|
EARTHTRUST & SEA LIFE PARK HAWAII'S HUMAN-DOLPHIN COMMUNICATION PROJECT (draft) RED = new PINK = not yet formatted Earthtrust has taken on the challenge of communicating with dolphins by the end of the coming decade...
The John Lilly Legacy More than any other man in history, John Lilly is responsible for revealing information about dolphins to the general public and stopping people in their tracks and getting them to think about the minds of these interestingcreatures. This increased awareness has caused people to re-assess how they treat dolphins. Earthtrust's Project Delphis was founded and based on this principle: study the dolphin mind, and reveal it to the global public. Let this longterm educational program help man make more informed decisions about how to treat dolphins, especially ecologically in the face of modern hightech fisheries, and also solidify legislative and practiced gains so that they become permanent. In the 1950's John Lilly studied dolphin communication (Lilly 1961b). One of his techniques was to place dolphins in separate tanks (Lilly 1961a, Lilly and Miller 1961). At Project Delphis we have replicated this famous experiment, resulting in our scientific paper "Communication in bottlenose dolphins (Tursiops truncatus) over an acoustic link between two tanks." This paper contains the "syntactic whistle hypothesis" and other concepts that are the seeds to the human-dolphin communication effort outlined here. In the 1970's and 80's, Dr. Lilly worked specifically on human-dolphin communication, attempting to teach dolphins a consensus sound-based system in which various laboratory-generated sounds stood for each letter of the alphabet. The hope was that the dolphins' great acoustic capabilites would allow them to become accustomed to hearing these sounds in recurring groups (words), and then associate these sound groups, or words, with certain contexts (the meaning of the words), and finally to gain some facility in utilizing these sound groups in a communication system with people (English). A modern day replication of this system is already underway by Project Delphis. It should be noted that the advances in computers today make the possibilities in this technique of research much greater than that of 10-15 years ago. What we propose here is a carefully developed combination of Dr. Lilly's method and the dolphin's own communication system, in an attempt to create a consensus language between us that is easier for the dolphins to relate to than English. Dr. Lilly visited Project Delphis in 1994. But it was in 1997 that Dr. Lilly and his webmaster, Jim Suhre, went to hear a talk given by Dr. Marten on the dolphin cognition research at Project Delphis, that a collaboration began to develop. Dr. Marten covered many topics in that talk, all the way from the project's pioneering scientific documentation of dolphin self-awareness to the harsh realities of the tuna-dolphin fishery. (Dr. Marten is himself is a retired National Marine Fisheries Service tunaboat propoise observer.) The subject that led to the collaboration is the underwater touchscreen for dolphins developed at Project Delphis (with Carroll Touch), which enables the dolphins to control a Powermac 8100 AV computer, kindly donated to them by Apple Computer. Dr. Lilly had been urging Jim to go into dolphin communications research. When the talk was given on the underwater touchscreen, the forces of synchronicity were at work. Jim, being a programmer, and audio-visual creator, was fascinated by the system. His prior experience with interactive interface design, for advertising and CD-Rom projects, gave him the skills to contribute right away. Ken and Jim hit it off at the conference, and the rest is history. The touchscreen had been a hit with baby dolphins and a group of males initially at the Delphis lab when the touchscreen was unveiled. But then the normal rotation of dolphins through the tanks at Sea Life Park led to a test group of very old females. Getting these older dolphins to take interest of the touchscreen proved to be a difficult task. The problem looked like making a video arcade in an old ladies nursing home. As outlined in our methodological paper "An Underwater Touchscreen for Dolphins", Jim's super-dynamic programming resulted in what might be considered one of Project Delphis's biggest accomplishments (although there is some pretty heavy competiion for this honor): turning these aged females, who had no interest at all in the touchscreen, into enthusiastic and adept touchscreen users. This success proved the value of adaptability and creative thinking for the task at hand--with the technology available to allow such adaptability. Bridging a gap to uninterested dolphins, using no reward whatsoever, except for the reward of the communication itself, was certainly building a foundation for language, and shows the promise of Project Delphis. The amount of data now generated nightly at the lab has created a new enthusiasm and a new understanding of the possiblilites of communication. We have a real chance for breakthrough, and will also generate a ton of data while trying. The project proposed here owes much, therefore, to John Lilly. Dr. Marten and Jim Suhre, along with a capable Project Delphis staff, hope to carry the same torch forward into the new millennium with the same aims: learning about dolphins, communicating with them, and watching out for their safety and survival in a commercial world that often cannot and will not take into account the "humanity" of these wonderful creatures. It is interesting to note that Dr. Ken Norris has been an informal advisor in Project Delphis since Dr. Marten came to it in 1990. The research at Project Delphis has the unique character of being able to claim full support from both of these cetacean research and conservation pioneers. The one thing they would both agree on is the urgency in dealing with the dangers that dolphins face in the oceans today. The Urgency of the Conservation Issue Exploring Dr. Lilly's conjectures has taken on greater urgency in light of the global war on dolphins. Now computational power and programming expertise has become available to make real these goals; yet simultaneously this race of beings is being destroyed. Earthtrust has been the global leader in programs to halt dolphin deaths from driftnets and drive kills; one of the top few groups working to end deathspurse-seine nets, and its DNA research has now uncovered a huge new market for dolphin meat in japan as "mock whale meat". As the preeminent dolphin-saving organization on the planet, it has come to the conclusion that advocacy alone cannot save the dolphins. What is needed is a "sea change" in the way these cognitively advanced animals are perceived by the global community. It is this belief which has focused the talents and intellects of many on the challenges posed by John Lilly many years ago. Participating in this project could save hundreds of thousands of dolphins. Although our program is designed to have an impact on serious issues, it does this by stressing positive qualities of dolphins and is very upbeat. (See Appendix for more information on the need for dolphin conservation.) What is needed now at the Project Delphis laboratory to carry the Lilly legacy forward is an upgrade in technology. A variety of computers are needed in the lab: upgrades for new touchscreen work and for sound recognition interactions with the dolphins. Three decades of advancements in computer hardware and software allow us to finally put to practical testing some of John Lilly's theories in human-dolphin communication: audio processing materials (sound recognition, duplication, classification) for fast dynamic screen animation, video processors, and vast databank machines to handle the large amount of data that is being generated. EARTHTRUST & SEA LIFE PARK HAWAII'S HUMAN-DOLPHIN COMMUNICATION PROJECT Earthtrust has taken on the challenge of communicating with dolphins by the end of the coming decade... Human-dolphin communication is the Holy Grail of dolphin research. It is the new developments in sound processing in computers described here, plus our ten years of experience conducting interactive research with the dolphins as equals without food reward, includiing experience in allowing the dolphins to control a computer using sound recognition hardware and software, and five years of the dolphins using their own underwater touchscreen to control a computer, that make us feel it is possible to achieve the goal of commicating with these very interesting sapient beings by the year 2010.
ABSTRACT WHO Earthtrust is a non-profit conservation organization dedicated to wildlife preservation. -Sea Life Park Hawaii works in co-operation with Earthrust, providing the setting for the research work. -Dr. Ken Marten is an internationally known dolphin behaviorist. He is also an expert on the tuna-porpoise problem, having served as National Marine Fisheries Service porpoise observer on tunaboats. -Don White is president of Earthrust, known worldwide for bringing the world's attention to the problem of driftnets and taking that to an international moratorium, and for developing the DNA technique of checking species of whalemeat in Japan. Illegal species of endangered whales (e.g. humpback & blue) are found, and, recently much of the meat tests out to be dolphins. -Additional Staff: Dr. Fabienne Delfour (cetacean cognition), Philip Goyal (sound recognition interface), Bruno Megessier (touchscreen programming), Jon Ross (bioacoustician), and Jim Suhre (touchscreen programming). All participate in experiemental design. WHAT Earthtrust is rising to the challenge of communicating with dolphins in a tangible, documentable way, within the next decade. The goal is to establish a consensus two-way communication system, much like has already been done with the great apes. WHEN Now. The Earthtrust research team is already working on it and has accomplished some of the initial steps. The research will continue until the goal of communicating with dolphins is reached. WHERE Earthtrust's state-of-the art dolphin cogntion research laboratory at Sea Life Park Hawaii (See Project Delphis at www.earthtrust.org for lab photos and some of its past research.) WHY The reason is threefold: 1. Science: People have been trying to communicate with dolphins for decades. Much will be learned about the dolphin mind along the way. 2. Conservation: The idea behind the project is to educate the global public about the cognitive characteristics of dolphins so that they can enter this into their decisions about how to treat the animals ecologically. Congress lifted the embargo on tuna caught by killing dolphins; dolphins are being killed once again by the millions and could be extinct in the next few decades. The conservation situation is analagous to the American buffalo. People weren't out there on the plains seeing them die and suddenly they were just gone. Similarly, sometime sooner than we think, people will be scractching their heads saying "didn't there used to be dolphins out in the ocean?" 7 million of the dolphins in the Eastern Tropical Pacific are already documented to have died in tuna nets. National Marine Fisheries Service data suggest that they have been taken down to approximately 40% of their initial population levels. Extinction lurks in these dolphins' future. Is a logistically enforceable solution possible? Yes: without speedboats to chase dolphins to exhaustion for capture, tunaboats cannot catch (and therefore kill) dolphins. (See Appendix for more information on the need for dolphin conservation.) 3. Environonmental enrichment: This work offers the dolphins at Sea Life Park Hawaii fun things to do. The paradigms developed can be used in other oceanaria, providing environmental enrichment for the dolphins there, or even with other species, for comparative psychology research. HOW A consensus language will be developed in steps. Each step is itself meant to be a project of scientific worth. Earthrust has developed an underwater computer touchscreen for dolphins, and has experience letting the dolphins drive a computer through sound recognition hardware & software. We have developed research techniques that do not utilize food rewards, treating the dolphins as equals, and procuding untainted, objective results. We will make a flexible sound recognition interface and work with it in combination with the underwater touchscreen to teach the dolphins and humans consensus sounds (from the dolphins' repertoire) and their meanings, and simultaneously study their own communication system with each other, and as much as possible, inject their own sounds and,when possible, meanings, into the consensus communication system to "dolphinize" it as much as possible.
Scientific topics along the way are: -
The Earthturst laboratory is the world's leading laboraotry in self-awareness research on dolphins. It is also the leading laboratory in studies in the wild of the natural history of the dolphin species killed in tuna purse seine nets. We have documented social bonds in wild spinner dolphins off Oahu existing for more than twenty years. The entire ensemble of cognitive and natural history research, self-awareness and its functions in complex social systems, longterm social bonds in the wild, better understood communication systems, and even communicating with dolphins, all provide a tour de force argument for stopping the dolphin holocaust. As we have done with our research over the last nine years, the scientific and the conservation messages will be put out to the global public in the form of televison programs on our work.
BACKGROUND OF INTERSPECIES COMMUNICATION RESEARCH HUMAN-APE COMMUNICATON RESEARCH Breakthroughs in the second half of the twentieth century in the language communication studies with apes came about with major shifts in methodology. The animal subject was treated more like an equal, sometimes even living with the researchers. And rather than trying to teach English to the animal or trying to communicate with it in its own imperfectly-understood communication system, an unambiguous intermediate consensus language was used. The Gardeners are the ones who began interspecies communication research in the 1960's with Washoe, a young chimpanzee (Gardner and Gardner 1969). This was followed by additonal human-ape communication research by Francine Patterson (1981 gorillas Koko and Micheal), Savage-Rumbaugh (1986, 1990 chimpanzees), Fouts (chimpanzee), and currently Savage-Rumbaugh et al. (1994 pygmy chimp Kanzi). The efforts by the Gardeners with Washoe, and Patterson with Koko and Michael, utilize Amercian Sign Language accompanied by English words. All of the above are true interactive two-way communication systems. Human-ape communication research is easier than human-dolphin research, because apes are our clostest relatives. We can rely on intuitive reactions to their behavior, and we can often tell how they are responding to experimental work: whether they are agitated or content, whether they are interested or not, whether thay seem to understand something, and what keeps their attention and interest to keep the research going.
HUMAN-DOLPHIN COMMUNICATION RESEARCH In dolphin research, few intuitive advantages are present. In a tank with more than one dolphin in it (which is always the case for these obligately gregarious animals), it is difficult to even detect which dolphin gives a vocaliztion, let alone what mental state the animal is in. Background Research Demonstrating That There is Communication Content to Vocal Excanges Between Dolphins: This research falls into two categories: 1. Two-tank experiments aimed at demonstrating that there is an exchange of information with vocalizations in a highly controlled situation, and 2. Normal ethological studies of dolphin vocal behavior integrated with a description of their social behavior. Two-tank experiments: John Lilly is the pioneer of this methodology (1961a,b,c). In his early papers in the 1960's he documents the dolphins' vocalizations in this kind of experiment. The next to work this kind of experiment was Bastian (1967), who demanded that the dolphins in tank A communicate information for a food reward task to be completed by the dolphins in tank B. Gish (1979) conducted a quantitative analysis of voaclizations in the two-tank situation. The most recent experiment using this technique was our own, in which a back-and-forth exchange of 327 whistles took place over the acoustic link, leading to the syntactic whistle hypothesis (Marten and Goyal, submitted), providing the structural and conceptual framework for the phoneme appraoch outlined below. Communication and Social Behavior: This subject has been extensively studied by researchers, starting with Essapian and McBride and Hebb (1948) at Marine Studios in Florida in the 1940's and '50's, and continuing with the Caldwells (1965), Tyack (1985, 1989, Tyack and Recchia 1991), Reiss and McCowan (1993, 1995) and Wells, Tyack and Singh in the wild. Despite this extensive attention to bottlenose dolphin communication research, it is amazing how poorly understood it remains. Some of the major reasons are the difficulties of cetacean compared to primates mentioned above. The problem of determining which dolphin is vocalizing was tackled and solved, at least when specialized equipment is available, by Tyack (1985, 1991). At our laboratory we record in stereo, which gives information for determining which dolphin voaclizes, although it is not always definitive. We also record on a third channel through a bat detector, to keep track of the ultrasonics in their vocalizations (set on "divide frequencies by 8"--giving a 160 kHz bandwidth, covering the entire hearing range of bottlenose dolphins). Human-Dolphin Interactive Research Systems Command Systems: In these systems, a consensus command system is established. It is not a two-way communicaton system, but an elaborate vocal or visaul set of commands (one way) that the dolphin must execute for food reward. Batteau: Wayne Batteau (1967) working for the U.S. Navy, set up a sound translator that responded to a voice command by translating it into an equipment-generated artificial whistle or other arbitrary sound. This sound was then used as the command cue to command the dolphin to do various conventional training activities, such as moving an object from one place to another, using food rewards. Herman: Lou Herman, working at the University of Hawaii's Kewalo Basin Marine Mammal Laboratory in the 1970's, '80s, and '90's has set up the most elaborate system of dolphin commands, incorporating grammatical features such as verbs, nouns, and direct and indirect objects (Herman et al. 1984, Herman 1986). Thus, a command such as "move the surfboard to the basket" can be given. One dolphin specializes in visual commands, the other in auditory commands. Schusterman: Ron Schusterman, working in the 1970's at Hayward State University and the 1980's and '90's at Long Marine Laboratory at the University of California at Santa Cruz, has conducted grammatical command research on sea lions similar to that done on dolphins by Herman (Schusterman and Kreiger 1984, followed by numerous publications by Schusterman and R. Gisiner). Both Herman and Schusterman and their colleagues have made many important discoveries about the cognitive characteristics of bottlenose dolphins and California sea lions with many ingenious and interesting experiments, often using variants the grammatical command system in their methodology. Two-Way Interactive Research Systems with Dolphins Again, John Lilly is the pioneer in this field. The co-habitation experiment conducted at his laboratory in the Virgin Islands by Margaret Howe took interactive research with dolphins to new limits. (Lilly 1967). And we have already alluded above to his pioneering research attempt to communicate with dolphins using computers at Marineworld Africa USA. Reiss and McCowan: Also at Marineworld Africa USA, Reiss and McCowan (1993) used an experimental paradigm in which the dolphins had a big "keyboard" in the water of a half-dozen buttons the dolphins can push. Each button gave out a sound when touched, and ellicited an activity by the researchers. For example the ball button gave out a tone, and a ball was thrown into the tank when it was pressed. An interesting observation from this research was that not only did the dolphins often mimic the sound from the keyboard, but they often gave the sound coupled with ball before touching the ball button. Epcott: Another device consisting of about a dozen 6-inch diameter PVC pipes the dolphin could stick its rostrum in to interrupt a beam and ellicit something from the researchers. However, I have never seen any scientific publications on this device. Earuthrust & Sea Life Park Hawaii: An interesting predecessor to our own research was that conducted by R. Stuart Mackay in the underwater viewing area that eventually became our own laboratory at Sea Life Park, Hawaii. He had a bank of filters which he used for basic classification of the dolphins' vocalizations, in an attempt to give them back vocalization-specific interactive human behavior. The most elaborate back-and-forth exchange device by far made for dolphin research was produced at our own laboratory, the underwater touschscreen for dolphins, which allows almost unlimited horizons of exchanges and exploration of choices made by the dolphins. This amazing piece of equipment will be used in the research proposed here.
Relationship of This Project to Past Research Over the last eight years, the Earthtrust Dolphin Research Laboratory at Sea Life Park Hawaii has come up with innovative interactive techniques, methodologies, and research tools for investigating the mind of dolphins. This includes a new methodology treating the dolphin as an equal has been developed in which no food reward is used, and all scientific interaction is done on the dolphins own volition. Using this new methodology, we discovered that dolphins treat television as reality like people do (Marten and Psarakos 1993), that dolphins appear to be self-aware beings like man and the great apes (Marten and Psarakos 1994, 1995a, 1995b), and that dolphins posess highly complex play (blowing underwater air rings and helices) that might even be a form of culture (Marten et al. 1996). The way our own appraoch relates to past research can best be summarized as follows. Our overall approach of interaction of equals without food reward, though novel for dolphin research, greatly resembles the formula for success in the breakthrough research on apes. We are using the basic "sounds have certain meanings" approach of John Lilly. We are substituting dolphin phonemes for English words (where in his system a word was a sound sequence of the letters of the word). We will use a concept similar to Batteau's, having an English word or idea translated into an equipment-generated sound that will be played to the dolphins. And we will take the pioneering work of Mackay to the highest level possible using modern sound recognition software and hardware to listen to and classify the sounds the dolphins give back to us in our interactive work. (We already did this, in collaboration with computer expert John McAfee, in the early '90's to the level technologically possible then, so we are experienced in this area.) We will be using concepts similar to Reiss in that the use of the consensus language sounds will generate interactions between the dolphins and ourselves related to the meaning of the sound. We will also be relying on concepts discovered by Herman and his colleagues with their different, but productive, command appraoch, relating to bottlenose dolphin congitive abilities, such as vocal mimicry (Richards et al. 1984), cognitive competence in the visual mode (Herman et al. 1989, Herman 1990), cross-modal transfer (Pack anbd Herman 1995), and genralizing and abstract conceptual abilities (Herman et al. 1994). So, in many ways the project proposed here hybridizes and builds on past research, both from our own laboratory, and from the research community at large. The new feature we feel could tip the balance is that the sounds we use in the consensus language will be the the sounds of the dolphins themselves, derived from research on their own vocal communication system.
SUMMARY OF THE OVERALL APPRAOCH TO HUMAN-DOLPHIN COMMUNICATION Language is the communication of information using symbols. We intend to create a language of censusus sound and visual symbols that have a mutually learned meaning between humans and dolphins. It should be emphasized that this will be a two-way communication system, in which the dolphins are free to express themselves to people. With this consensus set of agreed-upon symbols, or language, we intend to communicate with the dolphins and therefore learn much more about them than we possibly could without it. The dolphin touchscreen developed at our laboratory and computer sound recognition techniques will be the major tools used. The basic modus operendi in the auditory domain will be to create phonemes in a new "consensus language" for use between us from the dolphins' own sounds, and teach the dolphins a consensus meaning of these phonemes. Since the phonemes are from the dolphins' own sounds, the dolphins are familiar with the sound, and are used to using it for communication. We will be giving the sound a new meaning. The dolphins' touchscreen computer's sound recognition system will also be taught the phoneme so that it can be used in various programs with the dolphins on their touchscreen, creating various exercises and experiences that the dolphins and people can have together, with the main aim of teaching the meaning of the phoneme to the dolphins. This will also act as a test of the syntactic whistle hypothesis (Marten and Goyal, submitted). For example, over and under or up and down might be taught to the dolphins by visually demonstrating it in the lab or by exercises on their touchscreen. Each concept will be coupled with a given phoneme. We will go through various procedures to assure that both man and dolphin attribute the same meaning to each phoneme we both learn. It will be the ongoing building and use of these phonemes in various contexts andcombinations that will constitute the consesus "language". Much of the above is predicated on the dolphins mimicring the sounds used inthe consensus language. Mimicry of sounds is well documented at our lab aswell as in the scientific literature. Although the touchscreen and other visual interactions will be used for both species to learn the system, the actual consensus language will be auditory, tapping into the most highly developed part of the cetacean brain. We intend to employ as much of our knowledge about dolphin-dolphin communication in the consensus language as possible. The obvious place to search for the use of symbols in the dolphins' own communication system is to investigate the whistle name hypothesis. We have proposed a new paradigm for trying to understand dolphin whistles, the syntactic whistle hypothesis (Marten and Goyal) The work on the consensus language will not only draw on the communication research data used to formulate the hypothesis, but the actual work with the language should furthermore go a long way toward proving or disproving this hypothesis. The steps laid out in this proposal are the scientific steps along the way to accomplish this. The eyes of the world will certainly be on this effort, so an important part of the plan is to share the effort with the rest of the world. OUTLINE OF STEPS IN HUMAN - DOLPHIN COMMUNICATION RESEARCH The end-product of this research is a consenus "language" that is used for communication between the humans and dolphins who develop it. Although much of the language will have to be somewhat arbitrary, due to our ignorance about dolphin communication, a primary principle of the work is that whatever we can learn about dolphin-dolphin communication will be injected into the language, to make it as un-arbitrary as possible. In addition, dolphin-computer communication is already an integral part of the work, and some of its characteristics can be utilized to keep us on as objective a track as possible. This means that three research projects will actually occur in parrallel, dolphin -dolphin, dolphin-computer, and a human-dolphin communication studies. The contruction, development, and utilization of the human-dolphin communication system will rely on the dolphin-dolphin and dolphin-computer studies feeding into it. Besides these three major avenues of research, the entire program is broken into steps, with each step being a scientific project of its own. Thus, the research not only makes valuable contributions to the scientific literature as we go, but goes through a process of developing its own underpinnings as it goes. A good analogy would be the Gemini and Apollo missions that led to landing on the moon, or NASA's ongoing work towards a space station. What follows below is an outline of the major steps we plan on following to eventually communicate with the dolphins. In the experimental techniques we have developed over the last eight years, the dolphins are the guides. We follow their lead, and we can only go where they will take us. In this sense, the research ends up to be flexible and opportunistic. Some of the steps below may be blind allys, others may be productive, and, above all, the dolphins will show us new avenues we haven't yet thought of. We will take on the cognitive and sensory research paradigms as they arise. As in our past work, this makes the research extremely challenging--yet it ends up to be its most unique and valuable asset. Note: In our ongoing efforts to share this project with the public, displays of the research will be made at Sea Life Park, Hawaii. A possibility is a second, more demonstration-oriented laboratory in another tank complex called Whaler's Cove. It could also take the form of collaborative experiments between Earthtrust and Sea Life Park Hawaii training staff to share actual research sessions with the visiting public. If these events occur, they present interesting research possibilities for this project. This is why there are some protocols below involving a two-tank setup and training dolphins. (See Appendix for more on public displays and sponsor exposure.) In the outline below, bold with underlining denotes a topic that is discussed in more detail after the outline. The expanded sections after the outline have the same heading numbers and letters as in the outline. I. DOLPHIN - DOLPHIN COMMUNICATION A. Devise methods to deal with the problem of telling who is vocalizing
B. Investigate the Signature Whistle Hypothesis
C. Investigate the Syntactic Whistle Hypothesis (Marten and Goyal) D. Advanced Topics in 2-tank linkage (these can even be done with no lab at Whaler's Cove)
a. Co-operative task experiments 1) Determining specific dolphin meanings for particlar vocalizations a) Train dolphins in Whaler's Cove to Report Up and Down b) Require that the dolphins at Splash U give the answer, at first by observation of the Whaler's Cove experiments c) Remove the visual link between the two tanks, and record what sounds are communicated acrooss the acoustic link for up for down. If they do in fact exist, these should then be the dolphin vocalizations which mean "up" and "down" respectively. 2) Apply this paradign to other actions (meanings), determining the vocalizations correlated with each 3) Inject the vocalizations and the meanings learned into the consensus human-dolphin language Note: The hardest job in dolphin-dolphin communication research is to be able to understand what the dolphins are doing with each other to correlate vocalizations with context. Studying primates, we can use our own communication system and intuition to understand if an individial is fightened, angry, etc.--but distinguishing these states in dolphins is difficult, in fact usually impossible. The Advanced 2-tank experiments above are aimed at controlling the context variable, and relating the dolphins' own vocalizations to it--a novel approach to the best of our knowledge. II. DOLPHIN-COMPUTER COMMUNICATION In addition to dolphin-dolphin communication and human-dolphin communication, we will also work with dolphin-computer communication with no humans in the loop. Of course this is what we have some experience at already, especially with sound recognition and the underwater touschscreen work. By simultaneously studying dolphin-computer and dolphin-dolphin communications., we can more objectively lay the building blocks for dolphin-human communication. Dolphin-computer is particularly good for eliminating subjectivity and reality-checking dolphin-dolphin and human-dolphin communication work objectivley. Sound recognition hardware and software and a touchscreen-midi interface are two of the major tools we will use for co-ordinating the touchscreen and sound input/output in this effort. III. HUMAN-DOLPHIN COMMUNICATION A. Feed the Dolphin-Dolphin Communication and Dolphin-Computer results into this system B. From the D-D communication research, collect whistle phonemes for the language C. Develop a Sound Recognition system for the phonemes D. Utilize the Phonemes in communication exercises with the dolphins
E. Utilize the technique of demostrating and giving a sound, and encouraging them to mimic the sound for the appropriate context (as in Reiss 1993) F. Using the methods above, build the vocabulary of the language
G. Capitalize on the fact that a communication system allows more and more information to pass back and forth about "you", "me", "them", etc. for further investigations of self-awareness and self-vs. other strategies and behavior (See Appendix for scientific protocols for self-awareness research, even prior to achieving much in the human-dolphin communication componet.) H. Employ the language established for communication with the dolphins for additonal research
EXPANSION OF IMPORTANT COMPONENTS OF THE OUTLINE ABOVE I-C. The Need for a New Conceptual Framework in Dolphin Whistle Analysis The Signaute Whistle Hypothesis The signature whistle hypothesis has long been the conceptual framework from which dolphin whistles are studied. It has served us well, giving a platform and perspective for interpreting the multiplicity of whistling that comes from many dolphins engaged in rapid ongoing social behavior. Its major weakness is its uni-dimensionality: it slots the whistles into an idenification-related role without allowing for gross or subtle variations which might correspond to gross or subtle contextual differences and concomitant differences in meaning. The two-tank experiment (Marten and Goyal, Submitted.) puts the dolphins in the controlled situation of having to communicate strictly with sound, and thus to communicate every nuance and new shade of meaning with the sound alone. This puts greater demands on the whistles than in normal interactions involving the give-and-take of movement and the use of other modes of communication, such as vision and touch. This situation has revealed characteristics about the whistles that suggest moving the conceptual framework for whistle study to another level. In particular, the fine structure in the whistles in this experiment have moved us to propose a new hypothesis. According to the new hypothesis, whistle structure itself is broken down so that sub-units of the whistle might themselves have meaning. The strength of the data suggesting this hypothesis is the varied and fine structure of the whistles used by Maui and Puna in this experiment. The weakness in the hypothesis is that we have no richness of contexts to relate the whistle substructural variation to. One measure of the usefulness of a hypothesis is its testability. The test of this hypothesis would be the ongoing probing of whistle fine structure with fine changes in context. One of the best ways to control this experimentally is with future multi-tank experiments in which the interactions of the dolphins between the tanks are experimentally controlled, so that vocalization structure can be correlated with context and the requirements of the experiment in terms of what specific meaning needs to be communicated. The Syntactic Whistle Hypothesis The whistles produced by Puna and Maui in the two-tank experiment of Marten and Goyal (Submitted.) differ markedly from typical dolphin signature whistles. The latter's sonograms tend to be sinuous, continuous and lacking fine detail (Caldwell and Caldwell 1965, Tyack 1985, Tyack and Recchia 1991, plus numerous sonograms from our own laboratory), whereas the former's are generally rather linear, discontinuous and possess fine detail. In addition, the sharp rises and falls in frequency seen at the beginning and end of their whistles is consistent with the function of packaging information for reliable identification by the listener, and the discontinuities in frequency within whistles is consistent with the function of clearly separating the whistle-segments from one another. These considerations lead us to tentatively conclude that the function of the whistles produced by Maui and Puna differs from that of signature whistles. In particular, the whistles they produce are well-suited to syntactic encoding of information, and we hypothesise that these whistles might indeed convey synactically-encoded information. Specifically, the whistle-segments may, roughly speaking, be functioning as phoneme-like primitives, with the meaning of the whistle determined by the particular sequence of these primitives. Thus, although we lack behavioral data that would allow us to conclude that meaningful information was transmitted during the conversation, we hypothesise that such might indeed have occurred. There is a marked difference between Puna's whistles, which are relatively stereotyped, consisting mostly of a monotonic rise, and Maui's whistles, which appear to vary much more than Puna's, as if he were trying to communicate much more information. One caveat is that there is potential in the mother's whistles for microscopic modulation (from the viewpoint of sonograms), or changes that are subtle to us, but adequate and understandable for communicating information for dolphins.
I-D-3-a Co-operative Task Experiments Methodology: Train dolphins in Tanks A & B in a co-operative task. (A could be Whaler's Cove and B Splash U, but they should be interchangeable.) Let the dolphins of tank B be dependent on getting the right answer from dolphins in Tank A who are presented with the target. Dolphins in Tanks A & B get rewarded only if the dolphins in Tank B give the right answer. Start the training by letting the dolphins of Tank B see the target in Tank A on their video link. Move it out of view and let the dolphins of Tank A figure out for themselves how to give the dolphins in Tank B the right answer. If they can't figure out how to do it, help them by "training them to do it" (below). In the final step, remove all visual commumication between the tanks, leave an acoutic link only, and have the reward require that the dolphins in Tank B be able to select the right choice from alternatives on their monitor. We could train the 2 dolphins in the two tanks using the same procedure, each of them can see the other practising. The following step would be to remove the targets from B and let him see Dolphin A from the audio-video link. An alternative is to train A then B. They can see each other getting the reward, second step: you train A then B but A gets the reward after B answered properly (the audio-video link is still there), third step: remove the targets from B and wait for A to send the info to B in order to get rewarded. This pardigm can also be used for the deception and empathy experiments described in the Appendix.
"Training Them To Do It" This would take the form of having an inter-touchscreen communication system between the tanks in which the dolphins in Tank A conveyed the target info to the dolphins in tank B, and could watch the B guys get rewarded for giving the right answer. The communication system would be faded to an acoustic link demanding detail. A concrete example would be as follows. T=0: you present a target X in tank A, T+1= dolphin A touches target X or the TS where the targets are pictured and dolphin B in tank B has to touch it again on TS B to get the reward or dolphin A can be seen by dolphin B when is practising and dolphin B gets the reward when he pointes out the same target ? T+2= dolphin A can see everything and gets the reward.
III-B. From the D-D communication research, collect whistle phonemes for the language We are already conducting this step. It amounts to picking out the structural detail in whistles from our two-tank communication experiment (see "Sytactic Whistle Hypothesis Section above), and incoorporating them into average unit whistles. Thus, the dolphins will hear sounds they know they use for communication, but because of our limited understanding of meanings in contexts for dolphin-dolphin communication, a new meaning will be assigned to this familiar communicative sound. It is equivalent to an alien taking English words, and using them in communication with us, but with new meanings assigned to the words, unfortunately a necessary step, since they have received our language and know its elements, but don't know what the elements mean.
III-C. Develop a Sound Recognition system for the phonemes Acoustic Interface for Human-Dolphin Communication The kind of interface we will establish with the dolphins for the human-dolphin communication component of the work is a uniqure and challenging system in itself, involving both the underwater touchscreen and a computer sound recognition system, both of which have already been worked with at our laboratory (Marten, Submitted.). Again, this is a high-tech attention-getter that should lead to excellent demonstation material and possibly increase the visitation to Sea Life Park as well. Our central aim is to create a two-way interactive acoustic interface between a dolphin and a computer (or human). We envision constructing artificial dolphin whistles that mimic in their structure many of the recurrent features of real dolphin whistles, but whose meaning is conveyed wholly syntactically. That is, we will form a set of primitive (phoneme-like) elements which we will combine in various ways to form artificial whistles. To be very specific, an early goal would be to allow a dolphin to control an object on a computer screen : if it produces whistle A, the object moves up; if it produces whistle B, the object moves down. It should not be hard, I imagine, to devise a way to "mark" the artificial whistles so as to allow their easy discrimination from real dolphin whistles. The whistle for "up" might contain just two phonemes, for instance. I envisage these phonemes would be separated from each other by some kind of discontinuity in pitch (that we see commonly in dolphin whistles) which, I imagine, would make for relatively easy automated detection. Matlab is a very flexible and easy-to-use system with built-in DSP routinesthat would make everything from the sound-sampling to the sonogram generation to be fairly straightforward. We could write critical routines in assembler if the compiled-C from Matlab were to run too slowly. The crucial difficulty seems to me the pattern-recognition routines. One of ourideas is to treat the sonogram much like an EEG trace and pick out relevant features (such as discontinuities, rapid rise and falls of pitch, flat straight-line segments etc,), and then pick out patterns of these features as higher-level primitives (our phonemes). Dolphin-Human Interface Whistle Recognition and Translation: Software Whistle recognition: We will create an initial program that will first filter the whistle to a more basic component, while stepping through the output comparing it to an established whistle. Filter: track the maximum amplitude of the whistle creating a simplified wave with no overtones, thus dealing with just a curve. Evaluation: directly after each extraction of the filtered sample it will be compared with a file in the database. We will leave a tolerance for variation and follow a Markov or similar process to determine a match. Example: Five Files ABAC, ADACFE, ABBD, ABACD, CCGG Evaluation: First sample: E-> no match create new file New Evaluation: first Sample: A -> Norrow down to the first four Second Sample: B (AB) -> eliminates two leaving 1,3 & 4 Third sample: A (ABA) -> eliminates 3 leaving 1&4 Forth sample C (ABAC) -> still could be both if more data is present. Fifth sample D (ABACD) if this is all of the data we have a match with four. We will determine our allowance for varience (for example, if we do this at the sample level of 1/44,100Õs of a sec. Then we will have to look at a bigger curve meaning that ABA and AAA and ACA would be very similar. as Would BAB and CBC). We will evaluate pitch and other possible variables. If the file is not matched it sould then be added to the database and be able to be labelled. We will start by using 10 known whistles that we have and use those whistles and a few new ones to test the system. We have already pursued such a task to the third order. The operations to do per sample are: A-D conversion, FFT, extract maximum amplitude, compare it to a database, make a desicion, and then go on to the next sample. A way to define the beginning and the end of a whistle will be devised. When the sound recogntion protocal is elaborated to phonemes, the whistles will be broken into their component phonemes. Care must be taken to work with the entire sound, because, for example, similar whisltes with similar phonetic content might have a different beginning or ending. When these steps in phoneme and whole sound recognition have been accomplished, we will then work on using the sound with the dolphins, employing the touchscreen, training and practice for dolpins and humans, demonstrating meanings, and establishing vocabulary. Will will then break the overall research into steps that correspond to the various separate experiments presented here. These experiments, self-awarenness (being done currently), two-tank audio, two-tank video, two-tank with touchscreens in place, empathy, and deception, will all be conducted on a timeline in which each can be its own scientific accomplishment, scientific paper, or televeision program, yet all are moving towards the final goal of two-way communication between man and dolphin. III-D. Utilize the Phonemes in communication exercises with the dolphins Whistle Construction And Translation: Software After the sound recognition software has been established we will create a program that can call upon whistle components and sequence them in a predetermined order. This program will be part of the recognition program. Two examples are evaluating if the dolphin is mimicing a given sound, or using this protocol for training dolphins and humans a given sound. After the methods are established for the consensus language, then we will have a palate to choose from either in real time or called upon and automated in an experiment that would collect the data. This can be done using the dolphin touchscreen in high level language Director, or even contructing other experiments that can be programmed in C. Categories are needed for the sounds and their meanings and contexts: nouns, verbs, and actions that are similar. Experiments can then plug in the sound and play it in the experiment, yet incorporate the ability to add new sounds and define them quickly.
SIGNIFICANT ACHIEVEMENTS TO DATE (March 1999) The most three most important achievements to date for launching this project are: 1. We have conducted significant research in dolphin-dolphin communication: the most important for this project is the 2-tank communication experiment. This will supply us with dolphin communication material in a controlled situation. This is the source of our initial generalized whistles and phonemes. 2. We have developed the underwater touchscreen and worked with it for a number of years now. We have the ultimate interactive tool at our fingertips and ready to go. And the dolphins are in good touchscreen-working trim. 3. In collaboration with computer expert Joh McAfee, we have already executed a sound recognition system that the dolphins used to control their monitor. Developing the current sound recognition system is a matter of updating our hardware and software to take advantage of the high computational power available today. WHAT WE ARE DOING NOW (March, 1999) We are currently working on a multiplicity of fronts: The Human-Dolphin Consensus Language Our current work at the laboratory is in phase one of the human-dolphin communication project. Our central aim is to determine whether or not dolphins can accurately mimic the microstructure of whistles artificially constructed to resemble dolphin whistles recorded in our laboratory. In addition, we are simultaneously underway with the development of sound recognition software that will allow real-time detection and analysis of artificial dolphin whistles. Sound Recognition System We have obtained a donation from Matlab of 1 year of use of approximately $20,000 worth of their software. As Matlab will not have ongoing support on Mac, we have also just purchased a new PowerPC: a Compaq K-6 with 64 meg RAM and 8 Gig hard drive. We are starting to work with it and getting used to Matlab. Making the Project into an Attraction at Sea Life Park Hawaii We are working on phase 1: making the lab visible to the visiting public and creating the intial demonstrations in the underwater viewing area just outside the lab. Conservation Earthtrust is at this time conducting a number of research and advocacy programs for the dolphins, as it has since its incorporation in '76. Specifically: * It is an active member of the Dolphin Safe/Fair Trade Campaign fighting to prevent a massive rEturn of dolphin-sEts to the waters of the Eastern Tropical Pacific (ETP). Earthtrust has been crucial in congressional testimony and succeeded in getting some provisions cut out of the "dolphin death bills"before passage. It is now engaged in contacts and meetings to use that legislative victory to create a new catagory of "dolphin saving" tuna in the marketplace, based on licensing, accreditation, and contract law, which could render the recent gutting of US dolphin-safe laws moot. (Earthtrust is the only conservation organization ever to bind major tuna firms by contract to an independent standard). * Earthtrusts's landmark forensic DNA analyses of whalemeat in Japan uncovered what is one of the biggest threats to dolphins in the world today: the mislabeling of their meat as "whalemeat" in the Japanese market. This makes individual dead dolphins worth as much as US$3000 apiece as it defrauds (and slightly poisons) the Japanese consumers. Fully 1/3 of the whalemeat sold in Japan in 1997 was really dolphin meat, probably amounting to tens or hundreds of thousands of dolphin kills. Trends in the data beginning in 1993 show a steady rise, so the percentage of "whalemeat" that is really dolphin is probably now approaching 50%. (there is a time lag due to the time and expense involved in doing the DNA analyses). Earthtrust has launched an ambitious program to - in a positive and non-confrontive way - make this an issue within Japan, and stop this mislabeling before huge imports of dolphin meat become a staple of Japan's commerce. * Since its landmark Driftnet victories at the U.N., Earthtrust has created and coordinates the "DriftNetwork", an international collaboration of individuals and organizations dedicated to identifying rogue driftnet vessels and shutting them down; as well as detecting the existence of bilateral driftnetting agreements which allow Japan to driftnet within the EEZ's of other nations, so these nations may be educated to disallow this destructive activity. This involves interfacing with the U.S. Coast guard, Taiwan enforcement vessels, and myriads of contacts around the world. * In addition to coordinating the Forensic DNA Alliance for Wildlife, a growing coalition which seeks to expand the use of adapted DNA technology to enforce endangered species treaty protections; Earthtrust was a founding member of Species Survival Network and continues on its board to co-direct that international organization in its work through CITES. *Our longterm wild spinner dolphin research project is an immersive Jane Goodall style observation study of the family patterns and associations of the two species (spinner and spotted dolphins) killed in the largest numbers by the tuna purse seine fishery. Our results are useful for bringing attention on television to this high mortality and the tuna porpoise problem in general, as well as providing documents for legislation and lobbying providing information on how the disruptive effects of the fishery affects the natural history of the dolphins who manage to survive. The study is also an excellent source of study material of dolphin-dolphin vocal communication in the wild, which could prove useful for the project proposed here. *In non-dolphin projects, Earthtrust continues to be a leading organization in the whaling fight; its DNA exposes have revolutionized the IWC's scientific standards and have drastically cut down the amount of endangered-species whale meat showing up in Japan. PERSONNEL Dr. Ken Marten. Principle Investigator. Dr. Marten received his Ph.D. training in the primate behavior research group of Dr. Thelma Rowell, University of California at Berkeley. His Post-doctoral training was at the Naval Ocean Systems Center, Hawaii. Dr. Marten is currently Director of Research at Earthtrust's internationally known Dolphin Behavior Research Laboratory at Sea Life Park Hawaii. Prior to this he served on the faculty at the University of California, Santa Cruz, where he was Director of Long Marine Laboratory's Marine Acoustic Services Laboratory. He is an internationally known marine mammal biologist, with numerous scientific publications (see Literature Cited for a selection of publications releveant to this project). Jim Suhre. Computer Artist-Programmer/Virtual Reality Pioneer. Assistant to John Lilly. Webmaster/Creative Director for the John Lilly' Homepage (http://www.garage.co.jp/lilly). Works on touchscreen programming and experimental design. Philip Goyal. Ph.D. Physics candidate, University of Hawaii, CCRMA (Center for Computer Research into Music & Acoustics) Stanford, sound recognition. Philip works at the core of experimental development of the human-dolphin communication component. He is a central part of theory and experimental design for the project. He is currently putting together the sound recogntion system utilizing Matlab, and conducting interactive experiements testing the syntactic whistle hypothesis. Don White. Don White is president and CEO of Earthturst, and provides the executive management portion of the project. Mr White's training includes a BS in Geophysics from IUBloomington and extensive cross-disciplinary technical expertise acquired in subsequent decades as a scientist and innovator. Don is a valuable participant in experimental and equipment design. Earthtrust brought the world's attention to the problem of driftnets, and was the major player in bringing about an international moratorium on driftnetting. Earthturst has innovated DNA analysis for checking what cetacean species are marketed as whalemeat in Japan. In 1997 Congress lifted the tuna embargo, thus re-opening the dolphin killing fields in the Eastern Tropical Pacific tuna fishery. Earthturst reamains one of the two major conservation organizations trying to help save the millions of dolphins slaughtered there. Dr. Fabienne Delfour. Ph.D.Cognitive ethology in marine mammals. Dr. Delfour will be one of the main researchers working with Dr. Marten on the project. Teaching experience in Animal Biology and Physiology, Neurobiology, Ethology and Psychology at UniversitŽ Paul Sabatier (France). Presentations on Marine Mammals in International Scientific Meetings but also slide shows in different schools. Member of the European Cetacean Society for several years. Bruno Megessier. Masters degree. Computer scientist. Will be programming the dolphin touchscreen in high-level language Director, and can also help with internet programming in Java. Jon Ross. Bioacoustician; Publicist. Jon Ross will be managing the component of the project aimed at releasing our work to the media. He is the manager of the Earthturst media department, and runs Earthtrust's digital video media studio.
BUDGET FIRST YEAR COMPUTER, VIDEO, AND INTERNET EQUIPMENT
MAJOR ADDITIONS FOR LARGER GRANTS (wish List)
See You-See Me Internet Server $15,000 Web/Connectivity Personnel $30,000 High Speed Telephone Lines $6,000 Avid Random-Access Digital Audio-Video Database System $100,000 Kiosk Display in Sea Life Park $15,000 Additional Research Personnel $150,000 Additional "Videophone" Touchscreens 4 @ $10, 000 $40,000 Mobile Touchscreen Technology (installation on ship) $100,000 SUBTOTAL FOR ADDITIONS $456,000
PROMOTIONAL PLAN Sponsorship for the research presented here is being sought with large corporations, so public relations for promoting visitation to SeaLife Park and the labs will not be done by Earthturst, but by professional promoter, as the corporations will be demonstrating thier computer equipment or project, or whatever. This should attract a lot of people to the Park. Big corporations will be approached for participation in the project. It will put their wares on display and allow them to do some useful conservation. Possible compnaies are: Motorola makes cable modems, tons of other stuff (3Com - their competitor) Hewlett-Packard - make the RR servers Microsoft creates some of the software for cable modems Apple - quicktime creator of one of the big streaming-mediaapplications Real audio/video streaming media creators Macromedia and any other development tools by big firms IBM voice recognition products Road Runner (national)
LITERATURE CITED Aubauer, R., & W.L. Au. Phantom echo generation: a new technique for investigating dolphin echolocation. Journal of the Acoustical Soc. of Amer., 104 (3):1165-1170. Bastian, J. 1967. The transmission of arbitrary environmental information between bottlenose dolphins. In: R.G. Busnel (ed.), Animal Sonar Systems, Vol. III, pp. 803-873. Batteau, D.W. 1967. Man/Dolphin Communication : Final Report. U.S. Naval Ordanance Test Station, China Lake, California. Contract No. N00123-67-C-1103. Caldwell. M.C. And D.K. Caldwell 1965. Individualized whistle contours in bottlenosed dolphins (Tursiops truncatus). Nature 207:434-435. Connor, R.C., & Norris, K.S., 1982. Are dolphins reciprocal altruists? Am. Natural., 119(3): 358-374. De Waal, F., 1996. Good Natured. Harvard University Press, Cambridge, Mass. 296 pp. Delfour, F. and K. Marten. Submitted. Behavior of killer whales (Orcinus orca), false killer whales (Pseudorca crassidens), Atlantic spotted dolphins (Stenella frontalis), and California sea lions (Zalophus californianus) in front of a mirror, and its implications for self-recognition. Journal of Marine Mammal Science, to appear in Ken Norris Memorial Volume Nov. 1999. Gardner, R.A. and B. Gardner. 1969. Teaching sign language to a chimpanzee. Science 165:664-672. Gish, S.L. 1979. Quantitative analysis of two-way acoustic communication between captive Atlatic bottlenose dolphins (Tursiops truncatus Montague) . Doctoral disseration. University of Califnornia, Santa Cruz. Herman, L.M. 1986. Cognition and language competencies of bottlenosed dolphins. In: Dolphin Cognition and Behavior: A Comparative Approach, Eds. R.J. Schusterman, J.A. Thomas, and F.G. Wood. Erlbaum. Hillsdale, New Jersey. pp. 221-252. Herman, L.M. 1990. Cognitive performance of dolphins in visually guided tasks. In: Sensory abilities of cetaceans: laboratory and field evidence, eds. J.A. Thomas and R.A. Kastelein. Plenum. New York, NY. pp. 455-462. Herman, L.M., J.R. Hovancik, J.D. Gory, and G.L. Bradshaw. 1989. Generalization of visual matching by a bottlenose dolphin (Tursiops truncatus): Evidence for invariance of cognitive performance with visual or auditory materials. J. Exp. Psychol.: Animal Behavioral Processes 15:124-136. Herman, L.M., A.A. Pack, and A.M. Wood. 1994. Bottlenose dolphins can generalize rules and develop abstract concepts. Marine Mammal Science 10:70-80. Herman, L.M., D.G. Richards, and J.P. Wolz. 1984. Comprehension of sentences by bottlenosed dolphins. Cognition 16: 129-219. Lilly, J.C. 1961a. Man and Dolphin. Pyramid Books. New York, NY. 191 pp. Lilly, J.C. and M. Howe. Co-habitation with bottlenose dolphin. Lilly, J.C. and A.M. Miller. 1961b. Sounds emitted by the bottlenose dolphin. The audible emissions of captive dolphins under water or in air are remarkably complex and varied. Science 133: 1689-1693 (May 26, 1961). Lilly, J.C. and A.M. Miller. 1961c. Vocal exchanges between dolphins. Bottlenose dolphins "talk" to each other with whistles, clicks, and a variety of other noises. Science 134: 1873-1876 (Dec. 8, 1961). Mackay, R.S. Filter recognition system for dolphin sounds and driving a toy submarine with dolphin vocaliztions: Bateson's Bay, Sea Life Park, Hawaii. Cetology. Marten, K. Submitted. An underwater touchscreen for dolphins. Journal of Marine Mammal Science, to appear in Ken Norris Memorial Volume Nov. 1999. Marten, K. In press. Ultrasonic analysis of pygmy sperm whale (Kogia breviceps) and beaked whale (Mesoplodon sp.) clicks. Journal of Marine Mammal Science, to appear in Ken Norris Memorial Volume Nov. 1999. Marten, K. and P. Goyal. Submitted. Communication in bottlenose dolphins (Tursiops truncatus) over an acoustic link between two tanks. Journal of Marine Mammal Science, to appear in Ken Norris Memorial Volume Nov. 1999. Marten, K., D. Herzing, K. Allman, and M. Poole. Submitted. The dolphin sonic prey stunning hypothesis: predation in spotted dolphins, bottlenose dolphins, and killer whales, and the effects of low frequency impulse sounds on fish. Journal of Marine Mammal Science, to appear in Ken Norris Memorial Volume Nov. 1999. Marten, K., K.S. Norris, P.W.B. Moore, and K.A. Englund. 1988. Loud impulse sounds in odontocete predation and social behavior. In: Animal Sonar: Processes and Performance. Ed. P. Nachtigall. Plenum, N.Y. pp. 567-579. Marten, K. and S. Psarakos. 1992. Do dolphins perceive television as a representation of reality? Earthtrust Chronicles (November issue), p.8. Marten, K. and S. Psarakos. 1994. Evidence of self-awareness in the bottlenose dolphin (Tursiops truncatus). In: Self-Awareness in Animals and Humans: Developmental Perspectives. Eds. S. Parker, M. Boccia, and R. Mitchell. Cambridge University Press, pp. 361-379. Marten, K. and S. Psarakos. 1995a. Using self-view television to distinguish between self-examination and social behavior in the bottlenose dolphin (Tursiops truncatus). Consciousness and Cognition, vol. 4, pp. 205-224. Marten, K. and S. Psarakos. 1995b. Marten and Psarakos Commentary Response. Consciousness and Cognition, vol. 4, pp. 258-269. Marten, K. and S. Psarakos. In press. Long-term resightings of wild male spinner dolphins (Stenella longirostris) seen together in 1978-79 and again in 1995-98 off Oahu, Hawaii. Journal of Marine Mammal Science, to appear in Ken Norris Memorial Volume Nov. 1999. Marten, K. , K. Shariff, S. Psarakos, and D.J. White. 1996. Ring bubbles of dolphins. Scientific American, vol. 275, number 2 (August issue), pp. 82-87. McBride, A.F. abd D.O. Hebb. 1948. Behavior of the captive bottlenose dolphin (Tursiops truncatus). Journal of Comparative Physiology and Psychology 41:111-123. Pack, A.A. and L.M. Herman. 1995. Sensory integration in the bottlenose dolphin: Immediate recognition of complex shapes across the senses of echolocation and vision. Journal of the Acoustical Society of America 98:722-733. Patterson, F.G. 1981 . The Education of Koko. New York, Holt, Rinehart, and Winston. Powers, J. 1979. National Marine Fisheries Service SOPS/79/28. Richards, D.G., J.P. Wolz, and L.M. Herman. 1984. Vocal mimicry of computer generated sounds and vocal labeling of objects by a bottlenose dolphin, Tursiops truncatus. Journal of Comparative Psychology:98:10-28. Reiss, D. 1988. Can we communicate with other species on this planet? In: Bioastronomy - The Next Steps, Ed. G. Marx. pp. 253-264. Kluwer Academic Publishers. Reiss, D. and B. McCowan. 1993. Spontaneous Vocal Mimicry and Production by Bottlenose Dolphins (Tursiops truncatus): Evidence for Vocal Learning. Journal of Comparative Psychology vol. 107 no. 3: 301 Reiss, D. and B. McCowan. 1995. Whistle Contour Development in Captive-Born Infant Bottlenose Dolphins (Tursiops truncatus): Role of Learning. Journal of Comparative Psychology vol. 109 no.3 : 242 Savage_Rumbaugh, E.S. 1986. Ape langauge: from conditioned response to symbol. Columbia University Press, New York. Savage-Rumbaugh, E.S. 1990. The Lana Project. Wiley, New York. Savage-Rumbaugh, E. S., et al. 1994. Kanzi : the ape at the brink of the human mind. Wiley, New York. Schusterman, R.J. and K. Kreiger. 1984. California sea lions are capable of semantic comprehension. Psych. Record 34:3-23. Tyack, P.L. 1985. An optical telemetry device to identify which dolphin produces a sound. J. Acoust. Soc. Am. 78:1892-1895. Tyack, P. 1989. Whistle repertoires of two bottlenose dolphins, Tursiops truncatus: mimicry of signature whistles. Behavioral Ecology & Sociobiology 18: 251-257. Tyack, P. 1991. Use of a telemetry device to identify which dolphin produces a sound. In: Dolphin Societies: Discoveries and Puzzles, eds. K. Pryor and K.S. Norris. Univ. of Calif. Press, Los Angeles. pp. 319-344. Tyack, P.L. And C.A. Recchia. 1991. A datalogger to identify vocalizing dolphins. J. Acoust. Soc. Am. 90:1668-1671.
APPENDIX 1. Conservation 2. Demonstration Material at Sea Life Park Hawaii 3. Advanced Topics in Self-Awareness 4. Dolphin Echolocation Target Returns and Communication
1. INFORMATION ON THE NEED FOR CONSERVATION: DOLPHINS BEING DRIVEN TO EXTINCTION IN THE TUNA FISHERY As has been the case with much of our research in the past, this project will carry a high profile in the media. The true mission behind the project is to save the dolphins in our oceans from extinction. All media releases will contain information and lessons in helping the global public to move towards stopping the massive slaughter of dolphins currently occurring in the modern purse seine tuna fishery. The dolphin species of the deep ocean in the Eastern Tropical Pacific face a perilous existence. Utilizing the symbiotic relationship between spinner dolphins, spotted dolphins, and yellowfin tuna, tuna purse seine boats catch tuna by setting their lethal, mile-long nets around dolphins rather than on the tuna (often catching and killing only dolphins with no tuna). In 1997 Congress passed a law lifting the embargo on tuna caught by killing dolphins. The National Marine Fisheries Service has divided each species into stocks (interbreeding populations). Unexploited (pre-purse-seining) and percent population sizes have been estimated to determine the level of depletion of a given stock. Estimates of status in 1979 for the eastern spinner dolphins was 27 % of the original population, and 42% for the coastal spotted dolphin (Powers 1979). This suggests that extinction is possible within the next few decades, with increasing sets around diminishing numbers of dolphins, and no reasonable controls. Contrary to the popular belief that dolphins are killed by accident in the tuna fishery, in fact dolphins are chased, corralled, encircled, bombed, and kept from leaving the nets -- intentionally, not accidentally -- since if the dolphins escape out of the net, the tuna usually go with them. Thus, the dolphin deaths are neither incidental nor accidental, but are brought about by executing a premeditated and well-known procedure, with full knowledge of the inevitable outcome. The results of our study of wild spinner dolphins off the island of Oahu (Marten and Psarakos, in prep.), documenting the twenty-year social relationship between identified, known individual spinner dolphins, and all the long-term bonds and relationships of the individuals in that study, show that the high mortality experienced by these animals in the tuna fishery is being experienced by individuals whose natural history and way of life is based on living and working together as a cooperative unit for a lifetime. The destruction of this social order combined with the massive mortality this species now experiences in the tuna fishery, certainly must tear apart the very biological fabric these animals use to survive. People have speculated for decades that the remaining individuals who survive the high mortality of the tuna holocaust might be some kind of self-aware beings that could be suffering. Our research on self-awareness in dolphins (Marten and Psarakos 1994, 1995a, 1995b), and our research cited above on the longterm social relationships in spinner dolhins, validates this fear. Action must be taken immediately to stop tuna fishing by setting nets on dolphins if these dolphins are to escape extinction. The situation is analagous to the American buffalo. People aren't out in the deep ocean and don't see or understand what is going on. In a few decades we will all be scratching our heads saying, "What happened to the dolphins out in the ocean?" The answer will be simple: we killed them all. 2. CREATING DEMONSTRATIONS OF THE PROJECT AT SEA LIFE PARK HAWAII TO SHARE THE RESEARCH WITH THE PUBLIC The display section of this proposal is aimed at putting the scientific research that we conduct on display at Sea Life Park for visitors. This makes new Park displays, as well as educates visitors about dolphin research and conservation. It not only creates displays for Sea Life Park, but by placing accompanying poromotional material, it also puts any of our sponsors on diplay too. There are various levels of demonstration. Sea Life Park and Earthtrust should settle on a level that is comfortoble for both parties. The first level is demonstrations outside our Splash U lab for the visiting public during the day when we are not there. The lab window can be left open for visitors to see it, and the demonstrations given below can be made. The second level of demonstration is making demonstrations outside of the Splash U area. This can take the form of a demonstration-type lab at Whalers Cove, and various possibilities of kiosks and demonstration displays outlined below. A demo lab in Whaler's Cove has the scientific advantage of enabling us to have a two-tank link between Whaler's Cove and Splash U. This can first be acoustic, then video can be added, and finally a second touchscreen added at Whaler's Cove to enable sophisticated two-tank research. At the highest level of co-operative demonstration, there could be a large Earthrust display area, with the visitors sitting right between the two tanks that are linded, as described below. There could also be co-operative food reward experiments done, with the training staff doing the training for the experiments--thus putting the research on display in yet another way. Making The Earthrust Dolphin Lab At Splash U Into An Attraction Background Information: The Earthrust dolphin research and conservation laboratory is located in the underwater viewing area of the tank of the "Splash U" attraction at Sea Life Park. "Splash U" is the attraction where the visitor can pay extra to have a personal experience with the dolphins under staff guidance, being able to touch the dolphins, participate in a training session, and learn more about them. It is located in an area that used to be "behind the scenes", but is now open to the public. We refer to this laboratory area as "Splash U", and the rest of Sea Life Park that has always been open to the public, as "the Park." Earthtrust staff conduct research on the dolphins in the late afternoon and evening after Splash U and the Park have closed. Up to the present time, the laboratory has been sequestered, out of view. The laboratory has a large window facing the public underwater viewing area. The window has traditionally been covered in the past for security reasons, with a closded venetian blind blocking the view into it. Step 1 will be to always leave this window open, with the inside of the lab clearly visible to the public. Step 2 will be to develop a series of displays during 1999 as time and funds allow (including funds in addition to those from The sponsor), which will enhance the visitors' learning experience about the laboratory and what it does. Under corporate sponsorship, this is also an excellent chance to enhance the sponsor's image in the community. Devolopment of Science & Conservation Demonstration Material for the Public Each demonstration area or phase given below will be a specific area with a large sign giving the sponsor's role in the project. As more elaborate and interactive demonstration material is developed in the pases below, we will continue to devise new and clever ways to promote the sponsor's role in the program. Phase 1 - Hydrophone and Realtime Sonogram Display: This display will allow the visitor to see how a scientist examines dolphin sounds. As has been traditionally the case when the public has access to the area just outside our lab, the hydrophone will be set up for listening to the dolphins. In addition, a monitor will be present displaying a realtime sonogram of the dolphin sounds. In addition, there will be an Exploritorium-type button for people to push allowing them to look at a realtime sonogram of their own sounds to learn how the display works. When the button is released, it will revert to dolphin sounds on speaker and realtime sonogram display. If Sea Life Park wants it, the microphone can be acoustically linked to the tank via low-amplitude piezo-film (cannot get loud, so won't spook the dolphins), making it into an interactive system allowing the people to watch the dolphins' reactions to their sounds. This can be recorded in case anything of scientific interest occurs. Video: The second demonstration that can be set up is a VCR and TV, which can be located against the back wall in the underwater viewing area just outside the lab. Sea Life Park staff can play a short videotape describing the project. We can use the 13 minute professionally made Project Delphis video. There is also an approximately 7 minute video about the project in Japanese. We will also re-connect the old BHP-supported hydrophone to provide the dolphins' sounds to the public. A button can also be provided to hear the dolphins' ultrasonic emissions thorugh a bat detector (which compresses the ultrasounds down into the human hearing range). Sponsorship signs will be prominantly displayed. Phase 2 - Touchscreen: An in-air touchscreen is being sought from longtime supporter Carroll Touch (developer of the underwater dolphin touchscreen). We will program it to be an interactive demo in which the visitors can choose things from the screen that: a. tell them about the project, both its science and its conservation, and b. give them programs to run just like the dolphins use. They can play the part of the dolphin and get a feel for how the touchscreen in used in the research. The touchscreen will be programmed to also contain promotional information about the sponsor. Phase 3 - Touchscreen Interaction with Dolphins: We will devise interactive programs between the visitor touchscreen and the dolphin touchscreen. When taken together with the cooperative venture we have entered into with Ocean Road Runner high-speed internet service, this will raise the possibility of such terminals in the park, in waikiki, and anywhere education and The sponsor promotion are desired. For this phase, additional funds to match the sponsor's will be sought. Phase 4 - Demos for the Park: Demos for the Park will require planning and obtaining significant funding. The possibilities range from implementing phase 1-3 - type materials (above) in a kiosk in the park, to establishing a demo-typed lab in Whaler's Cove that has its own dolphin touchscreen in it for inter-tank dolphin interaction with the Splash U lab terminal, as well as any human terminals that might be desirable in the park as demos. These terminal areas will be well signed to show that the sponsor is sponsoring them. Sponsor promo material will also be programmed into what visitors see on their terminals. The plan is to raise additional funds, beyond the sponsor's sponsorship, to allow us to progress as far as possible in the phases outlined above. Phase 1 is inexpensive and could be initiated with a few hundred dollars. In the final form of phase 4, a cool, shaded resting area could On either side of the seating area would be a bigscreen TV video screen, one showing a camera feed from the Splash U tank, the other a camera feed from the Whaler's Cove tank. Hi Fidelity sound will also be played from each screen from the hydrophone in each respective tank. This concept inserts the visitors between the tanks, and allows them to eavesdrop on inter-tank communication, as well as watch what the dolphins in each tank are doing while it is occurring. They can become armchair scientists for a few minutes. This kind of demonstration will keep the crowds occupied and happy, yet allow the researchers in the Whalers Cove lab the privacy they need to do productive research. The demo can be updated and changed to incorporate interesting features of the research as they occur. Another demonstration could be a young Indian Hill Mynah bird who also has an audio video link to one of the tanks. Chances are this bird will pick up a lot of dolphin communication--interspecies communication without humans as one of the species! It could even serve as an information area if the collaborative developmental psychology paradigm involving babies at language-learning age and communicating dolphins is ever seriously pursued. Phase 4 in its fully developed form would by itself make Sea Life Park a world attraction and sponsors' names household words. The special case of a new demo lab at whaler's cove Earthtrust has been approached by curator Paka Nishimura to consider creating a demo lab in the ship Essex at Whaler's Cove. If this entire space is available, it could be an attractive addition to the park, introducing many new interesting scientific and educational possibilities. Earthtrust staff have evaluated this alternative, and consider it a good one. However, it can only be accomplished if the Splash U laboratory is retained as the lab where basic, uninterrupted scientific research can continue. The most valuable asset of the entire Earthtrust science and conservation program is its high scientific credibility. Maintaining the Splash U lab is necessary for safeguarding this most valuable asset. There is no real scientific dislpay right now at SLP. One of the themes of the new lab will be consonant with a theme in much business development in Hawaii now: high technology, the ocean, and the enviornment. The project presented here should make Sea Life Park a technical focal point for the world, and draw all kinds of P.R The touchscreen is a unique tool. SLP is actually quite fortunate to have people working on it there. People frrom all around the world are interested in our experiments with the touchscreen (resulting in TV programs, many visiting scientists, etc.), as well as our studies in other areas such the dolphin ring blowing culture at Sea Life Park (Marten et al. 1996), the wild spinner dolphin study (Marten and Psarakos, in prep.), and our research on the sonic predation hypothesis (Marten et al. in prep.). As for the scientific temporal relations between the two labs: the Splash U lab will be about 1 year ahead of the Whaler's Cove lab. Depending on what projects we want to put on display, may be less than 1 year.
Some of the Science and Demonstration Possibilities from the Whaler's Cove Demo Lab: The existence of two laboratories introduces exciting new possibilities in science involving dolphin-dolphin communication between the two tanks, co-operative experiments between the two tanks, and interesting and exciting scientific and educational possbilities associated with integrating the scientific work with the dolphin training and shows. A. A 2-tank acoustic link between the two tanks will be the first order of business. We have already done this with hard-wire methodology between the Splash U tank and the Nursery tank (Marten and Goyal, in prep.) This new linkage will be accomplished with radio-frequency broadcasting of the signals between tanks. 2-tank video link and touchscreen control can be added as they are developed. The public can see and hear what is going on in the other tank, can see and hear dolphins communicating to each other in another tank. This is unique.What we are about to do nobody has done before, and would certainly be of great interest to visitors to Sea Life Park, and should contribute to increasing gate revenues. B. The potential for combining the science from our lab with the food reward shows should be explored. Food reward means having SLP staff working with us. They can make a show of our experiment and people can see the trainer being part of the experiment. People can see real-time scientific experiments. We should be able to put this together in a way that Sea Life Park will like and in a way that it will really happen. Other Laboratories: The technological and methodological developments of the project could be expanded to other facilities , who can undertake pieces of theoverall research. Dr. Delfour is ready to do it in Europe, and she knows several places there that are potential collaboration sites. Conclusion: The only way a scientific demo lab like the one that will be made in Whaler's Cove can exist and have true sienctific credibility is if it is associated with conducting true cutting edge science. To put something on display we need time to do it so we need to keep the Splash-U lab to work on real scientific projects and work on the displays we are going to show the public. The lab at Splash U will provide all of this. 3. ADVANCED TOPICS IN SELF AWARENESS One of the scientific project areas that arises in the process of doing the human-dolphin communication research is follow-up work on advanced topics in dolphin self-awareness, the scientific area in which we have made the largest contributions to the scientific literature. Utilizing The Results Of Siganture Whistle Research To Conduct A Final Set Of Experiments To Finish The Documentation Of Self-Recognition In Bottlenose Dolphins A. Use of the signature whistle B. How do the dolphins use it? 1. Study of the realtion of whistles to various behavioral contexts C. Use Signature Whistles to Explore Self-Awarenss -- Background: the dolphins we work with are sometimes exposed to a mirror, and are also sometimes allowed to wath themselves of television. This, plus their normal life around reflective underwater windows, assures that they have had plenty of exposure to visual images of themselves. 1. Using the whistle collections above, run experiments using the Self-Awareness-Touchscreen Paradigm a. Teach dolphins an association between an artificial sound and a geometric symbol on the touchscreen b. Run the dolphins through the exercise of now letting them select visuals of various individual dolphins when played signature whistles, to see if they associate visual representations of inidividual dolphins on the touchscreen with their respective signature whisles. c. Test Phase: Present the dolphins with their own whistle, and, putting a visual of themselves amongst the choices, see if they associate their own whistle with a visual representation of themselves. D. As an interesting demo for SLP visitors, conduct the food reward experiment presented by Marten & Psarakos in their Response to Commentaries in the journal Consciousness and Cognition: let 2 dolphins be trained together, reporting whether a removable icon is on "my harness" or "the other dolphin's" harness, moving the harness until eventually the answer has to be given by looking in a mirrior. Functions of Self-Awareness With the final work done on the existence of self-awareness per se in bottlenose dolphins, the most interesting thing that can be done is to explore the functions of self-awareness. Self-awareness evolved for me-you behaviour in a complex social matrix, and allows the individual possessing it to devise behavioral strategies using other individuals as models, as well as using itself as a model for how others will behave. A. Empathy 1. Train the dolphins in a given task, such as a visual task 2. Handicap a dolphin, such as by having it wear soft rubber eyecups designed specifically for dolphin wearing comfort in research 3. Assess whether the other dolphins help the handicapped dolphin B. Deception 1. Train the dolphins in a co-operative task 2. Present the opportunity for individuals to cheat and get the reward themselves, instead of sharing it with the others, and see if it does cheat NOTE: 2-tank acoustic, video, and touchscreen linkage could make these experiments excellent ineed, as in I-C in the outline in the proposal. Dolphin-Dolphin Communicaton and Self-Awareness Marten and Psarkos (1994, 1995a) have shown that dolphins test postive in the experiments designed to test self-recgnition in primates. Additional research can now be done in the very cutting edge research area involving the functions of self-awareness rather than just its existence per se, such as empathy. We would like to explore whether empathy causes epimeletic behavior (Connor & Norris, 1982) by testing if dolphins help another dolphin whose vision is blocked with eyescups to accomplish a desired result. This might be similar to the kind of empathy hypothesized by De Waal (1996) when a chimpanzee consoles another chimp after a traumatic social encounter. Deception (co-op task exp's between the tanks): we would also like to extend this kind of research to include deception of the kind reported by Savage-Rumbaugh (1990) and discussed in the papers of Robert Mitchell. An example would be seeing if the dolphins cheat in tasks in which one dolphin is required to tansfer a reward to a second dolphin who is performing the task. These experiments are very interesting for people watching them: people can see dolphins doing something different , and they can see SLP staff also participating to these kinds of activities. 4. POSSIBLE EXPERIMENTATION TO TEST THE HYPOTHESIS THAT DOLPHINS MIGHT COMMUNICATE ECHOLOCATION TARGET RETURNS If there are two laboratories for a two-tank setup, and if we do have a working relationship with the training department at Sea Life Park, a simple elegant test can be conducted to test whether dolphins can communicate echolocation target returns. It is the same paradigm desribed in the text of this proposal having the dolphins in tank A report what they see over an acoustic link to the dolphins in tank B, and they are all rewarded if the dolphins in tank B give the right answer. The task can be structured to require more and more detail, and final tests given that would vitrually require using the target return to get the right answer. Whether a target return was used to communicate the answer could be investigated using the paradigm that Aubauer and Au (1998) present in their phantom echo generation research. |
||||||||||||||||||||||||||||||||||||||||||||||||||||