The term “technological determinism” is believed to have been coined by American economist and sociologist Thorstein Veblen. Taken on its own, determinism is the doctrine that all events occur as a result of a direct, prior cause. Accordingly, at the broadest possible level, technological determinism is the conceptual framework in which social, cultural and historical phenomena are primarily the outcome of technological causes, in particular technological advancement.
As with many philosophical views, there is no single, universal definition. And as Bimber puts it, “Until we are able to agree about what precisely we mean by this concept, we are unlikely to resolve the question of whether technological determinism is a useful lens through which to interpret history.” 
The various positions within this school of thought have been vaguely distinguished into two versions characterising the intensity of the role technology plays in determining society. Proponents of the “hard” view hold every social activity is solely and directly driven by technology, which is itself an independent force. The assumption is that technology undergoes perpetual advancement according to some given logic and specific stages of technology development give rise to specific social change. This change is independent of society’s desires and preferences. This theory is neat and tidy, but it is also not widely accepted because of its lack of flexibility and generality.
The “soft” version accommodates this exception. It argues that technology is still the fundamental force for effecting social change by creating conditions favourable for certain social activity to take place. However, this “disposition” or “tendency” could be rejected with the collective conscious will of society. Hence it distinguishes itself from the “hard” version in that it allows for the possibility of human intervention.
Bimber goes against the grain from this standard model of classification and identifies three common interpretations of technological determinism. First is the “normative account”, which argues that efficiency and productivity have become the primary ends in which technology is advanced and which society functions, excluding it from the human domain and its associated elements.
The second interpretation is the “nomological account”. Supposing that the scientific laws are discoverable in a sequential order, technology, being the recombination of existing artefacts and the manifestation existing scientific knowledge is also subject to this trajectory. For example, it is unlikely that motor vehicles would be built without the engine having been previously invented. More importantly is the extrapolation of this mechanistic train of thought to the assertion that certain technological developments necessitate a particular form of social organisation and subsequent technological development such that an inevitable path ensues. In this view, only a narrow definition of technology as artefacts is admissible. Broad definitions, such as technical practices bring the human dimension back into the picture.
The third interpretation draws attention to the fact that the effects of technological development are sometimes impossible to predict even among participants key in the development. On this account, one could argue that technological development is at least partially autonomous because it has outcomes that are independent of human control. Here the idea of technology as an autonomous agent of change is brought into play because for technology to be given credit for having brought about social change, technology itself must first be independent of social causes. Otherwise you end up with a theory where society influences itself through indirect means (technology) 
Bimber advocates Cohen’s criteria for judging the three interpretations, namely that it should be both technological and deterministic. On this basis, Bimber argues that only the nomological interpretation stands up to both tests. The normative account is not technological because the root agent of change is social norms rather than technology. It is not deterministic because this preoccupation with efficiency and productivity is itself a product of a varying culture and time, and not entirely external to men’s locus of control. Bimber claims that the unintended consequences interpretation also fails because it amounts to indeterminism. Here I disagree. The uncertainty principle pose a fundamental limit on the accuracy of measurements, rendering accurate predictions impossible. But the concept of causality is still an incontrovertible axiom, and the universe may still yet be deterministic. In any case, Bimber points out that the theory does not attribute the nature of unpredictability to an inherent quality of technology (it may be the fallibility of the human intellect that fails to predict the consequences), and thus it fails on that basis.
This criteria does not admit the distinction between “hard” and “soft determinism because the “soft” version also fails on both accounts. Thus, if we were to admit Cohen’s criteria, we would be in a very tenuous position. Intuitively, technology does not exist in a vacuum and social forces simply weigh too much on the development and adoption of technology to be dismissed so simply.
Why then should we accept his definition of technological determinism? Bimber suggests that it is attractive “on grounds of semantic clarity and integrity” and consequently “this standard of clarity is surely more useful than one in which the theory of change is hidden by less precise language”. Just as a peanut is neither a pea nor a nut, this line of argument could be construed as pedantic and splitting hairs. At the very least, it is unhelpful for furthering the discussion. Thus for the purposes of breaking this impasse, I propose that we momentarily abandon this hardline approach to avoid throwing out the baby with the bath water. Rather, we will take the pragmatic approach and examine the more mainstream “soft” determinism put forward by Heilbroner. I am not debating whether this is in fact technological determinism, I am merely supposing that it is, and examining how good it is as a tool for elucidating twentieth century history.
Heilbroner’s view can be summarised into two fundamental tenets. Firstly, technology develops in a determinate sequence of steps. Secondly, that this advancement of technology effects a social “evolution”. Heilbroner  makes his case for the first proposition with three pieces of evidence. Firstly, the “simultaneity of invention”. In many instances, independent parties have made similar discoveries within a short span of time of each other. This is the case for the invention of calculus and the telephone. The existing knowledge and prevailing technical expertise provides the basis and foreshadow of what is to come. Hence the discovery about to be made is merely the inevitable next logical step; something that must be obvious or at least imaginable to more than a single individual. This brings us to the second and third point which is the “absence of technological leaps” and the “predictability of technology”. Heilbroner makes the point that technology advances at its frontiers (direction) in small incremental steps (magnitude) that seems to make progress appear predictable.
Heilbroner limits his discussion of the second proposition to the influence of technology on the political and economic aspects of society. He suggests that a certain technology such as factory automation imposes a particular characteristic constraint on the makeup and organisation of the social structure. He proposes a normative argument where economics is the intermediary which links technology and society. He explains that with the principle of economic maximisation presupposed, changes in technology are brought to bear on the economic system, and thus the effect on social order can be quantified and analysed. But he rightly concedes that the extent to which technology affects society sociologically are more difficult to ascertain.
How useful then is such a theory of technological determinism? Firstly, technology has meaning and purpose only if thought in relation to humans. Hence technological determinism is especially useful in the twentieth century because of technology’s increasing influence. It ties two important strands in history together and actively brings regularity and order in history in a simple and general manner. That certain civilisations undergo periods of technological stagnancy and even regression does not undermine technological determinism though such societies may be less open to systematic analysis than when it is guided by an economic imperative. In fact as Edgerton suggests, it seems that a lack of innovation in technologically-backward societies point towards the limits technology imposes, and thus it could be understood as being more determined. 
Secondly, technological determinism is useful because it is in some sense more enduring than social institutions. This is similar to the practise for historians to infer the lives the ancients led based on the tools they used and the structures they built. Cultures and social conventions are temporal and evolve with the times. But technology provides a background with which to put matters in context, because it appears to travel in a fixed path, especially in hindsight. So while technological determinism may not be useful when viewing the twentieth century from the twenty-first, it certainly might in the thirtieth.
This view is elaborated by Misa who states that the longer the time frame, and the broader and more aggregate the topic in question, the more technological determinism appears plausible. Those who conduct studies in the micro-level will find a variety of societal forces and agents at work obscuring the importance of the role of technology.  To illustrate this argument, society can consciously reject a technology due to cultural preferences. But while this affects the rate of adoption, it is possible to make a case that this is in fact merely delaying the inevitable. The underlying assumption is that the cultural context changes but technological benefits are absolute. As long as mankind is composed of rational beings, social evolution will eventually favour the adoption of the beneficial technology. This is applicable for example in the study of genetically-modified foods.
Lastly and perhaps less importantly, technological determinism is useful insofar as it is simple. Therefore, despite obvious difficulties in the theory, the idea is prevalent in popular imagination and political rhetoric. It is useful to think that the internet gave rise to an information revolution, or that the contraceptive pill gave rise to a sexual revolution, particularly for the masses who if such abstractions did not exist, would be ignorant of these issues at all.
Nevertheless, several problems persist. Perdue uses several agricultural examples to demonstrate that technology itself does not account fully for the variances in historical outcomes.  In fact, applying social factors in single-factor models are also not satisfactory. In theory, we would like a theory in which social and technical factors are not only independently considered, but the mutual interaction, reciprocity, feedback, and equilibrium of the evolving relationship emphasised. This well summed up by Winston Churchill, “We shape our buildings; thereafter they shape us.” But such a theory is difficult if not downright impossible to formulate. In Heilbroner’s words, “A recognition that the technological structure is inextricably entwined in the activities of any society does not shed light on the connection between changes in that structure and changes in the socioeconomic order”.
Misa’s exposition on the difficulty of constructing a “middle-level” approach offers us a solution. He suggests that macro studies resist integration with the micro-level because its definition as an independent force rejects the effects of micro-forces. In micro-studies the multitude of agents involved swarms the obvious importance of technology; sometimes it is simply omitted from the narrative. This parallel to entropy that offers us a way out. Entropy is useful as a macroscopic empirical principle, but as far as we know it does not result directly from any of the underlying physical laws such as quantum dynamics. Therefore, as a scientific model, technological determinism is merely a model, a simplified, reduced representation of the vagaries of a complex underlying reality. As a partial theory analogous to relativity or quantum mechanics, technology determinism accounts for many of the observed social change in the twentieth century, especially when considering the major themes of industrialisation, urbanisation and the growth of wealth. To give technological determinism any less credit, or to expect it to be the sole basis of all human activity without exception would be overly presumptuous.