Existential Graphs

Origin. Charles Sanders Peirce invented the existential graphs in 1896 — Murphey's dating, cited by Zeman and repeated by Ma and Pietarinen — as the successor to the entitative graphs, published in "The Logic of Relatives" (The Monist, 1897). Peirce called the graphs his chef d'oeuvre and worked on them until his death, but printed almost nothing: the principal lifetime publication is "Prolegomena to an Apology for Pragmaticism" (The Monist 16(4), 1906, 492–546), and the main corpus reached print posthumously in the Collected Papers, volume IV, paragraphs 4.372–584 (1933). The modern metatheory begins with J. Jay Zeman's dissertation "The Graphical Logic of C. S. Peirce" (University of Chicago, 1964); Don D. Roberts, The Existential Graphs of Charles S. Peirce (Mouton, The Hague, 1973), is the standard study of the system and its manuscript development; Sun-Joo Shin, The Iconic Logic of Peirce's Graphs (MIT Press, 2002), re-reads the rules.

Models. A proof system in which formulas are drawings and inference is licensed redrawing. The sheet of assertion represents the universe of discourse; a blank sheet asserts nothing, scribing two graphs side by side asserts both, a cut — a closed curve — denies what it encloses, and in Beta a heavily drawn line of identity asserts the existence of an individual and, by its continuity, the identity of whatever its ends attach to. The permissions depend on the geometry of enclosure alone, above all its parity: areas inside an even number of cuts behave positively, areas inside an odd number negatively. The rules therefore apply at any depth of nesting, with no privileged outermost position — the format that deep inference made systematic a century later. Three systems share the apparatus: Alpha for propositional logic, Beta for first-order logic with identity, Gamma — unfinished — for modality and for reasoning across a book of sheets rather than a single one.

Formalism.

Alpha syntax: the sheet of assertion SA; atomic graphs scribed on it; juxtaposition in one area = conjunction; the cut = negation of its contents. A cut within a cut (the scroll) reads as implication: outer area antecedent, inner area consequent. Blank SA = truth; an empty cut = absurdity.

Alpha transformation rules (Peirce, as reconstructed by Zeman and Roberts): Insertion: in any oddly enclosed area, any graph may be scribed. Erasure: from any evenly enclosed area, SA included, any graph may be erased. Iteration: a graph may be copied into its own area or into any area enclosed by all the cuts enclosing the original. Deiteration: the exact converse — any graph that iteration could have produced may be erased. Double cut: two cuts with nothing between them may be drawn around any graph, or removed, in any area. A proof transforms the premise graph into the conclusion graph by these permissions alone; a theorem is derivable from the blank sheet.

Beta: spots for predicates; the line of identity fills their blanks, branches to assert identity, and quantifies. Reading (Peirce, CP 4.458): a line whose outermost part is evenly enclosed asserts something; one whose outermost part is oddly enclosed refers to anything there may be — existential and universal force set by parity, with scope set by depth. The rules extend: lines of identity may be joined in odd areas and severed in even ones; iteration and deiteration carry line attachments; a double cut may be drawn across lines of identity, which pass freely through the space between the cuts.

Gamma: the broken cut, read "possibly not" of its contents (CP 4.516), with Peirce's three rules: in a broken cut already on SA any graph may be inserted; an oddly enclosed broken cut may be filled in to an ordinary cut; an evenly enclosed ordinary cut may be broken. Beyond the broken cut: tinctured regions (1906) and the book of sheets, announced as the extension to possible universes and never pulled together into a settled system.

Symbols.

Symbol	Unicode	Name	Meaning
∧	U+2227	Conjunction	Reading of juxtaposition in one area
¬	U+00AC	Negation	Reading of the cut
→	U+2192	Implication	Reading of the scroll, a cut within a cut
∃	U+2203	Existential quantifier	Reading of a line of identity evenly enclosed at its outermost part
∀	U+2200	Universal quantifier	Reading of a line oddly enclosed at its outermost part
SA	—	Sheet of assertion	The surface; blank asserts nothing
cut	—	Cut	Closed curve denying its contents
broken cut	—	Broken cut	Gamma's modal cut: possibly-not
line of identity	—	Line of identity	Existence plus identity, drawn as one heavy line
scroll	—	Scroll	The two-cut figure read as implication
Alpha, Beta, Gamma	—	The three systems	Propositional; first-order with identity; modal, unfinished
CP	—	Collected Papers	Citation convention: CP 4.372–584 is the graphs' main corpus

Metatheory. Zeman (1964) proved the equivalences: the theorems of Alpha map one-one onto the theorems of the classical propositional calculus — "alpha itself may be considered a complete classical CPC" — and Beta matches the classical first-order calculus with identity; his Gamma chapter shows that broken-cut systems differing only in the restrictions on the broken-cut rules are equivalent to the modal logics T, S4, S4.2 and S5, so the fragment Peirce specified underdetermines its modal logic. On Gamma as a whole Zeman reports that Peirce "was not able to pull the system together as he wanted to". Roberts (1973) reconstructed the system's development from the manuscripts, including the tinctured graphs of 1906. Shin (2002) restated the transformation rules symmetrically by area parity and argued that the traditional reading obscures the graphs' iconicity, licensing multiple readings of one graph; Ma and Pietarinen (Review of Symbolic Logic 9(2), 2016, 209–237) answer that her argument for multiple readings is circular. Brady and Trimble gave Alpha a categorical interpretation (Journal of Pure and Applied Algebra 149, 2000, 213–239); Dau gave the graphs and their concept-graph descendants a mathematical treatment (The Logic System of Concept Graphs with Negation, Springer 2003; "Some Notes on Proofs with Alpha Graphs", ICCS 2006). The deep-inference reading is now explicit: Ma and Pietarinen analysed Alpha's proof theory ("Proof Analysis of Peirce's Alpha System of Graphs", Studia Logica 105(3), 2017, 625–647) and built GrIn, a deep inference system formulated in terms of existential graphs, sound and complete for Heyting algebras and equivalent to the sequent calculus G3ip (Logique et Analyse 245, 2019, 73–114).

Applies to. The history of logic: a complete graphical formulation of first-order logic with identity from the 1890s. Diagrammatic reasoning, where Shin's work made the graphs a central test case for logic without a linear syntax. Conceptual graphs (Sowa 1984), the knowledge-representation formalism descended from them. Deep inference, whose parity-conditioned rules at arbitrary depth the graphs anticipate. Recent propositional formula simplification built on EG transformations (Theory and Practice of Logic Programming, 2025). Peirce scholarship and the philosophy of notation.

Limitations. The notation scales badly: cuts nest, lines of identity branch and cross, there is no abbreviation or definition mechanism, and the parity and nesting the rules consult must be recounted as cuts accumulate. Insertion admits an arbitrary graph into any odd area, so derivations have no subformula property and proof search is unbounded. Beta is purely relational: no function symbols and no individual constants, so terms must be paraphrased through predicates and lines. Gamma is unsettled: rules exist only for the broken-cut fragment, and those rules fix no single modal logic. And the system did not transmit: invented in 1896 and almost entirely unpublished in Peirce's lifetime, it reached print in 1933 and comprehension only with Zeman and Roberts, a half-century in which the notation of modern logic was fixed without it.