When Doing Is Saying

The Theory of Behavioral Implicit Communication

Cristiano CastelfranchiISTCnr Roma, ThΨ-- Theoretical Psychology Group 1

<castelfranc@unisi.it>

WORK IN PROGRESS (uncorrected text)

1. Introduction

I will introduce, define and illustrate in this work the phenomenon of usual, practical, evennon-social behaviors contextually used as messages for communicating. Behavior can becommunication without any modification or any additional signal or mark.I will call this form of communication without specialized symbols: Behavioral - ImplicitCommunication (BIC).“Behavioral” because it is just simple non-codified behavior.“Implicit” because – not being specialized and codified – its communicative character isunmarked, undisclosed, not manifest, and thus deniable. In BIC communication is just a ‘use’and at most a ‘destination’, not the shaping ‘function’ (Conte and Castelfranchi, 1995 Ch. 6)2.Normally communication actions are on the contrary special and specialized behaviors (likespeech acts, gestures, signals, …).Behavior can be communication without any modification or any additional signal or mark.BIC is a very important notion, never clearly focused, and very frequently mixed up withother forms of communication (typically the so called “non-verbal” or “expressive” or “extra-linguistic” or “visual” communication). It has been the source of a number ofmisunderstanding and bad definitions, like the Batesonian claim that all human behavior iscommunication (see 2.1). Sometime, sub-categories of it have been identified (likeStigmergy), but without any reasonable characterization of the general notions and providingwrong definitions (see 2.5).This ill-treated notion on the contrary is really crucial both on the theoretical and on thepractical side: 1 OMLL project – Eurocores . I’m grateful to my students Anna Galli, to Francesca Giardini and EmilianoLorini (which are also collaborator in the OMLL project) and Luca Tummolini for their pretious researchesand discussions on various issues of this subject. I also thanks Isabella Poggi for our discussions and for hercompetent explanations on gestures and NVB; Maria Miceli and Rino Falcone for comments.2 I might call this form of communication simply “behavioral communication” (non-non-verbal), but I preferto use also the term “implicit” (although not so clear and rather abused) because it makes more evident

- the unmarked character of the signal (the vehicle, the “significant”), and- the special relationships between BIC and human implicit agreements and tacit negotiation on whichthe whole “Social Contract” holds.

I will use the term “explicit” communication – as opposed to BIC – in a very specific and rather well-definedway. “Explicit” in this context means “manifest”; it cannot be denied that it is a communication act/signal,since it is conventional, artificial (or selected by evolution), shaped by and for communication, it is aspecialized act/object.This is why I do not call “explicit” communication the meta-communicative BIC (section xx), although in asense the intention to communicate is explicit, is disclosed not hidden: the source intends that the addresseeunderstands that s/he is willing to communicate (“make the other understand that…”). However, if there is nospecial and specialized unambiguous signal of such a meta-communication also this intention is not reallyexplicit, it is just inferred and can be denied.

A) It is crucial for the whole theory of social behavior: coordination, control, social order(norms keeping) instauration, identity and membership recognition, social conventionsbuilding, cultural transmission, deception, etc.Frequently it is very important in its more sophisticated and advance form of meta-communication that we will call meta-BIC.A lot of social control and collaboration monitoring and coordination, are based on thisform of communication and not on special and explicit messages (communicationprotocols).

B) Even for the theory of linguistic communication BIC theory is fundamental:- because behavioral communication is involved (like ‘non verbal communication’) in

a lot of pragmatic inferences: part of the meaning of the sentence is inferred from(and conveyed by) the attitude and the behavior of the speaker;

- because it is possible to derive from BIC the Gricean meta-communicative characterof linguistic communication and in fact of any artificial and conventionalcommunication system; (see xx)

- because BIC is the basis for meaning ‘negotiation’, and for any form of linguisticconventional ‘negotiation’ which very rarely is explicit; (see xx)

- because (as I will claim) this form of emerging and spontaneous communication with nospecialized signal is one of the forerunners and premise for the evolution andacquisition of language (section xxx; castelfranchi, Siena; Giardini and Castelfranchi,2003).

C) BIC theory is also relevant for quite serious practical problems and in severalapplication domains, like:

- the problem of social order and social control in MAS, CSCW, virtualorganizations (Omicini, ….; Tummolini, …..)

- friendly and natural HM interaction [Castelfranchi & Giardini, 2003]; Observation- interaction design (Castelfranchi, 2005)- social exchange and its rules- learning and teaching- monitoring- redundant and expensive communications- efficient, simple and spontaneous forms of coordination (ex. mobile agents Gulyas,

99).??In this perspective I will for example criticize the message sending paradigmdominating CSCW, HCI, HRobotI, MAS, etc. I claim that BIC will play animportant role also in artificial agents, in HCI, and in computer mediated interaction.TOGLIERE?

Although this form of communication is some sort of social glue and is ubiquitous, I will startprecisely by denying that all human behaviors (also in front of others) be “communication”, inorder to arrive to a heuristic and discriminative notion and to a clear cutting definition ofcommunication. The crucial point will be that the fact that something is a sign, that somethinghas some meaning (means something for Y) is not enough for having communication.

I will analyze implicit behavioral communication in: 3

3 Except for the example of a defensive mechanisms based on acting in order to reinforce a belief (section ) Iwill not deal at all with neurotic and psychosomatic symptoms (signification for the doctor) that are

RIORDINARE e COMPLETARE- commitment satisfaction (for example in exchange) (section xx),- monitoring or surveillance (section xx, and conventions; conclusion),- norm and role obedience; then in “organization” (section xxx);- deception through behavioral communication (ex. bluff and simulation) (??);- tacit “negotiation” about meaning, rules, etc. in linguistic interaction (xx);- tacit creation of expectations and entitlements (section xx); till- the formation of true conventions and of tacit agreements for delegation, exchange or

cooperation;- imitation;- the origin of language and gestures.

2. Towards a Definition of Behavioral Communication: when behavior iscommunication and when is not.

2.1 Are we damned to communicate? Signification vs. Communication

A famous thesis of Palo Alto psychotherapy school [Wat67] was that:"It is impossible do not communicate", and "any behavior is communication" in socialdomain. In this view, a non-communicative behavior is a nonsense. 4This claim, which is now again very trendy -even in “artificial” interaction models like inMAS- is wrong or, if you prefer, too strong.5 It gives us a notion of communication that isuseless because is non-discriminative at all.In order to have communication having a "recipient" which attributes some meaning to acertain sign (or exhibits a specific and appropriate reaction to it) is not a sufficient condition[Eco75]. I will claim that we cannot consider as communication any information/sign arrivingfrom X to Y, unless it is aimed at informing Y. For example, Y can observe X -X not beingaware of this- and can understand a lot of things about X, but X is not communicating with Yor informing Y about all those things. In other words, a ‘source’ is needed which sends onpurpose such a "message" to the receiver. A teleological (intentional or functional) "sending"action by the source is necessary. The source has to perform a given behavior in order theother agent interprets it in a certain way, receives the message and its meaning.Is for example an escaping prey “communicating” to its predator/enemy its position andmove?Watzlawich’s overgeneralization cannot avoid considering communication to the enemy thefact that a predator can observe the movement of the prey, or that in a hidden way the enemy

frequently claimed to be messages. I do not want deal with them because I disagree about the theory ofunconscious intentions and its is questionable the presence of a communicative “function” of them (Castelfrr) since such a communication seems to be quite unsuccessful, more the failure than the success reproducesthe attempt (if any).4 “…..il comportamento non ha un suo opposto. In altre parole, non esiste un qualcosa che sia un non-comportamento o, per dirla ancora pi˘ semplicemente, non e’ possibile non avere un comportamento. Ora, sesi accetta che l'intero comportamento in una situazione di interazione ha valore di messaggio, vale a dire Ëcomunicazione, ne consegue che, comunque ci si sforzi, non si puÚ non comunicare.”. The authors are evenmore extreme on page 44: “ Per riassumere, si puo’ postulare un assioma "metacomunicazionale" dellapragmatica della comunicazione: non si puo’ non comunicare.”5 At the end of our long path through BIC –after introducing not only the ‘intentional’ BIC but also the‘functional’ non-intentional one – we will show that Watzlawich is practically right; his claim is only tostrong and based on invalid principles.

can follow our moves with binoculars. Although this information is certainly very relevantand informative for the enemy or predator, is not communication. The target is unaware or istrying to avoid that the information arrives to the enemy. The fact that the enemy can see isnot taken into account among the more or less motivating expected results of the action.Sometime one lets the other see but just because she cannot avoid this.In functional terms, receiving the information is functional (adaptive) on the side of thepredator and the species has developed such an ability, but it is not functional at all, is nonadaptive for the prey: “sending” that sign is not a functional (evolutionary) goal of the prey, isnot the evolutionary pressure on its side. To be communication the sign should have beenselected for an advantage of both sides (not necessarily the same), and more specifically itshould be an advantage for X that Y is able to catch the message. If Y just exploits for itsown advantage X’s signs (and even X’s messages to Z) we do not have “communication”from X to Y. It seems that some animals exploit the alarm signals of chimps [], but this is notcommunication towards them.It is even possible that the same behavior/signal be communication towards a given(conspecific) animal while is not communication when intercepted by another animal (forexample a predator or a parasite) since it is not functionally addressed to it [Castelfranchi,1980, 1998: Millikan, ].Consider another example: a pilferer entering a stock during the night and cautiouslyproceeding in a corridor. He does not notice that there is a working TV camera surveillancesystem and thus he does not know that there is a guard which is following him on a screen!For sure this observation is full of meaning and very informative for the guard, but for surenobody would say that the pilferer is informing or communicating to the guard about hispresence and moves. Consider also a pilferer escaping from the police but leaving on theground prints and traces of his direction (see later on Stigmergy). Are those signs (verymeaningful for the police) messages to it?In other words, the end of the source is determinant: sending the message must be eitherintentional or functional. 6

Normally communication actions are special and specialized behaviors (like speech acts,gestures, signals, …). However, what about the case in which the observed agent is aware ofthe interpretation of the other about its practical behavior, and intends the other to make such

6 My position is close to a clear statement of Bruno Bara (Bara, B. Pragmatica cognitiva, Boringhieri, 1999)in his definition of communication, where he claims that: “…it is impossible that A communicates somethingto B without having the intention to do so; otherwise it is just B that was be able to autonomously infer someinformation, but without A’s participation”. This is precisely my argument against Watzlawich (but adding tothe ‘intentional’ level the ‘functional’ one). Unfortunately, Bara takes another path, and seems to mixe upthis clear distinction between mere “signification” and true “communication” with the Gricean view ofcommunication, where not only there must be an intention, and thus a real “sending” and a real “message to”,but it is crucial – and intended- the recognition of such an intention by the addressee (some sort of meta-message, in fact).The first part of Bara’s sentence was in fact “in order to use the term communication all the agents mustmake explicit their conscious intention to take part in the interaction…..”. And he also clarifies that “Asdeparture point I will use the notion of communication provided by Grice…. …. (where) the intended effecton the receiver B is obtained, at least in part, thanks to B’s recognition of A’s intention to communicate”.Gricean characterization – altough enlightening and profound - in my view, is not general enough and doesnot apply to all the forms and layers of communication. In particular, BIC does not requires it; only meta-BICis at the gricean level.On the contrary, I will claim that gricean functioning – typical of human language and communicationsystems – is due to and explained by BIC: the meta-message is just and precisely BIC communication due tothe fact that also communicative specialized (linguistic) behavior, is just a kind of behavior subject to BICuse (section xx).

an interpretation? This is some sort of "implicit" communication, without any special action,but with a communicative intention. This is what I call BIC (in its intentional version).

To have a theory of this implicit form of communication we need to attempt to (partially)clarify in the mean time some basic issues like:

- the distinction and the relationship between “signification” (getting a meaning) and“communication” (getting a message);

- the nature of action, distinguishing passive (to let) and active (to make) forms (sectionxxx);

- the nature of the above mentioned teleological nature of the “message”, i.e. thedistinction between intentional communication and the “functional” one (§ 2.3);

- in case of intentions, we have to make clear the difference between true motives andside or passive intentions (xxx) MANCA.

2.2 Signification in Semiotics

We call “signification” – following [Eco, 1973; 1975; 1994] – the fact that a perceptual inputis endowed for the recipient/interpreter of some meaning; it is “interpreted”: it is a sign ofsomething else.Prints on the ground are for the hunter signs of the passage of a deer; smoke is the sign of afire; spots on the windows can mean "it is raining" (they are for Y signs of the fact that it israining).Clearly there is no sort of communication in those examples, while percepts and events areendowed with meaning in the mind of the “observer” and it is consistent and reasonable tocall them “signs”. We have here simple processes of signification. Notice that meanings arenot conventional but simply based upon natural perceptual experience and inference. Noticealso that the signal, the vehicle, the ‘significant’ has not been manufactured on purpose forconveying this meaning, has not be “coded” and “decoded” via some conventional artificialrule.Our view of the “signification” phenomenon is inspired by Eco’s notion but it is not identicalto it. Eco has the merit of introducing in a clear way such distinction between signification(and thus “sign”, “meaning”) and communication; and also to make clear that the formerprocess is independent from the latter, preliminary, and a necessary presupposition of thelatter: no communication without signification.However, Eco’s view seems aimed at restricting the theory to cultural phenomena. He wantsto limit “signification” to humans and their cognitive interpretation processes, and –moreimportant – he wants to stress very much the cultural nature of the “signification” process asbased on a set of conventional rules of correspondence between the perceived signs and theirmeanings.7

The role of “natural signs”, of non-convention-based semiotic process, of sign-creation inhumans is not well clarified, and even contradictory. Eco also mixes up – as everybody –

7 This restriction might be good for our attempt to distinguish BIC communication from other forms ofhuman communication; but it is a pity

- to miss the relationship with animal communication (and the possibility of a general theory),- to miss the interesting parallel of the evolutionary transition from mere perceived significant behavior to

specialized communication signals,- to lose the relationship with stigmergy (that in fact is just a sub-case of BIC).

“expressive” signs, non-verbal symbolic communication, and BIC (precisely what I’mattempting to disentangle). 8

The main point of disagreement seems the presumed necessary presence of conventional“codes” for signs and signification. For us, it is important to stress the non-conventional, evenidiosyncratic, spontaneous, and natural formation of signs and thus of communication; whilesemiotics is interested in stressing the system and code dependent nature of meaning. Aclarification of this issue is needed.

Is there any code under the BIC?It is not completely correct that in BIC there is no “code”, “coding” and “decoding”. Thisobviously depends on how strictly we define those notions. For example, Eco considers acode present and necessary in any sign and any signification process. This is too strong if wehave necessarily in mind conventional and artificial shared codes (due to culture) – assemiotics usually do. There is a sign and a signification process even when Y sees at distancesmoke and believes that there is some fire there, or when takes sun light as a sign that it isworm outside. However, if we consider a “code” also something that can be individually andun-conventionally developed, and we consider that to perceive means to select theinformation and to integrate it during the “recognition” process, and that recognition impliesat least some pattern matching with some simple pattern in the perceiver’s neuron-system, or -at a more complex level - implies a schema or even a category and concept to whom thepercept is “assimilated” (Piaget), we can clearly say that any real perceptual process impliessome form of “code” either inborn or individually developed or culturally shaped. If this istrue, in intentional BIC there is some role of this “code”.In fact X, the “sender”, must known (guess) the (“natural” or cultural) code of the addresseein order to plan that it will recognize from the stimuli this and this, and interpret it in such andsuch a way.Notice that it is not strictly necessary that they share such a code and the system of meaningsand the underlying ontology. It is enough that X knows Y’s code and ontology. Sharing is justthe simplest, natural, and efficient way for knowing the code of each other; and in fact it is themore common case, specially based upon the projection and ascription of our own schemataand meanings to the other, and to the confirmation of this due to the success of the attempt(section xxx). Thus when X behave while taking into account (also) how Y will recognize andinterpret its behavior, in a sense X is “codifying” its behavior for Y as a message to it,because X would not use that behavior (or not use it in such a way) if Y would not perceiveand interpret it in a given way. But saying this is different from asserting that there is alexicon of behaviors specialized as signs by evolution, convention, or artifice.

2.3. Goal-Governed and Goal-Oriented Agents: Intentional Communication vs.Functional Communication

Since the crucial component for the notion of communication is the finalistic nature of the actof “sending” the message, the purposiveness from the point of view of the source, we shouldclarify a bit this issue since we do not want ascribe intentions and mental state to any animal 8 Although I accept Eco’s claim that not every inference is a semiotic act, and not every semiotic processimplies inferences (a criticism to Peirce’s view), I do not believe that this hard problem (which inferencescreate signs, and why) has been solved, and that the solution simply remains in conventions. This is thesolution only for advanced and stabilized communication systems (languages). It is also very lacking in Eco– in that period - the theory of teleonomy, both on the side of “intentionality” and – even more - on the sideof functions.

(for example to insects) although we want to have a general notion of communication andcoherently use this notion for animals. In order to do this we need two kinds of ‘finalism’, of‘teleology’ or ‘teleonomy’ (Mayr, 19xx).There are in fact two kind of goal-oriented systems and behaviors, the cognitive, intentionalones, and the merely goal-oriented without any internal anticipatory representation of theresult of the action, where the finalistic character of the behavior is merely in its adaptivefunction.

We define BIC as:

(In Order to (Performs X β) (Causes (Observes Y β) (Knows/Bel Y p)))

“X performs the behavior β in order Y perceives it and on such a basisunderstands/believes that p”. 9

This teleonomic notion we need, this "in order to", has two different meanings:- either the message is sent on purpose, intentionally by X, which should be for this acognitive purposive system, an intentional agent (in this case X believes and intends allthis);- or, the message is not intentional but simply functional; the sending behavior is notdeliberated but is mere goal-oriented behavior, either designed by some engineer, orselected by natural or artificial evolution, or selected by some learning.

Thus, we have two basic kind of communication:• Intentional (or better "goal-governed") communication; and• Functional (or "merely goal-oriented") communication.

Actions and social actions are possible, of course, also at the reactive level, among sub-cognitive agents (like bees). A definition of social action, communication, adoption,aggression, etc. is possible (and necessary) also for non-cognitive agents. However, also atthis level those notions must be goal-based. Thus, a theory of merely goal-oriented (not "goal-directed") systems and of implicit goals is needed.As we said, there is a form of BIC which is functional but not in evolutionary sense: thecommunicative import of the behavior is not biologically established (“inborn”). It is justemergent and self-organized by a cycle of reinforcement learning (Castelfranchi, 2001); it isan acquired function. I will call it learning-based BIC (and distinguish it in merely reactiveand proto-intentional) 10. Since I claimed that in BIC there is no learning of specialized actsfor communication, of special signs/signals, it is very important to make clear that what islearned in learning-based functional BIC is only its use and destination for communication, itscommunicative effect, not the very act which is already existing and preserves its pragmaticgoal. 9 I will call also µ the meaning conveied by the act/message β (§ xx)10 To be more precise BIC theory should be developed at 3 layers:FUNCTIONAL, with several sub-type: by evolution-selection; or by design; or by reinforcement learning basedon the effects (conditioned BIC);PROTO-INTENTIONAL, based on Anticipatory classifiers (proto-intentions) thus reinforcement learning butwith some anticipatory representation;INTENTIONAL that usually also presupposes an intentional stance and more precisely a “theory of mind” in theinterpreter, since the message bring by the action can be about the mind of the source: his intention, or emotion,or motives, or assumptions, etc.

In the rest of the paper – since I’m interested here in stressing the importance of BIC alsoamong cognitive agents, and especially among humans, even for the theory of linguisticinteraction but for sure as one of the basis of social structure - I will let aside functional BICamong reactive agents (except for discussing Stigmergy) and focus on intentional BIC, itsvarieties and diffusion, and mental conditions and steps necessary for it. Only in section xxx Iwill come back to a more general notion (non limited to intentional agents) by analyzinglearning-based BIC and in particular the pro-intentional form. I will also shortly discuss(section xx) BIC based on social functions in order to explain why a lot of human behaviors,beyond intentionality, are communication.

At the intentional level, the definition of BIC (that will be analyzed and justified on section3) is as follows:

in BIC the agent (source) is performing a usual practical action β, but he also knowsand lets or makes the other agent (addressee) to observe and understand such abehavior, i.e. to capture some meaning µ from that “message”, because this is part of his(motivating or non motivating) goals in performing β.

In sum, BIC is a practical action primarily aimed to reach a practical goal, which is alsoaimed at achieving a communicative goal, without any predetermined (conventional or innate)specialized meaning.

In § 3 I will carefully explain all these distinctions (to let vs. to make; motivating vs. non-motivating). But now it is more urgent to immediately disentangle BIC from the so called“non-verbal”, “extra-linguistic” communication, with whom it is always mixed up.

2.4 Why BIC is not “non-verbal”, “extra-linguistic” communication

Behavioral communication as conceptually characterized and well defined here is not thesame and has not very much to do with the so called non-verbal or extra-linguisticcommunication (NVC) although NVC is through some behavior or behavioral features, andBIC is for sure non-verbal and extra-linguistic.Usually, the few of BIC that has been identified has been actually mixed up with the formercategory that is never well defined. and is rather confuse both conceptually and descriptively.Consider for example a quite recent definition of NVC: "The process of sending and receivingwordless messages by means of facial expressions, gaze, gestures, postures, and tones ofvoice… Also included are grooming habits, body positioning in space, and consumer product(e.g., clothing cues, food products, artificial colors and tastes, engineered aromas, mediaimages and computer-graphic displays). Nonverbal cues include all expressive sings, signalsand cues (audio, visual, tactile, chemical, etc.) which are used to send and receive messagesapart from manual sign language and speech." (Givens, 2003). Such a definition, including abroad notion of gestures, body positioning, etc. is not able at all to disentangle BIC and ingeneral practical actions and their “meaning” from non-verbal specialized signals.11

11 The same mixture holds in other definitions that embrace facial expressions, gaze, tones of voice,… andposture body movements, position in a group, etc. G. W. Porter (1969) ha diviso la comunicazione nonverbale in quattro ampie categorie:

Fisica: fa riferimento al tipo di comunicazione personale. Include espressioni facciali, tones of voice,senso del tatto, odore e movimenti del corpo;

The situation is not much better with the alternative notion of “Non Verbal Behavior” (forexample Ekman and Fiesen 1969) that also seem unable to be discriminative. 12

Being so badly characterized non-verbal communication could include also BIC, but this isnot an advantage at all. It prevents from identifying a very important spontaneous pre-conventional and (in humans) not-inborn form of communication that has its own explanatorypower and foundations, and cannot be mixed up with mimicry, emotional signals, orconventional gestures (like those of a policeman regulating traffic).In fact, non-verbal and extra-linguistic communication refers to specific and specializedcommunication systems and codes based on facial expressions and postures, specific gestures,supra-segmental features of voice (intonation, pitch, etc.), etc. that communicate specificmeanings (Poggi, xxxx) by specialized, recognizable (either conventional or universal) signals(EiblEibelsfeldt, xx). This special codes are not BIC as we characterized it: usual, practical,even non-social behaviors contextually used as messages for communicating.BIC is not a “language”, is not a communication code, although is a way for communicatingand a form of communication. Any (verbal or non-verbal) “language” has some sort of“lexicon” i.e. a memory of a list of (learned or inborn) perceptual patterns specialized as“sign”: where “specialized” means either conventional and learned as signs, as possiblemessages for communication, or built in, designed for such a purpose (function) by naturalselection or engineering.BIC does not require (in its intentional version) any specific learning or training13, ortransmission; it simply exploits perceptual patterns of usual behavior and their recognition.One might object that in any recognition process there is some interpretation and meaning, so,where is the difference? The difference is not so much in the receiver/interpreter (except forthe more or less “natural” or “conventional” nature of the meaning association process) and inthe mental part of the “signification” process; the difference is in the sign itself and in thesender.

Estetica: ha luogo attraverso le espressioni creative: riguarda la capacit‡ di suonare uno strumentomusicale, di ballare, di dipingere e scolpire;Segnica: Ë un tipo di comunicazione meccanica, che include l'uso di bandiere segnaletiche, corni, sirene,il saluto militare;Simbolica: questo tipo di comunicazione fa uso di religione, status o simboli di costruzione dell'io.

La CNV, inoltre, viene ripartita da alcuni ricercatori in due grandi categorie: statica e dinamica. Della CNVstatica fanno parte quegli elementi che non mutano durante il corso dell'interazione, ad esempio i vestiti ed ilcolore dei capelli. In se stesso questi sono elementi facilmente modificabili da un giorno all'altro, ma nondurante il corso di un dialogo convenzionale. Lo stile dei vestiti, ad esempio, viene usato da particolari gruppiper esprimere valori come la coesione e l'orientamento nei confronti dell'autorita’, quello che e’ successonegli anni '60 per gli hippies, nei '70 per i punk, e negli '80 per i romantici. Esempi di identit‡istituzionalizzata attraverso uniformi o divise ci vengono forniti dai poliziotti, dai giocatori di football, daimonaci con i loro proverbiali abiti. I vestiti assegnano un significato a chi li porta. Come si vede l'esempiodei vestiti si trova sullo spartiacque tra la comunicazione comportamentale (BIC) e quella non verbale. Comeabbiamo precisato quest'ultima nozione e’, infatti, mal definita e include fenomeni che sono comunicazioniimplicite, o possono esserlo. Vestirsi e’ infatti un comportamento vero con reali scopi pratici (ad esempio,ripararsi dal freddo), ma puo’ diventare anche comunicazione, se il mio abbigliamento e’ studiato o fattoapposta per trasmettere agli altri determinate informazioni (ad esempio: sono ricco, ho certe idee,ecc.). Puo’anche essere una semplice via di significazione, se mi vesto in certo modo, ad esempio indosso una pelliccia,per ripararmi dal freddo, mentre altri inferiscono che lo faccio per mostrare la mia ricchezza.12 Moreover, I wouldn’t necessarely consider BIC non-verbal behavior, since also verbal behavior (speachacts for example) are behavior and can become – at a secondary level - BIC! (section xx)13 In BIC forms where the communicative effect is not intended, learning is necessary for the practicalbehavior to acquire this additional function in non accidental, systematic way. But the behavior per se is notlearned; it is presupposed.

As for the sign, it only has the parasitic use of sign, it has not been shaped, selected,designed for this; its morphology and existence does not depend on its use as a message(Conte and Castelfranchi, 1995, Ch.6).As for the source, in mere signification, it does not “send” the sign, there is not a “messagefrom” the source, there is no end/telos on the side of the source (neither intention norselective function); in BIC use there is a message but it is questionable, there is no markand shape of the vehicle as a message.

BIC is an observation-based, non-special-message-based, unconventional communication,that exploits simple side effects of acts and the natural disposition of agents to observe andinterpret the behavior of the interfering others. 14

Consider for example the difference between BIC and gestures in the NVC theory tradition.There is a “system of symbolic gestures” with their associated (conventional orunconventional) meanings, gestures that are “symbolic” precisely because they are made forconveying meaning. For example, the Italian gesture with the special meaning “he’s on mystomach” “I cannot bear that person” (Poggi, 2001).BIC gestures are just gestures, they are non-symbolic15 but practical: to drink, to walk, toscratch oneself, to chew. They represent and mean themselves and what is unconventionallyinferable from them (like the agent’s intentions and beliefs). 16

Of course, the border between a specialized non-verbal sign and a non-symbolic behaviorused for communication, is fuzzy; there is a continuum there17 in terms of iteration of use, ofconventionalization, of an evolutionary process, of the passage from mere use, to destination,practice and habit, and to a true “function” (either natural, or socio-cultural, or artificiallydesigned) (see section xx).Frequently, from the empirical point of view the border and the transition between BIC andnon-verbal (expressive or symbolic) communication is rather fuzzy and frequently they arealso used in a mixed way. Both by his practical behavior and by his mimicry (face, postureand voice) X communicates something to Y. For example, when X is biting Y, he iscommunicating his rage and aggressive disposition both with face and voice but also with the 14 Obviously in functional BIC or in repeated and habitual (pre-ritualized) BIC there is a specialization of thatbehavior for communication; however the function did not (yet) morphologically shape a specific non-practical sign.15 It would be very clear and simplifying claiming that they are unconventional (as in fact they are), but this isnot clear cutting because of ethological communication signs in humans and other animals, that for sure arenon-conventional although they are specialized as signs, as messages (although several of them for surederive from preexisting non-communicative behaviors that received a secondary function and specialize forthis, through a process of exaptation, simplification, emphasis, and “ritualization” (Huxley, ; Tinbergen, ).16 BIC is what Posner and Serenari (2000) call “the degree 0 of gesture” when the gesture with a practicalfunction not yet symbolic start to be used for communication. I do agree with this view and with theirinteresting remarks on this transition and its ambiguity; I only object about the term “gestures” that does notcover well the entire behavior (like walking, entering, to go outside, eating, sleeping, studying, bombing, etc.etc.). All behaviors and their outcomes and traces can be used for communicating; gestures are just a sub caseparticularly interesting for their easiness to specialize into symbolic conventional signals.Also symbolic gestures are gestures, behaviors, thus they can be interpreted at a meta-level as BIC. Forexample emotional facial expressions with their specific meaning (like “I’m surprised” “I’m interested” “I’mdelighted”) can have a more indirect, inferential (not stably associated as its lexical meaning) meaning like:“I’m following you”; or “I’m moved”, “I’m an emotive guy”, etc.17 The idea of a continuum between non-symbolic or not completely “lexicalized” gestures and fullysymbolic, specialized, conventional ones, but perhaps also between practical actions and symbolic gestures(unavoidable given the mess of the notion of VNC and NVB) is topical in that literature. Kendon (1980introduced the notion of a "gesture continuum", from gestures more communicative, symbolic, to lesscommunicative. This gesture continuum is present in McNeill (1992) and in fact is also close to or inspiredby Lyons (1972) idea that NVC varies as for its degrees of “linguisticity”.

action of biting by itself. For example, when I want to stay alone, both my practical behaviorof not being involved, of doing my own affairs, and my expression and posture bring mymessage.Analogously, the border between (meta)BIC and ritualized, conventionalized communicativebehaviors is fuzzy. In fact during the process of ritualization for shaping an explicit (seedefinition) coded sign, not necessarily β completely looses its practical component (seesection on BEC).Similarly, the act of “handing or offering” is a borderline act between metaBIC and a fullycommunicative conventional act with the same goal structure of a speech act. X wants that Ytakes the object w (Goal 1), and intends (Goal 2) that Y understands that he wants this, andplans to achieve Goal 1 also through Goal 2; and also wants (Goal 3) that Y understands thatX wants that Y understands. The transparence of X’s (speaker) communicative intention is ameans for achieving it. This is also true for the metaBICs, but in a conventionalized sign thistransparence is related to some feature of β, to some shaping of the act/object making it arecognizable (undeniable) sing (this is why I call it manifest). ‘Handing’ it is not simply to putw close to Y, within her reach so that accidentally and without understanding she takes w. Iwants that you understand that I makes w closer to you in order to communicate to you that Iwant that you take it, and you take it also because I want (offer) you to do so. This messagenature seems also to characterize (shape a bit) the way in which the act of making w close toY is performed (for example making sure that you notice it); the typical act of‘handing/offering’ (for example a bit different from the act of ‘giving’ that might alreadypresupposes that you are taking or intend to take). QUI????

2.5 The Stigmergic Over-generalization

The notion of Stigmergy comes from biological studies on social insects, and more preciselythe term has been introduced to characterize how termites (unintentionally) coordinatethemselves in the reconstruction of their nest, without sending direct messages to each other.18

“When they start to build a nest, termites modify their local environment by making little mud balls and placing them onthe substrate; each mud ball is impregnated with a minute quantity of a particular pheromone. Termites deposit their mudballs probabilistically, initially at random. However, the probability of depositing a mud ball at a given location increaseswith the sensed presence of other mud balls and the sensed concentration of pheromone. The first few random placementincrease the other termites' probability of putting their loads at the same place. By this blind and random game littlecolumns are formed; the pheromone drifting across from neighbouring columns causes the tops of the columns to be builtwith a bias towards the neighbouring columns, and eventually the tops meet to form arches, the basic building units.Finally, as the influence of other stigmergic processes comes into play (e.g. processes involving water vapour and carbondioxide concentrations, and modulated by the presence of the queen), the whole complex and highly differentiated neststructure is produced, with the royal cell, brood nurseries, food stores, air circulation, communication and foragingtunnels, and other areas all contained within one of the largest non-excavated structures built by any creature except man”19.

Stigmergy is derived from the root "stigma" (outstanding sign) and "ergon" (work), thusgiving the sense of "incitement to work by products of work". It is essentially the production

18 Il primo a parlare di comunicazione stigmergica è stato il biologo francese P.P. Grassé P.P. Grassé, Lareconstruction du nid et les coordinations iner-individualles chez Bellicositermes natalensis et Cubitermes,sp. La teorie de la stigmergie: Essai d'interpretationdes termites constructeurs. Ins. Soc., 6, 41-83, 1959 (1959), in seguito ai suoi studi sul comportamento delle termiti nella costruzione del loro nido.19 "From Local to Global tasks Stigmergy and Collective Robotics", R. Beckers, O.E. Holland, J.L.Deneuborg, Artificial Life IV, Rodney A. Brooks and Pattie Maes (eds.), 1996, pag.181.

of a certain behaviour in agents as a consequence of the effects produced in the localenvironment by previous behaviour 20.This presentation of Stigmergy is not able to discriminate between simple signification andtrue communication, and between prosocial and antisocial behavior. They would for examplecover prey-predator coordination and our pilfer (unintentionally) leaving footprints veryprecious for the police21. In order to have “communication”, it is not enough that an agentcoordinates its behavior with the behavior or thanks to the traces of the behavior of anotheragent. This is an over-generalization.Stigmergy is also defined as “indirect communication through the environment”; 22 to me sucha definition is rather weak and unprincipled.First, a lot of usual communication and even linguistic messages are directed towardsunknown or unspecified addressees.Second and more important, isn’t an utterance propagated through the environment as energy?Isn’t a letter or a book a physical environmental sign? Any kind of communication exploitssome environmental ‘channel’ and some physical outcome of the act; any communication is“through the environment”! This cannot be the right distinction.The real difference is that in Stigmergic communication we do not have specializedcommunicative actions, specialized messages (that unambiguously would be “direct”messages because would be just messages); we just have practical behaviors (like nestbuilding actions) and objects, that are also endowed with communicative functions. In thissense communication is not “direct” (special communicative acts or objects) and is “via theenvironment” (i.e. via actions aimed at a physical and practical transformation of theenvironment).From our perspective, Stigmergy is communication via long term traces, physical practicaloutcomes, useful environment modifications, not mere signals. Moreover, in insects (andsome simple artificial agents) Stigmergy is a functional form of behavioral communication,where the communicative end cannot be represented in the agent’s mind (intention) but it is afunctional effect selected by evolution or built in by a designer. This means that Stigmergy isa sort of “innate” and trace-based Behavioral Communication (see later) and, in our point ofview, very effective especially for sub-cognitive agents.Stigmergy is just a sub-case of BIC, since in fact any BIC is based on the perception of anaction that necessarily means the perception of some “trace” of that action in the environment(for example air vibrations).We restrict Stigmergy to a special form of BIC where the addressee does not perceive thebehavior (during its performance) but perceives other post-hoc traces and outcomes of it. To 20 R.Beckers, O.E. Holland, J.L Deneuborg, op. cit., pag. 181"From Local to Global tasks Stigmergy andCollective Robotics", R. Beckers, O.E. Holland, J.L. Deneuborg, Artificial Life IV, Rodney A. Brooks andPattie Maes (eds.), 1996, pag.181.21 For example “…Coordination of the agents' movements is achieved through stigmergy. The principle,initially developped for the description of termite building behaviour, allows indirect communicationbetween agents through sensing and modification of the local environment which determines the agents'behaviour” R.Beckers, O.E. Holland, J.L Deneuborg, op. cit., pag. 18122 Holland and Beckers (xxx) generalised this notion and distinguish between two forms:

a) cue-based stigmergy, where the agents coordinate with each other basically observing their behaviorsand patterns (like in birds’ flocking).b) sign-based stigmergy, where the agents create and leave special physical signs in the environment (likein termites).Here clearly the authors while attempting to generalize the stigmergic (trace-based) notion arrive ratherclose to the BIC concept. But this is the wrong way: from insect stigmergy to broader low level forms ofbehavioral communication. What we need is a principled and general theory of behavioralcommunication where stigmergy represents a specific case.

be true, perceiving behavior is always perceiving traces and environmental modifications dueto it; the distinction is just a matter of perception time and of duration of the trace. The onlydifference is that when we refer to communicating via “traces” we have in mind more longterm traces that persist also when the authors is no longer there: the “receiver” “observes” thetrace while could not observe the authors performing the action. But in order a trace-basedcommunication be Stigmergy it is necessary that the perceived “object” be also a practical oneand the originating action be also for practical purposes (like nest building). Stigmergy is notonly for insect, birds, or non-cognitive agents. There are very close examples also in humanbehavior. In animal Stigmergy is non intentional, but intentional forms of it are possible.Consider a sergeant that – while crossing a mined area –says to his soldiers: “walk on myprints!”. From that very moment any print is a mere consequence of a step, plus a stigmergic(descriptive “here I put my foot” and prescriptive “put your foot here!”) message to thefollowers. Consider also the double function of guard-rails: on the one side they physicallyprevent car from invading the other gangway and physically constrain their way, on the otherside they in fact also communicate that “it is forbidden to go there” and also normativelyprevent that behavior. This is what law theorist call xxxxx materialization of the norm: thenorm its “hardwired” since either the external or the internal condition for the agent’s doingdifferently are excluded; there is no possible decision of violating. In fact the communicativefunction is a parasitic effects of the practical act and of its long-term physical products. It isjust BIC and specifically Stigmergy.

3. Behavioral Communication step by step in cognitive agents: From non-communication to meta-communication via behavioral communication

Let us now characterize how BIC emerges from simple acts, how it evolves in stronger forms,and how it dissolves while loosing its original practical character to become a pure symbol.

First step:

i) Mere behaviour: An agent X is acting in a desert word; no other agent or intelligentcreature is there, nobody observes, understands or ascribes any meaning to this behaviour β. 23

Neither "signification" (see later) nor a fortiori "communication" are there. Next step:

ii) Signification: An agent X is acting by its own in a word but there is another agent Yobserving it which ascribes some meaning to this behavior β. There is in this case"signification" (X's behaviour is a sign of something, has some meaning for Y, informs it ofsomething p: for example p can be = "X is moving", "X is eating", "X is going there".), butthere is no necessarily "communication".'βx means to Y that p'

1) (Causes (Observes Y β) (Knows/Bel Y p))

As we know, to have "communication" (next step) we need at least a "source" which "sends"a "message" to Y. To have a "message" and a "source" some teleonomic, goal-oriented act ofsending is needed. In other words the disclosure of the information from X must not beaccidental or simply causal from its own side; it must be aimed at, functional to, in order to Yunderstand that p. (1) is no more sufficient, something like (2) is needed: 23 Although sometimes we use BIC and stigmergic messages with ourselves.

2) (Goal X (Causes (Observes Y β) (Knows/Bel Y p)))

or even better

(Causes (Goal X (Causes (Observes Y β) (Knows/Bel Y p))) (Does X β))that should correspond to

(In Order to (Does X β) (Causes (Observes Y β) (Knows/Bel Y p)))

Let us now consider more carefully that scenario (ii) (signification without and beforecommunication) from the cognitive point of view, and discuss different cases.

(iia) X does not know

Consider our example of a thief entering a stock during the night while ignoring that there is aguard which is following him on a TV screen.If we characterise the minds of X and Y, we have that:

Y's mind(Know Y p) - where p is "X is coming inside the building"? (Goal Y (Know Y p))

X's mind(Not (Know X (Know Y p)))(Goal X (Not (Know Y p)))

In this example, not only X is not communicating to Y, but he even has a anti-communicativegoal (Castelfranchi, Poggi, xx): the goal that if some Y is there, s/he does not know that p.This is why he is entering in a secret, cautious, and silent way.In our example this goal of X has failed, but X does not know about this (there is no "goalfrustration"). But this is a special, strong sub-type of our (ii). We may have simpler cases.Consider a soccer player which is playing alone in a park. He is kicking the ball and followingit; he ignores that somebody is observing him.

(iib) X’s awareness: “weak BIC”

Consider now that X knows about being monitored by a guard, but that he does not care at allof it, because he knows that the guard cannot do anything at all: telephones and alarms are outof order, and doors and corridors between the two parts of the building are blocked. AlthoughX knows that Y knows that p, and Y knows that p from and thanks to X's behaviour, X is not"communicating" to Y its presence. Neither he wants Y knowing this, nor he is acting (also)in order Y knows that p. He is simply doing his own job for his own reasons, this activity isperceived by Y and has a meaning for Y, and X knows this. That is all.

Y's mind(Know Y p) - where p is "X is coming inside the building"? (Goal Y (Know Y p))

X's mind(Know X (Know Y p))(Not (Goal X (Know Y p)))) & (Not (Goal X (Not (Know Y p))))

Y's understanding is here among the known but unintended effects of X's behaviour. Althoughperhaps being an 'anticipated result' of the action it is not intended by the agent.We assumed a neutral or indifferent unintended effect, but this is not necessary. We may havetwo cases.

- negative unintended effects.The agents expects these effects but he would like to not having them. Nonetheless hedecided to perform the action (the risk is not too high). In our example, my be the pilfererworries in fact about being detected by the guard (who knows if he would be able to dosomething?).

- positive but not motivating uninteded effects.The agent expects these effects and they are also positive for him, he like them; i.e. theysatisfy some goal of the agent. Nevertheless we claim that they are not intended, or moreprecisely their are not the true motivating intention of the action. They are just additional,non necessary (the action would have been chosen also independently from those effects. Icall these expected positive non-motivating effects: side-intentions.

In our example the pilferer might be happy and laughing about the guard being alerted andpowerless and angry.Let me call this "weak behavioural communication"

In the weakest forms of BIC the agent is not strictly acting in order the other understandswhat he is doing (this is not one of his motivating intentions) but he simply is aware of thispossible result of his behavior and lets this happen either because he likes it or because hecannot avoid it or doesn’t care of it. 24 He could avoid this -either by doing some additionalhiding action, or by renouncing to do in order do not been observed - but he decide of doingβ.This is a very weak form because there is no real intention, the “communicative” result is justa know (negative, indifferent or positive) side effect of the action.

To be really or more strongly BIC the communicative effect should be desired by the agent; itshould be at least a side-intention of him.This shows that is very relevant the theory of action results and of different form of intentions.

3.1 The notion of intended and motivating results

To have a theory of intentional BIC we need a theory of ‘motivating’ i.e. really intendedresults of an action.

FIGURA RESULTS 24 Notice that in order a predicted result becoming in some way intentional, X must have the possibility toavoid, prevent it. Then, when X decides of letting it happening (castelfr 98) this becomes some sort of"passive intention" of X. If this is a mere effect of nature or of other agents in the world (X is simply lettingthem to act), we have passive intentions; if these are consequences of X's own actions (i.e. the only way hehad to prevent them would be renouncing to perform that action) we have side intentions of the action.

We should distinguish at least between expected or anticipated effects/results of the actionand non-anticipated ones. The results can be either good (corresponding to some desire of theagent) or bad (against the agent’s desire). The intended results (towards whom the action isoriented) are necessarily both anticipated and good (desired). However, there might be alsogood anticipated results that are not intended in strict sense: they are not motivating theaction. The agent knows that there will be (also) such additional good results but he is notacting for achieving them. He will act also in case that there were not such additional goodresult.Motivating or truly intended results are both necessary and sufficient for intending the action.

Even more sophisticated distinctions are relevant, like the one between avoidable vs.unavoidable results (and consequently between passive and side)

Known non-motivating effects of βPOSITIVE NEGATIVE

UNAVOIDABLE additional acceptedexcept by droping β pos. consequences risksand its goal (passive-intentions)________________________________________________________

AVOIDABLE side-intentions calculatedthen chosen risksbut not motivating

Ex SoccerI know that my partner sees and understands that I going to receive the ball, and he lets it tome. It is coordinating with me as it would have been messages like: "let, let it to me!!" or "It'smine!". Actually I want and have to take the ball; I do not go there in order to communicate"it is mine", however I know this and I exploit this by relying on this effects on my plan; but Iwill do the same in any way (even if my partner would not understand). Passive intentions.Sometime I let the other (the ennemy) to see but just because I cannot avoid this (acceptedrisks).

In sum, what is relevant here is that:• the expected side-effects that I let happen could be either indifferent, or good (but neither

necessary nor sufficient) or bad but not so strong to renounce to the action.• a second step in behavioral communication is when I not only know that you will see and

understand, but I also like/desire this. This is a side or passive intention, but not yet amotive, a necessary or sufficient motivating effect.

• a third step is when I chose and perform the action also in order you see and understand.This is full behavioral communication since communicating is part of my motives(necessary).

• The last step, where communication is my only motive (necessary and sufficient), is apeculiar one; the practical action can loose its practical, original ends, and be justritualized or simulated (faked) (see later).

As we said: in BIC the agent is performing a usual practical action β, but he also knows andlets or makes the other agent to observe and understand such a behavior, because this is partof his (motivating or non motivating) goals.If the understanding of the other agent was planned and is one of the motivations of theaction, we have an "active" intention (since the behavior is produced for that purpose); if theagent didn't plan for this but is positively aware of this and let the other understand, we have a"passive intention" as the basis of implicit communication.

In Passive Communication: X lets Y become aware of p, to know p, to understand p(Let me call W the proposition: "Y becomes aware of p, knows p, understands p")

In Active Communication: X performs an action β in order that W (Y becomes aware of p,knows p, understands p)Thanks to passive intentions (awareness) a lot of behaviors in social environment become"communication" among the agents.

3.2 The full form of BIC

Thus, after the case of expected side-effects that are either indifferent or good (but neithernecessary nor sufficient) or bad but not so strong to renounce to the action; and after the casewhen the agent not only knows that the other will see and understands, but he alsolikes/desires this (a side or passive intention) but this is not yet a motive, a necessary orsufficient motivating effect, we have another step:

(iii) Strong or True BIC

X choses and performs the action also in order the other sees and understands. This is fullbehavioral communication since communicating is part of my motives (necessary).Consider now that the pilferer not only knows but intends that the guard sees it. There hadbeen alternatives, but that behavior has been chosen also for this expected result. The realrole of this pilferer in the plan is that of diverting the guard's attention or behaviour(inducing the guard to follow him abandoning his place). Now, really X's behaviour isfully communicative. In fact X on purpose performs it also in order the guard knows that p,he predicts and intends that his behaviour be a sign and has a meaning to Y and performs itin order to covey this meaning to Y. The behaviour is both a practical action for pragmaticends (braking the door and entering, etc.) and a "message". The fact that Y knows that p is"motivating" the action.We will call this "strong behavioural communication", the pragmatic behaviour whichmaintains its motivation and functionality acquires an additional purpose: to let/make theother know/understand that p.

The important point for fully understanding BIC (and the difference with the following meta-BIC) is that: we have here a fully intentional communication act, but without the aim(intention) that the other understands that X intends to communicate (by this act).(Castelfranchi, 2005-Ivrea)‘Intention of communicating’ and ‘communicating (this) intention’ are not one and the samething. Given the well-consolidated (and fundamental) Grice-inspired view of linguisticcommunication - that frequently is generalized to the notion of ‘communication’ itself - thesetwo different things are usually mixed up, and it is difficult to disentangle them; but they are

clearly different both at the logical and at the practical level.With a BIC message X intends that the other recognizes her action, and perhaps thatrecognizes and understands her practical intention motivating the action (eating; having thedoor closed; knowing what time is it; etc.), but X has not necessarily (at this communicativestage) the intention that the other realizes her higher-intention that Y understand this, that isher intention to communicate something to Y through that practical action: I want that Yunderstands that I intend to go, but not that I intend that he understands that I intend to go.(see also Note 6).

The last step, where communication is my only motive (necessary and sufficient), is a peculiarone; the practical action can loose its practical, original ends, and be just ritualized orsimulated (faked).

(iv) Beyond BIC: For communication only

Let us assume now that the behaviour β is intended and performed by X only for itsmeaning, only for making Y to know that p. There are no other pragmatic purposes. Thereare two sub-cases:- The behavior has just a communicative purpose, but there continue to be the originaleffects (that now paradoxically become "side" effects). (pre-ritualization of β; or fakedsimulation). 25

- Communication is the necessary and sufficient condition for executing β which alsolooses its practical aspect while not changing its appearance. The practical action is justsimulated (faked).

What Simulation isThe fact that the β has only a communicative goal means that it is a fake action! In fact, ifβ has no other goals apart from communicating to Y, Y will be deceived. The information µhe will derive from observing β (apparence), will be false (and β is aimed at this result). Xbehavior is just mimecring the original practical behavior; its model and denotation. It is just abluff.This is true in all cases where Y should not understand that β is only aimed at communicatingsomething to him. If there is an explicit meta-communication (“this is just apparence, justfiction”) or the contextual conditions and rules make this clear (like in a theater) we have nottrue deception but just acting. In bot case the behavior is a pure (iconic) symbol.

RitualizationAt this point behavior is also ready for ritualization, especially if is not for deception but forexplicit communication. Ritualization means that β can loose all its features that are no longeruseful (while were pertinent for its pragmatic function) while preserving or emphasizing thosefeatures that are pertinent for its perception, recognition and signification. The incompletenessand simplification of the act can be apparent, since it is not aimed at deceiving.It can be (and usually is) used out of its original context (decontestualization), thus is aninappropriate situation and way, since in fact it has just to ‘remind’ us the original act.After ritualization the behavior β will be a specialized communicative act, a specialized andartificial symbol (generated by learning and conventions) meaning “I’m doing β”. We are at 25 A nice case is a ‘simulated simulation’ (Castelfranchi e Poggi, 200) where X is doing something true but inorder Y believes that it is a faked, false action. For example X has a true orgasm but wants that Y believesthat she is simulating an orgasm.

step (iie).

3.3 Recapitulating BIC steps

Let us recapitulate those steps from emergence to dissolution:Let us call W the proposition: "Y becomes aware of p, knows that p, understands p" (p is therecognition of β and its interpretation)26

0. W is false, nothing like this happens;

1. W happens but X ignores that;

2. X knows that W, but W is not a goal of X (it is indifferent to X or even a negative sideeffect); X just lets W to happen

3. X knows that W, and W is a goal of X. It is not necessary for X’s action, is not (part ofX's intention) but is a positive side effect (a side-intention) of β.

4. X knows that W, and W is a goal of X, and is necessary but not sufficient: i.e. Xexecuted his action also for W (but not only for W).

5. X knows and intends W, and this is the necessary and sufficient condition for executingβ which continues to have its practical aspect. So the behavior has just a communicativepurpose, but there continue to be the original effects (that now paradoxically become"side" effects). (pre-ritualization of β; or meta-bluff, simulated simulation).

6. X knows and intends W, and this is the necessary and sufficient condition for executingβ which looses (in that case) its practical aspect but does not change its appearance. Thepractical action is just simulated (faked), β is just for deception.

7. Next step is full-ritualization of β (Castelfranchi e Poggi. Per una teoria dell’inganno):when the behaviour has regularly and only a communicative purpose, and no longer theoriginal use. Now that behaviour is specialized only for communication; it can beexecuted out of its original and suitable context; it changes its morphology or some of itspragmatic features and becomes a symbolic “gesture”; it is no longer a true BIC.

Some of those steps (like ritualization) require in fact also meta-comunication, and theemergence of meta-BIC. Steps are not linear; it is more a geneaological tree.

3.4 The Transition from Sign to Message-Sign: Convenzionalization and Symbolization

One should not mixe up the ‘conventional’ caracter of a sign 27 with it possible‘arbitrarieness’. A sign can be ‘conventional’ not being ‘arbitrary’ (and vice versa). The

26 Consider that β may be an action that changes the world and leaves traces. So Y may see those traces andunderstand p without observing the running behavior.27 If a sign is ‘conventional’ usually it is for communication; this is the reason for creating or evolving asignification ‘convention’.

‘arbitrarieness’ refers to the more or less ‘natural’ link betwee significant (vehicle) andmeaning, while ‘conventionality’ is about its more or less consolidated and possiblyinterpersonal use (the agreement, convergence between Source and Addresssee).Moreover, it exists a very important transition (that doesn’t have a specific intuitive names!)from the ‘sign’ as vehicle of a simple unilateral process of signification/interpretation and thevehicle of a communication process, when ther sign becomes a ‘message’. Not all signs areused in/for communication thus a sub-species name would be needed, and also a notion forthis process. 28

Te evolution from a mere ‘sign’ to a ‘sign-message’ (a sign used, destinated, or with thefunction of a message) 29 on the contrary is fundamental both in intentional terms, in learningterms and in functional terms.This transition is usually (and necessarily in the case of the establishment of a commonsymbolic communication system between X and Y) joined with the first transition from non-conventional to conventional-routinary.Let us carefully analyze those relations.In mere signification (disconnected from a true ‘language’), Y takes n for m, he interprets n asm, i.e. n signifies to Y m. This happens through an inferential process based upon correlationand coincidence (idexes) or similarity (icons) (or also more complex, like in medicaldiagnosis).This is the original associative-interpretative process; but when Y uses it habitually (foralways signifying him m), litle by litle the link becomes just automatic, routinary: a directassociation; independently of its origin, i.e. as if it was conventional and not naturallyderived.At the beginning there is the mediation of a ‘natural’ process of interpretation ofevent/feature n as m, an inference or association between two events (ex. fire – smoke), or themediation of a specific learning process;

but at the end there is just a direct automatic link: the original and grounding mediation isjumped.

Consider for example the difference between the activity of the police that reconstructs mypresence and activity in a given environment on ther basis of my fingerprints (sign, index,reasoning), compared with the case of fingerprints preassociated and used as an automaticrecognition sign for the access in a protected building.In my view, this is what Peirce catches and intends while introducing the distinction between‘icons’, ‘indexes’ vs. ‘symbols’. This is the process of routinization, conventionalization(especially when socially shared) and in a sense the creation of a ‘code’ of symbolicautomatic correspondences between vehicles and meanings.

28 Is this what Peirce means by “symbol”? It seems that it refers to the first process: the conventionalization(independent on more or less iconic and natural origin of the sign). TROVARE29 On the distinction between ‘use’, ‘destination’ ‘ function’ see Conte and Castelfranchi, 1995, Ch.6.

COGNITIVEINTERPRETATION meaningsign

AUITOMATICASSOCIATION meaningsign

This process of routinization, automatization, shortcut, and convenzionalization is not thesame of the process of passign from a simple sign to a sign-message, when for example Xsimply decides to use n in order to spontaneously signify to Y m. X can definitely exploit anew process of Y, an inference, a reasoning, a diagnostic index, not something routinary andwell established between them, or something routinary but not as a message, for example therecognition of an action. This is typically the case of intentional BIC on new behaviors. Xmight change everytime the way to make Y to understand (from her behavior) m.Obviuosly enough if there is an interest in communicating in a systematic way and habitually,it is convenient for both to converge on the signs that have been successful, and thus tostabilize on them and repeatedly use them. In such a way the signs also convenzionalize andbecomes association-routines (for X and Y).Using them becomes a tacit convention between them (a shared code) although notnecessarily those signs are of an arbitrary, ‘innatural’ origin, but icons or indexes.

In conclusion, signs become routinary and automatic already in their systematic signification(also non social) use, while become conventional in their consolidated use for communication.The second process implies the first one (but not viceversa) 30

In conclusion, there are:non-routinazed & non arbitrary signs:

for example: “this glass looks spotted, it must be already used”non-routinazed & arbitrary signs

for example: “in order to recognize my glass (in the party) let me put it here”routinazed & non-arbitrary signs

for example: “John is here! I recognize how Bob is barking”routinazed & arbitrary signs

for example: “when I put something here it means that I have to bring it to the office”

Arbitrary signs can only emerge (a part from Communication + learning) in a sort of‘convention’ with ourselves, some idiosyncratic established ‘rule’. Although conceptuallydifferent, one might claim that arbitrary signs (except if innate) necessarily are conventional(or conditioned from outside).

4. From BIC to meta and explicit BIC

4.1 Meta-BIC

In meta-BIC, there is a meta-communication, typical of higher forms of communication likelanguage. BIC meta-message is as follows: "this is communication, this is a message not justbehavior; it is aimed at informing you".Frequently BIC has such a high level (Grice’s way) nature. 31 This is the typical Grice’s viewand definition of communication, especially fitting human linguistic communicatioin (but not

30 Except if we want to consider communication to my self and creation of a convention with my self thesystematic use of a sign: for example smoke for fire in habitual fire surveillance and detection.31 This might induce to consider it as the full and more complete form of BIC (that however continues to bewithout specialized messages!).

only).32 Gricean communication is a higher form that is much richer, having a well identifiedadditional super-stratus of reflective beliefs and goals. 33

For example, the act of giving or handing is not only a practical one, but is a meta-communicative act where X intends that Y understands that she is putting something closer toY in order Y (understanding that she intends so) takes it. But also in teachnig by example, orin team coordination a lot of actions are in fact strongly Grice’s way communicative. Forexample in soccer a "pass" or a "lob".

Meta-BIC is a more “explicit” form of communication, since the intention to communicate iscommunicated, or manifest. However, I do not want to consider BIC even at its meta-levelreally “explicit” communication, since the fact that it is a communication act remainsarguable and doubtful since β remains a normal and ambiguous behavior. When the act is justan arbitrary and conventional symbol, or a clearly ritualised act, the source cannot deny that itis communicating by emitting such a behavior.

4.2 Cooperation in Behavioral Communication

It is important to track the evolution of Cooperative relations along this path fromsignification to meta-communication and conventional signs via BIC.Clearly there is no cooperation at all in mere signification, where unilatyerally and for hisown purposes or functions Y observe (perceive) X’s behavior (traces) and recognize it, reactsto it in a specific way or interprets it as a meaningfull sign. Y’s goal is to perceive and readX’s behavior, while X has no goal about this.Not necessarily while starting communication cooperation entra in scena. In fact we consideralready Communication when X ‘sends’ on purpose (either intentionally or functionally) thesign of her behavior to Y. At this point Y’s reaction or interpretation is a goal of X, but it isnot necessarily cooperative with the goal of Y’s of acquaring knowledge (from X’s behavior).In fact the sign can be on purpose deceptive. In a sense X is cooperating with the goal of Y ofperceiving/observing her behavior (or the relevant world) or better is exploiting it, profiting ofsuch a goal, but in order to damage Y’s super-goal (motivating the former) of knowing thetruth. So, for example we cannot say that the predator deceiving the prey by a camufflage(simulating a rock) or the prey deceiving the predator by hiddening itself thanks to a mimecry,are coperating with the other. This holds both for specialized ethological sygnals and for BIC.

When BIC is non for conflictual ends it starts in fact to be cooperative, but in a very specificway. X and Y have a common goal but they do not have a shared plan and do not adopt thegoal one of the other, nor adhere to the other’s expectation.In BIC there are two goals/functions meeting each other:

i) the communicator's goal: X’s behavior has the goal or function that Y"understands", recognizes, and comes to believe that p (and this holds from step (iii) §2.3)34

32 See aldso Posner, along the same line: Communication is “characterized as a primary action performed bymeans of a secondary action aimed at making the addressee understand the sender’s intention to perform theprimary action.” p. 252. “The essential feature of communication lies in the sender producing a secondarysign process” p.250. Posner will probably accept as “communication” only our meta-BIC.33 As we will see we try to derive, to let emerge the reflective, gricean message and level from specialized,conventional, recognizable messages precisely by re-applying BIC to them (see x.x).

34 This definitioon is in cognitive terms, but this is in contrast with ther functional extra-mental version of BIC.

ii) the interpreter's goal: Y has the goal/function of interpreting X's behavior in orderto give it a meaning (and this holds from step (ii) § 2.3). 35

These are the two faces of “sign” when sign is for/in communication (see also Posner, 1993CIT). However, those goals in BIC can be independent from one the other.It is even possible that they (X and Y) both succeed, but their "cooperation" is accidental.They do not really have a "common goal" (Conte and Castelfr 94) which presupposes mutualknowledge. In this case either both ignore the goal of the other, or just X is aware of Y's goaland is exploiting it (weak delegation/reliance).Since, in step (ii - Signification) X does not know that Y wants to understand her behavior;while in step (iii) Y does not know that X is communicating to him through it behavior a.Thus Y has not the goal of: "understanding what X means by β"; that is the real common goal36 of higher form of communication (like linguistic communication) on which usually X and Ycooperate for a successful communication (Meijers, ). The communication has beensuccessful only when X and Y achieve this goal: that Y understands precisely what X intendsto communicate to him (Castelfr e Parisi, 198x).

In animal communication there are two cooperating functions: there is the sending activity ofthe source which is aimed at informing the other; and there is the attentive, perceptual,cognitive activity of the receiver, that is aimed at being informed. For example, the male ofthe firefly send the light message in order the female "understands 'there is a male here'" andthe female actively perceive and process that message understanding it (and in order tounderstand it).

The situation changes if/when Y understands that X intends Gx (to communicate); now eitherhe also adheres to X’s goal, cooperates with him (although X can ignore this) or heconsciously exploits this goal of X (of being observable/ed) for his own goal Gy (to observe).Meta-communication exploits precisely this new – recursive - BIC signification of X’sbehavior: the fact that Y realizes that X intends to signify (i.e. is communicating). X thus usesthis new sign effect as a goal.Very different is in the meta-BIC relation where Y not only understands that X intends Gx (tocommuncate) but also understands that X wants that he understands this, and thus also adoptsthis goal and adhere to this implicit request.

Thus in meta-BIC Y knows that X is communicating. Therefore he has a special form of goal(b) (by adhesion (Conte e Castelfranchi, 1995) and/or convergence); the goal of catchingwhat X is communicating:

b’) goal of Y to understand what X's intends to communicate, to understand which is themeaning in X’s mind.

The agents in such away arrive to cooperate in strict sense (like in linguistic exchange), andthe two goals (a) and (b) become complementary, convergent and functional to each other;that is X and Y have the same goal and they know the goal of each other.

(Bel y (Intend x (Understand y s))) ==> (Intend y (Understand y s))

35 As we said in the case of the enemy, or predator (signification) there is only the unilateral second goal: xdoes not have the goal or has the opposite one (that y does not knows, understands, etc.)36 “Delegated” by the source which counts on the addressee for this, and held by the adresse also by“adoption” of the source’s goal or – even better – by “adhesion” (castelfranchi, 2000)

4.3 Explicit Communication

Explicit communication is when β is a specialized action: i.e. its intended and relevant, butalso conventional and usual results are precisely (i): to make the other understand that p, to beused as a message to Y, to make the other understrand that it is a message-sign. These resultsmake β a recipe in memory.Thus when X performs β the other understands that X has the cited intention (i). Y is entitledto understand/interpret this; X cannot deny this. No more ambiguity. When the relationshipbetween sign and meaning is artificial, arbitrary (Saussure) then it is absolutely clear thecommunicative nature and intention of the act (and unambiguously meta-communicates it);while conventional but non artificial communication remains ambiguous.Given the sending of behavior/object specialised for communication when one uses it thisbehaviorally means: “I’m communicating, I intend to communicate”; that is Grice’s meta-message! Communication is explicit when I cannot deny that I’m communicating, since Inecessarily and undeniably meta-communicate precisely because the emitted behavior isspecialised for communication and while using it I implicitly declare my intention tocommunicateThe act is qualified and recognizable as a communication act.Now, since also a communication act is a behavior it follows that as a behavior it can be usedat the BIC level, i.e., it will communicate about X and itself: “I’m communicating”. Thusthere is a potential meta-communication (via BIC) in any communication act. On such a basiswe predict/derive Grice’s view on linguistic act. (see ection xxx).

More precisely, BIC use of a symbolic act is only the origin and the basis of the meta-message. Precisely because of the conventional nature of the message, now also the meta-communication (“I’m communicating” “I intend to communicate”) is no longer behavioralcommunication as we defined it, is no longer inferred. It is part of the “official” establishedmeaning of the speech act (Grice, ).As we said, Posner (1993) adopts a gricean definition of communication: “The essentialfeature of communication lies in the sender producing a secondary sign process (i.e.,performing a secondary action at a secondary level)” (p.250). This formulation rises a ratherimportant issue:

It is necessary an additional, special act, sign for communicating the secondary intention?The secondary meta-message?Have I, for example, to wink, or to change voice, or to add a special gesture?

I deny this. Just one and the same act or gesture can be simple BIC or meta-BIC; obviously ifthere is no special sign of its meta-communicative character the interpreter can never be sureof the sender intentions.

Behavioral explicit communication (BEC); a mixed caseThere are behaviors in between BIC and a specialized non-verbal sign. For example, the actof “giving a wave” with the hand or a hanky is clearly a symbolic, mainly cultural andconventional one (its “bye-bye” meaning), but its practical purpose to make X more visibleand attract Y’s attention holds and supports the symbolic goal. Both components arenecessary but one is just from practical action and preserves its practical effect.Analogously, tending the open hand for asking and receiving something is both a practicalaction for taking and a specialized sign for request.

We assume that the evolution of the symbolic gestures of “pointing” and of “grasping” insmall children follows the same rooth, and is case of Behavioral explicit communication(BEC). (see section on the origin of language).

4.4 Functional cultural BIC (back to Watzlawich)

Let’s spend a few words for explaining that as there is merely ‘functional’ behavioralcommunication at the biological level (and this is also the origin of several specializedcommunication signals in animals), analogously it is possible to have non-intentional butequally teleonomic forms of communication in humans due to social functions andunconscious use based on learning and reinforcement.Some of the ‘messages’ we send to the others with our actions or objects are not intended.They only have a communicative function due to cultural practices, and the long selection ofsocial functions (Castelfranchi, 2003).For example, our dresses send a lot of messages. Not only they have a lot of meanings and aresigns from the point of view of an observer (signification), they are used and reproduced alsobecause of their sign effect. We send a lot of messages by our dresses, and part of them aresurely intentional; but we are not aware or we do not necessarily intend all those messages,like: “I’m a sophisticated persons”, “I’m a sport lover”, “I’m from the higher class”, “I do notcare of stupid things like elegance”, “I’m a professor”, etc.In this perspective, if we accept that there are a lot of signs in our social behavior that are notmessages at the intentional level but could be messages at the functional level, havingacquired the social function of signifying to the others - as I said - Watzlawich results to bepractically right! A large part of human acts are communication.But it remains that he is right only accidentally: he has not the right conceptual instrumentsfor supporting his claim.

5. Ubiquitous BIC: Uses and Functions of BIC (and metaBIC) in Social Life

There are so many and ubiquitous uses of BIC (or of its mutually aware form) in socialinteraction and structure that it is impossible to systematize them. However, let us provide theexemplification of some of the main uses and functions of it just to let the reader perceivehow intrusive it is and how relevant (and how it should not be mixed up with NVC). I willorganize them around the more salient meaning µ of the act/message. I will also underlinyngsome stigmergy (eiher intentional or functional) in humans, i.e., communication via “traces”,long term observable modifications of the physical environment as outcomes of practicalactions.

5.1 µ = “I’m able” or “I’m willing”

The most frequent message sent by a normal behavior is very obvious (inferentially verysimple, given an intentional stance in the addressee) but incredibly relevant:

(as you can see/constatare) I’m able to do, and/or I’m willing to do; since I actually did it(I’m doing it) and on purpose.

There are several different uses of this crucial BIC message.

Skills demonstration in learning, examines, and testsWhen Y is teaching something to X via examples and observes X’s behavior or product tosee whether X has learned or not, then X’s performance is not only aimed at producing agiven practical result but is (also or mainly) aimed at showing the acquired abilities to Y.More in general, doing the same action a of a model, imitating, is the base for a possible tacitBIC message of X, for the possible use of the action a as a message to Y: “I’m doing thesame”. But for this specific additional conditions are needed:

i) X performs a (imitates Y)ii) Y observes and recognizes (i), and forms the meaning m “X is doing a/like me”iii) X knows that (ii)iv) X intends that (ii)v) X performs a also in order (ii) (that is because of (iv) & (iii))

In this case a is a real (successful) message to Y.When and why should X inform Y about imitating him? Especially when Y has the goal that(i).This is true for any “examine” and “text”: the behavior or product under examination is aBIC. This is clamorous and pretty clear in the “domanda-esame” (xxxxx) (ex. “where wasNapoleon born?”) where the usual presupposition of any question (the speaker does notknows the answer and wants to learn about) is violated: the speaker already knows theanswer, he just wants to know whether the addressee knows it or not.Frequently, the student or the candidate is also communicating about (showing) her/hiswillingness (to learn or to be subject to control, or to accept examination).Notice that in teaching by example and showing (based on imitation) also the behavior of theteacher is a BIC; its message is: “look, this is how you should do”. Usually this is also joinedwith expressive faces and gestures (and with words) but this is not the message I’m focussingon.

In general, if showing, displaying, exhibiting and demonstrating 37 are intentional acts they arealways communication acts

Warnings without wordsThis is a peculiar use of exhibition of power that deserves special attention. Mafia’s “warning”, monition. The act (say: burning, biting, destroying, killing) is a true actand the harm is a very true harm, but the real aim of this behavior (burning, killing, etc.) iscommunicative. It is aimed at intimidating, terrifying via a specific meaning or threat: “I cando this again; I could do this to you; I’m powerful and ready to act; I can even do worst thanthis”. This meaning - the “promise” implicit in the practical act - is what really matter andwhat induces the addressee (that not necessarily is already the victim) to give up. Thepractical act is a show down of power and intentions; a “message” to be “understood”.The message is “if you do not learn, if you will do this again I will do even worst”.The same do nations: consider for example the repeated reactions of Sharon after terroristattacks in Israel; it is not only a revenge, it is a message: “do this again and I will do this(bombing) again”; the same holds for terrorist bombs. The military objective is irrelevant ormarginal. This is just a horrible way of communicating. Perhaps it would be bettercommunicating via words and diplomacy.Is all this “expressive - non verbal - communication”? Bombing is bombing (not particularly“expressive”), and can be unintentional (by mistake and accident), or intentional just fordestruction and/or mere revenge or material prevention, but it can (also) be a message,possibly without any different features at all.

BIC messages to myselfA che serve mostrare, esibire a se stessi?? (problema rilevante per il ‘gioco simbolico’E’ una forma di RIPROVA; e’ la “prova provata” che ho xuna sorta di maggiore evidenza, di verifica empirica, che deriva proprio dall’ “agito”: non e’pura rappresentazione autogenerata, non e’ fantasia ed immaginazione, desiderio o cose chemi racconto, non “me la racconto”; ci sono le prove dei fatti che arrivano dal mondo esterno,dalla percezione e dall’agire, nonche’ dagli altri (conferma/riconoscimento sociale).E’ come nella negazione supportata da un comportamento comprovante: non sonoomosessuale, corteggio le donne!!Da un lato e’ costruzione di una realta’ non solo interna ed illusoria; dall’altro e’ lacostruzione di una realta’ socialmente condivisa.

mostrare forz, abilita’, abilita’ etc.

(Subspecie di “dimostrativo”talvolta il messaggio e’ a se stessi: omosex corteggiare donne perconsolidare una credenza)

5.2 µ = “ I did it”, “I’doing it”

Another typical meaning of BIC is simply “I did it; I did so”. This is very relevant in severalhuman interaction where a given behavior of X is expected by Y. Consider for example, I

37 ‘Showing’ can be unintentional but intention seems ineherent in the very semantics of ‘Exibiting’ and‘Demonsrtrating’.

child showing the mother that he is eating a given food, or a psychiatric patient showing to thenurse that he is drinking his drugThis message is particularly important in the satisfaction of social commitments, expectations,and obligations. For example, when being invited at dinner our finishing the food and‘cleaning the plate’ means “I finished it, I liked it”, as the guest wish and expects. (See later).But it has other uses.For example, for serial synchronization in coordination and collaboration: if the action of Y ina common plan presupposes the previous accomplishment of the act of X, and the coordinateis based on observation, then the act of X means: “done! its your turn”.

The satisfaction of social commitments and obligations (see later)Psychiatric patient shows to the nurse that he is drinking his drugThis is another kind of demonstrative act, not basically aimed at showing power and abilities,or good disposition, but primarily intended to show that one have done the expected action.Thus the performance of β is also aimed at informing that β has been performed! This isespecially important when the expectation of X’s act is based on obligations impinging on X,and Y is monitoring X’s non-violation of his duty. Either X is respecting a prohibition, orexecuting an order, or keeping a promise.A second order meaning of the act can also be: “I’m a respectful guy; I’m obedient; I’mtrustworthy”, but this inferential meaning is reached trough the first meaning “I’m respecting,obeying, keeping promises”.

We do not usually realize the extent on which social order is created and preserved throughthose messages (Castelfranchi, 2003), for this reason I will devote an entire section (4.) toBIC based social order and normative stuff.

5.3 µ = “I’m doing” -BIC for Coordination

In coordination it is not so important the fact that I intend to do (and keep my personal orsocial commitments – which is crucial in cooperation) or the fact that I’m able and skilled, itis more relevant communicating (informing) about when, how, where I’m doing my act/partin a shared environment where we interfere with each other, so that you can coordinate withmy behavior while knowing time, location, shape, etc. . (Castelfranchi, 1998; Castelfranchi2004b).

Clearly in order to coordinated with a given event or act Ev X should perceive it or foresee itthanks to some perceptual hints, 'index' or sign. In other word usually it is an intrinsicnecessity of Coordination activity that of observing and interpreting the word in which X isacting pursuing its goals, and in particular observing Ev.In social coordination X must observe the other Agents' behaviors or traces for understandingwhat they are doing or intend to do. In sum coordination is based on observation and - moreprecisely - 'signification'.A large part of Coordination activity (and social interaction) is not simply base onObservation and Signification but is BIC-based.For example, clearly enough in mutual coordination not just Signification is needed but trueBIC. Actually, since X wants that Y coordinates his behaviors observing and understandingwhat she is doing, she is performing her action also with the goal that Y reads it, i.e. she iscommunicating to Y - through her action - what she is doing or intends to do. But let’s moresystematically examine this.

In unilateral Coordination:Non-BIC-based Unilateral: Y coordinates (adapts) his own behavior to the interferingbehavior of X, who does not perceive at all or does not care at all of those (reciprocal)interferences. In this case X’s behavior is highly significant for Y (signification) but is notcommunication since X does not know or does not care of the fact that Y is observing her andinterpreting her behavior.BIC-based Unilateral: only Y coordinates (adapts) his own behavior to the interferingbehavior of X, but X knows and intends this, although she does not want to coordinate herown action with the other. X’s behavior is communicative.In bilateral (symmetric-unilateral) Coordination: both Y and X coordinate their own actionon the actions of the others but they ignore or do not intend that the other do the same. Againthere can be no communication at all, but if one of the agent acts also in order the otherperceives and understands what s/he is doing, there is BIC.In mutual Coordination: both X and Y wants the other to coordinates with his/her ownbehavior and understands that s/he intends to coordinate with the other's behavior. As we said,mutual coordination, based on symmetric intentions and mutual awareness (shared beliefs)entails and requires BIC: each coordination act (adaptation of the behavior) is a message tothe other.Let us draw some conclusions on this point.Coordination is possible without any communication both in human and artificial societies( C a s t e l f r a n c h i , 1 9 9 8 ; s e e a l s o F r a n k l i n ,http://www.msci.memphis.edu/~franklin/coord.html38). This is an important statement againstcommon sense. However, usually coordination exploits communication.Since BIC is i) a very economic (parasitic), ii) a very spontaneous, iii) a very practice andrather effective form of communication just exploiting side effects of acts, traces, and thenatural disposition of agents to observe and interpret the behavior of the interfering others, arather important prediction follows.One can expect that agents acting and perceiving in a common world will use a lot of BIC andwill spontaneously develop it.Actually a very large part of communication for coordination in situated and embodied agentexploits reciprocal perception of behavior or of its traces and products; i.e. it is just BIC. Evenmore, (second prediction):Both in natural and in social systems a lot of specialized (conventional or evolutionary) signsderive from BIC behaviors that have been ritualized.This kind of observation-based, non-special-message-based communication should be muchmore exploited in CSCW and computer/net mediated interaction, in Multi-robot coordination,in Human-robot coordination, in MA systems (see § 6.).

5.4 “I conform; I agree”. Imitation-BIC as convention establishment and memeticagreement

Imitation (i.e. repeating the observed behavior of Y – the model) has several possible BICvalences (we already saw one of them). The condition is that Y (the model) can observe (be informed about) the imitative behavior ofX.We can consider at least the following communicative goals:

38 However, Franklin seems to miss the difference between ‘no communication’ and ‘tacit/behavioral communication’.

a) In learning-teaching via imitation. X communicate to Y “I’m trying to do like you;check it: is it correct?”

b) In convention establishment and propagation. “I use the same behavior as you, Iaccept (and spread) it as convention; I conform to it”.

c) In imitation as emulation and identification: “I’m trying to do like, you I want to beand to behave like you” you are my model, my ideal”

d) In imitation as membership: “I’m trying to do like, you I want to be and to behave likeyou; since I’m one of you; I want to be accepted by you; I accept and conform to youruses (see –b)”.

This BIC use of imitation is really important and probably the first form of memeticpropagation through communication.

X interprets the fact that Y repeats its innovation as a confirmation of its validity (goodsolution) and as an agreement about doing so. Then X will expect that Y will understandagain its behavior next time, and that Y will use again and again it, at least in the samecontext and interaction. (See section on linguistic conventions, xxx).

6. Stigmergy in humans: some nice example with deontic compents

6.1 Leaving the coat on the seatA very nice example of (stigmergic) BIC is the use of leaving some hat, coat, bag, or otherobject to mark a seat, a place is already “taken”, not free. This is a sign, deliberately used formeaning (signaling) this. It's communication. But for communicating we simply use an usualobject in its usual practice: putting a beg on a seat; leaving a coat in a place. Since peoplederive from this "trace" the fact that "this seat is already in use by somebody that ismomentary absent but will be back" and we know that, we use on purpose this as a BICmessage.Let's notice that:This diffuse social practice has become a "convention": in certain circumstance a coat, a hatetc. left on a seat means this; they are an entitled request, a prescription to "not seat here"based on a shared expectation that everybody knows this convention, understands themessage, and conforms to it. It is a nice case of convention whose natural and spontaneousorigin is transparent since the meaning of the message remains basically un-conventional,understandable even without knowing about the convention.The fact that we deal with a convention explains the normative (prescriptive) character of themessage (and the reaction in case of violation).It is remarkable the fact that this BIC message - although now conventional - did not lost itspractical original effect and the action has not changed its shape or form because of aritualization. Thus it is a (rare) case of conventionalized, ritualized behavior for BIC functionnot loosing its BIC character (remaining a fully practical and recognizable action/object).Last, notice that the obstacle to use the place is not really physical, material (one could movethe object) but is deontic and communicative.

BestsellersWhile buying a book (for your own pleasure- we in fact leave a strange trace in theenvironment: we modify the number of sold copies. This changes the position of the book inthe bestsellers list, and this is an information (intentionally sent by the publisher or by thebooksellers to the potential clients) that will be taken into account by other persons. Iscommunication, although your act just remains the practical act of buying a book, with its

practical intended effect for you. You do not intend in this case to communicate anything atall, but in fact in that market your behavior has acquired such a parasitic (exploited)communicative function

6.2 Parking marksA beautiful example of stigmergic communication with normative (prescriptive andpermissive character) is the use of painted blue or white lines on the ground for car parking,delimiting the car area and indicating their disposition: either in form of a comb, or parallel tothe side-walk. Those lines not are just signs and instructions: "you are allowed to park; andshould park in this position", but they also have a practical and physical function. They arenot merely messages; in fact, they cannot be replaced by a simple poster illustrating theprescribed car disposition in that street. They also have the practical function of visualreference point in the maneuver to be used during the act of parking. So we put in ourphysical environment - for coordinating our actions - physical object that are at the same timemessages: precisely like termites, but with an additional deontic character.

Parking without marks: Prescribing by doingIo arrivo in una via privata che non conosco, e devo parcheggiare. Se vedo le macchineparcheggiate a pettine parcheggio a pettine; se le vedo messe lungo il marciapiede parcheggioin quel modo. Interpreto le "tracce" lasciate come segno e messaggio (prescrittivo) di unaconvenzione vigente tra gli abitanti; di una tacita norma (niente cartelli, scritte, portieri). Miconformo perche' non voglio (mi aspetterei) osservazioni critiche, dispute.Si noti che nel parcheggiare in quel modo gli abitanti (ed io stesso) hanno anche uno scopopratico e lo soddisfano (minimizzare l'intralcio al passaggio o massimizzare le macchineparcheggiabili); ma al contempo hanno uno scopo comunicativo. Siamo difronte ad un caso diBIC, precisamente stigmergy; e si tratta di BIC normativo: il messaggio di una normaconvenzionale.Il messaggio puo' essere intenzionale (consapevole) o meramente funzionale (la condotta simantieme anche perche' la comunicazione funziona).

6.3 BIC Soccer: actions as BIC + Stigmergic Communication Through the BallI know that my partner see and understand that I going to receive the ball, and he lets it to me.It is coordinating with me as it would have been messages like: “I’m going/intend to take theball”,"let, let it to me!!" or "It's mine!". Actually I want and have to take the ball; I do not gothere in order to communicate "it is mine", however I know this and I exploit this by relyingon this effects on my plan; but I will do the same in any way (even if my partner would notunderstand).

COMMUNICATION

OBSERVATION

One and the same move, observed by the adversary - who perhaps gets the same meaning(“I’m going/intend to take the ball”) – is NOT comunication but only signification. As wealready stressed an escaping prey is not “communicating” to its predator/enemy its positionand move.Although this information is very relevant and informative for the enemy or predator, is notcommunication. The prey is unaware or is trying to avoid that the information arrive to theenemy. The fact that the enemy sees is not taken into account among the more or lessmotivating expected results of the action. Sometime I let the other see but just because Icannot avoid this (Passive intentions).Usually they are just ‘signs’ not ‘messages’, except when they are faked or in a fewcircumstances where I intend that my adversary understands what I’m doing

7. Some particularly nice cases

7.1 Exit and VoiceThe consumer’s behavior is certainly very significant, meaningful for the producer or for theseller. In fact, if the consumer or, more in general, the client “exits” i.e. abandon the relation,such a ‘refusal’ means something to the other part (signification but not yet communication).If the client is aware of this and also intends the other perceives its refusal we have a case ofBIC. The extreme case of this is boycotting as a form non-verbal (behavioral) protest wherewe have metaBIC and the communication aim is definitely overcoming the aim of not buying.39

Consider that the non-boycott form (and sometime even the boycott) is a form of non-intentional cooperation, since the consumer behavior with its carried information is crucial forthe other part adjusting the offer. This cooperation is accidental in the mere non-communicative exit, and usually beyond intention in the BIC forms. When the consumerpasses to Voice, to explicitly expressing its discontent and requests, one achieves a differentlevel of cooperation and of “negotiation”.

7.2 What matters is “to be there”Presence, the act of being there is frequently a message.

• Presence as testimonianza. In certain circumstance to be there is an act “testifying”solidarity end/or membership: “I’m with you; empathic with you; one of you; likeyou”.

• Presence as status symbol, of VIP membership. To be there, to appear in that,situation, compagnia or place (for example in a TV talk show or in a “salotto”) means“I’m one of them; I’m important”.

This is one of the case where Mc Luan is right 40: the content of what I say in the talkshow isirrelevant, what matters is the showing, i.e. the content of the message is the message itself,the fact of being talking there.

7.3 Social status and respectVery frequent are behaviors (and their traces) as messages (intentional or functional) of socialpositions and relations.

39 Similarly the “exit” behavior (non attention, moving away) of the tourist can be a message sended to theguide.40 McLuan’s paradoxical claim that in the mass media “the content is the message itself”.

To let the other wait in beyond the door on the waiting room before ricevere him in ownsoffice is a message stressing respective power position (“I’m the boss you are in a lowerposition”); the same meaning has letting the other standing in front of you that are sited.On the contrary to arredare and make decent if not luxurious (plants, pictures, etc.) rooms forpublic and reception – for example in services for mental health or for poor people - is amessage of respect towards them “I recognize that you have rights and dignity; we respectyou”

7.4 The puzzle of 3 wises and 3 hats

The three wises test: One of them must choose the righ color of his own hat and has only onepossible attempt; they are not allowed to turn their head or talk to each other (otherwise theywill be killed). They are only allowed to speak if they have the answer. They know that twoand only two hats have the same color (either black or white).For sure A is not in condition to know the color of his head. The same holds for C since heobserves one black and one white. However, the fact that C is not able to speak/answer is avery significant sign for B: it means that C sees two hats of different colors (otherwise hewould be able to answer!). On this basis B should be able to answer. But also C knows this,thus surely his silence is an impatient message to C: “Don’t you understand why I’m silent?!Don’t you understand from my silence that I cannot speak and why, and what I see!?”. C’ssilence is a communicative silence, a case of BIC, and the problem can be solved only thanksto this sign and message.

7.5 Silence as communicationIt is very well known that silence can be very eloquent. In general, doing nothing, abstainingfrom an action, is an action (when is the result of a decision or of a reactive mechanism), thusit can be – as any behavior – aimed at communicating via BIC. The meanings of silence orpassivity are innumerable, depending on the context and on the reasons for keep silence (ordoing nothing) that the addressee can ascribe to “sender”; for example, indifference “I’m notinvolved, I do not care”, or disprezzo, or sopportazione “I do not rebel”, or - as answer to a“domanda esame” – “I do not know”, etc. …..or agreement. Tacit consent and agreement areso important in human interaction and society that I will devote a section to it.

Conclusions of section

Frequently several of those meanings of the BIC coexist. Consider for example acasseur/’black block’ behavior.At the same time, their concrete act of broking a windows of a bank (mixed with hisexpressive behavior) communicates:

A B C

- to the authorities “I intend to fight against banks domination of the worlds”, “I do notrespect your norms”, “I do not fear you”;

- to the peers and himself “I do not respect norms”, “violating is possible”, “do as I do,follow me” “I keep my promises”, “I’m one of you”;

- and perhaps to the girls and himself: “look how audacious I am, I’m skilled and strong”;- to the others pacific demonstrants: “you are cowards and integrated”.

A single act brings several messages/meanings in parallel to different addressees preciselybecause do not put those messages in words (with their relatively fixed conventionalmeanings) but exploits the various subjective interpretations of the behavior by the variousobservers. 41

8. BIC basement of Social order

Let me now focus on a very special area of BIC use in society. Its privileged role in socialorder, in establishing commitments, in negotiating rules, in monitoringh correct behaviors, inenforcing laws, in letting spontaneously emerge conventions and rules of behaviors.To begeneral let me say that

if there is a Social Contract at the basement of society this Social Contract has beenestablished by BIC and is just tacitly signed and renewed.

There should be some pre-conventional ground and means at the basis of conventions andtheir establishment, either inborn or spontaneously emerging.

BIC for CoordinationIn coordination it is not so important the fact that I intend to do (and keep my personal orsocial commitments – which is crucial in cooperation) or the fact that I’m able and skilled, itis more relevant communicating (informing) about when, how, where I’m doing my act/partin a shared environment where we interfere with each other, so that you can coordinate withmy behavior while knowing time, location, shape, etc. (Castelfranchi, 1998).

• Coordination can be simply unilateral:Non-BIC-based Unilateral: Y coordinates (adapts) her own behavior to the interferingbehavior of X, who does not perceive at all or does not care at all of those (reciprocal)interferences. In this case X’s behavior is highly significant for Y (signification) but isnot communication since X does not knows or does not care of the fact that Y isobserving him and interpreting his behavior.BIC-based Unilateral: only Y coordinates (adapts) her own behavior to the interferingbehavior of X, but X knows and intends this, although he does not want to coordinate hisown action to the other. X’s behavior is communicative. 42

41 Obviously also sentences can have multiple indirect meanings, purche’ non-manifested and inferential(implicit). It is rare and limited the case of deliberated ambiguity: two litteral meanings intentionallyexploited by the speaker for different addressees (Vincent and Castelfranchi).

42 A nice example in Buskens and Royakers (2002) where they present a general use of coordinating andcooperating through signaling intentions by acting (they even define “commitment” as an act that signal anintention): “…. starting to cross the road is the commitment signaling the intention of the pedestrian to go firstwithout the consent of the car driver”. In my view, the message is not only declarative (and commissive) but alsoa request or command to have the precedence; the pedestrial relies on the driver’s tacit assent to this request, notsimply on an automatic breaking the car.

• Coordination can be bilateral:in the sense of symmetric-unilateral coordination; both Y and X coordinate their ownaction on the actions of the others but they ignore or do not intend that the other do thesame. Again there can be no communication at all, but if one of the agent acts also inorder the other perceives and understands what s/he is doing, there is BIC.

• Coordination can be mutual:Both X and Y wants the other to coordinates with his/her own behavior and understandthat s/he intends to coordinate with the other her/his own behavior. Mutual coordination,based on symmetric intentions and mutual awareness (shared beliefs – Colombetti)entails and requires BIC: each coordination act (adaptation of the behavior) is also amessage to the other.

Let us trarre some conclusions on this.Coordination is possible without any communication both in human 43 and artificial societies(Castelfranchi, 1998; see also Franklin, xx). This is an important statement against commonsense. However, usually coordination is multilaterally intentional and exploitscommunication.Since BIC is i) a very economic (parasitic), ii) a very spontaneous, iii) a very practice andrather effective form of communication just exploiting side effects of acts and the naturaldisposition of agents to observe and interpret the behavior of the interfering others, a ratherimportant prediction follows.

One can expect that agents acting and perceiving in a common world will use a lot of BICand will spontaneously develop it 44.

Actually a very large part of communication for coordination in situated and embodied agentexploit reciprocal perception of behavior or of its traces and product; i.e. it is just BIC. Evenmore, (second prediction):

both in natural and in social systems a lot of specialized (conventional or evolutionary)signs derive from BIC behavior that have been ritualized.

This kind of observation-based, non-special-message-based communication should be muchmore exploited in CSCW45 and computer/net mediated interaction (see section xx ??), in Multirobot coordination, in Human-robot coordination, in MA systems.

3.2 µ = “ I did it”

The satisfaction of social commitments and obligationsThis is another kind of demonstrative act, not basically aimed at showing power andabilities, or good disposition, but primarily intended to show that one have done theexpected action. Thus the performance of β is also aimed at informing that β has beenperformed! This is especially important when the expectation of X’s act is based on

43 This is pretty clear (although not explicitely captured) for example in Heath and Luff (1992), where also theobligation-fulfilment tacit messages play an impressive role.44 A very nice example of accidentally emergent cooperation and non-explicitely-negotiated tregua is that citedby Axelrod (84) and Gambetta (89). The “vivi e lascia vivere” cooperation among ennemy soldiers in the trinceeduring the first mondial war: coinciding bilateral stopping firing during the rancio interpreted and used as asegnal of tregua.45 See Note 2, and our criticisms to CSCW systems in Castelfranch and Falcone (199?) especially section 2“From saying to doing in action workflow”.

obligations impinging on X, and Y is monitoring X’s non-violation of his duty. Either Xis respecting a prohibition, or executing an order, or keeping a promise.A second order meaning of the act can also be: “I’m a respectful guy; I’m obedient; I’mtrustworthy”, but this inferential meaning is reached trough the first meaning “I’mrespecting, obeying, keeping promises”.

At the interpersonal level:Consider an exchange based on a contract (formal or informal), and the execution of thiscontract: X gives to Y the bene w, and Y gives X some money. The aim of those act istwofold: on the one side what matter is the material result: the transfer of a given resourcefrom X to Y and from Y to X. On the other side, those action are messages aimed at declaringto the other “I did it: I satisfied my commitment/obligation towards you”. More precisely: by“giving” or porgendo w to Y, X is meaning “I give you what I promised and my obligation isextinguished”; by taking, receiving w or money Y tacitly declares: “I’m receiving what I wasexpecting; I’m satisfied”. In face to face interaction also gazes are important to control thatyou realize that I give you w/money, and to see whether you see that I saw it. But gazeinteraction is not necessary. If I expect that you do something for me, when you leave it onmy desktop this also means “I did it, here it is”.In general to satisfy a commitment means that the other believes that the commitment hasbeen satisfied. If we just do the act but the other does not realizes this and continue to pretendand wait for it, there is something wrong and incomplete. Full social commitment satisfactionrequires the other’s acknowledgment.Thus the goal of the act is double: to realize p in order the other knows that p has beenrealized. Usually the achievement of the secondary goal is not due to special additional(verbal) messages but to their observable performance of the very act or of its results. Thusany act of commitment satisfaction is a potential intentional BIC. 46

At the collective level. When I respect a norm I pay some costs for the commons andimmediately I move from my mental attitude of norm addressee (which recognized andacknowledge the norm and its authority, and decided to conform to it) while adopting themental set of the norm issuer and controller (Conte and Castelfranchi, 1995): I wants theothers to respect the norm, pay their own costs and contribution to the commons. While doingso I’m reissuing the norm, prescribing a behavior to the others and checking their behavior(expectation). In the mean time I want the others (and the authority) - that might sanction meby punishment or bad reputation or might simply violate the norm - know that I respected thenorm. (Castelfranchi, 2003). Thus my act is aimed also at informing them which are supposedto monitor my behavior and have expectation (prescriptions) about it.Any act of respecting social norms has the objective “function” of communicating to theauthority and to the others such a respect, and through this has the function of maintaining,reinforcing and diffusing the norm thus reproducing the social normative order (this is why itis a “function” in strict sense; Castelfranchi, 2001). Frequently this BIC message is also anintention of the performer.

46 The communicative valence of an obligation-fulfilling act is pretty clear – although not explicitly stated – forexample in Boella and Lesmo where they list different reasons for violating obligations and say “.. In particularthis may happens if some of the actions do ensure that the normative agent become aware of the fulfillment sothat she will probably apply the sanction anyway ….. There is some plan which does not fulfill the obligationbut which induces the normative agent to believe otherwise. (p.120). Noticed that in the last case we havedeceptive communication. As we claimed when dedceptive the BIC is only and entirely communication, itspractical valence is absent, faked.

To be precise the meaning is twofold: on the one side “I obey, you have not to sanction me”;on the other side “Do as I do, norms must be respected”.As an example of this behavioral message let me propose Socrates’ drinking the poisonalthough his friends and fellows prepared the evasione and incoraggiavano him to escape.Socrates wants to drink in order to teach us and his fellows that norms (although iniquitous)must be respected: the content of the message, the conveyed meaning of the act is itsmotivation, its reason. No semon could be more eloquent than his act?

Also the act of violating a norm can be a communicative act, either intentional or evenfunctionally. For example, typically the intention of communicating norm violationcharacterizes the “provocative” attitudes of adolescents (where the message is addressed boththe group of peers and to the hostile adults); but is also true for many form of social devianceendowed of proud, and for revolutionary movements like civil disobedience: Gandhi’s publicstrappare of special passport for Indians in South Africa was mainly if not only ademonstrative i.e. a communicative act, more efficacy than any word and making wordssuperfluous. 47

Paradoxically violating norms could have a communicative function (independent on theintention of the author) since it can be reproduced in time by its own feedback effectsalthough socially bad (Castelfranchi, 2001) : since I violate you violate; since you violate Iviolate.

In sum, respecting or violating norms is what common sense call “to give the good (respect)or the bad (violation) example”. When giving the “example”, to be the “model” is intentional(or functional) clearly X’s behavior is a BIC (message µ: “this is how one can and shoulddo”) or a metaBIC (µ: “imitate me; do as I do”).

4.1 Fulfilling Social Commitments as BIC

A Social-Commitment of x to y of doing a (Castelfranchi, 94), in order to be really (socially)fulfilled, requires not only that agent x performs the promised action a, but also that the agenty knows this.Thus, when x is performing a in order to keep his promise and fulfil his commitment to y, healso intends that y becomes to know this.If he is not planning for a specific and explicit message to inform y about his fulfilment (“Idid a”) - like in the boring classical CSCW systems (Castelfranchi and Falcone, ) or in theContract Nets- y is relying on y’s observation of his behavior (since there is some sharecontext) or in y receiving information about some result of his action (for example, y willreceive the promised merchandise that x is sending to her).In other words, x’s performance of a is not only a practical action but it is also an intentionalbehavioural communication to y with the meaning “I did it”.Notice that what is important for exchange relationships or for social conformity, is not that xreally performed a, but that y (or the group) believes so. 48

x can try to deceive y for defeating her of for free riding in society. In this case,

47 This is frequently true of BIC: BIC can be more effective, more eloquent than word because it also puts in actthe intention (it is not simply a declaration or a promise), makes more realistic and perceivable the results;provides an example; BIC can make words superfluous because is more immediate and economic, the exploitedside effect of an act that one should comunque do.48 Of course, that x performs a is important from the moral and legal point of view: x didn’t really fulfil hisobligations if he didn’t do a

- either y declares with a special message that he did a, while this is false (his message isa “lie”);- or he only seemly performs a, his action is only a simulation, is a faked and mimeticaction: mere communication not a real pragmatic action. In other words of the twointentional effects of a:

1 - to perform a, to produce p, to satisfy the S-Commitment; and2 - to induce y to believe that 1,

only the second motivation has remained, and then 2 is a deceptive motive. The action is a“bluff” aimed at deceiving. It is just a “symbolic” action.

In sum:

• (if there are no explicit and specific messages) any act of S-Commitment fulfilment isalso an implicit communication act about that fulfilment.

3.5 Norms obedience and convention conformity as BIC

The same holds for norms obedience and convention conformity. One of the functions ofnorm obedience is the confirmation of the norm itself, of the normative authority of the group,and of conformity in general, thus one of the functions of norm obeying behaviors is that ofinforming the others about norm obedience.At least at the functional level x’s behavior is implicit behavioural communication.Frequently, x either is aware of this function and collaborates on this 49, thus he intends toinform the others about his respect of norms (or conventions), or he is worrying about socialmonitoring and sanctions or seeking for social approval, and he wants the others to see andrealise that he is obeying the norms.In both cases, his conforming is also an intentional behavioural/implicit communication to theothers.Of course, x can also simulate his respect of the norms, while hiddenly violates them. 50

In conformity to conventions the behavior is less intentional (Lewis' restriction)Conformity to convention is more routinary and automatic; they are our habits, they do norequire a conscious decision. Thus, although clearly there is an informative aim of thisbehavior (since the whole convention is based on mutual assumptions and expectation aboutthe others’ conformity), this aim is usually a function of the behavior not an intention; itbecomes an intention in cases that I want that people notice that I’m following thatconvention.

3.6. BIC in establishing Commitments

To establish a Social Commitment (Castelfranchi, 1994) two moves are essential: the move ofX that creates his obligation and Y’s right and expectation (“the promise to do”), whichbasically is goal-adoption (and its communication to Y); but also Y’s move of consent,accepting X’s “help”. Also this consent must be mutually believed.Without such (often implicit) agreement (which creates a reciprocal – altough asymmetric -Commitment) no true S-Commitment of x to y has been established. 49 As I argued in (Castelfranchi, 2001) intentionality does not eliminates the possible functional character of thebehavior.50 Consider the famous gossip about Neapolitans wearing - after the Italian law making obligatory car securitybelts - a T-shirt with a black transversal strip in order to optically deceive policemen.

This does not implies necessarily explicit communication between x and y; we meancommunication by apposite message sending. Both X’s committing act and Y’s consent actcan be (and frequently are) tacit, i.e. via BIC.

One can characterize the principle of implicit committing as follows:

IFthere is mutual knowledge between x and y about an expectation of y about an action of

x (where an expectation is a believe about a future state or action plus a goal about thesame state or action);

andx intends to do that action also because he knows about y 's expectation;andx does not explicitly deny his intention, does not contradict y 's expectation;

THENx implicitly takes a S-Commitment to y for that action;y is entitled by default to consider x S-Committed to her.

Principle of implicit consent in Social CommitmentAt the same time by relying on x and by letting x believe so and believe that he iscommitted to y, y is implicitly accepting x’s commitment, and agreeing about it.

This is how tacit committing works when –let’s say – the iniciative is by X that has someexpectation on Y’s help. But also the opposite starting point is possible: X does not considerY and her possible help, while Y take the initiative of “offering” her help to X. Sometimesalso this spontaneous commitment (based upon simple goal-adoption not goal-adhesion)works without explicit messages (speech acts or symbolic gestures). For example, X isrunning towards the target Z which is on serious danger, but observe Y that – looking alsotowards X – is anticipating him and reassuring him; X stop running (consent) and wait whilerelying upon Y’s intervenction.

This mutual "understanding" among the agents is necessary and intended. S-Commitmentsalways requires communication but not necessarily explicit; on both side tacitcommunication is enough.

3.7 Tacit agreement about Cooperation in Exchange

In exchange the agent have at least two common goals:

- one is that of negotiating -during the negotiation phase- and trying to rich an agreementsome deal;- the other is that of exchanging (after the negotiation phase) (if both are sincere)

In fact any agreement is an agreement for cooperation (in a broad sense); any agreement isaimed at a goal-adoption and if successful produces a common goal (the agreed one). Relativeto this goal the agent cooperate also in exchange. This cooperation is -of course- subordinatedto the selfish and private motives of the two actors.

• the agreement about the specific deal is achieved explicitly,• the agreement about having an exchange and negotiating is usually just a tacit

agreement.

It is based on the following inferential steps and on some BIC moves.

If I ask/propose to Y “would you sell me z for 5000 lira?” this implicitly also meansand communicates :“I conditionally intend to buy x” and “I’m willing to negotiatewith you, if it is the case”.From Y’s answers “10000 lira!”

sinceif Y would not intent to negotiate,

- either he did not respond to my offer, or- he had explicitly rejected not only the specific deal but my offer to buy (“I donot sell x”)

sinceB did not refuse, since he counter-offer,

I can infer that:he is negotiating and intends to negotiate, and ascribes me the intention tonegotiate;implicitly he complies with my intention to negotiate; there is an, implicitagreement (I know that he knows that I know...) and we are now cooperating innegotiation.

3.8 The act of trust as a message for creating trust

A lot of social relationships require more than explicit words and declaration the‘demonstration’ of a given attitude, feeling, or decision. It is when we say “I make him tounderstand it” (glielo ho fatto capire) or “I attempted in any way to make him understandthat!”: this always refers to behavior as message.In particular trust relationships require and exploit BIC. It is basically true that trust createstrust (and distrusts creates diffidence) but this means that the act of X of trusting Y forsomething, of relying on Y and delegating a ‘task’ to him, is for Y a sign that “X trust him”.Moreover, X is aware of this and – since she is relying on Y – she also wants (at least instrong delegation) (castelfr Falcon) that Y understands that she is counting on him and trusthim, that is X’s act of trust is also a (weak or strong) BIC message to Y..There is in particular a nice phenomenon of Trust dynamics studied by Cast E falc where thefact that X trust Y increases Y’s trustworthiness (for example Y’s care or persistence) and Xknows this and exploits this for making Y more trustworthy. In this case the BIC message ofthe act of trusting is absolutely necessary, is the presupposition of the dynamic phenomenonand of X’s calculated plan.As claimed in sociology (Gambetta, ) there is in social relations the necessity of activelypromoting trust; “trust starts by being open to experience, by acting as if one trusted theother” and “ la concessione della fiducia - that generates precisely that behavior that seems tobe its logical presupposition- is part of a strategy for structuring social exchange” (Mutti,1987).In sum, there is a circular relation, and more precisely a positive feedback, between trust inbilateral delegation-adoption relations (from commerce to friendship). That -in cognitive

terms- means that the (communicative) act of trusting and eventually delegating impacts onthe beliefs of the other (“trust” in strict sense) that are the bases of the “reliance” attitude anddecision producing the external act of delegating.

5. “Qui Tacet Consentire Videtur” 51 From BIC to Tacit Agreements and Conventions

Let me now generalize a bit those examples and cases while estrapolating the general BIC-based rule for implicit agreement or tacit consent: silenzio assenso.

Inaction as BIC in tacit consentIf x knows (believes) that he can do something (a’) in order to prevent the event or the actiona”, and x decides of not doing a’, i.e. of letting e/a” happening, his inaction, his omission of a’is an action a (a “passive” action).Now suppose that from x’s not-doing a’ (i.e. from his a) y infers that p (that x decided of notpreventing a”), and suppose that x knows and wants this understending. At this point x’sinaction is a communicative behavior, is BIC to y that p.This form of implicit communication (silence as consent) is particularly important not onlyfor establishing “tacit agreements” but for the emergence of tacit social “conventions” .The socio-cognitive process of this is as follows.

I believe that you know that I’m doing (intend to do) β (that you see and understandwhat I’m doing), i.e. my behavior is BIC informing you that I’m doing β

• I believe that you could oppose to this (either practically, by stopping me or creatingobstacles, or expressing your opposition, disagreement, etc. or even strongly -if youhave the authority for this- by prohibiting me of doing β);

• I believe that if you was against my doing β you will oppose to this by some actionand communication, while if you do not oppose, if you omit to oppose to my actionthis mean (by abduction) that you agree that I do it, or at least that you weaklypermit me, let me to do it.

• I believe that you believe that I believe all this and that I’m seeking for a sign ofyour attitude (i.e. that I’m “reading” your behavior -action or inaction- as a sign andcommunication of your attitude of agreement or disagreement). Thus my β is also arequest: I’m informng you about my intention and act but also asking you forpermission (agreement).

• You know all this and you decide of say/doing nothing and letting me doing. Youknow that your inaction is interpreted by me as a assent/consent. Since you decide ofnot preventing such an interpretation, your inaction is an implicit communication ofyour assent: you take the responsibility for my interpretation.

Tacit agreement are based on a lot of common knowledge or at least of shared beliefs but alsoon two implicit communication acts: a tacit conversation.

51 “If one keeps silent apparently consents”

Conventions are characterized by mutual expectations about the behavior of the member ofthe community and conditional adhesion to those expectations as a member. When and howare those conventions negoitiated and their related “contract” subscribed? Convenmtions areestablished and maintaned by communication but – in true conventions – such acommunicatyion is always implicit:

• the attempt, the innovation is a silent request for approval, consent and imitation;• the iteration or imitation is a message of consent and of request of confirmation;• the emerging expectation which is not only a prediction about the others’ behavior but

a prescription is communicated tacitly (just by doing, by monitoring, by reacfting todeviance);

• the conformity behavior is a message of obedience, of report, and of prescription tothe others; etc.

All the behaviors instauring and maintainging a social convention are BIC where doing issaying.

6. BIC basement of Institutional Order: “Count As”, Institutional Actionsand Empowerment as Based on BICommunication

True Institutional Empowerment (the Count-As empowerment) is different from interpersonalempowerment. The compliance of a third party is strictly necessary: the public, the peopleinvolved in the institution.The efficacy of the conventional institutional act in fact presupposes a tacit agreement orconsensus of people in front of it.

People (P) must recognize X’s act as a special one and acting on such a basis; actually isthis that gives the act its special effect. If X’s action Ax counts as action Ai of theinstitution Ist, people must act “as if” Ai has happened.

This means that any performance of a ‘count as’, institutional act is necessarily a BICcommunication act to the participants and the ‘public’, aimed at informing them that a ‘countas’ act has been accomplished and must be taken into account as such.It is a sort of self-realizing expectation: since and until people expect that Ax counts as Ai, itcounts as Ai. They must (conditionally) believe or at least “accept” (Meijers, 2002) that this istrue and that the others believe/accept as they do and will act accordingly.The effectiveness of the count-as effect passes through the minds and the consequentialbehavior of people.Thanks to P compliance X is really empowered. P obviously do not recognize this role; theyare simply believed to acknowledge what already exists, but in fact they are creating it thanksto this acknowledgment.Any Count-as effect (convention) and any true institutional empowerment is due to acollective acceptance of the fact, and to a diffuse or to collective intention of actingaccordingly (Tuomela, 1999, 2002).

People empowering the institution (the Leviathan)The fact that one respects the authority, decides not to oppose and rebel to it, the fact that onesurrenders to its symbolic force, etc. makes it an “authority”.The fact that one accepts the conventional and artificial effects of the institutional actions,gives them such an effect. And this behavior is spreading (as a message and as a behavior)and self-confirming: since one does A, others do the same, and vice versa.The more people acknowledge the authority and follow the institutional prescriptions themore the institution becomes strong and has Power over the individuals.In a sense, it is the arrested guy who - by surrendering not to the agent’s private strength butto his institutional force - gives to the policeman (by giving to the institution, and vice-versa)such a force.In a sense, he is surrendering to his own alienated force/power (Marx).

However, not the whole social (and societal) reality is “acceptance”-based, a collectiveconstruction; the conventional result of some explicit and organizational, or diffused and tacitagreement and pact.Part of social reality is merely emerging and self-organizing in an “objective” way; it is given,independent of human awareness, decision and even acceptance.

7. BIC and Language

Has we already said there are at least 4 important relationships between Language and BIC:- the role of BIC in the origin of language- the role of BIC in the origin of the meta-communicative nature of language;- the role of BIC in linguistic ‘implicatures’ and pragmatic inferences;- the role of BIC in meaning and linguiustic rules ‘negotiation’.Let us examine the last three issues while postponing the first one to next Section.

7.1 BIC for Meaning & Linguistic Rules Negotiation

We assume that:- Imitation in learning and teaching is BIC based (section xx) and this also applies to

imitation-based language learning and spreading

- Communicative 'conventions' and linguistic conventions and rules are 'negotiated',changed, and transmitted by BIC.

We will first consider a (real) toy example, and them generalize a ‘BIC-based negotiationprocess’.

An example: linguistic negotiation (terms, meaning, syntax, …)To name X I use the new term bbb (for example to call Amedeo I introduce the name“Amed”) with my hearer H:

1. my hearer understands (I infer this from her answer or reaction)2. my hearer does not protest/discuss

I interpret H’s non-protesting/discussing as an implicit acceptance (at least passively and forthe moment) of my use; and -more than this- of an implicit behavioral communication of suchan acceptance (in not reacting H is communicating me “OK, I let you use this term”)This is some sort of weak “implicit acceptance” of my use of bbb by H. When I will useagain bbb with H I will expect (believe + want) that

- H understands again,- H will not protest/discuss- H knows about my expectations.

In strong implicit acceptance, H re-uses herself the term bbb (in the same occasion orlater). In doing so H expects that:

- I understand,- I do not protest/discuss- I know about these expectations.

There is now a true implicit convention, a tacit agreement about using bbb (at least betweenus and in similar contexts) We can distinguishing two phases.

• One is a tacit negotiation and produces weak implicit acceptance;• The other is active reuse and produces a true convention.

• If somebody else listen to us in using bbb, or H uses bbb with other people, the newterm is spreading around and a diffuse collective linguistic convention is establishing.

use of anew term

under-stands?

to clarifyorchangeortry again

Explicitnegotiationand explanation

DISAGREEMENT

EXPLICITAGREEMENT

FAILNO

YES

I will use it again;I expect that he expectsthat I use it againetc..

TAC IT

NEGOT IAT ION

protests?YES

NO

YES

NOdoes heimitate?

CONVENTION

Unilateral useandweak, passive acceptance

7.2 BIC and the origin of meta-communication in Language:If “Saying is Doing”, since “Doing is Saying” = Saying2

We claimed at the beginning that it is possible to derive from BIC the Gricean meta-communicative character of linguistic communication and in fact of any artifactual andconventional communication system. The argument is as fiollows:

1. Any correct use of a specialized semiotic means, i.e. shared coded communication system(SCCS), entails Y’s (the observer/addressee) possible recognition of the fact that X (thesender) is communicating and intends to communicate.

Thus:

2. X can predict and intend this effect and can use SCCS and her act also in order Ycomprehends that X intends to communicate:

- there is a potential BIC meta-message (“This is a message”, “I’m communicating”) inany use of a SCCS.

The problem is:Why and for which purpose should X intend to let/make Y recognize/comprehend hercommunication goal?There must be some additional facilitation/advantage for the communication itself or for itsper-locutionary goal G.

3. We claim that the reasons and advantages are the following ones:

a) a stronger focalization, attention, activation of Y on the fact that a communication isrunning and there is a code and decodification process;

b) an Adoption by Y of the goal of X that “Y understand what I mean”, and thenCooperation on this share goal;

c) Y wonders why X wants to communicate, which is the goal G for communicating(implicature);

d) on such a basis Y does what X desires (believing, answering, doing, …) also because Xwants so and communicates that wants so (Adhesion to G).

6. BIC Forms, Evolution and Dissolution

QUADRO DA OMLL

6.1 Intentional, proto-intentional (learning-based), Functional BIC

BIC theory must be developed at 3 layers:

FUNCTIONAL

either by evolution-selection; or by design; or by reinforcement learning basedon the effects (conditioned BIC)

PROTO-INTENTIONALbased on Anticipatory classifiers (proto-intentions) thus reinforcementlearniong but with some anticipatory representation

INTENTIONALIntentional BIC presupposes an intentional stance and more precisely a “theory ofmind” in the interpreter, since the message bring by the action is about the mind of thesource: his intention, or emotion, or motives, or assumptions, etc.

6.2 From Use and Destination to Function

As I said in BIC the signal, the ‘significant’, i.e. the sent and perceived vehicle of the meaningis non market, uncoded (see below), non specialized as a signal. It is usual behavior just“used” for communicating or at most “destinated” to communication, but not shaped by sucha “function”. It is relevant clarifying such a terminology, and the difference between use andfunction with the intermidie notion of destination.

BRANO DA VECCHIO PAPER FUNCTION

Notice this notion of function means “shaped for and by” and does not completely overlapwith the other notion where “function” is opposed to “intention” (section xxx) that referes toan effect E of a feature or behavior B that has selected and manitained B through areproduction process. 52 This is a self-organized, emergent result, and it can also involve -usually involves – the shaping of B, i.e. of its forms and physical and processual properties. 53

The former ‘function’ can be a designed and artificial one (in that case we have an “artifact”).

In intentional BIC we have that a practical behavior is simply occasionally used, withoutmodification, (also) as communication signal; this use is paraxitic of the practical purpose.When, on the contrary, we have a real (either verbal o non verbal) language the signs arespecialized, their ‘shapes’ (form and perceptul properties) are due (by design or by selection)to their use that now is their specific function as (natural or atificial) ‘tools’. They arespecilised for meaning.

52 See for example Millikan charectirization as…. (dal file di Mirolli), and Castelfranchi 2001.53 But perhaps not necessariliy, the function might exploit the effects and maintaing previous ‘forms’.

There are intermidiate situations where a simple use becomes habitual, recurrent andprevalent (destination as a sign), its communicative effect is memorized as one of its possiblefunctionalities and perhaps the practical import becomes marginal or unrelevant like in theritualization process. When this process altera the form of the behavior, we have a marked,specific sign and the use has become its function (in the sense 1). When the practical use hasdisappeared and only the communicative use remains, even without any alteration of the form(that is rather implausible), in this case too we have a ‘specialized’ sign, used andunambiguously recognized as such, with no practical but merely communicative “function”(but in the scond sense 54).In non-intentional, functional (evolutionary) BIC the practical behavior has acquired a doublefunction, a double motivation for existing. If the original practical purpose is obsolete –although perhaps the morphology remains identical (rare) – we got a specialized sign, anormal form of communication.

================================

Cosa e’ un rito o cerimonia:

E’ un atto fondalmentalmente segnico-convenzionale.Un messaggio behaviorale con una procedura specifica in un contesto stabilito. Esso significasempre due cose: cio’ che l’atto deve significare, piu’ il fatto che si svolge il rito. Cioe’ e’BIC + meta-BIC grazxie alla sua convenzionalizaazione.A che serve il rituale ed il rito?Cosa e’ ‘celebrazione’? (commemorare, mostraree rendere omaggio, insegnare, ecc. funzionimemetiche: propogare, confermare, ravvivare memoria, ribadire identita’)

54 Not as “shaped by/for” but as existing in force of. The beahior would no longer exist as for its originalpractical purpose, but exist only thanks to its communicative purpose.

APPENDIX Related work

Posner

My analysis, although developed along another path, for different arguments and objectives,is really quite close to the deep and systematic approach already proposed in (Posner, 1992;2002). However, on the one side, Posner’s analysis is too subtle for my purposes,distinguished categories are beyond my needs. On the other side, paradoxically he misses orpasses over how important is the category that we are trying to enlightening: mere behavioralcommunication.The main differences are the following ones:

- Posner is aimed at providing a very good and analytic theory of intentional significationand communication, with their cognitive ingredients 55; he does not care too much ofgeneral notions for both reactive and cognitive systems. Thus, we do not have a notion ofsign, or a notion of communication for low level animals. He limits those notions tobeliefs and intentions owners. 56 We, on the contrary, – having a general theory of goal-oriented, teleonomic features and behaviors (internal and external goals) – are interestedin a possibly unified basic notion of sign, of communication, etc.; also in order to accountfor stigmergy as communication and as a sub-category of BIC.

- The notion of communication he provides not only is intentional but is already endowedwith the gricean meta-level, meta-message 57. We, on the contrary, want a notion ofcommunication not restricted to typical human linguistic communication. We need amore basic but reasonable notion. Gricean communication is a higher form that is muchricher, having a well identified additional super-stratus of reflective beliefs and goals.Moreover, we try to derive, to let emerge the reflective, gricean message and level fromspecialized, conventional, recognizable messages precisely by re-applying BIC to them.

- A marginal point is that we reject the proposed term “gestures” as any act, any signexpressing an intention. This would be a definitely dangerous term since

- it does not naturally cover behaviors like sleeping, walking, be silent, eating, etc.,and

55 For the same kind of caracterization of linguistic communication, but less systematic, see also Castelfranchi e Parisi(198x).56 In fact in a more recent paper (Posner, 2002) where the author analizes ritualization also in animals, it treatsthe development of behavior used for communicating in intentional terms.

“If an animal always uses the same behavior pattern in pursuing a certain physical purpose, then otheranimals sooner or later start to assume that this purpose is pursued whenever the behavior pattern occurs.The execution of this behavior pattern thus becomes a sign for them which indicates this purpose (cf. Eibl-Eibelfeldt 1979:82). This is the starting point for all further semiotization of behavior. When the animalconcerned notices that its behavior reveals its purpose, then it can respond by making its behaviorintransparent in order to mislead enemies or competitors, or by making its behavior especially transparent sothat other members of the same species will not find it difficult to make the appropriate assumption. If thelatter occurs, then one can say that the animal is utilizing that behavior as an instrument for conveying itspurpose to the others.”

This is Ok to us; but we claim that the analogous exists also at the non-cognitive level: insects for example;behavior can be a sign a a message without intentionality.57 Communication is “characterized as a primary action performed by means of a secondary action aimed atmaking the addressee understand the sender’s intention to perform the primary action.” p. 252. “The essentialfeature of communication lies in the sender producing a secondary sign process” p.250. Posner will probablyaccept as “communication” only our meta-BIC.

- it risks to induce confusion with “gestures” as expressive, symbolic non-verbalcommunication, while precisely a distinct category is needed. 58

Schelling

SIGNALING theory and BICBic theory has been fully understood and anticipated bySchelling in his theory of ‘signaling’ (mainly appllied ineconomics):“Players have to understand each other, to discover patterns of individual behavior that make eachplayer’s action predictable to the other; they have to test each other for a shared sense of pattern orregularity and to exploit clichès, conventions, and impromptu codes for signalling their intentions andresponding to each other’s signals. They must communicate by hint and by suggestive behavior.Two vehicles trying to avoid collision, two people dancing together to unfamiliar music, or membersof a guerrilla force that become separated in combat have to concert their intentions in this fashion, asdo the applauding members of a concert audience, who must at some point “agree” on whether topress for an encore or taper off together” (Schelling 1960, p.85).Scelling also fully underrstood how social coordination (and conventions evolution) is mainly basedon this for of communication. 59The basic theory of this form of communication howeverf has not been developed in general andabstact terms, in its various facets, and apllied in all the relevant domains where it shpould be applied.The only criticism I have to muovere to Scelling penetrating notionand analysis is that he seems tomix up ‘signification’ and ‘communication’ (is the hint, the signal just perceived, interpreted or issended, is a message?) and this make the theory quite weak. Moreover he does not make clear thedistinction between intentional vs functional behavioral communication.

58 On the other side Posner’s distinction is more subtle than our; for example, he differentiate when the actinforms about internal states of the source and the intention of the source. We put together, for the moment,informing about the action itself, informing about the internal state, informing about its reasons (beliefs andintentions), informing about other inferable aspects.59 “Players have to understand each other, to discover patterns of individual behavior that make eachplayer’s action predictable to the other; they have to test each other for a shared sense of pattern orregularity and to exploit clichès, conventions, and impromptu codes for signalling their intentions andresponding to each other’s signals. They must communicate by hint and by suggestive behavior.Two vehicles trying to avoid collision, two people dancing together to unfamiliar music, or membersof a guerrilla force that become separated in combat have to concert their intentions in this fashion, asdo the applauding members of a concert audience, who must at some point “agree” on whether to pressfor an encore or taper off together” (Schelling 1960, p.85).

Concluding remarks: Applying BIC

As I said BIC theory is also relevant for quite serious practical problems and in severalapplication domains, like:

- the problem of social order and social control in MAS, CSCW, virtualorganizations. One cannot believe that social order will be created and maintainedmainly by explicit and formal norms, supported by a centralized control, formalmonitoring, reporting and surveillance protocols, etc. Social oder will mainly beself-organizing, spontaneous and informal, with spontaneous and decentralizedforms of control and of sanction. In this approach BIC will play a crucial role.Sanctions like the act of excvluding, avoiding the bad guys will be messages; thesame for the act of exiting (section xx). The act of monitoring the others behaviorwill be a message for social order; the act of fulfilling committements, obeying tonorms, etc. (as we already saw) will be all BIC acts. We should design thosesystems allwoing for this perspective; for example allowing not onlòy messagesending but thepossibility for reciprocal observation of behaviors or of theiroutcomes.

- friendly and natural HM interactionThe same holds in this case. Machine collaboration and iniciative should be based

upon the popssibility to “observe” and understand what we are attempting to do,and in anticipating or correcting it. The same holds in Human-Robot interaction. Itis nopt synpliy a matrert of specialised messages (words or gestures (ex.CITARE)) this seems rather unnatural. Also expressive NVC sygnal (faces,emotions (Breazeal) are not enopugh. Even befor this one should provide therobot – for example for coordination with humans in a physical environmeny – ofthe ability to interpret human movements, understand them and react orcoordinate apprpriately. Atr this point the hunman action in presence of the robotwill be performed also for its understanding, i.e. as a BIC message to it.Analogously the human should be in condition to monitor what the robot is doingand to intervenie on it by adjusting its autonomy (xxx). Art this point the robotbehaviors – infront of the human – might become message for approval, help,coordination, and so on.

APPENDIX on BIC for CSCW and MAS: “Less Chats, Please!” For Silent Cooperationand Agents

ActionWorkFlow Limitations & Suchman’s misdirected criticisms

1. Implicit Vs ExplicitI disagree with Suchman criticisms to CSCW systems and approach attacking speech actstheory and classification; although I agree about the serious limits of such an approach. Mymain claim (see ICSA-NOTE 1) is that the most important cause of "oppression", non naturalinteraction, and rigidity in those systems is not the existence of a classification per se, but its

explicitness: the fact that the user is obliged to make explicit to herself (and to the machine)before acting the kind of action, classifying it in a very constrictive and simplistic taxonomy(if the taxonomy-menu would be richer it should be impractical for decision making; if it isshort, it is very brutal).In natural interaction it is possible to not explicitly state the kind of communicative act youare performing, its content, time, (addressee), all your commitments, and declarations ofcompletion.What is a really boring “discipline” is to be obliged to be explicit and to do your job twice:one by action (execution) one by word (report about your intended or performed actions) !!

There is also a misunderstanding about explicitness and its advantages. We can see thisconfusion for ex. in Winograd and in Blair.• Winograd quite persuasively argue about the necessity of "razionalized typology", of

"standardised regimes", of "regular patterns". Then he concludes as follows: "The use ofexplicitness make possible coordination of this kind that could not be effectively carriedout without it" (p.194).

But "explicitness by the user" and "structuration or standardisation" are not the same thing!Winograd did not argue about why user's explicitness is needed. This is neither because ofspeech act theory, nor of standardisation.

• Blairspeaks about "communication-awareness". He admits that "Workflow implementationrequires that we become aware of how we work together" (p.121), but:

1) even perhaps implicitly admitting that this is in contrast with the fundamental requirementof cognitive ergonomics of HCInterfaces ("trasparence": the unawareness of the medium), heseems willing to present this forced awareness as an advantage of workflow technology. Thisclaim is not obvious at all; it should be demonstrated.

2) he is incredibly mixing up the necessary awareness needed to design and implement thosesystems, and the awareness required to the user during the useIt is clear that to design CSCW systems an explicit knowledge and awareness of the processesand the acts is necessary. However this has nothing to do with the fact that this awareness beimposed to the users of such systems.When he says that "Increasing awareness is the crux of successful implementation" isambiguous: is awareness a condition sine qua non to use such a technology (which perhapsimprove the efficiency of administrative work), or the awareness is the real advantage of thattechnology and what improve efficiency ?? Two quite different claims!

This confusion is very close and related to the confusion criticized by Sperber and Wilson intheir book on relevance:

Paragraph in between example (118) and (119)However Sperber & Wilson do not consider a third possibility (perhaps the mostpsychologically plausible): that classification and categories could be real mentalentities/process, involved in recognition end/or production without being consciously applied,without an explicit reasoning by the subject. For example, perhaps, we have in mind and usesomething similar to grammatical rules described by grammaticians, in our language parsingand generation, but of course this is an implicit knowledge, we are not aware of them, we arenot able to verbalize such rules.

Communication Vs Understanding: the need for Implicit CommunicationCooperative activity requires both communication and execution of practical actions. But,there is a special relation between conversations and actions: they should be coherent, theyshould match with each other. You cannot support cooperative work obliging users toexplicitly describe what they are doing and without understanding the relationships betweenwhat they discuss about and what they are doing. You cannot support cooperative workobliging users to send explicit message to understand each other.To cooperate in natural conditions humans do not use only communication (sendingmessages), they use implicit knowledge and the interpretation of the behavior of the partners.They are able to understand each other without explicit communication.One should in part ascribe the artificiality of CSCW systems based on Speech Act Theory tothese causes.We claim that in CSCW development system's capability to understand actions taking placewill be fundamental. In particular, the capability of recognising plans, intentions and actionswill play a fundamental role.Critical judgements about the CSCW systems based on the analysis of conversations andSpeech Act Theory have been growing more and more recently; in particular, limits such asunnaturalness and coercion [SUC94] made their appearance. These limits can be brought backto different reasons (for instance Suchman herself ascribe them to the inadequacy of SpeechAct Theory). In our opinion, the inner artificiality of these systems should not be ascribed toits background theory, but to the fact that the user is obliged to explicit each time (and, as aconsequence, to classify) every act the user himself intend to do. This obligation is veryunnatural, the same holds for the obligation to declare explicitly the accomplishment of thetask and to erase the commitment from the “agenda”.How should we cope with this problem? In our view, a quite hard way, that cannot be easilyavoided, consists on the gradual introduction of the comprehension capability into the system.In general, we can say that the HM interaction systems are necessarily inclined to forcehuman interaction and cooperative work in the machines’ own schemes. On one side, humanbeings are coordinated in their actions according to two instruments which are equallyfundamental: message sending and the observation of the others’ actions. On the other side,machines just interact by means of the apt message sending.Therefore the problem consists of enabling the user of the HM system not to be adapted tothis machine characteristic, we must find ways and forms to introduce into machines thecapability of receiving and understanding the information about what has be done into theinteraction.The development of multimedia technologies will contribute to the solution of this problem,nevertheless the increase of the capability of the system not only to transfer but also tounderstand the actions taking place will be fundamental. The capability of recognizing plans,intentions and actions will play a fundamental role.

Strange consequences would follow from this approach. For example if I know that my agent or computer system is monitoring me,and that it say nothing, doers not react or prevent me when I attempt to do something (say toaccess a file; to delete a file) I’m entitled to believe that there is a tacit consent, an implicitagreement, a form of authorisation to do so.

SECTION on BIC and the origin of Language

I will also argue that this form of emerging and spontaneous communication with nospecialized signal is one of the forerunners and premise for the evolution and acquisition oflanguage (section xxx).For the evolution of conventional/cultural communication systems BIC is preconditionbecause those conventions are rarely explicitly negotiated. Usually the agreements aboutmeanings and rules (upon which their sharedness relies) are achieved through tacitcompromises; that is trhough the use of some BIC message (silence (xxx), imitation (xx), andothers).

SEE C. Castelfranchi: "When doing is saying: implicit communication beforeand without language and gestures" [abstract] [presentation]http://www.unisi.it/ricerca/dip/fil_sc_soc/dot-sc/evcoimes.html

All cultural languages are convention-based and their sygnal have conventional effects due to“rules” for codifying and decodifying.The question is: how are those conventions established among the “speakers”? This clearlyrequires communication for rule negotiation or transmission. But, if this communication toolfor convention –establishing is a language, we are in a paradox, in an infinite regression.After the existence of a sophisticated language L we can use L itself speaking about itself, byusing its symbol also at a metalevel, and explicitely “negotiating and explaining. Howeverthis cannot explain the origin of L and its spontaneous, non-explicit teaching.Our claim is that either a language is imborn, and its “negotiation” has been done inevolutionary terms (mutations, selection, genetic transmission) or necessarily somo othercommunication device is used for developping L. To avoid an infinite regress or an inbornlangiage we need BIC, i.e. the communicative exploitation of a non-conventional, nonspecialized communication able to establish tacit agreements and shared conventions. PerhapsBIC is not enough, but for sure it’s a spontaneous and effective basis.

References

R. Beckers, O.E. Holland, J.L. Deneuborg,"From Local to Global tasks Atigmergy andCollective Robotics", sta in Artificial Life IV, Rpdney A. Brooks and Pattie Maes editori,1996

Bara, B. Pragmatica cognitiva, I processi mentali della comunicazione, 1999, BollatiBoringhieri Editore S.p.a., Torino

Boella, G. and Lesmo, L. Deliberate normative agents. In R. Conte and L. Dellarocas (eds.)Social Order in Multi-Agent Systems. Springer Verlag, Berlin, 2001

Buskens, V. and Royakkers, L. Commitments: A Game-Theoretic and Logical Perspective.Cognitive Science Quarterly, xxxx Hermes, The Nederlands,

Cristiano Castelfranchi SILENT AGENT From Implicit Communication to Tacit Agreements- European Project ALFEBIITEA Logical Framework For Ethical Behavior Between Infohabitants In The InformationTrading Economy Of The Universal Information Ecosystem.IST- 1999-10298.

Cristiano Castelfranchi, intervento all'interno del Seminario sul gioco, Siena, 7 novembre2000, sta in HYPERLINK "http://www.media.unisi.it/ingioco/archivio/07nov00.htm"http://www.media.unisi.it/ingioco/archivio/07nov00.htm (scaricato il 25/05/01).

Castelfranchi, C. and Falcone, R. “To Say and to Do”. Virtual actions in the structure andrecognition of discourse plans with regard to practical plans. ????? xxx.

Cranach, Mario von, Klaus Foppa, Wolfgang Lepenies and Detlev Ploog (eds.) (1979), HumanEthology: Claims and Limits of a New Discipline. Cambridge GB: Cambridge University Press.

Eco, U. (1976), A Theory of Semiotics. London, Indiana University Press.

Umberto Eco, Segno, Enciclopedia filosofica, ISEDI, Milano, 1973

Umberto Eco, Six Walks in The Fictional Woods, President and Fellows of Harvard College,1994, Harvard University, Norton Lectures 1992-1993.

Ekman, P. e Fiesen 1969, "The repertoire of Nonverbal Behavior", Semiotica

Eibl-Eibesfeldt, Irenäus (1979), „Ritual and Ritualization from a Biological Perspective«. In: Cranachet al. 1979: 1-55 and 81-83.

Stan Frankling, Coordination without communication, HYPERLINK "http://www.msci.memphis.edu/~franklin/coord.html"http://www.msci.memphis.edu/~franklin/coord.html

David B. Givens "Nonverbal Dictionary of gestures, signs and body language cues". Spokane,Washington: Center for Nonverbal Studies Press, 2003

Grice, P. (1957), “Meaning”. Philosophical Review, 66: 377-88.

Grice P., Studies in the way of words, Harvard University Press, Cambridge, MA; 1989.

Heath, C. and Luff, P. Collaboration and Control. Crisis Management and MultimediaTechnology in London Underground Line Control Rooms. Journal of Computer SupportedCooperative Work, Vol.1. n. 1.pp. 24-48.

Kendon (1980): "Gesture and speech: two aspects of the process of utterance". In M.R.Key,ed., Nonverbal Communication and Language. The Hague, Mouton, pp.207-227

Lyons (1972)

McNeill (1992): Hand and Mind

Poggi, I. (2001) The Italian gestionary. Meaning representation, ambuguity, and context. In C.Mueller and R. Posner (Eds.) Proceeding of the Symposium on the SEmantics and Pragmaticsof everyday Gestures. Berlin Verlag Arno Spitz. Berlin.

O. W. Porter, "Non Verbal Communication", Training and Development Journal, 23, 1969,pp.3-8.

Posner, R. (1992), Believing, causing, intending: The basis for a hierarchy of sign concepts in thereconstruction of communication. In R.J. Jorna, B. van Heusden, R. Posner (eds.) Signs, Search, andCommunication: The Semiotic Aspects of Artficial Intelligence. Berlin, de Gruyter.

Posner R. (2003), Everyday Gestures as a Result of Ritualization T.R. Research Center for Semiotics,Technical University of Berlin.

[Schelling, 60] Schelling, T., The Strategy of Conflict. Cambridge, Harvard University Press, 1960.

========================R.Barthes, Elementi di semiologia,Einaudi, Torino, 1996.

K. Bloom, Moore-Schoenmakers, N. Masaka: "Nasality of female and male speakers;Language and gender bias in adult favorability ratings", Journal of nonverbal Behavior, n∞.231999

K. Bloom, D. J. Zajac, J. Titus: "The influence of nasality of voice on sex-stereotypedperceptions", Journal of Nonverbal Behavoiur, vol. 23, n∞4, inverno 1999

Eric Bonabeu e Guy Theraulaz, Sciami intelligenti, sta in "Le Scienze, edizione italiana di

Scientific American", numero 381, maggio 2000, Le Scienze Spa, Milano, 2000

J. Burgoon, D. Butler, W.G. Woodall, Nonverbal communications: the unspoken dialogue(2nd edn), McGraw-Hill Publishing Company, New York, 1995.

Cristiano Castelfranchi, From Implicit Communication to Tacit Agreements and Conventions.Fulfilling Social Commitments and Obeying Norms as Implicit Communication, nonpubblicato.

Cristiano Castelfranchi, IMPLICIT AGREEMENT on Cooperation in Exchange, nonpubblicato.

Cristiano Castelfranchi, Journal of Pragmatics: From Calci to Tacit Agreements, thoughBehavioural and Stygmergic Communication. Non-Communicative Behavior, BehaviouralTacit Communication, True Communication For Explicit and Implicit Agreements; FromImplicit Communication to Tacit Agreements, non pubblicato.

Cristiano Castelfranchi, "No more cooperation, please! In search of the social structure ofverbal interaction", in Ortony A., Slach J., e Stock O., eds., Artificial Intelligence andcognitive science perspective on communication. Springer, Heidelberg, 1992.

Cristiano Castelfranchi, Passive and Behavioural Communication, non pubblicato.

Cristiano Castelfranchi, Isabella Poggi, Bugie, finzioni, sotterfugi: per una scienzadell'inganno, Carocci, Roma, 1998

Cristiano Castelfranchi, Rapporto tecnico, Istituto di psicologia del C.N.R.,1992

Cristiano Castelfranchi, SILENT AGENTS: From Implicit Communication to TacitAgreements (National research Council - Institute of Psychology. Division of ìA.I., Cognitiveand Interaction Modellingî), all'interno di: European Project ALFEBIITE: A LogicalFramework For Ethical Behavior Between Infohabitants In The Information TradingEconomy Of The Universal Information Ecosystem. IST- 1999-10298, conferenza tenuta aTrento il 16 marzo 2000.

Cristiano Castelfranchi, Silent Cooperation: Through Tacit Understanding and ImplicitAgreements, non pubblicato.

Cristiano Castelfranchi, Understanding Coordination Introduction: ontology and conceptualfoundations for MAS; non pubblicato.

S.Cavicchioli, La Sirena. Analisi semiotiche intorno a un racconto di Tomasi diLampedusa,Bologna, Clueb, 1996.

Rosaria Conte, Cristiano Castelfranchi, La societ‡ delle menti, Azione cognitiva e azionesociale, UTET Libreria, Torino, 1996. Titolo dell'edizione originale Cognitive and socialaction, University College, London, 1995.

G.Di Caro and M. Dorigo, "An adaptive multi agent routing algorithm inspired by antsbehavior", in Proceedings of PART98- 5th Annual Australasian Conference on parallel and

Real-Time Systems, pag.261-272, Springer-Verlag, 1998

G.Di Caro and M. Dorigo, "Ant colonies for adaptive routing in packet-switchedcommunications networks". In A. E. Eiben, T. Back, M. Schoenauer, and H. P. Schwefels,editors, Proceedings of PPSN-V, Fifth International Conference on Parallel Problem Solvingfrom Nature, pag. 673-682. Springer-Verlag, 1998.

G. Di Caro and M. Dorigo, "AntNet: A mobile agents approach to adaptive routing",Technical Report 97-12, IRIDIA, UniversitÈ Libre de Bruxelles, 1997

G.Di Caro and M. Dorigo, "AntNet: distributed stigmergic control for communicationsnetworks", Journal of Artificial Intelligence Research (JAIR), 9:317-365, Dicembre 1998. Avalaible ad HYPERLINK "http://www.jair.org/abstract/dicaro98a.html"http://www.jair.org/abstract/dicaro98a.html.

G.Di Caro and M. Dorigo, "Mobile agents for adaptive routing", in Proceedings of the 31stInternational Conference in System Sciences (HICSS-31), volume 7, pag.74-83. IEEEComputer Society Press, 1998.

Marco Dorigo, Gianni Di Caro e Luca M. Gambardella, Ant Algorithms for DiscreteOptimization, UniversitË Libre del Bruxelles, Tech. Rep. IRIDIA/ 98-10, scaricato da: HYPERLINK "http://iridia.ulb.ac.be/dorigo/ACO/ACO.html"http://iridia.ulb.ac.be/dorigo/ACO/ACO.html, scaricato il 09/07/2001

M. Dorigo and L. M. Gambardella, "Ant colonies for the travelling salesman problem",Biosystems, 43:73-81, 1997

M. Dorigo and L. M. Gambardella, "Ant colony system: A cooperative learning approach tothe traveling salesman problem", IEEE Transactions on Evolutionary Computation, 1(1):53-66, 1997L. M. Gambardella and M. Dorigo, "Solving symmetric and asymmetric TSPs by antcolonies", in Proceedings of the IEEE Conference on Evolutionary Computation, ICEC96, pg.622-627. IEEE Press, 1996

Gergen Kennet J., Gergen Mary M., Psicologia sociale, Il Mulino, Bologna, 1996.

P.P. GrassÈ, La reconstruction du nid et les coordinations iner-individualles chezBellicositermes natalensis et Cubitermes, sp. La teorie de la stigmergie: Essaid'interpretationdes termites constructeurs. Ins. Soc. ,6, 41-83, 1959

Algirdas Giulien, Greimas, Del Senso II, Narrativit‡, Modalit‡, Passioni, Bompiani, Milano,1994, titolo oroginale: Du sens II Essais semiotiques, Edition du Seuil, Paris, 1983.

Gregory Hickok, Ursula Bellugi, Edward S. Klima, "Il linguaggio dei segni nel cervello", stain Le Scienze, edizione italiana di Scientific American, numero 394-giugno 2001, pag. 74-83

O.E. Holland, R. Beckers, The varieties of stigmergy: indirect interaction as a control strategy

for collective robotics, Robotica, Forthcoming, 1996

Xin Yao, Evolving Artificial Neural Network, in "Proceedings of the IEEE", IEEE Press,Piscataway, NJ-USA, September 1999

John Koza, Genetic Programming, Massachusetts Institute of Technology, 1992

F.Marsciani, A.Zinna, Elementi di semiotica generativa,Esculapio, Bologna, 1991.

Robert K. Merton, Teoria e Struttura Sociale, Il Mulino, Bologna, 1969. Titolo originale:Social Theory and Social Structure, Glencoe, Illinois, The Free Press, 1957. Traduzione diCarlo Marletti e Anna Oppo

G.Miller, E.Galanter, K.P.Pribram, Plans and Structure of Behavior, Holt, Rinehart & Wiston,New York, 1960.

Donal A. Norman, La caffettiera del masochista, Psicopatologia degli oggetti quotidiani,Giunti Editore, Firenze, 1997, Titolo originale, The psycology of everyday things, BasicBooks, Inc., Publishers, New York, 1998, traduzione di Gabriele Noferi

Donald A. Norman, Le cose che ci fanno intelligenti, Il posto della tecnologia nel mondodell'uomo; Feltrinelli Editore, Milano, Prima edizione in "Campi del sapere", aprile 1995.Titolo originale: Things that make us smart; defending Human Attributes in the Age of theMachine, Addison ñWesley Publishing Company, 1993, traduzione di Isabella Blum

Talcott Parsons, Il Sistema Sociale, Edizioni di Comunit‡, Milano, 1965, titolo originale: TheSocial System, The Free Press, Glencoe, 1951, traduzione di Amedeo Cottino

O. W. Porter, "Non Verbal Communication", Training and Development Journal, 23, 1969,pp.3-8.

Stuart J. Russel, Peter Norvig, Intelligenza Artificiale, un approccio moderno, Libreria UTET,1997, edizione italiana a cura di Luigica Carlucci Aiello.

E. J. Seavers, R. M. Dalston et al.: "A study of nasometric values for normal nasal resonance",Journal of Speech and Hear. Res., n∞ 34, 1991

R. Shoonderwoerd, O. Holland, J. Bruten and L. Rothkrantz, "Ant-based load balancing intelecommunications networks", Adaptive Behavior, 5 (2):168-297, 1996

R. Shoonderwoerd, O. Holland and J. Bruten, "Ant-like agents for load balancing intelecommunications networks", sta in Proceedings of the First International Conference onAutonomous Agents, pg. 209-216. ACM Press, 1997

Paul Thagard, La mente, introduzione alla Scienza Cognitiva, Angelo Guerini Associati, Spa,Milano, 1998. Titolo originale: Mind, introduction to Cognitive Science, MassachussetsInstitute of Technology, 1996. Traduzione di Anna Maria Marchini.

Paul Watzlavich, Janet Helmick Beavin, Don D. Jackson, Pragmatica della comunicazioneumana, 1971, traduzione di Massimo Ferretti, Casa Ed. Astrolabio-Ubaldini Editore, Roma.

Titolo originale: Pragmatic of human communication: a study of interactional patterns,pathologies, and paradoxes, 1967, W.W. Norton &Co., Inc., New York.

Terry Winograd, Fernando Flores, Conoscenza e Calcolatori, EST, Edizioni Scientifiche eTecniche Mondadori, Milano 1987. Titolo originale: Understanding computers and cognition.A new foundation for design, Ablex Publishing Corporation, Norwood, New Jersey, 1986,traduzione di Fabrizio Butera, Sonia Balance.

DARE A VEDERE CHE p P e’ falso!!

NON DARE A VEDERE che p P e’ vero !!

BIC (e stigm) e’ non solo cooperativo puo’ essere anche verso il nemico, ma solo quanto holo scopo che tu creda o sappia.

4 casi- l’informaz al nemico arriva mio malgrado (non so, non volevo ma falliscxo, non posso

evitarlo)- - irrilevante- voglio che lo creda ma e’ falso (simulare, bluff, fingere) e’ BIC o deriva da BIC- voglio che sappia (adotto il suo scopo, e’ coop a basso livello) ma per portarlo fuori

strada, es. attirarlo in un agguato.