Type I error is defined as the error of rejecting a null hypothesis when it is true. In this case type error would be the error of inferring that the bags are less than 12.0 ounces (less than claimed value) but in reality its weight is either 12.0 ounces or even greater.
Level of significance refers to the probability of type I error, that means a fixed probability, in statistical hypothesis testing, of wrongly rejecting a null hypothesis Ho, when it is true. It is represented by .
As evident from the given problem, the investigator had a doubt that the claimed weight of the potato bags is greater than the actual weight. To verify the authenticity of this claim, he collected some 30 bags and found the mean of those bags, which came out to be 11.9 ounces. Although the mean weight came out to be lesser than the claimed one. But the real question or logic behind hypothesis testing is that we want to ascertain that whether it would be appropriate to consider the difference of 0.1 ounces from observing 30 bags with the standard deviation of 0.4 a 'significant' one and infer this difference as on the entire population.