A number of points have been raised in this exchange that seem to call for further comment. 1. Re Adam McHugh's second contribution;: I agree that. until recently many, probably most, economic theorists, and members of the general public, took technology for granted. (We have a long section in Chapter 1 of our book suggesting reasons for this neglect.) The exceptions to are mainly found among economic historians, such as Mokyr and Landes, and general economists who have concerned themselves explicitly with technological change, such as Schumpeter, Nate Rosenberg and Chris Freeman. One interesting observation along the line of his interest in text book treatments is that few such books bother to define technology even when they discuss it. For example, in the Index to the 10th edition of my UK introductory text book (Lipsey and Crystal 2004) "technology" has 23 lines of references but NOT ONE for a definition. Neither does a definition appear in the 17 page glossary of terms! I now begin to wonder how many of those who have done major empirical and theoretical work on technology have bothered to define what they were studying. (Thanks for setting me a minor research task. I will explore this interesting question in the near future.) I agree with Mc Hugh that technology is implicitly defined in most texts in the way he states and that this definition probably stems from Hicks' treatment in Value and Capital (although, of course, it may have earlier antecedents, and it is not the definition my colleagues and I use). 2. To Robin Neill: I accept, and use, Marshall's terms "short run" and "long run" when doing comparative statics with technology constant. But my colleagues and I model continuous movement driven by endogenously generated technological change. For this reason, we are careful to use the term "long term growth" and never "long run growth". However, I disagree with Neill when he says "Once these [technologies] change we are in economic history not economics." I agree that for a full understanding of the causes and consequences of technological change, we need to know a lot of history. But it is also possible to theorize about such changes in what is recognizably "economics". (For examples, see the models of endogenous growth by Romer and Lucas, the models of General Purpose Technologies (GPTs) in the book edited by Helpman, and the models in Chapters 14 and 15 of our new book,) 3. To Donald Coffin: Yes, it is customary to follow Schumpeter in distinguishing invention, innovation and diffusion. But, contrary to his view, these are not related linearly, but in a system of mutual interdependence. For example, new technologies rarely, if ever, diffuse in unchanged form. As they move to new uses and to new places, they typically have to be altered to fit the local situation leading to further invention and innovation. 4. To Prabhu Guptara and Sameul Bostaph, let me repeat the first line of our definition: "Technological knowledge, technology for short, is the set of ideas specifying all activities that create economic value." Such knowledge gives the specifics of the goods and services that can be made, how to make them, and how to organize their production and distribution. In common with most theorists of the subject, we distinguish between technological knowledge and the things in which that knowledge is embodied, such as plant and equipment, infrastructure, firm location and organization. These are elements of what we call the "facilitating structure". Given new technological knowledge, no alteration in output and hence economic value is possible until that structure has been altered. Typically its various elements adjust with different time lags. The great advantage of making this split between technological knowledge and the artefacts that embody it and are used to create economic value is that it resolves the "Solow paradox". If one works with an aggregate production function with a form such as Y =Af(Labor, Physical Capital, Human Capital), changes in technology and changes in Total Factor Productivity (TFP) are identical; you cannot have one without the other. This gives rise to an apparent paradox when changes in one are observed without corresponding changes in the other. But given that new technological knowledge typically becomes embodied in the various elements of the facilitating structure with various lags, some of which can be decades long (as was the case when embodying the unit drive electric motor in factories and reorganizing those factories to fully realize the potential of this new type of machine), there is no paradox. For example, accelerating technological change can be combined with decelerating changes in TFP until the facilitating structure has been well adapted to the new technologies. Indeed, as well as arguing this verbally, this is what sometimes actually occurs in our mathematical models of endogenous, GPT-driven growth. 5. Pat Gunning poses an interesting question as to how far one should go in defining technological knowledge to include knowledge that covers activities beyond the sphere of production (product, process, and organization). As Tony Brewer notes, the other activities mentioned by Gunning are important but are not usually included in studies of technology and technological change. Our definition includes knowledge of all of these activities. But definitions are our servants, not our masters and for some purposes the knowledge that contributes to these other activates might well be treated separately, rather than as one category of 'knowledge of how to create economic value' as we do. In practice, we spend most to our time on those activities that design and produce goods and services rather than those that distribute and sell them. Richard G. Lipsey