Plausibility, measured

An interactive tool built from the English edition of Meaning and Distance by Michel Magnen — translated by Fay Guerry — and applied first to Baudelaire's Correspondences.

About this tool. The texts of this page draw directly on the English translation of Michel Magnen's book by Fay Guerry. The vocabulary (plausibility, collision, channel, shouldering, oddity, buffer, framework…) is the translator's own. The poem below is shown in its English translation; you may switch to the French original at any time using the button. The full work — in 9 parts — is available in the English edition of the book.
Correspondences
Nature is a temple where living pillars
Let forth at times confused words;
There man passes through forests of symbols
Which observe him with familiar eyes.
Like long echoes which mingle in the distance,
In a dark and profound unity,
Vast as the night and as the light,
Perfumes, colours and sounds answer each other.
There are perfumes as cool as the flesh of children,
Sweet as oboes, green as meadows,
— And others, corrupt, rich and triumphant,
Having the expansion of infinite things,
Like amber, musk, benzoin and incense,
Which sing of the transports of the mind and the senses.
La Nature est un temple où de vivants piliers
Laissent parfois sortir de confuses paroles ;
L'homme y passe à travers des forêts de symboles
Qui l'observent avec des regards familiers.
Comme de longs échos qui de loin se confondent,
Dans une ténébreuse et profonde unité,
Vaste comme la nuit et comme la clarté,
Les parfums, les couleurs et les sons se répondent.
Il est des parfums frais comme des chairs d'enfants,
Doux comme les hautbois, verts comme les prairies,
— Et d'autres, corrompus, riches et triomphants,
Ayant l'expansion des choses infinies,
Comme l'ambre, le musc, le benjoin et l'encens,
Qui chantent les transports de l'esprit et des sens.
Baudelaire, The Flowers of Evil, 1857 — translated by Fay Guerry

§1From word to paradox

Michel Magnen's book opens on two intuitions. First: a substantial distance between the words of a single discourse, in the absence of any formal link from one to the other, erases the relations of meaning between them. Second: if two clashes of meaning, fundamentally close, are put forward in the same discourse — one presented forcefully while the other is just outlined — the more vigorously conveyed one lends the second one increased force.

From these two ideas the author develops a rigorous measurement system — borrowed in spirit from the calculus of probabilities, though distinct from it — which makes it possible to evaluate, for each juxtaposition of meaning in a text, its strength and its plausibility. This application unfolds these notions step by step and lets you compute them yourself.

To analyse a text, Magnen first establishes elementary objects of extreme precision. Each receives a strict definition that isolates exactly what is being spoken of — and what must be avoided to prevent confusion.

Foundational notion
The case

A case is an isolated word given at a precise location in the text. Whenever the same word recurs elsewhere, that recurrence forms a new case.

Example: “perfumes” occurs at line 8 and again at line 9 — two distinct cases, indexed perfumes¹ and perfumes², even though their meanings are nearly identical.
Reference appearance
The buffer

A buffer consists of a tenacious appearance which can be used in order to understand a text. There are three kinds: ideas (a certain definition), rules (an indisputable rule of grammar), and objects (a real thing described by the text, the observation of which helps to understand it).

Example: In “Nature is a temple where living pillars”, the grammar requires that “where” refers to “temple”. This is a rule-buffer.
The strange
The oddity

An oddity is a significance which provokes an astonishment impossible to overcome by any reference to a buffer.

Example: “Nature-temple”, “living pillars”, “forests of symbols that observe”, “perfumes which answer each other”, “corrupt perfumes”… so many oddities.
The elementary decision
The arbitrationb(A–E) or d(A–E)

An arbitration is the elementary decision: is it preferable to bring together (written b) or to separate (written d) two cases of the text? The order of the terms is unimportant: b(Nature–temple) and b(temple–Nature) state strictly the same thing — which is what distinguishes an arbitration from an ordinary attribution.

Reading: b(Nature–temple) = “to understand Baudelaire, it is better to bring together ‘Nature’ and ‘temple’ than to oppose them”.
Central object
The tension

A tension consists of a pointer made to describe an oddity — an irreducible arbitration. It is written with an initial r to single it out.

Example: rb(corrupt–incense) — incense is traditionally pure (the priests who use it adopt that view), so calling it “corrupt” is an oddity. Asserting the association in spite of the strangeness constitutes a tension.
A front = a word indispensable to carry the meaning, countable in telegraphic style. Articles, prepositions, and conjunctions are not. Every term of a broad tension counts as a front.

§2Measuring the plausibility of a tension

Here lies Magnen's own contribution. Once a tension has been identified, how can one estimate numerically the extent to which the author actually intended this association of meaning — as opposed to an illusion projected by the reader?

The author defines four factors of vacillation (the floating of the tension). Each is the numerical counterpart of a cause of weakness of the significance. The total vacillation is their product; its own plausibility — the channel — is the inverse of the vacillation.

# Vacillation of a tension vacillation = t × s × m × w # Channel (strength of significance) = inverse of vacillation channel = 1 / ( t × s × m × w )

The four factors

t — Rank

The rank measures whether the arbitration position (to bring together or to separate) is unambiguous or reversible.

  • t = 1: only one position yields a broad arbitration. The text excludes the other. Example: rb(Nature–temple) — the text affirms this union; separating would be absurd.
  • t = 2: both positions (b and d) yield broad arbitrations. The tension is rivalled by its opposite.
s — Internal remoteness

The internal remoteness measures the distance between the two terms in the text, counted in fronts.

If the tension possesses an obvious framework (grammar or syntax renders the relation undeniable), then s = 1 automatically.

Otherwise: s = 2 + (n / 10), where n is the number of fronts between the two terms (telegraphic style).

Example: For rb(corrupt–infinite), one counts 5 fronts between the two terms (rich, triumphant, expansion, things, infinite) — so s = 2 + 0.5 = 2.5.
m — Left oscillation

The oscillation measures whether the term possesses a plateau — that is, a rival, more commonplace significance allowing the clash of meaning to be side-stepped.

  • m = 1: the left-hand term has no plateau through which the tension could be avoided.
  • m = 2: the term has a plateau (figurative sense, secondary meaning, etc.) which can act as a diversion.
Example: In rb(answer–perfumes), “answer” has a figurative sense (to correspond, to echo) which softens the strangeness of attributing it to perfumes. Therefore m=2.
w — Right oscillation

Symmetrical with m, but for the right-hand term.

  • w = 1: no escape plateau for this term.
  • w = 2: a plateau lets the tension be side-stepped.
Example: For rb(corrupt–incense), “incense” can also mean “praise” (the figurative sense, as in “to incense someone”), which softens the paradox — a praise can indeed be corrupt. Therefore w=2.

§3Worked example: rb(corrupt–incense)

Let us apply all this to the sonnet's most emblematic tension. Baudelaire affirms that certain perfumes are “corrupt” and lists incense among them — which, for the religious tradition, is precisely the embodiment of purity. A paradox to be measured.

FactorValueJustification
t (rank)1The text affirms without any hesitation that the incense is corrupt. Separation is not an option.
s (internal remoteness)1Obvious framework — the passage clearly links the terms: “and others, corrupt, rich and triumphant… amber, musk, benzoin and incense”.
m (left oscillation)1“rich and triumphant” rules out reading “corrupt” in a purely physical sense (organic decay). No plateau.
w (right oscillation)1The figurative sense of “incense” (praise) is far away in this passage dominated by concrete perfumes.
# Channel of rb(corrupt–incense) vacillation = t·s·m·w = 1·1·1·1 = 1 channel = 1 / vacillation = 1

A channel of 1 is the maximum possible value: rb(corrupt–incense) is a perfectly plausible tension, solidly anchored in the poet's thought.

§4When one tension shoulders another

Magnen's major theoretical contribution lies further still. He observes that tensions are not isolated: certain neighbouring ones reinforce one another. A vigorous tension can sustain a fragile one through the effect of shouldering.

The tandem
Shoulderingh′/c

When two collisions (broad tensions) share a common term, they may form a tandem. The shouldering received by one is given by:

shouldering = channel_of_neighbour / c where c is the exterior remoteness between the two tensions
Example: rb(corrupt–temple) receives from rb(corrupt–incense) (channel = 1) a shouldering of 1/8.4 ≈ 0.119, which makes it interpretable even though it is, in itself, very fragile.
Total plausibility
The network

The network of a collision is its total plausibility: the sum of its own channel and of all shoulderings (and enhancements) it receives from other tensions in the text.

network(collision) = channel + Σ shoulderings + Σ enhancements

It is the network — not the channel alone — that truly measures the strength of an interpretation. A weak, isolated tension may be taken for illusion; well reinforced by others, it becomes unavoidable.

§5From tensions to glosses

From the second part of the book onwards, Magnen extends the method to all texts — not only to those in which the author deliberately seeks to create paradoxes. The mechanism operates through the glosses.

Extended notion
The gloss

A gloss is a commentary linking two traces of the text (where a trace is either a term, or a pivot — a meaning-bearing carrier: punctuation, position, structure). Its weight is measured with the same tools as a tension. Magnen replaces the notion of tension by the broader one of gloss, and the channel by the gradient.

  • Problem gloss (written r): describes a difficulty, minor or grave, in the text.
  • Attenuation gloss (written v): comments on a problem gloss in order to soften it. The dams of joints (clamps, shelters, canvases…) belong here.
  • Neutral gloss (written o): plainly notes, without problem or attenuation. Comparable to a footnote.
Note: In the later parts (8 and 9, on pairs of images and pairs of interpretations), Magnen develops still more elaborate structures (erasers, wheels, flows, pavilions, modules) which exceed the scope of this interactive tool. The present application implements the notions of parts I–IV, where the essential of the method resides.

Note: This tool presents the method in condensed form, through its five pivotal concepts. The complete work of Michel Magnen, Meaning and Distance, deploys the theory in 9 parts (Paradoxes, Generalization, Influences, Analogies, Brief figures, Deliberate intrusions, Coherence, Pairs of images, Pairs of interpretations), each exploring a particular aspect of the plausibility of an interpretation.

Plausibility, measured

You enter a tension identified in a text. The tool computes for you the vacillation and the channel — and, if you supply further tensions, the shoulderings and the network.

Mode of use: The literary judgement remains yours. The tool faithfully computes Magnen's formulas from the assessments you supply (rank, obvious framework or not, presence of plateaus). If your assessments are sound, the result is sound.

The analysed text

You may edit this text to analyse any passage.

Compute the plausibility of a tension

The first word or group of words
The second word or group of words
Are we asserting the association (b) or the dissociation (d) of the two terms?

Recorded tensions

Tensiont·s·m·wVacillationChannelAction
No tension recorded yet.

Compute a shouldering between two collisions

A shouldering requires two collisions which share a common term and the same arbitration position. The formula: shouldering_received_by_h = channel(neighbour) / c, where c is the exterior remoteness.

If both collisions each possess an obvious framework and the grammar enforces their link, c = 1. Otherwise, c = 2 + (n/10) where n is the number of intervening fronts.

Compute the network of a collision

Enter the shouldering / enhancement values already computed, for example: 0.119, 0.008, 0.26