Nozick's 1971 article "On the Randian Argument," originally appeared in The Personalist (the house journal at USC which later became Pacific Philosophical Quarterly) and was subsequently reprinted in his collection Socratic Puzzles. It initiated a series of other articles, including contributions by past ARS presenters, Tibor Machan, Douglas Den Uyl and Douglas Rasmussen. While Nozick's article surveys a number of objections to "the Randian argument," the SEP entry focuses on one, and accurately names the most relevant Objectivist response to the objection:
Critics raise two objections to this argument. (i) It begs the question by assuming what is at issue, namely, that a non-living entity cannot be harmed (Nozick 1971). Unlike the robot of this example, real robots can be damaged or destroyed, not only by external events, but also by a failure to perform their functions well, that is, by their own actions or inactions. Hence they can, quite straightforwardly, be said to have values. (ii) Even if one were to accept that the concept of value entails the concept of life, one could consistently regard one’s survival as a means to a certain kind of life: a life of dedication to the greater glory of God, the common good, the environment, and so on (Mack 1984).
[...] The need for morality, according to Rand, is dictated by our nature as creatures that must think and produce to survive; hence we would need morality even on a desert island. There is, however, no duty to survive; morality is based on a hypothetical imperative: if you choose to live, then you must value your own long-term survival as an ultimate end, and morality as a necessary means to it. [...] If asked why the choice to live commits you to your own long-term survival rather than some other ultimate end (such as, for example, the greatest happiness of the greatest number (Nozick 1971), or becoming worthy of eternal life in heaven), the answer is: because any other ultimate end, if consistently adhered to, would lead to death.
I recently discovered that in 1977, Harry Binswanger, a longtime associate of Ayn Rand and then an adjunct assistant professor at Hunter College CUNY, wrote a letter to Nozick responding to this objection. Binswanger's letter fleshes out the response at the end of the passage above and connects it to his own scholarly work on teleology. He provides the following backstory:
After I mailed the letter to Nozick, some weeks passed without a response from him. So, I phoned him about it. He said he had received the letter, and that he had added it to a pile on his desk of papers that he meant to read. But, he added, it was a big pile and he might never get to my letter. He was quite open to hearing a rebuttal and even encouraged me to publish on the wider topic, which I eventually did (The Monist, January 1992).
With Binswanger's permission, the ARS is reproducing the text of his 1977 letter (PDF). I think it is of interest not only as a historical document, but as a substantive contribution to understanding Rand's actual position. Here is the most relevant segment of his reply, which addresses the question of why any other ultimate "end" disconnected (or alienated) from the needs of life could be a "value" or "goal" in name alone:
The fact that one's life must be sustained by a specific course of action to obtain needed items makes possible the concept of value, and all evaluative concepts. Without implicit or explicit reference to the alternative of life or death, the concept of "value" cannot be formed. Without such reference to life, there is no means of distinguishing between an action's goal ("that which one acts to gain and/or keep") and its mere effects. Goal-directed action itself cannot be conceptualized except by reference to the goal's beneficial relation to the agent's life. This is the point illustrated by means of the example of the immortal, indestructible robot. Say, for instance, the robot is constructed in such a way that it goes around gathering stones and putting them into piles. Is this goal-directed action? No, because how would this result of its action be distinguished from any other result (e.g., leaving footprints when it walks)? (We must abstract, of course, from the goals of the robot's designers.)
You are probably familiar with Hempel's example of the two concomitant effects of the heartbeat: circulating the blood and making heart sounds. On what basis can we say that the former is the heartbeat's goal or function, but not the latter? Partly on the grounds that blood circulation is beneficial to the organism's life, but the heart sounds are not.
Binswanger's thinking here about the distinction between goals and mere effects is no doubt reflective of the teleological theory he developed in his doctoral dissertation, which was later published as a book, The Biological Basis of Teleological Concepts (1990) and summarized in a pithy article for The Monist (January 1992), "Life-Based Teleology and the Foundations of Ethics."
I've long thought that Binswanger's work on teleology forged an important integration between Rand's value theory and the philosophy of biology, especially insofar as it reflected and even anticipated a variety of developments in mainstream work on teleology. His work is recommended not only to readers looking to understand this response to Nozick, but also the deeper aspects of Objectivist metaethics.