Kinematic Self-Replicating Machines

© 2004 Robert A. Freitas Jr. and Ralph C. Merkle. All Rights Reserved.

Robert A. Freitas Jr., Ralph C. Merkle, Kinematic Self-Replicating Machines, Landes Bioscience, Georgetown, TX, 2004.


5.1.9.E Passive Parts

A passive part represents the most primitive component that is handled by the replicator, as it replicates or manufactures non-self product. It possesses no onboard power or control. Examples might include carbon dimers if the replicator uses dimers as a principal carbon source for mechanosynthesis [2323], simple mechanical struts and gears if these are found in the replicator’s environment, and so forth. The simplest passive parts are the most primitive components of which the replicator is constructed, and which the replicator is capable of manipulating for purposes of fabrication or assembly.

E1. Passive Constituent Parts Count. Refers to the total number of primitive passive parts which the replicator must individually manipulate in order to make a copy of itself. If the simplest parts used are atoms, as in molecular manufacturing, then the “parts count” becomes simply the number of atoms in the replicator structure, or “atom count.”

E2. Passive Constituents Parts Scale. This item refers to the scale of the most primitive passive parts from which the replicator can be said to be comprised, and which the replicator must individually manipulate in order to make a copy of itself.

E3. Passive Constituent Parts Types. This item is intended to be a quantitative count of the number of distinct parts comprising the replicator. Of course, the number of distinct kinds of parts is not the sole measure of replicative complexity, just as reduced instruction set (RISC) processors are fundamentally simple but can still carry out very complex tasks. A nanodevice which replicated itself by manipulating atoms of all 92 natural elements could be said to have only 92 different “parts” (atoms), representing a fairly modest amount of parts diversity. The precise definition of “part” clearly depends on context. The Merkle-Freitas assembler (Section 4.11.3) may contain at most ~100 different parts types, and a modern jet aircraft contains more than 200,000 unique parts, each of them interacting with an average of 3-4 other parts [2523]. But in another view, for the case of the hydrocarbon nanofactory or molecular assembler (e.g., Section 4.11.3), as few as two “part types” might be said to be required (i.e., acetylene molecules plus one type of “vitamin” molecule).

E4. Number of Vitamin Parts. Vitamin parts are passive parts that cannot be manufactured by the replicator and must be provided from the outside in order for replication to be completed. In a “vitamin architecture,” key materials (physical components or parts) essential for replication are exclusively provided from without (Section 4.3.7). Notes biochemist Michael Adams [2524]: “Obviously, no life-form survives in isolation from its surroundings, but organisms vary considerably in their dependence upon their environment. Thus, humans require at a minimum ten or so amino acids, various minerals, an array of biological cofactors (vitamins), and a continual supply of O2 gas. Perhaps surprisingly, these same materials are also required by many microorganisms, although they typically differ from us in their ability to synthesize most, if not all, of the twenty amino acids as well as many, if not all, of what we term ‘vitamins.’ Like us, the vast majority of microorganisms require a fixed carbon source, which is usually a carbohydrate of some sort, although in some cases lipids, nucleotides, or various simple organic compounds are utilized. In contrast, some microorganisms are intensely dependent upon their environments. For example, some microbial parasites do not synthesize any amino acid or lipid, and only a few enzyme cofactors and nucleotides; rather, they obtain all of these compounds from their host.” The need for essential vitamin parts can serve as a valuable replication safety mechanism. For instance, one species of therapeutic gut-dwelling genetically engineered bacteria [2525] has had its gene for thymidine synthesis removed. The modified bacteria can survive only in the gut where thymidine is naturally available; outside the gut where thymidine is not available, the altered bacteria die in just 72 hours. One common safeguard used to provide total control over artificial cells is to make their lives dependent on chemicals that do not exist in the environment – withdrawing the critical chemicals would result in the death of the cells, particularly if they should escape into the environment. [3107]

E5. Replicator Nutritional Complexity. How many different sorts of parts can the replicator eat, and still successfully proliferate? This includes two closely related issues [414]: (a) minimal nutritional complexity, and (b) nutritional diversity or flexibility. The small-genome 0.17 mm3 marine bacterium Brevundimonas diminuta can only grow on a few kinds of sugars and amino acids [2526]; C. oligotrophus eats primarily hydrocarbons that number many in type, though the precise complexities of its metabolism are currently unknown [2527]; while the ~1 µm3 E. coli can consume very many substrates and has 285 [2528] or more (considering unknown sequences) different transport systems to accumulate them. M. Krummenacker [414] notes that the champions of minimal nutritional complexity in the biological world are probably the photoautotrophic cyanobacteria [2529], whose energy source is light and whose sole carbon source is CO2 – though these (and other microbes mentioned earlier) also require a number of trace minerals including metal ions, phosphorus, nitrogen and sulfur in addition to the carbon source. Genome size appears related to nutritional complexity [2527].

E6. Required Parts Preparation. According to Chang [793]: “Like other robotics researchers, Dr. Rodney Brooks, director of the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology, predicts the development of robots that assemble themselves, so to speak, out of ready-made parts.” Using pre-made parts is not “cheating”: Virtually all known replicators – including human beings – rely heavily on input streams consisting of “premanufactured parts” most of which cannot be synthesized internally. Indeed, the development of artificial replicators that do not require premanufactured parts seems unnecessary, more technologically challenging, and quite possibly a threat to public safety [2909]. Notes Merkle [793]: “Should we relinquish autonomous, self-replicating devices that can function in a natural environment? The answer is yes, that looks like a fine thing to relinquish.” And replication from more disordered starting materials is harder work and thus likely harder to engineer. If the length of its genome was an approximate measure of the difficulty of an organism’s function, then we might infer that the autotrophic flowering plants, many of which have surprisingly large genomes compared to animals of comparable size (Table 5.1), must have much greater difficulty manufacturing organic nutrients from less organized matter than do animals who must catch prey for food [20]. However, the situation is not so clear because certain DNA sequences are repeated hundreds or thousands of times, and apparently only a small fraction of the nuclear genome of plants with large genomes is transcribed – e.g., Robert B. Goldberg of UCLA estimates that only 1-2% of the tobacco nuclear genome is transcribed, based on hybridization of RNA to genomic DNA [2530].

E7. Passive Parts Complexity. A crude single quantitative measure would be the average descriptive complexity per part – e.g., the bits required to describe each part at the blueprint level of detail, summed over all part types, divided by the number of parts types. Alternatively, this measure could be reported as a histogram, e.g., showing the number of parts as a function of bits per part. A relative measure, such as the ratio of the average parts complexity to the total replicator complexity, might also be useful. Notes Aristides Requicha [2531] of the USC Laboratory for Molecular Robotics: “When you replicate something, the complexity of the components you are allowed to use is very important. Suppose, for example, that you replicate DNA by using a long strand as template to ‘catalyze’ the ligation of two others that are pretty complex. You may call this replication, but the problem is going to be in getting the two component strands. Where do these components come from? In other words, the final act of replication may not be the most interesting part of what is going on, it may be kind of trivial compared with making the components.”

E8. Parts Precision. This item refers to the precision with which parts must be fabricated (e.g., minimum feature size), as distinguished from the overall size of the parts themselves (E2). See related dimension E11.

E9. Multiple Utility of Passive Parts. See also B12 and F7.


E10. Parts Presentation. Do parts come to the replicator for assembly, or does the replicator go out and gather parts? Does the replicator have independent mobility? See also F8.

E11. Mode of Dimensional Specification. Digital specification of replicator parts is possible using such materials as LEGO® blocks, molecular building blocks, or especially individual atoms during molecular manufacturing. Analog specification relates to bulk machining and similar non-modular manufacturing processes. See related dimension E8.


Last updated on 1 August 2005