Dolphin researcher Denise Herzing speaks at TED: Do dolphins have a language, and can we crack the code of their communication?

Is this cool or what?

I’m going with “or what.”

Denise Herzing has been working with dolphins for 28 years. She must have spent tens of thousands of hours thinking about dolphin communication and interacting directly with them. So it astounds me that she doesn’t seem to recognize how trivial the stuff she’s doing is – that it’s fundamentally no different than what the dolphin trainers at marine parks do. It’s frustrating that after all these years our communication with dolphins is still at about the level we have with dogs. Think about it. Having a dolphin recognize three sounds that represent different toys is more or less the same as “sit”, “stay”, “heel”.

In her talk (between 6:00 and 6:40) Herzing puts up spectrograms of dolphin social vocalizations and human speech and remarks about how similar they are. She comments that mathematically the information content in these sounds is comparable to that of human speech. But then she just says, well, it’s hard to figure out what the dolphin vocalizations mean. And that’s the end of that. Yes, it’s hard – but so then you revert to using 3 push buttons to “communicate” ?

To be fair, the wearable device Herzing’s group has developed that allows a diver to recognize and generate sounds in the dolphins’ frequency range could be a tool that somewhere down the road could facilitate a higher level of communication. But if Herzing has a road map in mind of how to get from here to there she’s not sharing it with us.

A recent episode of Radiolab featured Denise Herzing’s work:

My friend Sherwin Gooch sent me the link to Herzing’s TED talk. In a subsequent email discussion, Sherwin wrote the following:

I think maybe the problem is that humans, on average, are so stupid. They insist on thinking they are the superior one in the conversation, and insist on trying to teach dolphins human-like language constructs instead of paying attention to what the dolphins are saying.

After reading John Lilly’s original “Man and Dolphin” book (and little else), I had no problem initiating conversations with dolphins. They know me by name. I am “5-5” to them. (click-click-click-click-click, click-click-click-click-click.) I chose my name so it would be simple, yet complex enough to avoid most accidental occurrences.

Lilly’s original research from the ’50s (?) (!) showed their hearing extended (and centered) in the ultra high frequency sound. Considering fourier decomposition, I realized clicks between hard materials should be audible to them. Immediately upon seeing them for the first time, I noticed they actually use clicks as one of their primary utterances. How much “research” did that require?

The first time I discovered there was a window in the “tunnel” beside the bottom of the performance tank in Redwood City, I clicked rhythms on the glass with a coin. The dolphins came to the window to say “hello” immediately. Then I started taking nylon-tipped drumsticks in my back pocket whenever I went to Marineworld. By playing drum rhythms on the glass, I was able to “speak,” but the dolphins could respond to me only visually by making movements on the other side of the glass. But even this was no problem for the dolphins. The one matriarch dolphin picked it up immediately. She remembered me as 5-5, and would always come to the window immediately when I showed up and tapped my name.

The human response: The installations all posted signs saying, “Don’t tap on the glass.”

I have had essentially the same experience with the dolphins in Golden Gate Park, the new Marineworld in Vallejo, and Disneyworld. It may be a stretch, but I am convinced that the dolphins in Disneyworld already knew about me before I contacted them. Maybe one had been shipped from California or something, or maybe they talked when housed together somewhere. I don’t know for sure, and I have to admit that I believe the path of financial least-resistance would seem to all but preclude such hypothesized movement and co-housing of captive dolphins; they can always just capture a new one.

The thing I am most impressed with is how fast their thinking is. It is like trying to talk to a child with an extreme case of ADHD. A particular dolphin sees something once, and, with few exceptions, they seem to “grock” it, and move on – not wanting to be bothered with having you repeat showing them the same thing twice. N.B. trying to train someone like you would train a dog is probably less than impressive (or interesting) to someone who picked up on what you were trying to “teach” them in the first iteration, figured out the whole “humans train dolphins for no apparent reason” game, then philosophized about whether and to what extent they choose to participate henceforward.

And now the “trainers” fall back to trying to teach them a handful of concepts linked to funny whistles. It must be doubly discouraging for them.

Why haven’t the researchers installed dolphin telephones, so they can talk to each other? All that’s required is back-to-back underwater mics and speakers, and echo cancellation circuitry based on the same technique as any conference phone. They could put the two ends of the connection in 2 different tanks initially until the captive dolphins got the hang of using it. Then they could put one end in the ocean near the home of a school of dolphins. Recording the conversations from inception would, no doubt, give great language data to study, and you would be guaranteed of no prior shared context between the communicators. Analysis of the progressing conversation should be straightforward. Of course, you always run the risk of the captive dolphins warning the free dolphins of the treachery of humans which might threaten the free dolphins’ continued cooperation.

Why don’t they at least give the dolphins TV to watch the way Penny Patterson does her apes? There is quite a bit of dolphin video available now, and it might help to avoid the depression dolphins sometimes slip into in captivity.

I’d like to blather on here about the progression of my conversations with dolphins, but we have already talked about most of it before. I just find it incredible that the researchers are so fixed in their concepts that, rather than produce an instrument with which they could initiate conversation, the development of a shared set of experiences, and the construction of a primordial shared language with the dolphins, they instead build a big-button machine that generates 4 utterances. If they would just build a tiny box with one button that generated an underwater “click” every time you pushed the button, you could begin an ab initio conversation.

Dolphins have such acute spatial awareness that I am convinced part of their inter-dolphin communication is based on location and orientation (movement) similar to bees. They tried to show me some of this, but it is just too hard for me to grasp how this communication works. Humans seem to naturally have “noun” and “verb” in our wiring. We can point to something, and that object is a ‘noun’. Acted-out motions are ‘verbs’. But 3 dolphins swimming in a 3D 3-phase pattern as if braiding a rope is neither noun nor verb, so I don’t know what it (or their other incredibly complex motions) mean. I know that as an engineer, trying to design circuits which behave in that complex manner is a challenge, so the fact that dolphins can construct that kind of 3D ballet in real-time is really impressive.

I note that the time required to progress from animal behavior to the singularity required only a few thousand years from cave paintings to singularity in humans. And only a few hundred years from printing press to singularity. Providing the dolphins with the use of a memory system which they could use to write and read would probably be a fruitful enterprise. Since it’s so hard to write (and read) under water, a start might be to simply put a sound recorder with record and playback buttons underwater. After they get the hang of recording a fixed-length segment and hearing it echo back, you could progress to allowing them to address and select pre-recorded segments.

But we live in interesting times, and much is in the offing.

I’ve also had thoughts about finding different ways for dolphins to communicate with each other. One dream I had back in 1982 when I was working on John Lilly’s project was to do something like an audio-based message board for dolphins – something where a dolphin could approach some underwater device that would record sounds from the dolphin and play back sounds other dolphins had recorded. I never pursued that idea, but I would love to see someone do it.