The theoretical probability of randomly choosing a vowel after removing the cards that spell "vowel" is:
\[
\boxed{\frac{1}{7}}
\]
There are 12 cats and 8 dogs, so the total number of pets is:
\[
12 + 8 = 20
\]
The probability \( P \) of randomly choosing a cat is:
\[
P = \frac{\text{Number of cats}}{\text{Total number of pets}} = \frac{12}{20} = \frac{3}{5}
\]
The theoretical probability that a randomly chosen pet is a cat is:
\[
\boxed{\frac{3}{5}}
\]
Two days later, there are half as many cats. The original number of cats was 12, so the new number of cats is:
\[
\frac{12}{2} = 6
\]
The new probability \( P_{\text{new}} \) of randomly choosing a cat is half of the original probability:
\[
P_{\text{new}} = \frac{1}{2} \times \frac{3}{5} = \frac{3}{10}
\]
Let \( D \) be the number of dogs two days later. The total number of pets two days later is \( 6 + D \). The probability of choosing a cat is:
\[
\frac{6}{6 + D} = \frac{3}{10}
\]
Cross-multiply to solve for \( D \):
\[
6 \times 10 = 3 \times (6 + D)
\]
\[
60 = 18 + 3D
\]
\[
60 - 18 = 3D
\]
\[
42 = 3D
\]
\[
D = 14
\]
The number of dogs in the pet store two days later is:
\[
\boxed{14}
\]