Since the measurement method we adopt is rather complex, its explanation will not be thorough on this English page. A more exhaustive treatment can be found in the Italian “Misurare quanto implicito c’è in un testo” section of this website, and, even more exhaustive (and in English), in our scientific publications listed in the bibliography section (cf. especially Lombardi Vallauri – Masia 2014 and Lombardi Vallauri et al. 2020).

The impression that one speech is more biased than another, or that an individual expresses themselves in a more biased or less “honest” manner than others, is common. However, as long as these are mere impressions, even if from highly qualified individuals (such as experts in language pragmatics), they have limited cognitive value. It is therefore useful to be able to go beyond impressions, especially in contexts where communication correctness is an important value, as is certainly the case with commercial advertising or political propaganda in a democratic country. A significant component of the honesty of a speech is constituted by the use of implicitizing strategies aimed at conveying doubtful contents. This is because such contents, which could hardly convince recipients through a clear and explicit assertion, can instead be transferred to their consciousness partially or entirely unconsciously by presenting them as implicit; that is, (a) by having the recipient imply them instead of directly asserting them, so that the recipient, having contributed to the formation of those contents, is less inclined to disown them than if they were entirely asserted by the sender (Lombardi Vallauri 2009a, 2016, 2019, 2021; Lombardi Vallauri – Masia 2014); or (b) by suggesting that it is information already known to the recipient, who therefore should not need to pay full attention to it.

The so-called Moses Illusion Test (Erickson – Mattson 1981) is now a classic in psycholinguistics and functional linguistics. It involved asking experimental subjects to respond to questions such as the following:

(1) How many animals of each kind did Moses take on the Ark?

The most frequent response was “two”, while the fact that Moses was mentioned instead of Noah as the savior of the animals from the flood went unnoticed. This was because such information was presented as presupposed; in fact, as Bredart & Modolo (1988) later demonstrated, the result of the experiment was quite different if the same information, instead of being presupposed, was included in the asserted part, and indeed focused, of the message:

(2a) It was Moses who took two animals of each kind on the ark

(2b) It was two animals of each kind that Moses took on the ark

In case (2a), where Moses is focused, subjects noticed the error much more than in case (2b), where Moses remains in the background and attention is directed towards the idea of two animals per species. In other words, in (2a) the sender instructs the recipient to diligently process the identification of Moses as the savior of the animals, whereas in (2b) it does the opposite; that is, it instructs them not to pay attention to that part of the utterance’s content.

All of this prompts a reflection: if those who know that it was Noah and not Moses who built the ark do not notice the error, what happens to those who do not know it? The most likely hypothesis remains, of course, that when the information is foregrounded, it is processed with greater attention, and when it is backgrounded, it is acknowledged, but examined less attentively. And what if this information is false? As observed since at least Frege (1892) (later Ducrot 1972, Rigotti 1988, Lombardi Vallauri 1993, 1995, 2016, 2019, 2021, Sbisà 2007), if asserted explicitly, it will be received with the necessary attention to recognize its disputability, and probably rejected; if instead it is presented implicitly, it may not receive the attention necessary to unmask it, and will transfer into the recipient’s set of beliefs even though they may potentially be able to see its falsehood, and even if they have every interest in doing so. But, precisely, that way of presenting the information has been chosen precisely so that they do not do so.

So, the impression that a text is biased can be validated, among other things, by identifying within it the presentation of false or debatable contents, which occurs through constructs that reduce the recipients’ attention. And it can be useful to try to transform the impression into a more precise measure, albeit within the limits set by the intrinsic non-quantifiability of meaning phenomena. In other contexts (Lombardi Vallauri – Masia 2014, Lombardi Vallauri et al. 2020), we have proposed a method for measuring implicatures with debatable content in extended texts.