Biomimetic signaling in Twitter

Biomimetic Signaling in Twitter

In Vonnegut’s futuristic dystopia, the Handicapper General uses a variety of handicapping mechanisms to reduce inequalities in performance. A spectator at a ballet comments: “it was easy to see that she was the strongest and most graceful of all dancers, for her handicap bags were as big as those worn by two hundred pound men.”  [ ]

Chapter 1 : The beasts move as one

Anselm: Have you noticed how it is that the beasts in nature can in some circumstances appear to move as one?

Socrates: Indeed. One has to look no further than to witness the swifts descend upon the Chapman chimney tower in Portland.

Anselm: Yes the swifts are a good example. They follow each other and swarm in patterns that clearly indicate an awareness of each other. There is some kind of signaling going on between them which coordinates their actions and makes them appear as if they are all components of a larger organism. As well in Portland many of the computer geeks themselves exhibit a similar swarming behavior using technology. I don’t think there is anything new under the sun. I propose that by observing how animals communicate in general that we can draw parallels to better understand the deeper roles that technology is manifesting.

Socrates: Certainly as it is in nature so is it in human behavior.  We are a part of nature and nature is a part of us and we cannot stand outside of that ‘wheel of life’ as the Buddhists like to say. Our cognitive wings let us fly at a greater elevation above the landscape but nature is grander than we can comprehend even so.

Anselm: Furthermore I’d like to focus primarily on Twitter.  Twitter itself is a very simple technology, letting people share brief messages in a public way about where they are and what they are doing. It crosses showing off with people watching.

Theodorus: Did you just say Twitter?! Nobody uses Twitter for anything serious. It is at best a distraction, at worst a plague.

Anselm: True Twitter may yet prove to be the digital age’s version of the sitcom. Nevertheless animal signaling theory has recently become popular among anthropologists as a way to study human communication [*]. And in that light I see Twitter as a form of biomimetic signaling. It is starting to emulate the kind of subtle, gestural and soft signaling patterns that we see in nature. If the metaphor holds true then by looking at nature we can gain insight into where our digital savannas are taking us.  [ * ]

Socrates: There may be a kernel of truth here. It is supported by others such as Rheingold who have already noted flocking behavior among the digerati.

Anselm: There’s an emerging degree of swarming, coordination, just in time planning, that makes a cohesive group appear to exist out of a series of autonomous individuals. I’ve personally witnessed people get rides from airports at midnight after the transit stopped, people collectively swarm to try track down a stolen bicycle, venue changes for meetings where nobody misses a beat, random get-togethers facilitated by a real time awareness. There is a kind of real time responsiveness not present with services such as email, the telephone, the classifieds or even newspapers and television mass media.

Theodorus: But if you look at how people tried to use Twitter for something serious – the Mumbai crisis [*] being an example – it had almost no impact. [ * ].

Anselm: Yes but where were television and other media at this time? They were not even reporting the event till hours later. Twitter is closer to a natural signaling pathway between peers. A near real-time operator-in-the-loop opportunity exists. There’s a possibility of not just consuming the big events but participating in them and understanding them. And that possibility is new.


Theodorus: You’ve been using the term ‘signals’. What kind of signals do you mean to speak of?

Anselm: By signals I mean not just say voice, or text, or the shriek of a falcon for that matter. Signaling includes any and all marks or signifiers embedded in the plainly visible world. For mice and men the world is decorated with placed hints that aid in navigation through it. Droppings, trails, marks on trees, the presence or absence of others, their haste or sloth. These all are aspects of a voice that can be used to communicate polyphonically across many media.

Socrates: Are all signaling mediums equal or are there divisions that we can apply?

Anselm: The choice of which medium to use depends on the goals but I propose a few core criteria namely 1) secrecy, 2) fidelity, 3) volume, 4) persistence. Creatures great and small often want a semi-public boundary of privacy where our messages are visible to specific peers. They may want to signal one message to predators or symbotic species, and a different message to other members. As well the choice of medium depends on the environment. Clearly for example the vibrancy or color of a rich coat of feathers is not visible at night – so to signal vigor requires perhaps an audio based medium.  There may be a desire to communicate clearly; more than simply “I am here” and the fidelity of the medium, it’s ability to clearly carry the message, may affect the choices. Volume itself; who can hear the message, may affect the medium. And finally some mediums afford a longer term persistence.

Socrates: What kind of mediums then do you see?

Anselm: I would roughly classify signaling mediums into 1) transient and 2) durable.

Socrates: Of the transient flavor then are there any different kinds of mediums? For certainly worms communicate differently than birds do they not?

Anselm: I would suggest that we define transient media just using the five basic senses of sight, sound, touch, smell and taste. A bird will flash the tip of its wing to signal that it is turning left or right. Carrion will circle over the site of a likely meal. Dinoflagelletes will bioluminesce in ocean tides (for reasons not entirely understood). Plants practice interspecies signaling using appearance, odor and taste. Grooming and nit picking instincts are employed by chimpanzees to broker peace.

Bioluminescent Dinoflagelletes surround a swimmer. Alpine berries signal readiness by color and taste.
Socrates: Then of durable signaling – are there any divisions that can be made? For is writing the same as leaving a message that a friend later repeats from memory? Bees can carry information about where nectar is by a ritualized spatial physical grammar; or in other words dancing. Human gossip networks spread awareness and information (accurately or inaccurately) quite effectively as well. How would you classify such distinctions?

Anselm: Of the durable flavor I would say there are not divisions but simply variations of a similar theme.  Any durable markings left persistently upon some medium. I admit it is worthy of note that marks, signals, scratches and the like can be imprinted on intelligent agents themselves but I don’t see that distinction as critical.

Theodorus: To be practical, where do you see evidence of signaling then in human communities?

Anselm: Well, Individual appearance and clothing style choices are strong indicators of social alignment. Using Portland as an example again personal tattoos are deeper signaling commitments. Individuals use private email for discretion but we also see the use of public blogging and indeed Twitter. Graffiti and advertising of course fit into this as well as standing on the corner with a bullhorn or dancing in a parade. Of course we know this from Marshall Mcluhan and others; how the mediums have a plasticity that changes in response to their loads, their equivalence and indeed their commonplace ubiquity. It is as if we are all constantly “Helen Kellering” our way through the world; adapting to new ways of being informed on an ongoing basis.


Socrates: Do you mean to indicate that all people are equal then?

Anselm: Yes. Let us start with the proposition that we’re all equals.  We are all people who attempt to communicate with each other.

Theodorus: Equal? I don’t see how I am equal with say large advertising firm such as Weiden-Kennedy. They have more money, more time, and more attention devoted to making their message heard.

Socrates: True. Communication arts are practiced daily by Weiden-Kennedy and by many others – even individual artists who exhibit at events such as First Thursday or Last Thursday or any of the numerous gallery art openings. These people put more resources into their messages than do most of the rest of us.

Anselm:  True. Some people do put a lot of effort into their messages. I suppose we call this advertising at a certain point. But advertising is more often hit than miss – nobody really understands the human mind yet.

Theodorus: Still one can scarcely say that these actors are equivalent. If some can devote more reasoning to their message then time itself becomes a barrier that makes us different.

Anselm: True. But it feels like we are just leaving the age of industrial broadcast media. I agree that one of the key characteristics of the previous empire was the ease with which our attention and our energy could be diverted for private gain. Advertisers successfully steered the flocking behavior of large segments of the population for worse. I am not certain that it will hold true much longer however.

Socrates: This is deeper than simply individuals. It is deeply woven into the legal and accepted definition of western culture itself.  Our concepts of the ownership of space, and appropriate use of space are largely owned by privatized interests. We accept that it is legal and accepted to place large billboards that capture our attention. We accept that it is illegal for an individual voice to “vandalize” equivalent space with graffiti. In fact almost all physical urban space is actually private. There is no real place to rest ones eyes or ones body. The illusion is provided as long as one doesn’t attempt to stand still.

Anselm: I must concur. I’ve often felt that if beauty is in the eye of the beholder then at least some portion of the value should belong to the beholder. Yet we see a tenacious externalized ownership of social objects even though these become part of the consumers cognition. It creates a space of false signals that reflect wholly accepted yet corrupted pseudo-truths. Baudrillard makes the point in Simulacra and Simulation that we don’t even know where truth is anymore in fact. This is not even a question of truth being relative and contextual but rather simply being arbitrary and unconnected to anything.

Theodorus: If you really believed this then you would actually do something about it.

Anselm: Certainly, but this implies a value judgment of deeply ‘right’ or deeply ‘wrong’. I’m not entirely certain industrial media is ‘wrong’ in the sense that it should be ‘stopped’. Yes it is ugly, and true all value is aesthetic, but the advertisers message is something we are now becoming inoculated against. In surmounting that obstacle we ‘the people’ have evolved. True many beautiful and transitory communication art forms do not exist that could exist. This is life. I do feel that attention economics is like real economics – unpredictable. Attention is that most rare of beasts, fooled once but learning quickly and innoculated quickly; in almost a Jungian manner.

Socrates: Not only are well funded interests able to make their own messages most loud but they’re also able to listen to us much better than we can listen to each other. One need look no further than ventures such as and to see how social signaling or sentiment tracking is a core part of day to day business analytics. Projects such as also deserve consideration. Even consider projects such as on the personal side.

Anselm: True. It’s clear that there is an inequity between participants and that there is some value, dollar or otherwise, in capturing attention. Let us acknowledge then that there are different kinds of participants ranging from small to large. In some senses the larger participants are the predators or at least symbiants, benefiting from the actions of the smaller predators. And perhaps they should be distinguished as such.

Growing Pains

Socrates: In what ways does Twitter permit signaling?

Anselm: I’ve proposed that Twitter is a form of biomimetic signaling, in that it emulates patterns in nature, but in fact to be more clear it is more akin to a nervous system that is incomplete. And in this light I would like to defend it not for what it is but for what it could be.

Theodorus: Why defend Twitter? It is just one more walled garden – another ‘latest internet craze’ as the BBC put it. There are many of these.  Do you mean to defend them all?

Anselm: I feel Twitter is one exemplar. And I feel that there is a backlash against technology in total even though technology is simply surfacing epiphenomena that exist anyway. For example human communities engage in a certain volume of channel maintenance; where any random traffic is sent across the channel simply to keep it open. This annoying bubbling of “social trivia” is often lampooned but it is critical to making sure the channel is there when it is really needed. Noise is a deliberate artifact of human behavior in general.  Knowing that somebody peed means knowing that they are alive and can hear you.

Theodorus: Nevertheless why Twitter? Why not FaceBook or why not other services foremost?

Socrates: Indeed I agree with Theodorus. Twitter is just one signaling mechanism. For example recently a bay area tsunami warning alert system became more visible after it failed for better or worse [ * ].

Anselm: Twitter is best known to my community so it serves best. And Twitter messages are closer to the atoms of social networks than Facebook messages. They are minimum sized “social objects”. Even the connections are one-to-one without any group concepts at this level. And messages themselves are just text, there are almost no special “powers” associated with a message such as you might find on other systems.

Socrates: Can you elaborate on this?

Anselm: In Twitter human agents have to type “rt” by hand to forward a message through the network – there is no special button called “rt” and there is no button called “thumbs up”. This makes Twitter simpler. What happens is that the weight of user needs shift into the grammar rather than through special features built into the framework. In the grammar one issues a message to say @anselm or issues a “leave” request. Like Rael Dornfest’s IWantSandy [ * ] project the burden is shifted into a more natural human dialogue. The same sense of a single input box is also visible in the new Firefox Ubiquity project [ * ]. This gives the environment a greater composability at some computational cost. In fact for this reason Twitter is a kind of universal solvent: it sits underneath other services. It is dissolving away other services that are too baroque.

Theodorus: Nevertheless Twitter itself has so many deficiencies. It doesn’t have group concepts – an idea that was already common in IRC and ICB a decade ago. It is hard to filter noise. It is a silo. It is crash-prone. It is hard to hear whole conversations. It is hard to have history. There is no way to subscribe to a geography. More deeply, it can be demoralizing – it is in many ways an ego game. And we don’t even know how the marketers are going to be exploiting this medium yet – or the spammers.  It seems like the only thing it can coordinate is a pillowfight – the minute it is used for anything dangerous the status-quo will turn it off. The biggest problem is that it is so full of noise that in order to consume it you need third party analytical tools.

Anselm: Granted. In these cases Twitter is the exemplar and it stands for the whole. It’s deficiencies are similar to deficiencies of other services and these separate pools will likely merge into a single view. My concern is that like all new things it has many weaknesses but they distract from where the future is leading. It is precisely to better analytical tools, but ones that are more evenly distributed that the future leads.


Theodorus: Can you provide more detail on where you believe the future is leading?

Anselm: In nature Biologists use the phrase “honest signals” and the phrase “dishonest signals” to distinguish between kinds of animal signaling. I feel that an idea of human curation could help improve the presence of “honest signals”. For example thomson gazelles engage in stotting when being chased by predators. They run in vertical bounds that actually make it easier to catch them – but at the same time the vigor of their jumps may indicate to predators that they are not going to be easy to take down. [ * ]. Stotting is hard to fake and therefore is an honest signal. If you could select for the human equivalent of the ‘stotting’ channel then you’d have access to a true read on a situation.

Theodorus: What is the benefit of more honesty?

Anselm: Insofar as critiques of Twitter we probably all acknowledge that the noise to signal ratio is indeed unbearable. Honesty in this sense is one part of a noise reduction strategy because it can be hard to fake and therefore expensive and therefore less common.

Theodorus: Well a private network would be even better – there would be no noise at all.

Anselm: Yes admittedly true. But being open points to an additional quality. Proponents of privacy argue that sharing information is a liability because predators or destructive forces in general can exploit the very communication pathways to find and take advantage of individuals. The term radio-silence derives from this fact. But an open communication network can be faster than predation on individuals. The network can signal very quickly, and the network itself can be improved by critical analysis. The word open is crucial because it has to be easy to access. Let us say that predators bring more computational power and analytics to the table. The flocks of lesser beasts bring a distributed network of sensors and computation to literally out-compute the situations in real time.

Anselm: As another parallel to nature consider frog cacophony. Frog cacophony in nature is designed to bewilder predators while allowing frogs to signal each other. The timing of these signals is crucial and in fact when airplanes fly overhead they disrupt the cacophony and allow predators to more easily pick out the location of frogs. Here the frogs are operating in the open but predatorial forces cannot easily exploit the signals. What is signal to the frogs is noise to the predator.

Theodorus: These signals seem awfully arbitrary.

Anselm: True, they are ritualized over time and usually side-effects of more pragmatic behavior. Narrowing the eyes and flattening the ears is a practical defense to protect the eyes and ears but it is now also a signal. Some of these signals are loud, visible to everybody, others are ‘conspiratorial whispers’ where the signaler and the receiver try to conceal the signals from third parties. There’s a concept that we see in stotting called Zahavi’s handicap principle – that in order to be honest a signal must be costly to the signaler. Peacock tails being an example. And of course we see mimicry of any signal; poisonous frog coloration being a good example of a predator defense.  [ * ]

Theodorus: What are predators in human systems?

Anselm: The term predator is perhaps loaded since it implies ‘bad’. I tend to think more of exploitive forces using energy for their own benefit but improving the fitness of the system. For example in nature mosquitos can smell carbon-dioxide and use that to find sources of blood. This is predative behavior that doesn’t necessarily kill the host but it’s an example of loads that the host carries. In human networks this could be anything from a brand such as Nike trying to create visibility for some arbitrary new style of shoe, to a banking institution trying to determine if they can get away with unusually high account fees. There are also large natural forces such as environmental change due to global warming, and the attendant pressures on food supplies and a resultant rise in father-knows-best style autocratic decision making.

Theodorus: Well, Twitter and like system’s don’t really deliver on these ideas. They are noisy and don’t do any particularily good job of making important facts available.

Anselm: There is a telios at work here. The future is drawing us towards bigger networks. The pressure is environmental, political and social; we need to “become bigger” because our social networks are larger, we are more mobile and indeed the problems we see are larger and swifter. Bigger means more noisy. But this isn’t by defacto an argument for smarter algorithms to search or cull data. People could indicate which media is worthy simply by the attention they pay to it. In human communities we can gaze in a direction and other people will follow our eye. That signaling behavior is unconscious but functional at steering attention.

Theodorus: How can this work on the Internet?

Anselm: We need an ability to formally “curate” which signals are worth listening to. I chose the word “curate” to imply a human mediated approach rather than a technology mediated one. There are millions of twitter channels, and hundreds of thousands of ideas, links, posts, articles pushed through the network every day. But this is no more challenging than the real world which has an equal density of objects. Many Twitterers are just noise, but some are specialists, there are outlets for utilitarian information and different kinds of communities define utility in different ways. Curation simply recognizes that there’s a matchmaking, ranking, scoring, categorizing and brokerage role that some participants in Twitter can perform for other participants.  A value chain of different kinds of participants can then emerge.

Theodorus: When you say formal curation what do you mean?

Anselm: The curatorial role is one of exploring the raw data and marking and categorizing worthy material. A curator needs to be able to filter the data by six criteria: 1) subject 2) location 3) time 4) trust network 5) novelty and perhaps also possibly 6) language. The actual interface, such as say offered by has to allow cleaving along these different criteria. And then it has to be possible to clump and aggregate results into buckets, and to up-score and downscore content.

Theodorus: So is this all about better search? Better search would make Twitter better?

Anselm: No, the search role is a curatorial and editorial role. Consider newspapers. A newspaper such as the New York Times has a staff of editors who in a sense curate what the readership is going to read. The readers flock to the newspaper based on their values, but they don’t pick the articles themselves. What’s missing in our system is a way to search the space from an experts vantage point in order to find the content that people will want to read.

Theodorus: But people have different criteria all the time. They move, their interests shift, they have new interests.

Anselm: Indeed. And subscription itself should be dynamic. You should be able to listen to a specific geography – and have that geography follow you around as you move [ * ] . The curatorial role is not restricted to the curator – it is just aimed primarily at people who want to put the energy in – and intended to benefit everybody. If you cannot set boundaries or filters then you can end up with something that is more intended to be serendipitous art than pragmatically functional such as [ * David Troy ].

Socrates: Such as it is in nature; birds can selectively listen to the channels they understand and ignore the irrelevant?

Anselm: Yes. The key draw is that Twitter and like systems could offer the potential to allow us to as a whole to react instantly and simultaneously to the signals of other agents if we could just find the emitters that were relevant. This isn’t just crisis response but every day opportunity. A well culled set of emitters can provide awareness about a specific topic extremely quickly to a wide community – including persons who are not explicitly listening to that emitter as a first order relationship at all times.

Examples of Curation

Theodorus: So how does one build such filters?

Anselm: Today there’s no way for the community to dynamically and collectively build the filters that we want. Twitter lacks group concepts so there is no natural way for clusters to emerge in an authoritative way yet. And searching in Twitter is just beginning to improve. But we can look at examples of primitive attempts to manually build such filters and we can use this as an example of how such curation might exist.

Theodorus: Ok, then, what are some examples of curated sets?

Anselm: There are collections such as Britta Gustafson mentions at , and the like. Of note here is one kind of manually built set of particular interest – location : . As well google readily returns a few more mainstream collections which I will repeat here for discussion:

Mainstream Green Twitterers [ * ]

Mainstream News Twitterers  [ * ]

Mainstream Tech Twitterers [ ]

(An older list of) Portland Tech Twitterers [ * ]

Mainstream Comedy Twitterers

Theodorus: That may seem like a lot of sources but the reality is that it doesn’t even to begin to reflect the diversity of values and interests that people have. Consider bicyclists, Baconists, Wiki fanatics, Data visualization folks, Artists, Musicians, Foodies, Furries, Parents. And as well in almost all of these cases there is a strong desire to filter geographically.

Anselm: True. Hand curation in this way is inefficient. That’s the whole point. There needs to be a way to do this using the power of the community.

Socrates: Twitter themselves have started to offer a suggestion service for people to listen to. What is needed beyond this? [* ]

Anselm: It is too hard to subscribe to individuals. I believe that there must be ways to subscribe to classes of signals en-masse. In mediums where group concepts are supported this is much easier.  The email mailing list ‘geowanking’ is a good example of an authoritative single subscription point that gives you a best overview of the entire social cartography scene. In a strong sense I see this a parallel to a better vision. If we can see the data better then we can choose which data we are most interested in.

The future

Theodorus: Let’s pretend that you had a team of people who would scour twitter for you and return to you exactly that set of twitters and say twitter people that exactly match your current interests – what real benefit or difference would that make for your everyday life?

Anselm: Well obviously as you state – a real signaling system will tell people what they need to know just in time. It would most likely tend to reflect real local concerns and local geography therefore. This is not entirely dissimilar from social cartography projects [ *  ] but with a much higher emphasis on a data driven approach. It is also not dissimilar from services such as Craigslist except for this aspect of being real-time. It’s just that with a phrasing around real time, curation and mobile access it would be a qualitiatively different experience.

Theodorus: Why don’t such solutions exist yet?

Anselm: is a good example of the non-trivial challenges. Ushahidi is an emergency response solution for crisis situations such as floods, earthquakes or conflict. The same problems that Twitter is encountering are evident in Ushahidi. Chris Blow and Kaushal Jhalla of Ushahidi have started on looking at ways to build a filtering system around these same data collection problem in fact [ * ]. Chris talks about the difference between “database barf” and human curated collections that are sensitive to subtle human concerns.

Socrates: Well, this has been an interesting discussion.

Anselm: Overall my hope here is to simply draw attention to numerous signaling parallels between human and animal populations. I hope that by thinking of digital media not as some kind of new space, but as a variation on existing spaces, that we can dispel some of the new age kind of response to new media and simply recognize it as just another part of our world.


  1. Skry
    Posted February 26, 2009 at 11:23 pm | #

    Wow, that’s a heavy thing to comment on. I’ll bravely add my feeble response anyway, but first, *applause*. Thanks, Anselm for another extremely thought-provoking article. I’ll be considering this for quite some time.

    It is interesting to see the tool ecosystem emerge around Twitter and the holes where more tools are needed. Now that the users have defined the tasks and use cases, the shape of the missing system begins to emerge.

    Maybe another useful idea would be the animal alarm sound. If Twitter had an urgent alert mechanism akin to the signal drum that could be heard above the noise, with a set of repeaters strategically positioned on the hilltops … .

    Ham radioland might also have some applicable ideas and practices to help with timely alert curation and propagation.

  2. Posted March 4, 2009 at 1:25 am | #

    What a great conversation Anselm, really thought provoking, and relevant. Thanks for putting this up!

  3. Posted March 18, 2009 at 10:34 pm | #

    That was an intense post, and the first speculative discussion between “dead” philosophers I’ve seen to date online. I think you win the Internet.

  4. Posted June 8, 2009 at 7:24 pm | #

Post a Comment

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>