/2) | 2nd ∇ → ∇ | – | |
all… except Smith | 1st | Δ → Δ | – |
∇ → Δ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B} ∪ {s} ≠ Ư | ||
2nd | Δ → Δ | A − {s} ≠ ∅ | |
∇ → Δ | – | ||
no… except Smith | 1st | Δ → Δ | – |
∇ → Δ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B} − {s} ≠ ∅ | ||
2nd | Δ → Δ | – | |
∇ → Δ | A − {s} ≠ ∅ | ||
the same… as… | 1st | ↓ | – |
∇ → Δ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B1} ≠ {xn: (x1,… xn) ∈ B2} | ||
2nd | Δ → Δ | ∀x1…xn−1.A ∩ {xn: (x1,… xn) ∈ B2} ≠ ∅ | |
∇ → Δ | ∀x1…xn−1.A − {xn: (x1,… xn) ∈ B2} ≠ ∅ | ||
3rd | Δ → Δ | ∀x1…xn−1.A ∩ {xn: (x1,… xn) ∈ B1} ≠ ∅ | |
∇ → Δ | ∀x1…xn−1.A − {xn: (x1,… xn) ∈ B1} ≠ ∅ | ||
different… than… | 1st | ↑ | – |
∇ → ∇ | ∀x1…xn−1.{xn: (x1,… xn) ∈ B1} ≠ {xn: (x1,… xn) ∈ B2} | ||
2nd | ∇ → ∇ | ∀x1…xn−1.A − {xn: (x1,… xn) ∈ B2} ≠ ∅ | |
Δ → ∇ | ∀x1…xn−1.A ∩ {xn: (x1,… xn) ∈ B2} ≠ ∅ | ||
3rd | ∇ → ∇ | ∀x1…xn−1.A − {xn: (x1,… xn) ∈ B1} ≠ ∅ | |
Δ → ∇ | ∀x1…xn−1.A ∩ {xn: (x1,… xn) ∈ B1} ≠ ∅ |
Rights and permissions
About this article
Cite this article
Chow, Kf. A Revised Projectivity Calculus for Inclusion and Exclusion Reasoning. J of Log Lang and Inf 29, 163–195 (2020). https://doi.org/10.1007/s10849-019-09292-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10849-019-09292-5