This is an electronic version of a poster presented at the 5th International Cognitive Linguistics Conference at Vrije Universiteit, Amsterdam on July 15, 1997.
One day, I overheard a fellow graduate student say that he was not interested in studying 'soft linguistics.' Although I'd never heard the phrase soft linguistics before, I could easily understand both what he was referring it and his attitude toward it: by using soft , he expressed his disdain for such topics as historical linguistics, sociolinguistics, and areas of semantics relying on 'folk definitions', and by using this phrase, he also implied a preference for 'harder' areas of linguistics, in particular those involving a high degree of formalism , especially through the use of mathematical or logical models, e.g., phonology and some brands of formal syntax and semantics.This chance remark started me wondering how hard and soft, adjectives whose basic senses describe the consistency of physical matter, had come to be used to describe sciences. I also wondered how hard had come have such a positive connotation and soft such as negative one.
The answers reveal something about metaphorical extension as well as cultural attitudes.
The student surveys and the leaner's dictionaries show two basic meanings for hard in modern English.For modern speakers, these senses are distinct, and they have two different antonyms: the antonym of hard1 is usually soft while the antonym of hard2 is usually easy.
- hard1
- (of surfaces or materials) firm, unyielding, solid: a hard bench, hard ground
- hard2
- difficult to do, requiring a lot of effort: a hard problem, a hard job
At the same time, the senses are clearly related. Just as hard surfaces and hard materials resist physical efforts to deform them, hard problems resist attempts to solve them and so require a lot of effort. Likewise, hard work is work that involves a lot of effort.
Both hard1 and hard2 are listed among the earliest senses of hard in the OED.
The basic meaning of soft in modern English seems to be the opposite of hard1: "yielding to the touch; not firm." Some examples of this use include soft groundandsoft leather.However, another important use of soft is describe things which are pleasant to the senses. As one student describes it, something that is soft is "gentle, smooth, relaxing, soothing and quiet." Examples include soft fur, soft breeze, and soft voice.
Many people feel a clear connection between these two senses; as one student wrote, "Soft is squishy; a soft object has a lot of give and is pleasant to the touch."
Although this characterization ('stubborn', 'raw') may sound rather negative, in fact, the phrase has rather positive connotations in most situations. Fact, as opposed to opinion, is something that is true, something that can be believed in, something that can be trusted not to change, and hard facts are the most fact-like of all facts. Hard facts are objective and thus must be accepted as true by anyone who is not soft headed. The positive connotations of hard facts indicate the value placed on objectivity and certainty in Western culture.
The prototypical hard facts reveal something about the nature of 'knowledge' in English speaking societies. The best facts are facts which are discovered through empirical observation, those which can be measured or quantified in some way and which can be verified by other observers. Opinions, which are subjective and based on a single person's experience, are distinguishable in this way from facts (unless, of course, the opinions are measured and quantified in poll, in which case, they might be taken as a kind of fact, although probably only as a 'soft fact').
Hard data are data based on hard facts, that is, data that can be measured and analyzed mathematically. The opposite of hard data is naturally soft data; however, there are many different types of soft data.
The soft sciences, in contrast, take as their subject matter things that are not easily studied through controlled experimentation or mathematical models. For example, history describes events in the past which cannot be repeated at will, and often historical data is not easily subject to mathematical modeling either because it is too vast or incomplete or both.
As another example, psychology is considered a relatively soft science because it is often impossible to do controlled experiments on human subjects, and because much the subject matter involves the mental processes which cannot be directly observed and which are often qualitative rather than quantitative in nature. (Imagine, for instance, a study of depression among college students. Depression is hard to define or measure in quantitatively). Subfields such as cognitive psychology, which can use controlled experiments and which often deals with quantifiable data (e.g., reaction time, number of correct responses) are considered 'harder'.
Within linguistics too, some subfields are felt to be harder than others. From certain perspectives at least, theories of syntax which rely on judgments of grammaticality, (which ideally involve a simple yes/no choice that can be independently verified) , are considered 'harder' than sociolinguistic theories which deal with topics such as speakers attitudes toward other people. (From other points of view, however, grammaticality judgments seem quite 'soft' in that they are subjective and, in fact, often are not verified by other native speakers. In such as view, corpus data is 'harder.')
First, it is used in the Amercian news media to describe a kind of political contribution. I first heard it used during the two years leading up to the 1996 U.S. presidential election. Under U.S. law, there is a restriction on the amount of money any one person, company, or PAC ('political action committee') can contribute to a political candidate in one year; however, there is no such limit an the amount that can be contributed to a political party. Soft money is money that is contributed to a political party allegedly for general party activities, but which is actually is intended to be used and is used to pay the campaigning costs of a particular candidate. In other words, soft money is 'sneaky' money contributed to candidates indirectly via a a legal loophole.
Political soft money is 'soft' in the same ways that soft data is soft. Soft money is hard to quantify: since there is no limit to the amount that can be contributed to a party, no record need to be kept on such general donations, making it difficult to confirm how much was received from a particular source. It is also hard to pin down the exact uses to which the money is being put; the distinction between promoting the general interests of a political party and promoting the interests of a candidate of that party is a qualitative distinction that cannot always be clearly made.
Second, the phrase soft money is also used in the academic world to describe money from received as part of research grants (in contrast to money received as salary). A researcher at a university or research institute whose position is funded mainly or entirrely through grants is said to be 'on soft money.' This kind of funding is 'soft' in that it is short-term (not enduring) and often not dependable.
The phrase hard money is an older than soft money, originally with the same meaning as hard cash (see hard8 above). Now, however, it is apparently acquiring new meaning parallel to the two meanings of soft money, being used to refer to political donations that are actually used for the purpose they were supposedly raised for or to money that is received as salary rather than through research grants.
In these phrases, hard has the meanings of hard4 and hard5 above, and it has strongly negative connotation. Hard drugs are drugs which have intense effects and which are for that reason considered dangerous: the prototypical drugs are also highly addictive, probably because long term drug use can result in damage to the health and because the craving for drugs may lead the user to commit crimes for drug money
The use of soft with drugs seems to result from a parallel with hard. Soft drugs, such as marijuana, are the 'opposite' of hard drugs in that they usually considered much less intense and much less dangerous. For this reason, the connotation of soft drugs is only slightly negative if at all. Some people, of course, consider soft drugs are marijuana to be totally harmless or even helpful, and for such people, the phrase may even have positive connotations.
The use of hard and soft to describe drugs is probably at least partially motivated by the use of hard to describe alcoholic drinks (hard cider, hard liquor, hard drink). Hard drink is intoxicating and sometimes addicting and harmful just as hard drugs are. This use also has an opposite in soft, namely soft drinks (non-alcoholic drinks).
The use of hard and soft to describe pornography seems to be an extension by analogy from the uses with drink and drugs. Hard pornography is more intense than soft pornography in that it is more explicit and/or violent; hard pornography is also believed by many be dangerous, by producing anti-social behavior in those who view it. Punning is probably also involved; as one student wrote, in the more explicit hard porn, you are more likely to see hard penises. Soft pornography is less explicit and less violent, and thus is felt to have weaker effects. As with soft drugs, the connotation of soft pornography is not especially positive, but it is much less negative than hard pornography.
Hardware itself, referring to the machinery of a computer, is an extension from an earlier meaning of 'machinery or equipment for a specific job' (e.g., military hardware). Probably because the hardware of a computer is useless without programs to tell it what to do, software was coined to refer to the other essential "part" of the computer. (Software seems like a better choice than the term which could be coined from the other opposite of hard, namely 'easyware.' 'Easyware' would not sound appropriate because computer programs are designed to solve hard problems. It might be appropriate, however, for a product that was especially easy to use.)
Another spinoff from hardware is in the phrases hard wired and hard-wiring. If a particular operating instruction is hard-wired into the computer, it is built into the mechanism (the hard wiring) of the computer itself and cannot be altered by the instructions of software.
Hard wired and hard-wiring have been extended to the fields of biology and psychology, where they are used to refer to instincts which are believed to be built into the structure of the brain rather than being learned through experience. This extension, of course, is made possible by the metaphor of the human brain as a computer, one that has become so commonplace that talk of the brain's hard-wiring is no longer felt to be an obvious analogy. In fact, I have found examples of soft-wiring used analgously to this sense of hard wiring , referring (I think) to systems of the nervous system which develop after birth. (Here's an example from an article in the Chicago Tribune on aging and paralysis:"Reversing such paralysis has always been thought to be impossible. But not anymore asscientists figure out ways to fix the brain's plumbing and soft-wiring, hoping to keep it operating at top efficiency." )
Finally, because computer software in a sense exists "inside" computers in electronic form, soft and terms containing soft are now often used to refer to things which have a 'virtual' existence. For example, a soft copy of a document is one that exists only in electronic form, as a web page or a word processing document saved on a disk. The opposite of a soft copy is naturally a hard copy, a copy of the document which is printed out on paper. This is quite an interesting use, since paper is not itself considered hard; a paper copy of a document is only hard in comparison to the electronic copy; because it is tangible, it seems to have a more enduring existence than the electronic version which appears to vanish when the computer is switched off.
Sometimes this sense of soft is found in phrases which already have other meanings. For example, although the main use of soft data seems to be that described above, I have found at least one example, a report on a fishery managment project, in which soft data seems to simply mean 'data stored in electornic form.' The data in this case is clearly 'hard' in that it is quantitative data whose accuracy is not in doubt. Another example involves the term soft-wiring used to describe a lighting system in which the lights can controlled by a computer. The soft apparently refers to the software whch runs the computer that controls the system.
The connotations of hard and soft, whether positive or negative, reveal a lot about social attitudes toward particular topics. The strongly positive connotations of hard science and hard data, for example, reveal the high value placed on objectivity in Western society.
| To the Top | To Victoria Muehleisen's Home Page |