a high-context negotiator's nightmare is loss of face. As listed above, there are many ways in which this might happen, and he or she will do everything in order to ensure that it will not happen. A high-context negotiator prefers to take as much uncertainty as possible out of the picture. Even failure to reach an agreement can result in loss of face, so he or she will try to foresee and plan every aspect of the ensuing negotiation in order to prevent failure. Cohen lists China and Japan as two different strategic examples for how this is done. Many Japanese negotiators engage in extensive information gathering, so that they know the positions ahead of time and can then adjust their own position to what they think will be realistic. Chinese strategy is the exact opposite. They make sure that the other party is quite aware of nonnegotiable positions ahead of time. They will only come to the table if these terms have already been implicitly accepted. Most likely, they will not come to the table if they think there is too much potential for humiliation and loss of dignity. The common denominator between these two positions, according to Cohen, is "avoiding a leap into the unknown" and diminishing the possibilities for loss of face ahead of time.

If we pander to China's desire to save face, we limit all functional options in curtailing the threat that they pose. This is an almost self-fulfilling prophecy in that China presets these disadvantageous conditions specifically to guarantee itself a more favorable outcome.

In October 2003, however, Motorola abandoned its MOS-17 plant, into which it had already sunk $1 billion, swapping it to SMIC for a 10% share in SMIC -- a deal that market-watchers estimated was a loss of about 90 cents on the dollar. In February 2005, Motorola sold its SMIC shares expecting to raise about $115 million.

Since 1986, the technology gap between U.S. and Chinese semiconductor manufacturing capacity has narrowed almost to zero. The current industry standard semiconductor fabrication dimensions are now around 0.18 and 0.13 micron line-widths, and Chinese wafer-fabs already produce DRAMS with these design rules. The current U.S. state-of-the-art is now 0.09 microns -- or 90 nanometers -- and at least one Chinese fab is said to be installing a 90 nanometer production line now. U.S. semiconductor manufacturers are now working on 65 nanometer design rules -- in concert with a French fab.

In February 2005, the Defense Science Board issued a report on "High Performance Microchip Supply" which -- to me at least -- seemed focused on the security challenge posed by the explosion in Chinese microchip design and production and th eimpact on America's strategic position. Alarmed by the leakage of U.S. technology to China, the DoD report even proposed bilateral Wassenaar-type agreements with Japan and Taiwan on SME exports to China. Incidentally, the DoD report also bemoaned the fact that Commerce Department microchip export rules are always out of date, and hence there is business pressure on the licensing offices to bend their own rules to "keep up with the times."

According to the Defense Science Board report, the strategic threat to the United States in the semiconductor sector is significant in two contexts: 1) the globalization of the microchip supply chain is draining production capacity from the United States and in a crisis it would be difficult to ramp up domestic output; 2) there is a real threat that microchip supplies from overseas -- particularly from China -- would be untrustworthy; that "opportunities for adversaries to clandestinely manipulate technology used in U.S. critical microelectronics applications are enormous and increasing."

In other words; not only is the Pentagon finding fewer and fewer sources for application specific integrated circuit microchips for highly classified defense applications (such as signals processing, encryption, guidance systems, etc.) but the US military already relies heavily on China for the unclassified laptops and PCs that are the bulk of the nervous system of our network-centric warfare doctrine. It is all well and good to say that the US simply won't buy Chinese-made computers for our military, but what happens when the global supply-chain means all laptops and PCs have some Chinese components in them?

Simply answering that 70 percent of China's advanced technology exports are made by non-Chinese companies is inadequate. As microcircuitry architecture becomes orders of magnitude denser than today, it becomes ever easier to hide lines that serve as Trojan Horse circuit designs, radio-frequency receivers and other "backdoors" to circumvent encryption, muddle signals, induce data failure and the like.

Are Chinese semiconductor firms capable of such chicanery? Chinese advanced technology companies have already proved themselves adept at down-loading and pirating tapeouts and masks that have been sent to contract fabs for mass production. And there are already several hundred semiconductor design labs in China -- sponsored and paid-for by foreign firms including America's top microchip corporations. While one American semiconductor design engineer told me this week that he did not think the Chinese designers he worked with were "smart enough" to handle the task of sabotaging circuit maps, he admitted that his Israeli colleagues were.

This is hardly reassuring. I suspect that US-sponsored semiconductor design labs in China lose engineers as they gain experience only to have them replaced by inexperienced engineers in need of new training. No doubt, experienced engineers are siphoned off by Chinese government, military and academic units to work on more advanced projects.