Shannon's circuit mapping

A relay contact can be open or closed.

That small fact is where I want to start with Shannon. His 1938 paper came from his MIT thesis. It puts those states beside the true and false values of logic.

I can follow that much without knowing electrical engineering. The part that catches me is what happens when the contacts connect. Their connections follow the same rules as combinations of true and false statements.

Put two contacts one after the other. Current gets through only if both are closed. That is AND: both must be true.

Put them side by side and either one will do. That is OR. These are the series and parallel connections in the table below.

That gives the comparison a job. A rule established in logic can help someone design a circuit on paper.

I want to understand how that happened, and why Shannon’s account travelled when others had reached the same structure.

I start Maethel here because this story lets me follow a match across fields and test its limits.

The usual story

It is easy to tell this as the start of digital computing, by one man. That assumes the useful idea began with him. The earlier accounts change what there is to explain.

A secondary history dates the thesis to 1937. MIT’s record of it says 1940. I am working from the 1938 paper.

The thesis did not invent digital computing.

Even the word bit is John Tukey’s, from a memo reportedly dated 9 January 1947. Shannon credits him in 1948.

What interests me is the method: express the desired circuit behaviour, then use algebra to design it.

How the relations line up

I read a closed contact as TRUE, because current gets through. This is the transmission reading: series means AND, and parallel means OR.

Here is the transmission reading throughout.

Relation	Logic	Circuit	Match and source
A value	TRUE	Contact closed	Match; Shannon reprint p. 2.
The other value	FALSE	Contact open	Match; Shannon reprint p. 2.
AND	Both must be true	Contacts connected in series	Match; Shannon reprint p. 2.
OR	Either can be true	Contacts connected in parallel	Match; Shannon reprint p. 2.
NOT	TRUE becomes FALSE, and FALSE becomes TRUE	The same relay’s break contacts, in place of its make contacts. Shannon pairs them so one set is open whenever the other is closed.	Match; Shannon reprint pp. 2-3.
Circuits outside series and parallel, including bridges	A general simplification claim needs its own proof	Beyond series and parallel connections	Limit: Shannon claims the simplest circuit only in some of these cases; reprint p. 1.

He discovered a mapping, something much stronger than metaphor.

The objects can be different while their relations stay intact. Once those relations match, algebra can carry a design through steps that preserve its behaviour.

Shannon writes it the other way round. He measures hindrance, meaning obstruction. His values are 0 for closed and 1 for open. He writes series as plus and parallel as multiplication.

Both readings hold because the algebra is self-dual. That means its rules have partners with the values and operations exchanged. Shannon writes, on reprint p. 3: “It gives each theorem a dual theorem”.

Where the structure appeared earlier

The earlier cases make me slow down. My account of these four comes from secondary histories. I have not read their originals.

Account	Match reported	Limits
Peirce, 1886 letter to Marquand	Both values, with series and parallel for the operations, in an electrical logic drawing	The available account does not establish NOT, a worked algebra or a simplification method.
Ehrenfest, 1910 review	Proposed applying logic to networks	No detailed relations shown.
Nakashima, Japan, 1934-36	All five relations	The route began with circuit cases; recognition of Boolean equivalence came later.
Shestakov, Moscow, 1935, published 1941	Relations and worked algebra	Priority relies on an unpublished manuscript; publication came in 1941.

A delay in publication says something about circulation, without itself disproving the relations.

According to Stanković and colleagues, Nakashima and Hanzawa recognised the equivalence to Boolean algebra in August 1938. I have not read the original Japanese papers. They reached logic by building an algebra from circuit cases. That is the opposite direction to taking logic and applying it to circuits.

According to the same history, Shestakov’s written account dates to January 1935, his thesis to 1938, and publication to 1941. The unpublished manuscript carries the early priority claim. I cannot settle that priority without reading it.

Nakashima, Shestakov and Shannon found the mapping within about three years, across three countries. I have only read Shannon’s. His is the one that travelled.

Who could use the idea

Drawing the story around Shannon alone leaves Peirce’s earlier values and connections unexplained. They were already in his 1886 drawing.

Have we drawn the system boundary in the wrong place?

Shannon’s work met a network that could use it. His paper appeared in a mainstream engineering journal, with a method expressed in equations.

Nakashima’s papers first appeared in NEC’s journal, initially without symbolic notation. He built the algebra from circuit cases. The secondary thesis history says he “was reluctant to accept the mathematical and abstract model”.

His work had an English version in May 1936, so language alone cannot explain the difference.

Shestakov’s account did not circulate for six years.

I need the wider boundary because a method reaching readers is part of what I am trying to explain. The accounts had different routes into use. A journal with readers belongs inside that explanation.

My biggest assumption is that those routes helped decide which account changed practice. I am treating circulation as an explanation to investigate. Evidence of comparable circulation and use for the earlier accounts would weaken that reading.

What would change my mind

First, I would look for a circuit whose behaviour the algebra predicts incorrectly within the stated scope. Shannon’s method covers circuits built from series and parallel links. Bridge circuits sit beyond those links, at the edge of the general simplification claim. The mapping fails only if the algebra gets a series-parallel circuit wrong.

Second, I would look for an equivalent design algebra in engineering practice before 1938. An earlier AT&T or Siemens design manual could overturn a simple before-and-after account. I have not checked those manuals. Nakashima already gives me reason to question that account.

Third, I would check Nakashima’s reported 1938 equivalence against the original Japanese papers. If it fails, I lose the evidence that the transfer ran in both directions. That would weaken the wider claim without disproving Shannon’s circuit mapping.

Where I land

I am keeping two things. The match has to hold in the relations. And a match nobody can pick up and use stays a footnote.

For the next comparison, I want the matched rows and the missing rows on the page before trusting it.

Sources and claim map

• Shannon, 1938 circuit paper: reprint pp. 1-2 support the method, values, series and parallel connections, and scope. Reprint p. 3 covers duality and the hindrances of make and break contacts of the same relay. These are Collected Papers pp. 471-473. Published in AIEE Transactions vol. 57, 1938; presented June 1938.
• Thesis history: secondary source for the thesis being written in 1937. MIT’s thesis record is dated 1940.
• Shannon, 1948 communication paper: reprint p. 1 credits Tukey for bit.
• Tukey memo record and Bit history: memo record and secondary report of its date. The memo was not read.
• Stanković, Astola and Karpovsky: secondary accounts of Ehrenfest, Nakashima and Shestakov, including the reported equivalence and dates.
• Schlaudt on Ehrenfest: secondary account of the 1910 proposal.
• History of Information on Peirce: secondary search extract for the letter, both values, and series and parallel operations. It does not establish NOT. The page was unavailable; the original letter was not read.
• The circulation explanation is my working reading. The three checks above are proposed tests, with the unread originals left visible.

This idea was developed through an ongoing dialogue with AI. AI was used to challenge the argument, locate related research and test alternative explanations. The conclusions are mine unless otherwise stated.