Motte-and-Bailey (fallacy)

Both, a fallacy of reasoning and of argumentation, in which concepts or positions are (intentionally or unintentionally) extended, narrowed, or exchanged in a way that distorts their meaning on the basis of ambiguous definitions.

This error is often exploited as an (unfair) discussion tactic, such as in the following example:

A: Homeopathy is quack, as it has no medical effect beyond the placebo effect.
B: That is not true at all, naturopathy is a recognized discipline of medicine that has often proven to be successful!

The term “naturopathy” is used as a generic term for a number of different alternative medical methods. This indeed includes various recognized treatment methods (e.g. hydrotherapy), but also some for which a benefit beyond a mere placebo effect is not demonstrable. Homeopathy belongs to the latter category.

The name refers to a type of medieval forticication that was originally named in French: “motte et bailey ”. It consists of a lightly fortified rampart (the “bailey”) that was combined with a more strongly fortified, usually tower-shaped castle (the “motte”), often built on a mound.

In the event of an attack, defenders could abandon the outer rampart, if they needed to, and retreat to the much more defensible tower, from where attackers could readily be held at bay with projectiles. This tactic is very similar to the rhetorical discussion strategy described above.

The essential aspect of this fallacy is that two positions are conflated and considered to be “identical” while they actually differ in important aspects. When used as a rhetorical diversion tactic (red herrings) it is trying to exploit the confusion by countering an attack on the weaker “bailey” position by instead retreating to the more defensible “motte” position.

From the opponent’s point of view, this kind is similar to a straw man argument, since they have to deal with a defense from from a position that they did not attacked at all.

This argumentation also has aspects of equivocation, since two actually different positions are named the same here.

A specific quirk of the Motte-and-Bailey as a rhetoric fallacy is that it is often in the interest of the opposing side to maintain the confusion, as this opens up the possibility to claim that a successful attack on the “bailey” also counts as a defeat of the “motte”.

For example, one could just turn the above example around:

Homeopathy has been proven to have no effect beyond the placebo effect.
This shows that naturopathy as a whole is nothing but quackery!

Again, this conclusion is invalid, because homeopathy and naturopathy are not identical.

Giving up positions that can no longer be held during a discussion and ( in most cases tacitly) retreating to those for which there are more solid arguments is a normal process and should not be misunderstood as an unfair strategy. On the contrary, it should be seen as a sign that the opponent is willing to listen to arguments.

It only becomes an unfair strategy when the actual indefensible position is not abandoned, but instead it is pretended that the positions are identical and that arguments for one position also support the other.

The characteristic feature of this fallacy is that two or more concepts or positions, which should actually be kept distinct from each other, are conflated. As with equivocations and other errors of ambiguity, the more abstractly and possibly vaguely these are defined, the more difficult it is to recognize such a situation.

The first step should therefore always be to insist on a clear and unambiguous definition and, if possible, a realistic example of the terms used. If it turns out that two different things may have been lumped together, it is a promising approach to insist in making the distinction clear.

Such an objection might be something like the following:

“But these are now two different things, which we should also clearly distinguish.”

In information technology, the term “blockchain” refers to a data structure in which blocks of user data are linked together by means of cryptographic hash values. This serves to ensure that once these blocks of user data have been created and published, they can no longer be changed without invalidating the entire rest of the chain.

A useful use case of such blockchains are so-called “distributed ledgers”, i.e. a type of account management in which information on payments and expenditures is shared with several computers in an untrusted network (i.e. on the Internet).

Simply put, the blockchain ensures that when the current block of transactions is accepted, all previous blocks and thus the transactions they contain are implicitly accepted as well. In this way, the blockchain ensures the consistency of all transactions that have taken place so far – and specifically it does so in a way that can be mathematically proven to be secure.

However, the distinction between the “blockchain” data structure and the applied form as “distributed ledger” is very important, as it does not follow from having a manipulation-proof data structure, that the code that is used to distributed it through the network is also safe and tamper-proof – or even that all participants receive and accept the same blocks.

In other words, even if the actual blockchain is secure against tampering, this does not mean that the same is true for the distributed ledger – not to mention other components of the blockchain infrastructure such as trading or mining platforms.

However, this distinction is all too readily blurred by using the term “blockchain” to refer to the entire system, even if it actually constitutes only one aspect of it. Such a broadening of the term, however, makes it difficult to scrutinize the security of the entire implementation.

In this scenario, the blockchain would correspond to the “motte”, i.e. the easily defensible position: Its cryptographic functions, which ensure data consistency, are demonstrably secure (at least within certain framework conditions). In contrast, the other components of the implementation represent the “bailey”, which attempts to benefit from the proven security of a blockchain, although it concerns separate subsystems whose security and reliability would first have to be proven. Such an approach therefore has characteristics of a confusion tactic.

TODO: More examples follow.

• Equivocation
• Cherry-picking
• Mereological fallacy
• Straw man argument

• Motte-and-bailey fallacy on Wikipedia
• Motte and bailey on RationalWiki